(ISSI Scientific Report Series 13) Eric Quémerais, Martin Snow (Auth.), Eric Quémerais, Martin Snow, Roger-Maurice Bonnet (Eds.) - Cross-Calibration of Far UV Spectra of Solar System Objects and The H

ISSI Scientific Report 13
Eric Quémerais
Martin Snow
Roger-Maurice Bonnet
Editors
Cross-Calibration
of Far UV Spectra of
Solar System Objects
and the Heliosphere
ISSI Scientific Report Series
Volume 13
For further volumes:

http://www.springer.com/series/10151
Eric Quémerais • Martin Snow
Editors
Cross-Calibration of Far UV
Spectra of Solar System
Objects and the Heliosphere
123
Editors
Eric Quémerais Martin Snow
LATMOS-IPSL Laboratory Atmospheric & Space Physics
Université Versailles-Saint-Quentin University of Colorado
Guyancourt, France Boulder, CO, USA
International Space Science Institute
Bern, Switzerland
ISBN 978-1-4614-6383-2 ISBN 978-1-4614-6384-9 (eBook)

DOI 10.1007/978-1-4614-6384-9
Springer New York Heidelberg Dordrecht London
Library of Congress Control Number: 2012956300
© Springer Science+Business Media New York 2013

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the mate-
rial is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting,
reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval,
electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter devel-
oped. Exempted from this legal reservation are brief excerpts in connection with reviews or scholarly analysis or
material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive
use by the purchaser of the work. Duplication of this publication or parts thereof is permitted only under the
provisions of the Copyright Law of the Publisher’s location, in its current version, and permission for use must
always be obtained from Springer. Permissions for use may be obtained through RightsLink at the Copyright
Clearance Center. Violations are liable to prosecution under the respective Copyright Law.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does
not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective
laws and regulations and therefore free for general use.
While the advice and information in this book are believed to be true and accurate at the date of publication,
neither the authors nor the editors nor the publisher can accept any legal responsibility for any errors or omissions
that may be made. The publisher makes no warranty, express or implied, with respect to the material contained
herein.
Printed on acid-free paper
Springer is part of Springer Science+Business Media (www.springer.com)

Foreword
Detecting and measuring the Sun’s ultraviolet emission absorbed by the Earth’s
atmosphere has been one of the earliest scientific objectives of space research and
continues, even more than 50 years after the launch of the first sounding rockets
and artificial satellites, to be of interest to several space research programs and mis-
sions. This is essentially because the electrical and chemical balance of planetary
atmospheres is directly determined by the strongly variable ultraviolet radiation
from the Sun below 300 nm. Of particular interest is the presence in the solar
spectrum of the intense emission of the resonance Lyman-α line of atomic hydro-
gen at 121.6 nm and its equivalent from neutral helium at 58.4 nm. These strong
resonance emissions from the two most abundant atoms in the universe are at the
origin of the interplanetary ultraviolet background, which results from light scat-
tering by these atoms present in the interplanetary medium and the heliosphere.
Measuring with the best precision this background emission allows a study of the
penetration of interstellar gas from the Milky Way into the solar system, thereby
offering a unique tool to study the dynamics of our solar system as it moves through
its mother galaxy.
Measuring ultraviolet radiation in absolute terms is one of the most delicate
and difficult tasks for experimenters involved in solar and planetary projects. In
most cases, ultraviolet instruments are well calibrated on the ground but, unfortu-
nately, optics and detectors in the FUV are very sensitive to contaminants and it
is very challenging to prevent contamination before and during the test and launch
sequences of a space mission. Therefore, ground calibrations need to be confirmed
after launch and it is necessary to keep track of the temporal evolution of the
sensitivity of the instrument during the mission. This is even more important for
solar instruments directly illuminated by the Sun’s radiation, although this book
is primarily concerned with instruments designed for planetary and interplanetary
missions. Even more difficult is to reconcile several measurements made at dif-
ferent times with different instruments calibrated with different techniques, in the
laboratory before launch or in orbit, using celestial objects selected as radiation
standards.
Although the International Space Science Institute (ISSI) is mostly involved in
the analysis of data obtained from space missions, it has naturally devoted many
of its working tools to the calibration of various datasets in an attempt to estab-
lish common standards against which they might ultimately and quantitatively be
compared. For example, the former ISSI Working Group on the calibration of So-
lar Heliospheric Observatory (SOHO) instruments (The Radiometric Calibration
of SOHO, A. Pauluhn, M.C.E. Huber & R. von Steiger Eds., ISBN 1608-280X,
2002), was a successful endeavour to compare on a standard scale the intensities of
various solar phenomena observed by SOHO with different ultraviolet and X-ray
detectors.
Over the past seventeen years of ISSI’s existence, heliospheric research has been
one main area of investigation by this institute, as is illustrated by the list of ISSI
books published so far in the ISSI Space Science Series. These publications are
V
VI Foreword
accessible through the ISSI web page. Indeed, several of the authors of this study
have been involved one way or another in these previous ISSI activities.
Back in 2008, the main author of this project, E. Quémerais, responded to ISSI’s
annual call for proposals to convene an international team to study the possible
ways to calibrate far ultraviolet instruments after their launch. His proposal’s goal
was to gather as much information as possible on the best approaches for calibrating
future experiments. The ISSI Science Committee in the course of its review process
did not consider that proposal appropriate for the “International Team” category.
An ISSI “Team” offers only a few opportunities to work over one to two years
with the outcome of the team’s effort published in the peer-reviewed literature in
one or several usually unrelated articles dedicated to the subject. With the strong
support of ISSI, however, the Science Committee approved the formation of an ISSI
Working Group dedicated to the establishment of a defined set of standards for the
calibration of ultraviolet instruments on heliospheric and planetary missions. This
book published in the ISSI Scientific Report Series represents the outcome of the
working group.
Hopefully, more heliospheric and planetary instruments will be launched in
the future, acquiring more data with better detectors and benefiting from better
calibration techniques. A natural response of ISSI to such progress would be to
consider this work as just the beginning of a long-term effort aimed at a more
precise understanding of the phenomena, which characterise the interaction of our
solar system with its home galaxy. Therefore, this book could be considered to be
the first of a series dedicated to such an ambitious goal.
Bern, Switzerland Roger-Maurice Bonnet

Contents
Foreword V
1 Introduction 1
Part I: Interstellar Hydrogen and the Sun 5

2 Distribution of Interstellar Hydrogen Atoms
in the Heliosphere and Backscattered Solar Lyman-α 7
3 Solar Parameters for Modeling the Interplanetary

Background 67
Part II: Interplanetary Hydrogen 139

4 Thirty Years of Interplanetary Background Data:
A Global View 141
5 Lyman-α Models for LRO LAMP from MESSENGER

MASCS and SOHO SWAN Data 163
6 New Horizons Cruise Observations of Lyman-α Emissions

from the Interplanetary Medium 177
Part III: Instrument Cross-Calibration 189

7 A New Catalog of Ultraviolet Stellar Spectra
for Calibration 191
8 Absolute Ultraviolet Irradiance of the Moon

from the LASP Lunar Albedo Measurement
and Analysis from SOLSTICE (LLAMAS) Project 227
9 Lyman-α Observations of Comet Holmes

from SORCE SOLSTICE and SOHO SWAN 255
Index 269
VII
—1—
Introduction
In this book we discuss the problems encountered for calibrating space-borne
instruments in the spectral range 50–300 nm. We also discuss the level of accuracy
that can be achieved, and review the different sources of emissions that can be
used in that range, i.e. the interplanetary background, stars, and solar system
bodies such as planets, moons, and comets. As most of the work presented here
is based on comparisons with older datasets obtained by experiments that have
been calibrated independently, we also analyse the consistency between existing
datasets. Whenever possible, we suggest a correction factor if an older dataset
is in disagreement with a more recent one. For example, a comparison of the
interplanetary background data acquired by the two ultraviolet (UV) spectrometers
on Voyager 1 and Voyager 2 with models and other more recent datasets shows that
it is indeed necessary to derive new calibration factors for these two instruments.
The interplanetary ultraviolet background emission was first observed at the end
of the 1960’s and is due to the backscattering of solar photons by interstellar hydro-
gen and helium atoms present in the interplanetary medium. The corresponding
resonance emissions (Helium at 58.4 nm and Hydrogen at 121.6 nm) can be observed
in every direction of the sky as long as the observer is sufficiently away from plan-
etary exospheres. For planetary missions, it is often possible to observe the inter-
planetary emission during their long cruise. During the last three decades, several
ultraviolet instruments have indeed observed these emissions. Hence, the possibil-
ity of using the interplanetary ultraviolet background to calibrate FUV instruments
in particular at the Hydrogen Lyman-α wavelength is of great importance. This
is a difficult task because the interplanetary Lyman-α emission is highly variable,
both spatially and temporally. The measured brightness strongly depends on the
position of observation in the heliosphere. It also depends on the direction of the
line of sight in the sky, mainly as a function of the angle between the line of sight
and the direction of the Sun. The interplanetary background emission also varies
with solar activity. Therefore, highly detailed models are necessary to represent all
these variations accurately.
The first section of the book (Chaps. 2 and 3) is logically dedicated to the
establishment of an accurate model of the spatial distribution of neutral interstellar
hydrogen in the heliosphere. Chapter 2 gives an overview of the main physical
phenomena that need to be taken into account for developing such a model based
on the interaction between the solar wind and the Local Interstellar Medium. It
presents a state-of-the-art 3D kinetic model of the interstellar hydrogen gas inside
the heliosphere based on the various inputs described in Chap. 3, in particular the
solar factors that are responsible for shaping the distribution of neutral interstellar
E. Quémerais et al. (eds.), Cross-Calibration of Far UV Spectra of Solar System 1

Objects and the Heliosphere, ISSI Scientific Report Series 13,
DOI 10.1007/978-1-4614-6384-9 1, © Springer Science+Business Media New York 2013
2 1. Introduction
hydrogen in the heliosphere. This concerns the solar Lyman-α flux and the resonant
radiation pressure force acting on neutral hydrogen atoms in the heliosphere, as well
as the process of photoionization of heliospheric hydrogen, and how these evolve
in time and with heliolatitude. Chapter 3 also deals with changes in the neutral
hydrogen ionization process by charge exchange with protons of the solar wind,
and also with changes in electron impact ionization, induced by solar activity. The
chapter considers their variations with time, heliolatitude and solar distance.
Based on these studies, the heliosphere can be divided into two regions. In
the inner heliosphere, the hydrogen distribution is dominated by the influence of
the Sun. There, the interplanetary background is modulated by the changes in
radiation pressure, solar wind flux and solar wind ram pressure. In the outer he-
liosphere, the hydrogen distribution is dominated by the influence of the interface
region between the expanding solar wind and the ionized part of the local inter-
stellar medium.
Chapter 4 compares the results obtained with models, such as those presented
in Chaps. 2 and 3, to the different datasets so far obtained in the outer helio-
sphere with the Voyager ultraviolet spectrometer (UVS), ALICE New Horizons,
and in the inner heliosphere Solar Wind ANisotropies (SWAN) instrument on the
Solar Heliospheric Observatory (SOHO), and the Space Telescope Imaging Spec-
trograph (STIS) on the Hubble Space Telescope (HST) instruments. Because these
datasets do not necessarily overlap, models are used to bridge the gaps in distance
or time. Combining these datasets with the model results allows one to derive
calibration factors that give a coherent picture of the photometric behaviour of
the various instruments. For example, the STIS calibration of interplanetary back-
ground observations is based on International Ultraviolet Explorer (IUE) stellar
observations as well as on those obtained with SPectroscopy for the Investigation
of the Characteristics of the Atmosphere of Mars (SPICAM) on Mars Express, as
described in Chap. 7. SWAN is calibrated at Lyman-α through direct comparisons
with interplanetary background measurements made by STIS. Therefore the SWAN
calibration ultimately derives from the IUE stellar measurements. The model of
Chap. 2 is used to fit the observations of SWAN and its extrapolation to the outer
heliosphere, allowing a comparison with the observations of ALICE New Horizons
and of the UVS-Voyager background measurements. This example illustrates how
all these successive operations lead to a coherent calibration ensemble originating
from the IUE stellar observations.
Chapter 5 presents a comparison of three datasets obtained in the inner he-
liosphere where the influence of the solar parameters is most important. Some
systematic discrepancies between two of the datasets (SWAN SOHO and MASCS
MESSENGER) can be alleviated by changing the calibration of one of the instru-
ments. This demonstrates that comparison of interplanetary background data can
be used to inter-calibrate UV instruments.
Chapter 6 presents a new set of observations shown here for the first time
and performed by the ALICE New Horizons instrument between 9 and 17 AU
from the Sun. The Voyager 1 and Voyager 2 spacecraft visited this region of
the heliosphere nearly 30 years before the New Horizons spacecraft. The inter-
comparison of the Voyager and New Horizons datasets is therefore important for
our current study because the interplanetary background is rather easy to model
Introduction 3
between 10 to 20 AU from the Sun. The hydrogen distribution at that heliocentric

distance is not strongly affected by the temporal variations of the Sun or by the
heliospheric interface.
The last section of the book deals with different aspects of the calibration of vari-
ous instruments. Chapter 7 presents a description of the catalogue of stars observed
with the absolutely calibrated SOLar-STellar Irradiance Comparison Experiment
(SOLSTICE) on the Solar Radiation and Climate Experiment (SORCE), including
comparisons to widely used databases of observations from previous missions such
as IUE. From a set of IUE reference stellar spectra, it is possible to derive the
calibration of an instrument through a direct comparison with observations made
with that instrument of the same reference stars. The case of SPICAM-UV on
Mars-Express is also discussed. Stellar observations of these two spectrometers are
compared to IUE observations of the same stars. Of course, the final accuracy
achieved through this process depends on the accuracy of the reference spectra.
Because SOLSTICE was accurately calibrated on the ground and its temporal
evolution closely monitored, it provides its own independent calibration, showing
very good agreement between the two methods and demonstrating that the White
Dwarf scale used to calibrate the HST UV instruments is in good agreement with
the SOLSTICE calibration between 130 nm and 300 nm.
The last two chapters concern ultraviolet observations of solar system bodies.
Chapter 8 deals with lunar observations performed by SOLSTICE. The Moon has
been shown to be an extremely stable radiometric reference for calibration and for
checking the long-term stability of sensors in orbit. As the majority of previous
measurements have been made in the visible using ground based lunar images,
SOLSTICE data allow extending the lunar spectral irradiance dataset to the 115–
300 nm range. Recently launched planetary missions have also made ultraviolet
observations of the Moon during Earth fly-by, and the SOLSTICE measurements
offer the possibility of calibrating the absolute response of both Earth and planetary
space missions.
The last chapter compares Lyman-α observations of Comet Holmes from SWAN
on SOHO with near-simultaneous observations from SOLSTICE. The comparison,
though promising, was somewhat affected by contamination from stellar radia-
tion. The mismatch in the respective fields of view of these two instruments also
makes detailed comparisons challenging. However, this chapter shows that care-
fully planned campaigns of observations of comets by different instruments could
provide good results for cross-calibration in the ultraviolet wavelength range.
Finally, as mentioned earlier, many other datasets do not appear in this work,
and these should be added in a future edition of this book, as more UV instruments
will provide new spectral measurements of stars and of solar system objects. To
name a few, it would be of great interest to add observations from the Pioneer
10 and 11 UV spectrometers, the UV instruments on Galileo and Cassini, the
SPICAV-UV instrument of Venus-Express, and hopefully many others to come.
One aim of this project is to track relevant data and to give links to the respec-
tive archives from where they can be accessed. The links are and will be provided on
the web page that was set up for this working group: http://bdap.ipsl.fr/fondue/.
Part I
Interstellar Hydrogen
and the Sun
—2—
Distribution of Interstellar Hydrogen

Atoms in the Heliosphere and
Backscattered Solar Lyman-α
Vladislav V. Izmodenov∗
Lomonosov Moscow State University, School of Mechanics and Mathematics,
Institute for Problems in Mechanics, Russian Academy of Sciences,
Moscow, Russia
Space Research Institute, Russian Academy of Sciences, Moscow, Russia
Olga A. Katushkina
Space Research Institute (IKI) Russian Academy of Sciences,
Moscow, Russia
Lomonosov Moscow State University, School of Mechanics and Mathematics,
Moscow, Russia
Eric Quémerais
LATMOS-IPSL,
Université Versailles-Saint Quentin, Guyancourt, France
Maciej Bzowski
Space Research Center,
Polish Academy of Science, Warsaw, Poland
Abstract
We review the modern concepts of penetration of interstellar atoms of hydrogen
into the heliosphere up to 1 AU. Before entering into the heliosphere the atoms
penetrate through the region of the solar wind (SW) interaction with the local in-
terstellar medium (LISM). In the interaction region the atoms can exchange charge
with both solar wind and interstellar protons disturbed in the SW/LISM interac-
tion region. Charge exchange results in a disturbance of the pristine interstellar
atom flow in the interaction region, and, therefore, the parameters of interstellar
gas inside the heliosphere are different from their interstellar values. This makes
it more difficult to determine local interstellar parameters from measurements of

8 2. Interstellar Hydrogen and Backscattered Lyman-α
the interstellar atoms inside the heliosphere, but, on the other side, opens possi-
bilities to study the SW/LISM interaction region remotely. This paper overviews
the main physical phenomena and modern models of the SW/LISM interaction
and presents a state-of-art 3D kinetic model of the interstellar hydrogen gas in-
side the heliosphere. The distributions of the gas parameters are compared with
the distributions obtained in the context of the classical hot model. Quantitative
and qualitative differences are discussed. The state-of-art model is employed to
calculate spectra of the backscattered Lyman-α radiation as they would be mea-
sured at 1 AU and the zero, first and second moments of the spectra. It is shown
that the SW/LISM interaction imprints in the spatial and velocity distribution of
the interstellar atoms are revealed in the intensities, line-shifts, and line-widths
of the distribution functions. A qualitative comparison of the model results with
SOHO/SWAN data are presented.
Introduction: A Brief Historical Review

The first evidence of the presence of interstellar atomic hydrogen in the inter-
planetary medium was obtained from rocket measurements of stellar ultraviolet
(UV) radiation, when a strong diffuse UV radiation at 105–122.5 nm was observed
instead of the expected stellar point sources. In the discussion of their rocket
night-flight results, Kupperian et al. (1959) interpreted this emission as due to so-
lar Lyman-α photons scattered by interplanetary H atoms. Similarly, Shklovsky
(1959) interpreted these rocket results, together with night time H alpha measure-
ments, in the same way. However, he mentioned an alternate explanation “..which
cannot be excluded for the time being,” in which this night time Lyman-α emission
would be produced by resonant scattering of H atoms linked to the Earth, in an
extended atmosphere that he called the “geocorona.” In order to discriminate the
sources, Morton and Purcell (1962) performed a night time rocket measurement of
diffuse Lyman-α radiation using an absorption cell technique (for a description of
the technique see, e.g., Bertaux and Lallement 1984). These measurements showed
that only 15 % of the observed scattered Lyman-α radiation was at a wavelength
shifted by more than 0.004 nm from the line center. 85 % of the absorbed radiation
was produced by transport of Lyman-α photons from the day side to the night side
by radiative transfer in an extended exosphere, while the remaining 15 % was highly
Doppler-shifted, possibly produced by precipitating magnetospheric protons, or of
extra-terrestrial origin. To produce such wavelength-shifted photons, the scatter-
ing gas should be either hot (to produce a broad spectral linewidth) or in motion
with respect to the Sun (to produce Doppler-shifted protons).
Patterson et al. (1963) proposed a scenario for the hot atomic hydrogen within
the heliosphere and obtained the first analytical expression for the distribution of
the atomic hydrogen in interplanetary space. This work was based on the concept of
the heliospheric interaction with the surrounding interstellar (or galactic) medium
suggested by Axford et al. (1963) and shown schematically in Fig. 2.1.
It was assumed in the model that the solar wind blows continuously and is spher-
ically symmetric outward from the Sun with a nearly constant, highly supersonic
Introduction: A Brief Historical Review 9
Figure 2.1: A sketch of the structure of the interaction region between the solar
wind and the interstellar magnetic field proposed by Axford et al. (1963). The solar
wind is supersonic in Region I, and the solar magnetic field lines form Archimedean
spirals co-rotating with the Sun. The termination shock is located at a heliospheric
distance S. Beyond the termination shock, in region II (the boundary shell), solar-
wind protons and interstellar neutral hydrogen atoms exchange charge, and dissi-
pative effects permit oppositely-directed solar wind magnetic-field lines to merge
and form closed loops. S∗ is the distance to the boundary between the solar and in-
terstellar magnetic fields. Blobs of the solar plasma and magnetic field detach from
region II and move out into region III (interstellar space), where they gradually
diffuse away
velocity. Hence, the solar wind density and dynamic pressure are directed radially
away from the Sun and decrease as ∼ 1/r2 . At the distance where the dynamic
pressure (magnetic pressure and/or thermal pressure) equals the pressure of the
local interstellar medium (LISM), the solar wind passes through a standing shock—
the termination shock (TS), beyond which it slows down and becomes subsonic.
Nevertheless, it was assumed that in the region beyond the shock front (that we
call now the inner heliosheath), the individual velocities of solar wind protons re-
main of the order of ∼ 300 km/s due to high thermal velocities. The energetic
protons undergo charge exchange with neutral interstellar hydrogen. This process
leaves low-energy protons in the transition region beyond the shock, and provides
an isotropic source of high-energy neutral hydrogen atoms. Some of these atoms
(now called energetic neutral atoms—ENAs) move through the termination shock
(TS) back toward the Sun.
Patterson et al. (1963) calculated the distribution of the neutral hydrogen atoms
in interplanetary space assuming that the TS is an isotropic source of the neutrals,
taking into account their velocity, the ionizing solar radiation, and the charge
exchange with the solar wind. It was also assumed that (1) the post-shocked solar
wind protons suffer charge exchange with interstellar neutrals just in a thin shell
beyond the TS, (2) half of the solar wind flux is returned into the heliosphere
in the form of fast neutrals. Later, Hundhausen (1968) showed that most of the
neutral hydrogen observable in the vicinity of the Earth does not undergo charge
exchange between solar wind protons and interstellar neutral hydrogen near the
shock boundary of the heliosphere, as assumed by Patterson et al. (1963), but in a
region far beyond the shock. Hundhausen’s model predicted much smaller atomic
hydrogen density near Earth. Therefore, the neutral hydrogen density near Earth
could then only be maintained in the model if the shock is near 5 AU.
An alternative approach, that actually became a commonly accepted paradigm
for at least two decades, was suggested by Fahr (1968a). We should point out that
modern concepts of the ENAs originating in the inner heliosheath (Gruntman et al.
2001) are very similar to those of Patterson et al. (1963) and Hundhausen (1968).
Fahr (1968a) pointed out that the assumption of the random motion of inter-
stellar hydrogen made by Patterson et al. (1963) is very unlikely because the solar
system itself has a velocity of 20 km/s with respect to the local standard of rest
of nearby stars, and also interstellar hydrogen clouds are moving. It was shown
(Blum and Fahr 1970; Axford 1972) that the concept of the Strömgren sphere of
ionized gas around a hot star fails for the Sun, in case that there is a relative bulk
motion of the interstellar atoms with respect to the Sun. The Strömgren sphere
is determined by equating the flux of ionizing photons emitted by the Sun to the
total recombination rate (Strömgren 1939). In case of the Sun, the Strömgren
sphere has a radius of about 1,500 AU. It was estimated (Blum and Fahr 1970)
that about 90 % of the cold interstellar hydrogen with a velocity of 20 km/s will
enter the heliosphere without being photoionized or charge exchanged.
Fahr (1968a) has calculated the heliospheric distribution of the number density
of the cold interstellar hydrogen taking into consideration the macroscopic motion
of interstellar gas and the solar gravitational field. Blum and Fahr (1970) took into
account the losses of interstellar atoms due to charge exchange with the solar wind
protons and due to photoionization by solar radiation. Therefore, in the model, the
cold interstellar neutrals enter the solar system along Kepler trajectories and suffer
losses caused by the EUV-ionization and charge exchange with the solar wind
protons. The charge exchange process produces secondary fast neutrals having
the velocity of the solar wind. These particles move radially outwards without
significant losses within the solar wind (Fahr 1968b), and cannot be observed in
Lyman-α because they are Doppler shifted too far from the center of the solar line.
According to the prediction of the cold model of Blum and Fahr (1970), the
number density of interstellar H atoms are larger in the direction toward the
interstellar flow, and, therefore, the observed Lyman-α emission should have a
maximum, and the maximum should be toward the direction of the interstellar
H atom flow. Extraterrestrial UV radiation (most likely Lyman-α) was measured
from interplanetary probes Zond 1 (Kurt 1966, 1967), of Venera 2, 3, and 4 (Kurt
and Germogenova 1967; Kurt and Syunyaev 1968), and of Mariner 5 and 6 (Barth
1970). These measurements proved that the 15 % highly shifted night time Lyman-
α found earlier by Morton and Purcell (1962) in their rocket flight was indeed
of extra-terrestrial origin. It was found in these interplanetary probes measure-
ments that the backscattered Lyman-α emission is distributed inhomogeneously
Introduction: A Brief Historical Review 11
and features toward the direction of the galactic center. Because of this specific
direction of the emission maximum, a new “galactic” source of the diffuse radiation
was suggested as an alternative to Patterson et al. (1963) scenario.
The dilemma of the galactic or interstellar origin of the diffuse radiation was
resolved by OGO-5 spacecraft measurements (Bertaux and Blamont 1971; Thomas
and Krassa 1971; Lallement 2001). The apogee of OGO-5 was out of the geocorona,
i.e. the geocorona emission could be ruled out at apogee. The spacecraft was spin-
ning so that the extraterrestrial diffuse Lyman-α radiation could be mapped over
the sky. A broad maximum in one direction and a minimum in the opposite direc-
tion were observed. To find the distance to the region of the maximum emissivity
of the backscattered Lyman-α emission two sky-mappings of the diffuse Lyman-α
were performed on board the OGO-5 spacecraft in September 1970 and April 1970,
when the Earth (and the spacecraft too) was at opposite positions in its orbit. The
difference in the position of the maximum of diffuse Lyman-α emission as seen in
the two maps was by about 30◦ . This proved that the emission region was located
at few AU from the Sun and the displacement of the maximum was due to the
parallax effect that would be negligibly small in the case of a galactic source of the
emission. Analysis of three sky-maps of the OGO-5 allowed to find the direction of
the interstellar H atoms approaching the Sun (Bertaux et al. 1972). Newer values
of the direction of interstellar hydrogen inside the heliosphere are (252.5◦ , 8.9◦ ) in
the ecliptic J2000 coordinates (Lallement et al. 2010).
Therefore, the theoretical prediction made by Fahr (1968a) and Blum and Fahr
(1970), i.e. that interstellar H atoms penetrate into the heliosphere up to distances
of a few AU from the Sun, was confirmed by Lyman-α measurements made by
OGO-5 spacecraft instruments in 1969–1970 (Bertaux and Blamont 1971; Thomas
and Krassa 1971; Thomas 1972), and the “galactic” interpretation was ruled out.
An upper limit of galactic emission of 15 Rayleigh (compared to the observed 300–
500 R of the extra-terrestrial emission) was derived from hydrogen absorption cell
measurements on board Prognoz-5 and 6 (Lallement et al. 1984).
Since the interstellar temperature is ∼ 104 K, the real situation is drastically
different from the cold model considered above. The cold model fails for the down-
wind direction at the axis of symmetry, where it gives infinite densities in case when
the solar gravitation is larger than the radiation pressure (μ < 1). The quantity μ
is defined as the ratio of the solar gravitational attraction to the repulsion due to
radiation pressure and will be used throughout this document:
μ = |Frad |/|Fg |. (2.1)
Actually, Danby and Camm (1957) were the first to solve the problem of the
motion of a cloud of gas particles in the gravitational field of a point mass. They
found an analytical formula for the velocity distribution function of the particles.
Later, the Danby–Camm formula was applied to the interstellar atoms moving in
the heliosphere and modified by including a loss function to take into account the
effects of charge exchange and photoionization. Details of the hot model formula-
tion will be given later in this paper (Lallement et al. 1985a; Izmodenov 2006).
The problem with finite temperature has been considered by Fahr (1971),
Thomas (1972), Feldman et al. (1972), Bertaux et al. (1972), Fahr (1974), Blum
et al. (1975), Meier (1977), Fahr (1978), and Wu and Judge (1979). Such a model
is called the classical hot model of hydrogen distribution. Very recently this clas-
sical hot model has been applied by Lee et al. (2012) to analyses of atom fluxes
measured at 1 AU by the Interstellar Boundary Explorer (IBEX). Measurements
of the backscattered solar Lyman-α radiation with a hydrogen absorption cell on
board the Soviet interplanetary probe Mars-7 (Bertaux et al. 1976) brought in-
formation on the spectral profile of the extraterrestrial Lyman-α radiation. More
accurate measurements with a hydrogen absorption cell flown on Prognoz-5, inter-
preted with the assumption of a Gaussian velocity distribution of H atoms, yielded
a temperature of (8.8 ± 1) · 103 K (Bertaux et al. 1977). However, a more accurate
interpretation of the spectral profile is only possible if the velocity distribution
function of the interstellar hydrogen is taken into account, even for the simple case
of μ = 1, in which atoms trajectories are straight lines.
In the 1980–1990s the classical hot model was widely used to interpret backscat-
tered solar Lyman-α measurements. However, very soon it became clear that effects
connected with the heliolatitudinal and solar cycle variations of the solar wind must
be taken into account.
It has been shown through modeling by Joselyn and Holzer (1975) that a non-
isotropic solar wind would strongly affect the distribution of atomic hydrogen in
the heliosphere, and that this could be observed in maps of backscattered solar
Lyman-α radiation. The signatures of the latitudinal variations were observed
in the backscattered interplanetary Lyman-α glow observations from Mariner 10
(Kumar and Broadfoot 1978, 1979; Witt et al. 1979, 1981), Prognoz 6 (Lallement
et al. 1985b; Summanen et al. 1993), Pioneer–Venus (Ajello et al. 1987; Lallement
and Stewart 1990), SOHO/SWAN (Bertaux et al. 1997, 1999), and by the Ulysses
GAS Lyman-α measurements (Pryor et al. 2003).
Lallement et al. (1985b) performed an analysis of Prognoz 5 and 6 measurements
by comparing the model results with the varying “a” parameter that is responsible
for the anisotropy of the solar wind. It was shown that the best agreement between
the data and the model is achieved for a 30–50 % decrease in the ionization rate
over the solar pole in comparison with the equatorial plane. The anisotropy of the
total ionization was studied in much more details by the SOHO/SWAN instrument
(e.g. Bertaux et al. 1997, 1999). Studies of the total ionization rate variations with
latitude and solar cycle were also performed by Pryor et al. (2003).
It follows from the classical hot model that due to net effects of solar gravitation,
radiation pressure, charge exchange and photoionization, the velocity distribution
of the interstellar atoms are disturbed at about 15–20 AU in the upwind direction.
In the downwind direction the “solar imprints” remain up to ∼ 100 AU. A typical
interstellar atom with a velocity of 20 km/s travels about 4 AU per year. The
time which is needed for the atom to pass through the region where the ionization
and gravitational effects are significant is therefore comparable with the 11 year
solar cycle. Since the solar radiation and solar wind flux both change during the
solar cycle, calculating the time-dependent variation of interstellar atoms within
the heliosphere requires that both be taken into account (Blum et al. 1993; Kyrölä
et al. 1994; Bzowski and Ruciński 1995; Ruciński and Bzowski 1995; Summanen
1996; Bzowski et al. 1997, 2002; Pryor et al. 2003; Quémerais et al. 2006).
The Solar Wind/Local Interstellar Medium Interaction 13
The variability of the solar factors exerts a significant influence on the hydrogen
density distribution within 10–20 AU from the Sun and is the most pronounced in
the downwind region. Ruciński and Bzowski (1995) have modeled the number
density of interstellar hydrogen in the frame of a time-dependent hot model. It
was shown that the departures of the density profiles from the stationary model
at different phases of the solar cycle are clearly visible up to 5 AU on the upwind
side (approximately the same occurs in the sidewind direction) and to 15 AU in
the downwind region. Further away from the Sun, the differences decrease and
practically vanish beyond 15–20 AU in the upwind and 50–60 AU in the downwind
directions, respectively.
Bzowski and Ruciński (1995) calculated the distributions of the H atom number
density by using a number of stationary models with “instantaneous” values of the
radiation pressure and ionization rate for the considered phase of the solar cycle.
Comparison of these distributions with the results of non-stationary model has
shown significant differences. Therefore, a time-dependent approach is essential to
accurately model the interstellar H atoms within the heliosphere.
It is clear from the historical review above that the development of the models
of the interstellar atoms inside the heliosphere is closely connected with the mea-
surements of the backscattered solar Lyman-α radiation, and the requirements of
the data analysis stimulated the theoretical model development.
It is important to underline that all of the models described above assume
that interstellar atoms have a pristine interstellar Maxwellian distribution func-
tion before they start to interact with the heliosphere. However, by the 1970s
(Wallis 1975) it was realized that before penetrating inside the heliosphere, inter-
stellar atoms of hydrogen pass through the so-called heliospheric interface that is
the region of the solar wind interaction with the charged component of the local
interstellar medium. At the present time, there is no doubt that the neutral and
plasma components interact in the heliospheric interface by charge exchange, and
these components should be treated self-consistently in the models by taking into
account their mutual interactions.
The Solar Wind/Local Interstellar Medium

Interaction
The heliosphere as a circumsolar volume where the properties of the medium are
primarily controlled by the Sun was first conceptualized by Davis (1955) prior to
both Parker’s theoretical prediction of the existence of the solar wind (Parker 1958)
and its discovery (Gringauz et al. 1960; Neugebauer and Snyder 1962). According
to the current paradigm, the boundary of the heliosphere is located at a distance
of ∼ 90–300 AU (where AU is the astronomical unit, or the distance from the
Sun to the Earth, equal to 149.6 million kilometers). The nature and position of
this boundary, as well as the structure and properties of the outer heliosphere, are
governed by the interaction between the solar wind and the interstellar environment
of the Sun. At present, there is no doubt that the Sun is moving inside the Local
Interstellar Cloud (LIC), one of several very diffuse interstellar cloudlets found in
the vicinity of the Sun (Lallement 2001). These cloudlets (also called the Local
e
au s
li op
ck
He
ho
ws
H Bo Termina
t io
n
sh
o ck
SW
LISM
Figure 2.2: Qualitative picture of the SW interaction with the LIC. The heliopause
(HP) is a contact (or tangential) discontinuity which separates the solar wind
plasma and the interstellar plasma component. The termination shock (TS) is
formed due to the deceleration of the supersonic solar wind. The bow shock (BS)
may also exist if the interstellar plasma flow is supersonic. Four regions are distin-
guished: the supersonic solar wind (region 1); the solar wind flow between the TS
and the HP (region 2 or the inner heliosheath); the disturbed interstellar plasma
component flow (region 3 or the outer heliosheath); the undisturbed interstellar gas
flow (region 4)
Fluff) are in turn embedded in the Local Bubble that is believed to be made of hot
(∼ 106 K) and tenuous (∼ 10−3 cm−3 ) gas with a characteristic size on the order
of 100 pc (Welsh 2009).
The LIC is partially ionized (Lallement 2001). Interaction of the charged com-
ponent of the LIC with the solar wind plasma gives rise to the formation of the
interaction region, which is often called the heliospheric interface or heliosheath
(Fig. 2.2). The heliospheric interface has a complex structure, where the solar
wind and interstellar plasma, interplanetary and interstellar magnetic fields, inter-
stellar atoms of hydrogen, galactic and anomalous cosmic rays (GCRs and ACRs)
and pickup ions play important roles.
To get some insight into the structure and the physical processes inside the
interface using remote observations, a theoretical model must be employed.
Theoretical studies of the heliospheric interface have been performed for more than
four decades, following the pioneering work by Parker (1961) and Baranov et al.
(1971). However, a complete theoretical model of the heliospheric interface has not
yet been constructed. The basic difficulty stems from the multi-component nature
of both the LIC and the solar wind. The LIC consists of at least four main com-
ponents: plasma (electrons and protons), hydrogen atoms, interstellar magnetic
field and galactic cosmic rays. The heliospheric plasma consists of particles of solar
origin (protons, electrons, alpha particles, etc.), pickup ions and energetic parti-
cle components that include, for example, the termination shock particles (TSP)
(Stone et al. 2005; Decker et al. 2005; Burlaga et al. 2005) and the anomalous cos-
mic ray (ACR) component. Pickup protons (or ions) are created by processes of
charge exchange, photoionization and electron impact ionization, and after being
ionized immediately picked up by the magnetic field. The pickup protons modify
the heliospheric plasma flow starting from ∼ 20–30 AU from the Sun. TSPs and
ACRs may also modify the plasma flow upstream of the termination shock and in
the heliosheath.
The first self-consistent stationary model of the interaction of the two-component
(plasma and H atoms) LIC with the solar wind was developed by Baranov and
Malama (1993). The interstellar wind was assumed to be an homogeneous paral-
lel supersonic flow, and the solar wind was assumed to be spherically symmetric
at Earth’s orbit. Under these assumptions, the heliospheric interface has an axi-
symmetric structure. The main physical process of this interaction is resonance
charge exchange (H atoms with protons), although the processes of photoionization
and ionization of H-atoms by electron impact can be important in some regions
of the heliosphere (for example, in the inner heliosheath or in the supersonic solar
wind). The significant effect of the resonance charge exchange is connected with
the large cross section of such collisions that is a function of the relative velocity
of colliding particles. However, it was discussed by some authors (Williams et al.
1997) that elastic H–H and H–proton collisions can be important in the problem of
the solar wind interaction with the local interstellar medium. This specific question
was discussed in detail by Izmodenov et al. (2000) and it was shown that the elastic
collisions are negligible.
The first self-consistent model of the SW/circumhelipspheric interstellar medium
(CHISM) interaction was developed by Baranov and Malama (1993). This is an
axisymmetric and stationary two-component model. The plasma component is
quasi-neutral and consists of electrons and protons. It is assumed that pickup
protons are assimilated into the plasma component immediately after ionization.
The plasma component is described as a fluid, and the Euler equations are solved
to get the spatial distribution of the plasma number density, np (r), bulk velocity,
Vp (r), and pressure Pp (r). The neutral component consists of hydrogen atoms and
is described kinetically. The two components interact by charge exchange. Pho-
toionization and electron impact ionization are taken into account in the model
as well.
The main results of the model can be described as follows. The interstellar
atoms strongly influence the heliospheric interface structure. The heliospheric
a b
Figure 2.3: Effect of the interstellar H atoms on the geometrical pattern of the in-
terface. (a) The heliospheric interface pattern in the case of a fully ionized circum-
heliospheric interstellar medium (CHISM), (b) the case of partly ionized CHISM.
Here BS, HP, and TS are the bow shock, the heliopause and the termination shock,
respectively. MD and TD are the Mach disk and the tangential discontinuity; RS
is the reflected shock that is formed in the case of a fully ionized plasma. These
results were obtained initially by Baranov and Malama (1993). Region 1 is the
supersonic solar wind, 2 is the inner heliosheath between the TS and HP, 3 is the
outer heliosheath between the HP and BS [From Izmodenov and Alexashov (2003)]
interface is much closer to the Sun in the case when the H atoms are taken into
account in the model as compared to the pure gas dynamical case as shown in
Fig. 2.3.
The termination shock becomes more spherical and the flow in the region be-
tween HP and TS becomes subsonic (the sonic lines disappear). The Mach disk
and the complicated tail shock structure, consisting of the reflected shock (RS) and
the tangential discontinuity (TD), disappear as well.
The supersonic flows upstream of the bow and termination shocks are modified
due to charge exchange with the neutral component. The supersonic solar wind
flow (region 1 in Fig. 2.3) is modified by charge exchange with the interstellar
neutrals. The new protons created by charge exchange are picked up by the solar
wind magnetic field. The Baranov–Malama model assumes immediate assimilation
of the pickup ions into the solar wind plasma. The solar wind protons and pickup
protons are treated as one-fluid, called the solar wind. The number density, velocity,
temperature, and Mach number of the solar wind are shown in Fig. 2.4a.
The effect of charge exchange on the solar wind is significant. By the time the
solar wind flow reaches the termination shock, it is decelerated by 15–30 %, heated
by a factor of 5–8, and loaded with the pickup proton component (approximately
20–50 %).
a b c
1.6
TS
1.5 10−1 HP
ρ HP
ρR2
ρER2E 1.4 ρ
BS LIC
1.4
1.3
10−2
TS
1.2 BS
1.2
1.1 TS
np, cm−3
1.0 1.0 10−3
0 20 40 60 80 100 250 300 350 400 450 500 100 120 140 160 180 200
R (AU) R (AU) R (AU)
1.2 1.0 TS
1.0 102 Vp, km/s

0.8
BS BS
0.8
0.6 101
0.6
V 0.4 V
0.4 TS 100
VE VLIC
0.2 HP
0.2
HP
0 −1
10
0 20 40 60 80 100 250 300 350 400 450 500 100 120 140 160 180 200
10 2.0
TS
T T 106 TS
8 1.8
TE BS TLIC
6 1.6
BS
105 HP
4 1.4 HP
2 1.2 Tp, K
0 1.0 10 4
0 20 40 60 80 100 250 300 350 400 450 500 100 120 140 160 180 200
2.0 100
25 TS
Mach number
Mach number
20 1.5
10−1
15
1.0 BS
BS
10
10−2
0.5
5 HP
Mach number HP
TS −3
0 0 10
0 20 40 60 80 100 250 300 350 400 450 500 100 120 140 160 180 200
Figure 2.4: Plasma density, velocity, temperature, and Mach number upstream
of the termination shock (a), upstream of the bow shock (b), and in the helio-
sheath (c). The distributions are shown for the upwind direction. The solid curves
correspond to nH,CHISM = 0.2 cm−3 , np,CHISM = 0.04 cm−3 . The dashed curves
correspond to nH,CHISM = 0.14 cm−3 , np,CHISM = 0.10 cm−3 . VCHISM = 25.6 km/s
and TCHISM = 7,000 K in both cases [From Izmodenov (2000)]
The interstellar plasma flow is modified upstream of the bow shock by charge
exchange between the interstellar protons and secondary H atoms. These secondary
atoms originate in the solar wind. This leads to heating (40–70 %) and deceleration
(15–30 %) of the interstellar plasma before it reaches the bow shock (BS). The
Mach number decreases upstream of the BS and for a certain range of interstellar
parameters (nH,CHISM np,CHISM ) the bow shock may disappear. The solid curves
in Fig. 2.4b correspond to a small degree of ionization in the CHISM (np /(np +
nH ) = 1/6); the bow shock almost disappears in this case.
The interstellar neutrals also modify the plasma structure in the inner helio-
sheath. In a pure gasdynamic case (without neutrals) the density and temperature
of the postshock plasma are nearly constant. However, the charge exchange pro-
cess leads to a large increase in the plasma number density and decrease in its
temperature (Fig. 2.4c). The electron impact ionization process may influence the
heliosheath plasma flow by increasing the gradient of the plasma density from the
termination shock to the heliopause (HP, Baranov and Malama 1996). The influ-
ence of interstellar atoms on the heliosheath plasma flow is important, in particular,
for the interpretation of kHz-radio emissions detected by Voyager and for analysis
of the heliospheric ENA fluxes.
Charge exchange significantly alters the interstellar atom flow. The atoms newly
created by charge exchange have the velocity of their ion counterparts in the charge
exchange collisions. Therefore, the velocity distribution of these new atoms depends
on the local plasma properties in the location of their origin. It is convenient to
distinguish four different populations of atoms, depending on the region in the
heliospheric interface where the atoms were formed. Population 1 are the atoms
created in the supersonic solar wind up to the TS (region 1 in Fig. 2.3), population 2
are the atoms created in the inner heliosheath (region 2 in Fig. 2.3), and population
3 are the atoms created in the outer heliosheath (region 3 in Fig. 2.3). The atoms of
population 3 are often called the secondary interstellar atom component. We will
refer to the original (or primary) interstellar atoms as population 4. The number
densities and mean velocities of these populations are shown in Fig. 2.5I as functions
of heliocentric distance. The distribution function of H atoms, fH (r, w H ), can be
represented as a sum of the distribution functions of these populations: fH =
fH,1 + fH,2 + fH,3 + fH,4 . The Monte Carlo method allows us to calculate these four
distribution functions. These distributions were presented by Izmodenov (2001)
and Izmodenov et al. (2001) at 12 selected points in the heliospheric interface.
As an example, the distribution functions at the termination shock in the upwind
direction are shown in Fig. 2.5II for the four introduced populations of H atoms.
It is seen from this figure that the distribution functions of all H-atom populations
are not Maxwellian inside the heliosphere, i.e. the fluid approach is not correct
for describing the motion of neutral atoms. Comparisons of kinetic and different
multi-fluid approaches show significant differences in the results (Izmodenov and
Alexashov 2005; Mueller et al. 2008).
Original (or primary) interstellar atoms (population 4) are significantly fil-
tered (i.e. their number density is reduced) before reaching the termination shock
(Fig. 2.5I-a). The outer heliosheath is the main “filter” for these atoms. Since the
slow atoms have a small mean free path (due to both larger charge exchange cross
sections and smaller velocities) in comparison with the fast atoms, they suffer larger
I II
a c
a c
Population 2
b d
b d
Figure 2.5: I. Number densities and velocities of the four atomic populations as
functions of heliocentric distance in the upwind direction. 1 designates atoms cre-
ated in the supersonic solar wind, 2 atoms created in the heliosheath, 3 atoms
created in the disturbed interstellar plasma, and 4 original (or primary) interstel-
lar atoms. Number densities are normalized to nH,CHISM , velocities are normalized
to VCHISM . It is assumed that nH,CHISM = 0.2 cm−3 , np,CHISM = 0.04 cm−3 . II.
Velocity distributions of the four atom populations—primary interstellar atoms
(population 4), secondary interstellar atoms (population 3), atoms created in the
inner heliosheath (population 2), atoms created in the supersonic solar wind (pop-
ulation 1)—at the termination shock in the upwind direction; wz is the projection
of the velocity vector on the axis parallel to the LIC velocity vector. Negative
values of wz indicate approach to the Sun. wx is the magnitude of the projection
of the velocity vector on the plane perpendicular to the interstellar velocity vector;
wR ,wθ are radial and tangential velocity components. All velocities are in km/s
[From Izmodenov et al. (2001)]
losses. This kinetic effect, called selection, results in a deviation of the interstellar
distribution function from a Maxwellian (Fig. 2.5II-a). The selection also results
in a ∼ 10 % increase in the primary atom mean velocity towards the termination
shock (Fig. 2.5I-c).
Secondary interstellar atoms (population 3) are created in the disturbed inter-
stellar medium by charge exchange between the primary interstellar neutrals and
protons decelerated in the vicinity of the heliopause. The secondary interstellar
atoms collectively make up the hydrogen wall, a density enhancement at the he-
liopause. The hydrogen wall was predicted by Baranov et al. (1991) and detected
in the direction of α Cen (Linsky and Wood 1996) by the Hubble Space Telescope.
At the termination shock, the number density of secondary neutrals is comparable
to the number density of the primary interstellar atoms (Fig. 2.5I-a, dashed curve).
The relative abundances of secondary and primary atoms entering the heliosphere
vary with the degree of interstellar ionization. The bulk velocity of population 3 is
about −18 ÷ −19 km/s. The sign “−” means that the population approaches the
Sun. One can see that the distribution function of this population is not Maxwellian
(Fig. 2.5II-b). The reason for the abrupt behavior of the distribution function for
wz > 0 is that the particles with significant positive wz velocities can reach the
termination shock only from the downwind direction. The distribution functions
of different H atom populations were calculated by Izmodenov et al. (2001) for
various upwind directions. The fine structures of the distribution functions of the
primary and secondary interstellar populations vary with direction. The direc-
tional variation of the velocity distribution reflects the geometrical pattern of the
heliospheric interface. The distribution functions of the interstellar atoms can be
a good diagnostic of the global structure of the heliospheric interface.
Another population (population 2) of the heliospheric hydrogen atoms are the
atoms created in the inner heliosheath by charge exchange with hot and compressed
solar wind and pickup protons. The number density of this population is by an or-
der of magnitude smaller than the number densities of the primary and secondary
interstellar atoms. This population has a minor importance for the interpretation
of Lyman-α and pickup ion measurements inside the heliosphere. Some atoms of
this population may probably be detectable by a Lyman-α hydrogen cell experi-
ment due to their large Doppler shifts (Quémerais and Izmodenov 2002). Recently
it was pointed out by Chalov and Fahr (2003) that charge exchange of these atoms
with solar wind protons may produce tails in the distribution function of pickup
ions that are measured at one or several AU from the Sun during quiet time periods.
Gruntman and Izmodenov (2004) showed that this population of H-atoms is a ma-
jor contributor to the density of interplanetary hydrogen at heliocentric distances
< 1 AU and could dominate in the downwind (interstellar wind) region under typ-
ical solar and interstellar conditions. Mass transport by heliospheric ENAs may
become especially important for determining the origin of the pickup ions attributed
to the inner source of neutral particles in the Sun’s vicinity.
Due to their high energies and large mean free path, a portion of the atoms
from this population penetrate upstream of the BS and modify the pristine inter-
stellar medium at large heliocentric distances. These atoms propagate freely in the
supersonic solar wind. Thus, these atoms are a rich source of information on the
plasma properties at the place of their birth, i.e. at the inner heliosheath. This
population of atoms is measured by the Interstellar Boundary Explorer (IBEX).
The last population of the heliospheric atoms are the atoms created in the
supersonic solar wind (population 1). The number density of this atom popu-
lation takes its maximum at ∼ 5 AU. At this distance, the number density of
the population is about two orders of magnitude smaller than the number density
of interstellar atoms. Outside the termination shock the density decreases faster
than r −2 , where r is the heliocentric distance (curve 1, Fig. 2.5I-b). The mean
velocity of population 1 corresponds to the bulk velocity of the supersonic solar
wind and is about 450 km/s. The distribution function of this population is also
not Maxwellian (Fig. 2.5II-d). The extended “tail” in the distribution function is
caused by the solar wind plasma deceleration upstream of the termination shock.
This “supersonic” atom population penetrates the interface and charge exchanges
with interstellar protons beyond the BS. The process of charge exchange leads to
heating and deceleration of the interstellar gas upstream of the bow shock and,
therefore, to a decrease of the Mach number ahead of the bow shock.
Since 2003 the standard Baranov–Malama model has been modified by adding
solar wind alpha particles and interstellar helium ions to plasma components
(Izmodenov et al. 2003). Up to now, helium ions and alpha particles are not
included in the alternative models of other groups.
To evaluate possible effects of both interstellar ions of helium and solar wind
alpha particles, Izmodenov et al. (2003) performed parametric model calculations.
It was shown that the heliopause and the termination and bow shocks, are closer to
the Sun when the influence of interstellar helium ions is taken into account. This
effect is partially compensated by additional solar wind alpha particle pressure
that was also taken into account in the model. The net result is as follows: the
heliopause, termination and bow shocks are closer to the Sun by ∼ 12 AU, ∼ 2 AU,
∼ 30 AU, respectively in the model taking into account both interstellar helium
ions and solar wind alpha particles as compared to the model ignoring these ionized
helium components. Despite the fact that the net effect of interstellar helium ions
and solar alpha particles is rather small (∼ 7 % displacement for the heliopause,
∼ 10 % the BS, and ∼ 2 % for the TS), it still can be important for the interpretation
and prediction of experimental data related to the inner heliosheath region.
Aleksashov et al. (2000) explored the problem of solar wind interaction with the
CHISM for the case where the interstellar magnetic field is parallel to the relative
Sun/CHISM velocity vector. In this case, the model remains axisymmetric. It
was shown that the effect of the interstellar magnetic field on the positions of
the TS, BS and HP is significantly smaller when H atoms are included (Baranov
and Zaitsev 1995). The calculations were performed with various Alfvén Mach
numbers in the undisturbed CHISM. It was found that the √ bow shock straightens
out with decreasing Alfvén Mach number MA = VCHISM 4πρ/BCHISM (i.e. with
an increasing magnetic field strength in the CHISM). It gets closer to the Sun near
the symmetry axis, but recedes from it on the flanks. By contrast, the nose of the
heliopause recedes from the Sun due to the tension of magnetic field lines, while the
wings of the heliopause get closer to the Sun due to magnetic pressure. As a result,
the region of the compressed interstellar medium at the heliopause (or “pileup
region”) is reduced by almost 30 %, as the magnetic field increases from zero to
3.5 × 10−6 Gauss. It was also shown that the H atom filtration and heliospheric
distributions of primary and secondary interstellar atoms are virtually unchanged
over the entire considered range of the interstellar magnetic field (0–3.5 × 10−6
Gauss). The magnetic field has the strongest effect on the density distribution of
population 2 of H atoms, which increases by a factor of almost 1.5 as the interstellar
magnetic field increases from zero to 3.5 × 10−6 Gauss.
Izmodenov et al. (2005), as well as Izmodenov and Alexashov (2005, 2006),
studied the problem assuming that the interstellar magnetic field (IsMF) is inclined
with respect to the direction of the interstellar flow. In this case, the SW/CHISM
interaction region becomes asymmetric and the flow pattern becomes essentially
three-dimensional. Here we present new results obtained within the framework of
this model for the interstellar magnetic field BCHISM = 4.4 μG and α = 15◦ , where
α is the angle between the CHISM flow velocity and the IsMF direction.
logarithm of plasma density, log10(cm-3)

a b plasma temperature, K
c Magnitude of magnetic field, Gauss
-3.2 -2.9 -2.6 -2.3 -2.0 -1.7 -1.4 -1.2 -1.1 0.1 0.4 9.3E+03 2.0E+05 5.5E+05 9.0E+05 1.3E+06 1.6E+06
600 600 2.0E-07 1.2E-06 2.2E-06 3.2E-06 4.2E-06 5.1E-06
600
400 400
400
200 200
200
X (AU)
X (AU)
X (AU)
0 0
0
-200 -200 -200
-400 -400 -400
-600 -600 -600

-200 0 200 400 600 -200 0 200 400 600 -200 0 200 400 600
Z (AU) Z (AU) Z (AU)
d Logarithm of (plasma+magnetic) pressure,

log [p+B2/8p], log (Pa)
10 10
e interstellar hydrogen atoms
normalized number density, n/n
lic
f interstellar oxygen atoms
normalized number density, n/nlic
-13.3 -13.0 -12.7 -12.4 -12.0 -11.9 -11.8 -10.7

600 0.05 0.20 0.35 0.50 0.65 0.80 0.95 1.10 1.25 0.05 0.20 0.35 0.50 0.65 0.80 0.95 1.10 1.25
600 600
400
400 400
200
200 200
X (AU)
X (AU)
0
0 X (AU) 0
-200 -200 -200
-400 -400 -400
-600 -600 -600

-200 0 200 400 600 -200 0 200 400 600 -200 0 200 400 600
Z (AU) Z (AU) Z (AU)
Figure 2.6: Isocontours of the proton number density (plot a), plasma tempera-
ture (plot b), magnetic field (plot c), total thermal plasma and magnetic pressure
(plot d), normalized hydrogen (plot e) and oxygen (plot f ) number densities in
the plane determined by the interstellar velocity and magnetic field vectors. The
streamlines of the plasma component (plot a), the magnetic field lines (plot c)
and “streamlines” of the hydrogen (plot e) and oxygen (plot f ) atoms are shown.
The termination shock and heliopause are shown as white curves [From Izmodenov
et al. (2009)]
Figure 2.6 shows the results of the model calculations in the xz plane. The xz
plane is determined by the Sun-CHISM relative velocity vector and the interstellar
magnetic field vector. The direction of the z-axis is chosen to be opposite to the
interstellar gas velocity vector. The plasma streamlines and isolines of the plasma
density are shown in panel a, the plasma temperature in panel b, and the total
(plasma thermal plus magnetic field) pressure in panel d. Panel c shows the mag-
netic field lines and isocontours of the interstellar magnetic field. The heliopause
and the termination shock become asymmetric with respect to the direction of the
interstellar flow due to the asymmetric pressure of the IsMF (panel c).
The interstellar magnetic field pressure pushes the heliopause and the termi-
nation shock towards the Sun as compared to the model without magnetic field.
Table 2.1: Positions of the TS in the directions of Voyager 1 and Voyager 2

B 0 2.5 2.5 5.0 4.375 4.375 2.5 2.5 1.25 2.5 2.5
α(B, V ) 0 15 15 15 20 30 45 45 60 90
Cross MT MT MT MT Steb. Steb MT MT MT MT MT

section
V1 98 99.9 97.4 92.2 96.0 93.3 91.5 85.8 92.1 84.5 83.1
V2 98 99.7 94.2 82.3 86.5 82.4 87.6 82.5 90.6 82.2 83.2
B is BCHISM , and α(B, V ) is α(BCHISM , VCHISM ). MT uses the charge-exchange cross-section
from Maher and Tinsley (1977), while Steb indicates the charge-exchange cross-section from
Lindsay and Stebbing (2005)
Quantitatively the effect of the IsMF on the location of the TS in the directions of
Voyager 1 and Voyager 2 is shown in Table 2.1. The TS distances in this table are
obtained in the framework of the stationary 3D kinetic-MHD model that is under
discussion, but corrected according to a kinetic-gasdynamic model of the interface
by Izmodenov et al. (2008) that takes into account solar cycle effects (Izmodenov
2009). It is seen from the table that the models with BCHISM = 4.375 μG and
α = 15◦ –20◦ are in good agreement with the actual distances of 94 AU and 84 AU
of the Voyager 1 and 2 crossings of the TS. This is the reason why this model is
presented here. It is interesting to note that using a multi-fluid approach for in-
terstellar H atoms allows us to estimate the distances for comparable observations
in other directions of the interstellar magnetic field and not just in the directions
measured by the Voyagers (Opher et al. 2009). The detailed discussion and com-
parison with models of other groups is out of the scope of the paper (Izmodenov
and Alexashov 2005).
The bow shock disappears for such a strong magnetic field because the Mach
numbers calculated with respect to the Alfvén and fast magnetosonic waves are
smaller than one.
Figure 2.6a also shows the streamlines of the plasma component. The stag-
nation point is located in the upper half of the zx plane and is displaced from
the z axis. It is important to note that the velocity vector of the plasma passing
through the region of maximum plasma density has a noticeable Vx component.
The secondary interstellar atoms which originate in the vicinity of the HP should
have the properties of the plasma in this region. Figure 2.6e presents the num-
ber density of this interstellar atom component. The maximum density appears
in the region between the HP and the BS. This is the so-called hydrogen wall, as
discussed above. It is seen in the figure that the maximum of the hydrogen wall
is also slightly shifted to the upper half of the xz plane and reflects the behavior
of the plasma distributions. The streamlines of the H atom component are also
shown in Fig. 2.6e. The streamlines were plotted based on the mean velocity field
distribution of the interstellar H atoms. The velocity vector V H determines the
direction of the averaged H atom flow. It can be seen from the figure that in
the heliosphere the velocity vector V H has a noticeable Vx component even very
close to the Sun. The effect of the shift of the direction of interstellar hydrogen
flow relative to the direction of the interstellar helium flow was observed by the
SOHO/SWAN H cell instrument (Lallement et al. 2005).
The measurements of the deflection of the interstellar H atom flow relative to the
direction of the interstellar helium flow is one of several observational constraints
on the magnitude and direction of the interstellar magnetic fields. Estimates per-
formed by Izmodenov et al. (2005) and Izmodenov and Alexashov (2006) showed
that the models with BCHISM = 2.5 μG, and α = 30 − 45◦ provide levels of de-
viation in agreement with the level obtained from analysis of the SOHO/SWAN
data. However, recent numerical results (Alexashov et al. 2008) showed that the
deflection angle is very sensitive to the charge exchange cross section. Results of
calculations with updated cross section from Lindsay and Stebbing (2005) show a
larger deflection angle as compared with the previous calculations which employed
the cross section from Maher and Tinsley (1977).
Other possible constraints on the direction and magnitude of the interstellar
magnetic field are:
1. Asymmetries of the TS toward Voyager 1 and Voyager 2 direction, which
were discussed before
2. Flows of the solar wind beyond the TS that are directly measured by Voyager
2 (Opher et al. 2009) and estimated from the anisotropy of the energetic
particle fluxes from Voyager 1 (Krimigis et al. 2011)
3. Analyses of the heliospheric ENA fluxes measured by IBEX (Heerikhuisen
and Pogorelov 2011).
More detailed parametric model calculations are required and the work is currently
in progress.
The Baranov–Malama model and more recent models described above assume
immediate assimilation of pickup protons into the solar wind plasma and consider
the mixture of solar wind and pickup protons as a single component.
However, measurements of the pickup proton distribution function on board
the Ulysses and ACE spacecraft show that the pickup proton distribution function
is non-Maxwellian. The observations show also that the velocity distribution func-
tion can be considered as isotropic (fast pitch-angle scattering) except some short
periods in the inner heliosphere when the interplanetary magnetic field is almost
radial. These data point to the absence of thermodynamic equilibrium between
the pickup and solar-origin protons, although their mean velocities are equal. In
the improved model of the solar wind/interstellar medium interaction developed by
Malama et al. (2006), the pickup protons are considered as a separate component
with thermodynamic parameters different from those of the solar wind. Since in the
solar-wind reference frame, the pickup proton distribution function is assumed to
be isotropic, the following angle-averaged distribution function can be introduced:

∗ 1
fpui (r, w) = fpui (r, v) sin θdθdφ . (2.2)
4π
Here, v is the individual velocity of a pickup proton, w = v − V is the pickup
proton velocity in the solar-wind reference frame (V is the solar wind velocity in
the heliocentric coordinate system), w, θ, and φ are the spherical coordinates of
vector w, and fpui (r, v) is the pickup proton distribution function. The kinetic
∗
equation for fpui (r, w) can be written in the form (Isenberg 1987):
∗ ∗ ∗ ∗
∂fpui ∂fpui 1 ∂ ∂fpui w ∂fpui
+V · = 2 2
w D + divV + S(r, w) . (2.3)
∂t ∂r w ∂w ∂w 3 ∂w
Here, D(r, w) is the diffusion coefficient in the velocity space, while the source
function of pickup protons S(r, w) reflects their birth and loss due to charge ex-
change, photoionization, and the electron impact ionization of H atoms. Malama
et al. (2006) assumed for simplicity that D(r, w) = 0 and the kinetic equation
(2.3) was solved together with the Euler equations written for the sum of all the
charged components and the kinetic equation for the interstellar hydrogen atoms.
The use of Euler equations for the sum of the charged components is justified by
the fact that all the components (solar protons and electrons, as well as pickup pro-
tons) move at the same velocity and that for all the components the distribution
functions are isotropic. In this case, the pressure p is equal to the sum of partial
pressures, while the effective pressure of the pickup protons is determined in terms
∗
of the distribution function fpui , so that:

4π ∗
p = pe + pp + ppui , ppui = mp w4 fpui (r, w)dw
3
To treat the passage of pickup protons through the termination shock we assume
the conservation of their magnetic moment. The analytical derivation of the mag-
netic moment conservation has been done by Alekseev and Kropotkin (1971) for
perpendicular and nearly perpendicular shocks. Later it has been shown numeri-
cally by Terasawa (1979) that the conservation of the magnetic moment exists for
all quasi-perpendicular shocks.
The reflections of pickup protons at the shock potential is not taken into ac-
count. A self-consistent solution of the problem just formulated was obtained by
Malama et al. (2006) for the case D = 0 corresponding to a quiescent solar wind
in which the magnetic field fluctuation level is low.
The thermodynamic non-equilibrium of the pickup and solar protons leads to
a thinning of the inner heliosheath region (Fig. 2.7a), which is attributable to the
decrease in the total pressure of the charged component. As compared with the
calculations performed in accordance with the corresponding equilibrium model,
the TS is displaced by 5 AU away from the Sun in the direction of the oncoming
flow, whereas the HP is by 12 AU closer. In the tail region, the heliocentric distance
of the TS increases by 70 AU.
Recently we expanded the Malama et al. (2006) model by including the pickup
proton component outside the heliopause. Indeed, energetic hydrogen atoms which
originated in the supersonic solar wind and the inner heliosheath can penetrate
through the heliopause and enter into the interstellar plasma. Outside the he-
liopause the atoms can suffer charge exchange with interstellar protons. As a
result new energetic interstellar pickup protons are created. Having an energy of
1 keV or more, these pickup protons have large mean free path as compared with
the thermal interstellar protons and can be considered as collisionless with a non-
Maxwellian velocity distribution. The pickup protons in the outer heliosheath can
charge exchange with interstellar H atoms. Eventually, new ENAs of the outer
heliosheath origin will be created in this secondary charge exchange. Similar to the
ENAs from the inner heliosheath, the ENAs originating in the outer heliosheath
can penetrate to small heliocentric distances and be measured there.
Figure 2.7b shows the calculated fluxes of H atoms at 1 AU obtained within the
framework of the new non-equilibrium model. The figure presents total fluxes of all
heliospheric ENAs (solid curve), fluxes of ENAs originated in the inner heliosheath
from solar wind protons (black dots), fluxes of the ENAs originated in the inner
heliosheath from the pickup protons (blue triangles), and fluxes of the ENAs orig-
inated in the outer heliosheath from the pickup protons (blue dots). The ENAs
originated in the inner heliosheath from the solar wind protons are the strongest
at energies of ∼ 0.02–0.2 keV. For this energy range the fluxes of the ENAs from
the outer heliosheath (blue dots in Figure) are smaller but still comparable. In
the energy range of ∼ 0.4–1.0 keV the ENAs originating in the inner heliosheath
from the pickup protons (blue triangles in Figure) dominate, while for energies
above 2 keV fluxes of the ENAs from the outer heliosheath (blue dots) become
higher than the other. Therefore, the model suggests that the ENAs with energies
above 1 keV may be of interstellar (outer heliosheath) origin. The ENAs of such
energies have been measured on board of several spacecraft. Upper limits of the
heliospheric ENA fluxes estimated from measurements on board of SOHO/CELIAS
(Hilchenbach et al. 1998; Hilchenbach et al. 2000), Cassini/INCA (Krimigis et al.
2009) and Venus Express (Brandt et al. 2009) are shown in Fig. 2.7b as green dots.
The figure shows that generally there is a good agreement between the model results
and the data.
Here we should also note that the calculations were performed in the frame-
work of a model that has no stochastic mechanisms for pickup proton accelera-
tion (Malama et al. 2006). Izmodenov et al. (2009) presented an extension of the
Malama et al. (2006) model by introducing a non-thermal population of pickup
protons in the interstellar medium. It has been shown that the interstellar pickup
protons form significant fluxes of ENAs due to charge-exchange, and those fluxes
exceed the fluxes of other ENAs at energies above ∼ 1 keV. These models predict
that the ENA fluxes have maxima near the upwind direction of the heliosphere and
minima at the flanks, though, of course, the position of the maxima can slightly
deviate from the upwind direction due to the effects of the interstellar magnetic
field and SW asymmetry (Izmodenov et al. 2009).
The first full sky maps of heliospheric ENAs measured by the Interstellar Bound-
ary Explorer (IBEX) spacecraft (McComas et al. 2009; Fuselier et al. 2009; Funsten
et al. 2009; Schwadron et al. 2009) show results that were entirely unexpected in
the frame of the previously discussed models. The first scan of the whole sky
showed that maxima of ENA fluxes form a long (∼ 250◦ –∼ 300◦ ) and narrow arc-
like feature called ribbon that was not predicted by any model prior to the IBEX
observations. Chalov et al. (2010) proposed a modification of the Malama et al.
(2006), Izmodenov et al. (2009) model that ignore the scattering of newly created
energetic pickup protons, but including the effects of ion transport for the pickup
protons generated in the region outside of the heliopause by charge exchange be-
tween the thermal interstellar protons and heliospheric ENAs. The results of the
Figure 2.7: (a) Strong discontinuity surfaces calculated taking into account the
pickup proton non-equilibrium (curves 1), and in accordance with the equilibrium
plasma model (curves 2). (b) Fluxes of H atoms at 1 AU calculated with the pickup
proton non-equilibrium model: total fluxes of all heliospheric ENAs (solid curve),
ENAs originated in the inner heliosheath from solar wind protons (black dots),
ENAs originated in the inner heliosheath from pickup protons (blue triangles),
ENAs originated in the outer heliosheath from pickup protons (blue dots). Upper
limits of the heliospheric ENA fluxes measured by SOHO/CELIAS, Cassini/INCA
and Venus Express are shown by the green dots [From Izmodenov et al. (2009)]
model yield a feature qualitatively similar to the IBEX ribbon. In future stud-
ies, the results of simulations will be quantitatively compared to the IBEX ENA
observations.
Note, however, that the more complex kinetic description of the supra-thermal
population of pickup protons has no direct influence on the primary and secondary
populations of the interstellar atoms which are created from the thermal protons
of the interstellar plasma. Nevertheless, an indirect influence is possible because
taking into account the pickup component properly would change the thermal
population too. Therefore, we do not expect that the effects of the non-equilibrium
plasma model would dramatically change the previous conclusion on the properties
of the primary and secondary interstellar atoms, but some effects are possible
mainly due to global dynamic effects (Malama et al. 2006) and need to be checked
in future.
Summary of the Section

This section reviewed a modern state of the art in the kinetic-MHD mod-
els of the solar wind interaction with the local interstellar medium developed
by the Moscow group (Baranov and Malama 1993). The multi-component and
non-equilibrium nature of both the solar wind and LISM requires development of
complex 3D kinetic-MHD models of the SW/LIC interaction region. The most im-
portant aspect of the model for the purpose of this paper is that upon entering the
heliosphere, the distribution function of interstellar H atoms is strongly disturbed
in the SW/LIC interaction region. This means that the properties of the interstel-
lar atoms inside the heliosphere are dependent on the structure of the heliospheric
interface, and, therefore, the distribution function of interstellar hydrogen can not
be used directly to determine the local interstellar parameters as it would be if
the population of these atoms were not disturbed in the heliospheric interface. At
the same time, this opens up a possibility to use the backscattered solar Lyman-α
radiation as a diagnostic of the properties of the heliospheric interface.
Advanced Model of the Interstellar Hydrogen

in the Heliosphere
Recently, Katushkina and Izmodenov (2010, 2011) developed a model that al-
lows to combine simplicity of the hot-type models of the H atom distribution inside
the heliosphere with the results of the global models of the SW/LIC interaction.
Briefly the model can be classified as a time-dependent 3D version of a classical
hot model with the boundary conditions at 90 AU taken from the global models
of the SW/LIC interaction. This section describes this model and presents some
results including both the global effects of the heliospheric interface and the three-
dimensional and time-dependent local effects in the vicinity (2–10 AU) of the Sun.
Description of the Model

Our advanced model describes the velocity and spatial distribution of interstel-
lar hydrogen inside the heliospheric termination shock. The effects of the SW/LIC
interaction are taken into account in the boundary conditions, which are chosen at
a sphere of radius 90 AU, centered at the Sun (see Fig. 2.8).
We consider only those atoms that penetrate into the heliosphere through the
heliopause, i.e. the primary and secondary interstellar atoms (atoms of populations
4 and 3). Atoms created by charge exchange in the inner heliosheath (population 2)
or in the supersonic solar wind (population 1) are neglected here. Neutral solar
wind (population 1) is not important for backscattering solar Lyman-α photons, be-
cause of their small number densities and a large Doppler-shift beyond the spectral
range of the Lyman-α line. Backscattered Lyman-α from the atoms of population
2 has been estimated by Quémerais and Izmodenov (2002). It has been shown that
at the Earth orbit, the hot component of the backscattered intensity represents
less than 5 % of the total intensity in the upwind direction. It reaches 15 % in the
downwind direction. Due to relatively small importance we neglect the population
2 in this section, but it can be added later without any further modification in the
model.
The solar gravitational force Fg and solar radiation pressure Frad influence the
motion of an atom inside the heliosphere. These forces counteract each other and
are proportional to 1/r 2 . The total force acting on an atom is:
Advanced Model of the Interstellar Hydrogen in the Heliosphere 29
Figure 2.8: Coordinate system used in the axisymmetric models: z is the axis of
symmetry directed toward the upwind direction, ρ is an axis in cylindrical coor-
dinate system that is perpendicular to the upwind direction; curve 1 is the outer
boundary in the advanced model of hydrogen distribution inside the heliosphere;
curve 2 is a schematic representation of the heliospheric termination shock [From
Katushkina and Izmodenov (2011)]
(1 − μ)GMs r
F(r) = mH (1 − μ)Fg = −mH · ,
r2 r
G is the gravitational constant, Ms the mass of the Sun, mH the mass of a H
atom, and r the radius-vector to a point. μ is the ratio of gravitational attraction
to repulsion from radiation pressure as defined in Eq. 2.1. In general, the parameter
μ depends on time (t), heliolatitude (λ) and the radial component of velocity (wr ).
The kinetic equation for the velocity distribution function of H atoms can be
written as:
∂f (r, w, t) ∂f (r, w, t)
+w· +
∂t ∂r
F(r, t, λ, wr ) ∂f (r, w, t)
· = −β(r, t, λ) · f (r, w, t) (2.4)
mH ∂w
Here f (r, w, t) is velocity distribution function of the hydrogen atoms, w is the
individual velocity of a H atom.
The right hand size of Eq. (2.4) represents losses of atoms due to charge ex-
change (H + H + ↔ H + + H) and photoionization (H + hν = H + + e). Electron
impact ionization is not taken into account. The coefficient β(r, t, λ) is the effective
ionization rate: β(r, t, λ) = βex (r, t, λ) + βph (r, t, λ), where βex and βph are rates of
charge-exchange and photoionization, respectively. It is assumed that these rates
of ionization decrease with distance from the Sun as ∼ 1/r 2 (where r is the helio-
centric distance), since these values are proportional to the number density of the
solar protons and flux of the solar photons, respectively. Therefore:
r 2 r 2
E E
β(r, t, λ) = (βex,E (t, λ) + βph,E (t, λ)) = βE (t, λ) ,
r r
where rE = 1 AU is the distance from the Earth to the Sun. The subscript E
indicates that the values are measured at 1 AU.
The functions μ(t, λ, wr ), βex,E (t, λ), and βph,E (t, λ) adopted in this model are
based on experimental data. Detailed descriptions of these functions will be given
below.
The kinetic equation (2.4) is a linear partial differential equation, and it can
be solved by the method of characteristics. It is known that the characteristics of
this equation coincide with the trajectories of the atoms. The distribution function
f (r, w, t) changes along a characteristic according to the equation:
df (r, w, t)/dt = −β(r, t, λ) f (r, w, t).
The solution of the kinetic equation can then be obtained by integration of the last
equation subject to boundary conditions:
t
f (r, w, t) = fb (r0 , w0 , t0 ) exp − β(r, t, λ)dt , (2.5)
t0
where fb (r0 , w0 , t0 ) is the velocity distribution function of hydrogen atoms at the

outer boundary (determined by the boundary conditions at 90 AU); r0 , w0 , t0 are
the position, velocity and time of the atom when the characteristic (i.e. the atom’s
trajectory) is crossing the outer boundary. The integration in the last equation is
performed along the trajectory of the atom.
Kinetic Properties of the Hydrogen Distribution at 90 AU from the Sun

As was discussed in detail in the section “The Solar Wind/Local Interstellar
Medium Interaction”, the velocity distribution function of interstellar hydrogen
atoms is disturbed after they cross the heliospheric interface region. These distur-
bances occur mainly due to charge-exchange between primary interstellar atoms
and decelerated and heated interstellar protons in the vicinity of the heliopause.
As a result, the velocity distribution function of hydrogen is not Maxwellian at
our outer boundary of 90 AU. In this section we will illustrate some kinetic non-
Maxwellian features of the hydrogen distribution at 90 AU resulting from the
Baranov–Malama (1993) model of the heliospheric interface. Since the model is
axisymmetric, it is natural to use a cylindrical coordinate system as shown in
Fig. 2.8, where the z-axis (upwind) is oriented towards the direction of the LIC
flow, and the ρ-axis is perpendicular to the upwind direction (i.e. the crosswind
direction). Also let us define the angular coordinate, θ, as positive in the clockwise
direction.
In the frame of this axisymmetric model we can identify at least five features
which will make the hydrogen velocity distribution function non-Maxwellian at
90 AU:
1. The existence of two populations of interstellar hydrogen atoms (primary and

secondary), which are considerably displaced from each other in the velocity
space (the secondary interstellar atoms have a smaller bulk velocity and a
larger temperature as compared with the primary interstellar atoms).
2. Spatial inhomogeneity of the hydrogen distribution at 90 AU. All moments

of the velocity distribution function (the number density, bulk velocity, and
effective kinetic “temperatures”) strongly depend on θ.
3. Anisotropy of kinetic “temperatures”, i.e. Tz = Tρ , where

Tz (r) ∼ f (r, w) · (Vz (r) − wz )2 dw
and
Tρ (r) ∼ f (r, w) · (Vρ (r) − wρ )2 dw.
4. The correlation coefficient is given by:

Kzρ ∼ f (r, w) · (Vz (r) − wz )(Vρ (r) − wρ )dw = 0,
especially for the secondary interstellar atoms.
5. The third moment of f (r, w) for the secondary interstellar atoms is not zero:

Kzzz ∼ f (r, w) · (Vz (r) − wz )3 dw = 0.
This means that this distribution function is not symmetric relative to its
maximum.
All these features are illustrated in Figs. 2.9 and 2.10. Figure 2.9 presents
moments of f (r, w) at 90 AU as functions of angle θ. These results were obtained in
the frame of the Baranov–Malama model with the following boundary conditions in
the undisturbed LISM: the number densities of protons and neutral hydrogen atoms
are np,LISM = 0.06cm−3 , nH,LISM = 0.18cm−3 , respectively; the relative LISM/SW
velocity is VLISM = 26.4km/s, and the interstellar temperature is TLISM = 6519 K.
For the inner boundary conditions at the Earth’s orbit the following parameters
of the solar wind were adopted: np,E = 6cm−3 , VE = 441.9km/s, and the Mach
number ME = 4.034. Note that in the 3D case, for example when the interstellar
magnetic field is taken into account, the hydrogen distribution at 90 AU becomes
asymmetric (due to three-dimensional structure of the heliospheric interface) and
the non-Maxwellian properties are even more complex.
Note, also, that the previous attempt to treat the distribution function as non-
Maxwellian has been done by Sherer et al. (1999). However, only the first and
third properties listed above have been taken into account in their model, while
Katushkina and Izmodenov (2010, 2011) studied all these non-Maxwellian features
of the velocity distribution at 90 AU (that hereafter we call the global effects of
the heliospheric interface). They concluded that these global effects are important
and should be included in any theoretical model of H atoms distribution inside the
heliosphere. Therefore, special boundary conditions for the velocity distribution
function at 90 AU are necessary.
a b c
d e f
Figure 2.9: Number densities (a), bulk velocities (b, c), correlation coefficient Kzρ
(d), third moment Kzzz , and kinetic “temperatures” Tz , Tρ and Tav (e, f ) of the
primary (blue) and secondary (red) interstellar atoms at 90 AU as functions of angle
θ from the upwind direction. Kzρ (d) and Kzzz are presented in dimensionless units;
kinetic temperatures Tz , Tρ and Tav are in Kelvins. Tav is an averaged temperature
defined as 3 · Tav = Tz + Tρ + Tϕ .
Boundary Conditions
In this paper we consider three models of heliospheric H atoms. One of the three
models (Model 3) includes the global effects mentioned above, while the two other
models are more simplified. We will compare the results of the three models to
identify and separate the effects of local and global phenomena on the distribution
of H atoms in the heliosphere. The models differ only by the boundary conditions
at 90 AU. A summary of the models discussed in this section is given in Table 2.2.
Model 1 is the one-component hot model that implies solution of the kinetic
equation (2.4) with the Maxwellian velocity distribution function at the outer
boundary. For the specific calculations performed in the frame of this model we
use parameters of the Maxwellian distribution, i.e., the number density (nH ), bulk
velocity (VH,z ), and temperature (TH ) calculated at 90 AU in the upwind direction
in the frame of the Baranov–Malama model. These values at 90 AU are:
nH = 0.54 · nH,LISM , VH,z = −0.79 · VH,LISM, TH = 15, 000 K.

Figure 2.10: Illustration of the asymmetry of f (r, w) of the secondary interstellar

atoms at 90 AU (for wϕ = 0) in the upwind direction. Panel a presents the results
of the Baranov–Malama model and shows an asymmetric f (r, w). For comparison,
the symmetric velocity distribution calculated using Eq. (2.6) is shown in panel b.
wr , wθ , wϕ are velocity components in the spherical coordinate system
Table 2.2: Summary of the heliospheric models discussed in this manuscript

Model number Model description
1 One-component hot model
2 Two-component hot model (Bzowski et al. 2008)
2b Two-component hot model without ionization
3 Advanced model (Katushkina and Izmodenov 2011)
3b Advanced model without ionization
3c Advanced model without ionization plus
flat solar Lyman-α spectrum
where nH,LISM = 0.18 cm−3 , VH,LISM = 26.4 km/s. These parameters correspond to
the mixture of the primary and secondary interstellar atoms. Note that the effects
of the heliospheric interface are treated very simply, assuming the parameters of
the Maxwellian distribution are different from the parameters in the LISM.
Model 2 is the so-called two-component hot model (Bzowski et al. 2008). In
this model, the kinetic equation (2.4) is solved separately for the primary and
secondary populations of interstellar atoms. The velocity distribution functions
are assumed to be Maxwellian for each of the populations and the parameters
of the distribution functions are calculated in the frame of the Baranov–Malama
model. These parameters for the primary interstellar atoms are:
nH,prim = 0.22 · nH,LISM , Vz,prim = −1.06 · VH,LISM , TH,prim = 6, 840 K.
The values for the secondary interstellar atoms are:
nH,second = 0.32 · nH,LISM , Vz,second = −0.63 · VH,LISM , TH,second = 18, 126 K.

As shown by Katushkina and Izmodenov (2010), model 2 leads to significant

discrepancies in the distribution of H atoms inside the heliosphere as compared
with the self-consistent Baranov–Malama model. This fact is related to the non-
Maxwellian behavior of the velocity distribution function of hydrogen atoms inside
the heliospheric interface (Izmodenov et al. 2001; Izmodenov 2001).
In model 3, we use the boundary conditions that take into account all of the
previously-mentioned kinetic effects at the outer boundary. This model will be
called the advanced model of hydrogen distribution inside the heliosphere. Com-
parison of this model results with the Baranov–Malama model gives very good
agreement (Katushkina and Izmodenov 2011) which confirms that all of the re-
quired kinetic effects are correctly implemented.
For the primary interstellar atoms, the distribution function at 90 AU is chosen
as the 3D normal distribution:
n 1 Dz
fs (r, w) = H · exp − (Vρ − wρ )2
ρ z − Vzρ
2 D D 2
(2π)3/2 · (D D − V 2 )D
ρ z zρ ϕ
1 Dρ
+ (Vϕ − wϕ )2 + (Vz − wz )2
Dϕ Dρ Dz − Vzρ 2
Vzρ
−2 · (V ρ − wρ )(V z − wz ) (2.6)
Dρ Dz − Vzρ
2
Here DR = mkH TR , Dϕ = mkH Tϕ , and Dz = mkH Tz . k is the Boltzmann constant

and mH is the mass of hydrogen atom. In this case all zeroth, first, and second
moments of the velocity distribution function are taken into account and calculated
in the frame of the Baranov–Malama model. However, for the population of the
secondary interstellar atoms such an approach does not give a complete agreement
with the Baranov–Malama model (Katushkina and Izmodenov 2010) because the
distribution function of this population is asymmetric with respect to its maximum
and has nonzero third moments, which are neglected in the 3D normal distribution
(see Fig. 2.10). That is why for the secondary interstellar atoms Katushkina and
Izmodenov (2011) used the velocity distribution function at 90 AU calculated using
a Monte–Carlo scheme in the frame of the Baranov–Malama model. It was shown
that 15–20 millions of unsplittable trajectories (Malama 1991) in the Monte–Carlo
code is enough in order to get acceptable accuracy of calculations. Therefore, in
this paper we will follow the Katushkina and Izmodenov (2010) approach, a 3D
normal distribution for the primary interstellar atoms and use the numerically
calculated velocity distribution function as the boundary conditions at 90 AU for
the secondary interstellar atoms.
Calculations of hydrogen distribution inside the heliosphere by means of our
advanced model consist of two separate steps. In the first step, a global heliospheric
interface model (the Baranov–Malama model or its modern versions, for example
Malama et al. 2006; Izmodenov et al. 2005) is employed, and the hydrogen velocity
distribution function and its moments at the boundary of 90 AU are obtained. In
the second step we numerically solve the kinetic equation (2.4) with the boundary
conditions obtained in the first step.
Among other advantages, this two-step procedure allows us to separate global
effects of the heliospheric interface, which are taken into account in the first step,
Figure 2.11: Parameters of the secondary interstellar atoms inside the heliosphere
as functions of heliocentric distance along the upwind direction (a) and crosswind
direction (b). Solid curves correspond to the results of the Baranov–Malama model;
dashed curves correspond to the results of model 2; dashed-dot curves correspond
to the advanced model 3. In these calculations μ = 1.258, βph,E = 1.16 × 10−7 s−1 ,
βex,E = 4.8 × 10−7 s−1 . To calculate the charge-exchange rate at the Earth orbit,
we applied averaged values of the solar wind number density (6 cm−3 ) and velocity
(440 km/s) known from measurements and the charge-exchange cross-section from
Lindsay and Stebbing (2005) [From Katushkina and Izmodenov (2011)]
from local effects (i.e. within the heliosphere), which can be explored in the second
step in details. These local effects may include time-dependent and 3D phenomena
as described in the next section.
At the end of this section we illustrate the importance of the non-Maxwellian
features at the outer boundary by comparing the results of model 2 and model 3
shown in Fig. 2.11. The results of the Baranov–Malama model are shown to verify
precision of model 3.
Figure 2.11 presents the number density and effective kinetic “temperatures”
of the gas along the upwind and crosswind directions. It is seen that there is a
very good agreement between the results of the Baranov–Malama model and the
advanced model of hydrogen distribution within 90 AU, while the two-component
hot model leads to significant discrepancies. Firstly, the two-component hot model
overestimates the number density of hydrogen everywhere in the heliosphere. This
is because the two-component hot model assumes constant parameters of hydrogen
at 90 AU, but due to the heliospheric interface effects the hydrogen parameters
depend on angle θ (e.g. the number density at 90 AU decreases with θ). Second,
there are qualitative differences in the behavior of temperatures Tz in upwind and

Tρ in crosswind between model 2 and model 3. These differences are caused by the
kinetic non-Maxwellian properties of the hydrogen distribution function at 90 AU
which were described above.
Concluding this section we summarize that: 1. Kinetic features in the velocity
distribution of H atoms in the outer heliosphere must be taken into account in
theoretical models of hydrogen distribution inside the heliosphere. 2. A good
agreement between the results of Baranov–Malama model and our advance model
(model 3) implies that all important kinetic effects are included in the advanced
hot model through the correctly chosen boundary conditions.
Modeling of Time-Dependent and Heliolatitudinal Effects

Both observations of the solar wind (McComas et al. 2008) and backscattered
solar Lyman-α radiation (Bzowski et al. 2003; Quémerais et al. 2008; Lallement
et al. 2010) have shown that the solar wind parameters vary with the solar cycle
and have a three-dimensional (3D) nature, i.e. they depend on heliolatitude. Since
the charge exchange ionization rate depends on the solar wind flux, then the 3D
and time-dependent effects may change the distribution of the H atoms in the
heliosphere (Bzowski 2003). In turn, the 3D and time-dependent nature of the
H atom distribution may effect the maps of backscattered Lyman-α radiation. In
this section we will describe the parameters and results of the 3D time-dependent
model of the H atom distribution in the heliosphere.
To calculate the velocity and spatial distribution of H atoms in the heliosphere
we employed the 3D time-dependent version of our advanced model (model 3)
with the boundary conditions specified above. To finalize the formulation of the
model we need to specify the photoionization rate βph,E (t, λ), charge-exchange rate
βex,E (t, λ) and balanced parameter μ(t, λ, wr ) as functions of time, and heliolati-
tude λ. The parameter μ is also a function of radial component of atom velocity
wr because the flux of photons illuminating (and forcing) an atom depends on the
atom’s velocity. Here we follow the approach used by Bzowski et al. (2013, this
volume).
To calculate the photoionization rate as a function of time in the ecliptic plane
we use the SOLAR2000 database (http://www.spacewx.com/solar2000.html; the
Solar Irradiance Platform (SIP), formerly known as the SOLAR2000 irradiance
specification tool, produces the variable, full solar spectrum in assorted spectral
formats for historical, nowcast, and forecast applications). The time resolution
is one day. In the calculations we neglect the angle between the solar equatorial
and ecliptic planes, i.e. we assume that from SOLAR2000 we get βph,E (t, λ = 0).
Magnitudes of βph,E (t, λ = 0) were adjusted from the Earth orbit to 1 AU and then
averaged over one Carrington rotation.
To calculate the charge-exchange ionization rate in the ecliptic plane we use
hourly data of the solar wind number density (np,E (t)) and velocity (wsw,E (t))
from the OMNI database (http://omniweb.gsfc.nasa.gov/; the OMNIWeb interface
provides access to the multi-source OMNI 2 dataset; the OMNI 2 dataset contains
hourly resolution solar wind magnetic field and plasma data from many spacecraft
in geocentric orbit and in orbit about the L1 Lagrange point). These hourly data
were used to calculate fluxes of mass, momentum and energy. The fluxes with
hourly resolution were used to calculate the averaged fluxes (over one day), which
were adjusted to 1 AU from the Earth orbit. Then these data were averaged over
one Carrington rotation. The charge-exchange ionization rate can be calculated as
follows:
βex,E (t, 0) = np,E (t) · wsw,E (t) · σ(wsw,E ),
where σ(wsw,E ) is the charge-exchange cross-section (Lindsay and Stebbing 2005).

The total ionization rate for λ = 0 is calculated as
βtot,E (t, λ = 0) = βex,E (t, 0) + βph,E (t, 0).
The heliolatitudinal dependence for the total ionization rate βtot,E (t, λ) has
been calculated using SOHO/SWAN data. A detailed description of how to get
the total ionization rate βtot,SW AN (t, λ) from SOHO/SWAN is given in Quémerais
et al. (2006). Overall, βtot,E (t, λ) is determined as follows:
βtot,SW AN (t, λ)
βtot,E (t, λ) = βtot,E (t, 0) · .
βtot,SW AN (t, 0)
Since the time-resolution of βtot,SW AN (t, λ) is sometime smaller than 27 days to

get the value at a given day we make a linear interpolation.
To calculate the parameter μeq in the solar equator plane (λ = 0) as a function
of time (for wr = 0) we use solar Lyman-α flux from the LASP Interactive Solar
IRradiance Data center (http://lasp.colorado.edu/lisird/lya/). From this database
we get the solar Lyman-α flux as a function of time with a resolution of one day.
Then these data are adjusted to 1 AU from the Earth orbit and averaged over
one Carrington rotation. We assumed that μeq (t, wr = 0) = 0.9 · Fsolar (t)/Fsolar,0 ,
where Fsolar,0 = 3.32 · 1011 ph/(s/cm2 · Å) is the total Lyman-α flux at the line
center. For the dependence of μ on the radial component of velocity an analytical
formula obtained by Bzowski (2008) was employed, namely:
FB (Fsolar (t), wr ))
μeq (t, wr ) = μeq (t, wr = 0) · ,
FB (Fsolar (t), wr = 0))
where FB is a function described in Bzowski (2008).

To get the heliolatitudinal dependence of μ we use the following expression from
Pryor et al. (1992):
μ(t, λ, wr ) = μpole (t, wr ) + cos2 (λ) · (μeq (t, wr ) − μpole (t, wr )),
where
μpole (t, wr ) = FB (Fsolar − ΔFsolar , wr )
and ΔFsolar ≈ 0.05 · 1011 ph/(cm2 s), this value is given from Pryor et al. (1998).
The resulting functions of βtot,E , βex,E , βph,E , and μ are shown in Figs. 2.12, 2.13
and 2.14.
Figure 2.12: (a) Time dependence of the total ionization rate (βtot,E ), charge-
exchange rate (βex,E ) and photoionization rate (βph,E ) at the Earth’s orbit in the
solar equatorial plane (λ = 0). (b) Time dependence of the radiation pressure
parameter μ for λ = 0 and wr = 0
Results of the Advanced Hot Model in the 3D Non-Stationary Case
Figures 2.15 and 2.16 show the number densities, radial components of the bulk
velocities Vr , and the radial kinetic temperatures Tr for the primary and secondary
interstellar atoms in the upwind direction as functions of time. The distributions
are shown for different heliocentric distances. As shown in the figures, the time-
variations of the number densities of the primary and secondary populations behave
qualitatively the same and both are anti-correlated with the time variations of the
ionization rate. For example, in 2002 at solar maximum there is a maximum
of ionization rate as shown in Fig. 2.12a and a minimum of the H atom number
densities. Note that charge exchange provides the main contribution to the total
ionization rate since the charge exchange rate exceeds the photoionization rate
significantly. The local maximum of the number densities in 2004–2005 is due to
the local minimum of the charge exchange rate. There is a monotonic decrease of
the solar activity and therefore ionization rate from 2005 to 2008 that results in a
continuous increase of the number density from 2005 to 2008. There is an abrupt
fall number density in 2009 caused by the increase in the charge exchange rate.
Figure 2.13: Total ionization rate as a function on time and heliolatitude
a b
Figure 2.14: (a) Parameter μ as a function of heliolatitude λ for wr = 0; (b)

μ as a function of radial velocity wr for λ = 0. Red curves correspond to solar
maximum conditions (2002), blue curves correspond to solar minimum conditions
(1996), green curves correspond to the 2009 solar minimum
a b
c d
Figure 2.15: Distribution of the number density (a, b), radial velocity (c) and radial
kinetic temperature (d) of the primary interstellar atoms along the upwind
direction as functions of time; different colors correspond to different heliocentric
distances: (1) r = 1 AU, (2) r = 2 AU (3) r = 3 AU, (4) r = 4.4 AU, (5) r = 7.4 AU,
(6) r = 10 AU
It can be seen from the figures that time fluctuations of all parameters are small
at 10 AU and beyond. Therefore, local effects become small for r > 10 AU in the
upwind direction, in agreement with Ruciński and Bzowski (1995).
Let us focus on the bulk velocities of the H atoms in the upwind direction.
Figures 2.15c and 2.16c show that the time-variation of the bulk velocities is very
similar to the time variation of μ(t), shown in Fig. 2.12b. Therefore, fluctuations
of the bulk velocities of the primary and secondary atoms are mainly determined
by the force changing the atom trajectories rather than the ionization processes.
However, a qualitative difference in the behavior of the primary and secondary
H atoms appears when the heliocentric distance increases. In Fig. 2.15c for the
primary population all curves cross each other at two time periods in 1998 and
a b
c d
Figure 2.16: Distribution of the number density (a, b), radial velocity (c) and radial
kinetic temperature (d) of the secondary interstellar atoms along the upwind
direction as functions of time; different colors correspond to different heliocentric
distances: (1) r = 1 AU, (2) r = 2 AU (3) r = 3 AU, (4) r = 4.4 AU, (5) r = 7.4 AU,
(6) r = 10 AU
2006. Approximately in these periods of time μ(t, λ = 0, wr = 0) ≈ 1 in the ecliptic

plane, i.e. solar radiation pressure balances solar gravitation, and the trajectories
of the atoms are straight lines. In this situation the bulk velocity of the gas would
be constant if the ionization processes were absent. However, due to the so-called
selection effect, the bulk velocity of the atoms would increase toward the Sun. The
selection effect is connected with the fact that the slow atoms have more chances
to be ionized before they reach the same heliocentric distance as the fast atoms.
It is simply because the slow atoms spend more time along their trajectory. As
a result of the selection, the maximum of the velocity distribution function of H
atoms moves towards larger velocities, and the bulk velocity increases. This effect
was discussed in Lallement et al. (1985a) and Bzowski et al. (1997).
When μ > 1, the effective force decelerates the atoms as they approach the Sun.
For μ ≈ 1.1, the effect of the deceleration due to the effective force compensates
the selection effect. In this case the bulk velocity practically does not change when
the atoms approach the Sun. In the case of μ > 1.1, as it was from 1998 to 2006,
the repulsive force decelerates the atoms and the bulk velocity of the gas becomes
smaller. The smallest bulk velocity was in 2002 when the repulsive force had its
maximum. Also, for μ > 1.1 and some fixed moment of time, the bulk velocity
decreases towards the Sun. In the case of μ < 1 the bulk velocity increases towards
the Sun.
For the secondary interstellar atoms the situation is slightly different. This
population is slower than the primary population, and approaching the Sun the
atoms are more effectively ionized. As a result, the bulk velocity of the secondary
interstellar atoms increases towards the Sun during the entire period considered
with the exception of the 2002 solar maximum when the bulk velocity did not
change with heliocentric distance.
Another qualitative difference between the primary and secondary populations
is in the behavior of the bulk velocity in the period from 2006 to 2009. As it was
mentioned before, the behavior of the bulk velocity for the primary populations
reflects time-variations of μ(t), while for the secondary population the bulk velocity
decreases from 2006 to 2008 and then rapidly increases in 2009. Such a behavior of
the bulk velocity can be explained by the strong decrease in the charge-exchange
rate before 2008. A strong increase of the bulk velocity in 2009 is due to the
increase in the charge exchange rate in 2009 and the corresponding increase of the
selection effect.
Now, let us focus on the kinetic temperature Tr , that is the second moment of
the velocity distribution function calculated for radial velocity component.
Figures 2.15d and 2.16d show that the temperatures of the primary and secondary
populations are qualitatively similar in behavior. For a small heliocentric distance
the time variations of Tr resembles μ(t). This means the radial temperature (as
well as the bulk velocity) is mainly affected by the resulting force. However, the
ionization processes are important as well. For example, the temperature decreases
towards the Sun for both the primary and secondary populations except near solar
maximum. Figure 2.17 presents the number density of both populations of inter-
stellar hydrogen as functions of time and heliocentric distance along the upwind
direction.
Figure 2.18 shows sky-maps of the number densities, radial velocities and kinetic
temperature Tr of the primary and secondary interstellar atoms at 1 AU calculated
for 1996, i.e. for the minimum of solar activity. The upwind direction has the
following heliographic coordinates: heliolongitude αHGI = 178.98◦ ± 0.5◦ and he-
liolatitude βHGI = 5.11◦ ± 0.2◦ . It is seen (Fig. 2.18a, d) that the maxima of the
H atom number densities occur at the upwind heliolongitude. For this upwind he-
liolongitude there are two maxima in the southern and northern hemispheres and
a minimum in the plane of the solar equator. This minimum in number density is
due to the maximum of the ionization rate near the solar equatorial plane. The
distributions of radial velocity, Fig. 2.18b, e, show minima near the upwind direc-
a b
Figure 2.17: Number density of the primary and secondary interstellar atoms in
the upwind direction as functions of time and heliocentric distance
Figure 2.18: Number density (a, d), radial velocity (b, e) and radial kinetic temper-
ature (c, f ) of the secondary (population 3, left column) and primary (population
4, right column) interstellar atoms, calculated at 1 AU as functions of heliolatitude
and heliolongitude in the heliographic coordinate system (HGI). These results are
obtained for solar minimum conditions (year 1996), when the heliolatitudinal de-
pendence of the total ionization rate is very strong. In HGI coordinates, the upwind
direction is at 178.98◦ in heliolongitude and at 5.11◦ of heliolatitude
tion. Similar minima in the upwind direction are seen in the maps of the radial
kinetic temperature Tr (Fig. 2.18c, f). However, the temperature maps also show
heliolatitudinal effects. The latitudinal effect is more pronounced for the secondary
population.
Application of the Hydrogen Distribution Models

for Calculation of the Backscattered Solar Lyman-α
In this section we will demonstrate how the model described above can be
applied to the analyzes of the backscattered solar Lyman-α and also, in the simplest
case, how the effects of the heliospheric interface are pronounced in the intensities,
line-shifts, and line-widths of the heliospheric Lyman-α radiation.
Modeling Backscattered Solar Lyman-α Profiles

The profiles of backscattered Lyman-α radiation I (r,ν,Ω) were computed for
anti-solar radial directions (Ω) for an observer located at 1 AU. Here r is a position
of the observer, ν is the frequency of the backscattered radiation, Ω is the line-
of-sight direction. We use the “self-absorption” approximation (Quémerais 2000).
In this approximation only singly scattered photons are considered (i.e. multiply
scattered photons are neglected), and the absorption of photons between the Sun
and the scattering point is neglected. As was shown in Quémerais and Izmodenov
(2002), the simplified “self-absorption” approach gives similar results as compared
with a full radiative transfer model. For the line-width of the backscattered pro-
file at 1 AU the difference between the full radiative transfer model and the self-
absorption model is less than 15 % for the upwind direction and becomes smaller
as the line-of-sight moves from upwind to downwind. In the downwind direction
two approaches give nearly the same line-widths. The simplified self-absorption
approach is sufficient for the purposes of this paper. As will be shown below, the
effects of the heliospheric interface are larger than the difference between the full
radiative transfer and self-absorption models.
The radiative transfer equation for I (r,ν,Ω) can be written as follows:
Ω · ∇I(r, ν, Ω) = −σν (r, ν)nH (r)I(r, ν, Ω) + nH (r)j(r, ν, −Ω). (2.7)
Here nH (r) is the number density of hydrogen atoms, σν (r, ν) is the differential
scattering cross-section that is proportional to the projection of the hydrogen dis-
tribution function on the line-of-sight, j (r,ν,-Ω) is the atomic emission coefficient
which measures the number of photons emitted by a hydrogen atom per second per
frequency unit and per solid angle. Note that scattered photons travel in the direc-
tion opposite to the line-of-sight direction, i.e. in −Ω. The first term on the right
hand side of Eq. (2.7) is the loss term due to absorption of the scattered photons.
The second term is the local source of emission due to the scattering process.
Equation (2.7) has the following solution in the self-absorption approximation:
∞
4π
I(r, ν, Ω) = 6 nH (r + sΩ) j(r + sΩ, ν, −Ω) e−τν (r+sΩ,r) ds, (2.8)
10 0
Backscattered Lyman-α 45
where s is the coordinate along the line-of-sight, τν (r , r) is the optical thickness
for scattered photons with the frequency ν calculated from the scattered point
r = r + sΩ for an observer located at point r. The atomic emission coefficient j
can be represented by:
j(r , ν, −Ω) = φ(ω) FS (r , νp ) σν (r , ν) . (2.9)
Here φ(ω) is the scattering phase function that gives the relation between the
directions of propagation of the photon before and after the scattering (Brandt
and Chamberlain 1959). FS (r , νp ) is the flux of the solar Lyman-α photons with
the frequency of νp at point r . The solar Lyman-α spectra obtained by Lemaire
et al. (1998) is used to calculate the solar Lyman-α flux at the Earth orbit at a
given frequency ν.
In the case when the line-of-sight is radial, there is a simple relation between the
frequency of the primary solar photon νp and the frequency of the backscattered
photon ν: νp = 2 · ν0 − ν, ν0 is the frequency at line center. Thus, if we know the
hydrogen velocity distribution function in the entire heliosphere, we can calculate
the profiles of the backscattered Lyman-α radiation.
Since we consider two different populations of the interstellar H atoms inside the
heliosphere and neglect multiple scattering effects, it becomes possible to calculate
profiles of the radiation scattered by each populations separately. To do this we
consider photons that were scattered by the primary and secondary interstellar
atoms independently. Optical thickness is calculated for the mixture of the primary
and secondary atoms, because a photon scattered, for example, by the primary
interstellar atom can then be absorbed by atoms of both populations.
We can calculate the following moments of the backscattered radiation profiles:
(a) Intensity measured in [R],
∞
Ilos (r, Ω) = I(r, ν, Ω) dν;
0
(b) Line-shift, expressed in km/s,

∞
0
u(ν) I(r, ν, Ω) dν
Vlos (r, Ω) = ;
Ilos (r, Ω)
(c) Line-width, expressed in degrees Kelvin,

∞
mH 0 (u(ν) − Vlos (r, Ω))2 I(r, ν, Ω) dν
Tlos (r, Ω) = .
kb Ilos (r, Ω)
Here, u(ν) = c (ν/ν0 − 1), mH is the mass of a hydrogen atom, kb is the
Boltzmann constant. The line-shift of the spectra is often called the line-
of sight velocity, and the line-width is called the line-of-sight (or apparent)
temperature. These integral characteristics of the backscattered Lyman-α
profile reflect the properties of the velocity distribution function of the H
atoms inside the heliosphere, but they do not coincide and should not be
confused with the bulk velocity and temperature of the gas far away from the
Sun.
Model Results for the Two-Dimensional Stationary Case

We have computed the backscattered Lyman-α profiles and their moments using
three models of the hydrogen distribution inside the heliosphere described above.
All calculations were performed for the anti-solar directions. For the 2D axisym-
metric problem considered here each line of sight is characterized by the angle θ
that is measured from the z-axis as shown in Fig. 2.8.
Figure 2.19 shows the intensities (left column), line-shifts (center-column) and
line-widths (right column) of the backscattered Lyman-α radiation at 1 AU as func-
tions of the line-of-sight angle θ. For models 2 and 3 the profiles of the photons
scattered by the primary and secondary interstellar H atoms were computed sep-
arately (plots b, c in Fig. 2.19). The total backscattered profile is shown in plot a
of Fig. 2.19. For model 1, which has only one component of atomic hydrogen, we
computed only the characteristics of the total Lyman-α radiation.
For the total radiation (row a of Fig. 2.19), models 1 and 2 lead to a systematic
increase in the intensities as compared with the intensities calculated for model 3.
Comparison of models 2 and 3 for the primary and secondary populations (plots
b and c) shows that the increase in the intensities is due to the secondary H atom
component (compare left-columns in the plots a and c). Models 2 and 3 agree
nicely for the primary H atom component.
Now, we will explore the main reason for the different results obtained from
models 2 and 3. The models differ only in the boundary condition. We can identify
two differences. The first one is the dependence of the boundary conditions on
the angle θ in model 3, which does not exist in model 2. The second difference
is the non-Maxwellian velocity distribution function at 90 AU in model 3, while
in model 2 the velocity distribution is Maxwellian. To explore which one of the
two identified differences mainly affects the spectral properties of the Lyman-α
radiation we performed an additional model calculation (model 2a). In model
2a we assume that the velocity distribution functions at the outer boundary are
Maxwellian for both the primary and the secondary populations of the H atoms.
However, the parameters of the Maxwellians (i.e. number density, bulk velocity and
temperature) are functions of the angle θ and they were calculated in the frame
of the Baranov–Malama model. Therefore, model 2a allows to separate the effects
of non-uniform flow of H atoms at 90 AU from the kinetic effects related to the
non-Maxwellian features of the velocity distribution function.
Intensities, line-shifts and line-widths calculated in the frame of model 2a are
also shown in Fig. 2.19a. It is seen that model 2a and model 3 produce very close
results in the backscattered Lyman-α intensities, although there is a difference of
∼ 20 Rayleigh for the upwind direction. It means that the main difference between
models 2 and 3 is due to the non-uniform flow of H atoms at 90 AU in model 3.
However, there is still a small difference due to the non-Maxwellian behavior of
the velocity distribution function at 90 AU (which is taken into account only in
model 3).
At first sight, it is not evident why the angular dependence of the H atom pa-
rameters at 90 AU would strongly influence the backscattered Lyman-α emission
measured at 1 AU. It is especially so, because the main emissivity region for the
backscattered Lyman-α radiation at 1 AU is located approximately at 2 AU for the
Figure 2.19: Intensities (left column), line-shifts (center-column) and line-widths

(right column) of the backscattered Lyman-α radiation at 1 AU as functions of
the line-of-sight angle θ from the upwind direction. Plots a (top row) are for the
total radiation scattered by both primary and secondary interstellar atoms. Plots
b (middle row) are for photons that were scattered by the primary interstellar
atoms. Plots c (bottom row) correspond to the radiation scattered by the secondary
interstellar atoms. Different curves correspond to three models of the hydrogen
distributions in the heliosphere: (1) is the one-component hot model (model 1); (2)
is the two-component hot model (model 2); (3) is our model (model 3) that takes
into account effects of the heliospheric interface; additional curves marked as 2a in
plots a,b,c correspond to the model of H atoms described in this section: model
2a which corresponds to model 2 plus the θ-dependence of hydrogen parameters at
the outer boundary [From Katushkina and Izmodenov (2011)]
a b
Figure 2.20: Contributions to the total number density (a) and radial velocity
(b) of atoms arriving from various directions. Curve 1 in plot (a) corresponds
to point 1 (r1 = 2 AU, θ = 0◦ ), and curve 2 corresponds to point 2 (r2 = 7 AU,
θ = 180◦ ). The contribution to the radial velocity was calculated only for point 2.
Various curves in plot (b) correspond to various models of hydrogen distribution:
the solid line is model 3, the dashed line is model 2 and the dashed-dot line is
model 1. All calculations are performed for a mixture of primary and secondary
interstellar atoms [From Katushkina and Izmodenov (2011)]
upwind line-of-sight and at 7 AU for downwind. From the simple (naive) consider-
ation one could expect that most of the H atoms in the regions close to the Sun
would arrive from upwind. However, Lallement and Bertaux (1990) have shown
(in the case of the hot model) that most of the atoms arrive into the vicinity of the
Sun not exactly from upwind, since the thermal velocities of hydrogen atoms at
the boundary are large enough compared to the bulk velocities. The same is true
for our model. To illustrate this we calculate the function n(ri , θb ) for two points
(i = 1, 2) which correspond to the maximum emissivity region. Point 1 is located
in the upwind direction at 2 AU from the Sun; point 2 is located in the downwind
direction at 7 AU from the Sun. The function n(ri , θb ) represents the contribution
to the total number density at a given point (point 1 for curve 1 in Fig. 2.20a and
point 2 for curve 2) from these particles which cross the outer boundary of 90 AU
at θ = θb . This function n(ri , θb ) is defined as follows:

n(r1 , θb ) = f (r1 , w1 )dw1 . (2.10)
Ω1
Here f is the velocity distribution function of H atoms. The integration is per-

formed over those w1 that correspond to the trajectories crossing the outer sphere
of 90 AU at θ = θb . Here for simplicity we assume a balance between the solar
gravitation and radiation pressure (i.e. μ = 1). This means that all atomic trajec-
tories are straight lines. In that case, the subspace Ω1 is a cone with an apex angle
Figure 2.21: Schematic picture of penetration of the hydrogen atoms from the outer
boundary to the vicinity of the Sun. Point 1 (r1 ) is located in the upwind direction
at 2 AU from the Sun. In the case of straight atom’s trajectories we consider only
those atoms that arrive to point 1 from the outer boundary at θb . The angle θ1
can be determined if we know r1 and θb [From Katushkina and Izmodenov (2011)]
equal to θ1 , as can be seen from Fig. 2.21. Let us introduce a spherical coordinate
system in velocity space. It means that we describe the velocity vector w by its
modulus w̃ and two spherical angles θ̃ and ϕ̃, i.e. orthogonal coordinates of vector
w can be represented as follows:
wx = w̃ · sin(θ̃) · cos(ϕ̃)
wy = w̃ · sin(θ̃) · sin(ϕ̃)
wz = −w̃ · cos(θ̃)
In spherical coordinates, dw1 = w̃12 sin(θ̃) dw̃1 dθ̃ dϕ̃. For velocities from the
subspace Ω1 , θ̃ = θ1 = const for the chosen value of θb and integration over θ̃ is
not needed. Hence, for our case Eq. (2.10) can be rewritten in the following form:
+∞ 2π
n(r1 , θb ) = f (r1 , w̃1 , θ̃ = θ1 , ϕ̃) w̃12 sin(θ1 ) dw̃1 dϕ̃. (2.11)
0 0
Now, in the case of μ = 1:
f (r1 , w̃1 , θ1 , ϕ̃) = fb (θb , w̃1 , θ1 , ϕ̃) · exp(−A(r1 , θb , w̃1 )),
where fb is the corresponding velocity distribution function at 90 AU, A is the loss

of hydrogen atoms along its trajectory from the outer boundary to point 1 due to
the ionization processes and θ1 = θ1 (θb ). fb does not depend on the angle ϕ̃ due
to the axial symmetry of the boundary conditions. Therefore Eq. (2.11) can be
represented as:
+∞ 2π
n(r1 , θb ) = fb (θb , w̃1 , θ1 (θb )) exp(−A(r1 , θb , w̃1 )) w̃12 sin(θ̃1 ) dw̃1 dϕ̃
0 0
+∞
= 2π sin(θ1 ) · fb (θb , w̃1 , θ1 (θb )) exp(−A(r1 , θb , w̃1 )) w̃12 dw̃1
0
= 2π sin(θ1 ) · g(r1 , θb ).
The function fb (θb , w̃1 ) decreases with θb for each given value of w̃1 , because at
90 AU Vz,H Vρ,H and the distribution function has a maximum in the upwind
direction. The loss-function, A(r1 , θb , w̃1 ), increases with θb for point 1 because
the length of the trajectory has a minimal value upwind and exp(−A) decreases
with θb . Hence, the function g(r1 , θb ) decreases with θb , and sin(θ1 ) increases with
θb for θb ∈ [0, π/2]. Therefore, n(r1 , θb ), that is a product of sin(θ1 ) and g(r1 , θb ),
should have a maximum at a certain θb between 0◦ and 90◦ .
Figure 2.20a shows n(ri , θb ) calculated numerically for point 1 in the upwind
direction (curve 1) and for point 2 in the downwind direction (curve 2). The figure
shows that the largest fraction of the interstellar atoms arrive at point 1 from
θ ≈ 15◦ and at point 2 from θ ≈ 55◦ . Therefore, contrary to expectations, most of
the H atoms reach the vicinity of the Sun not exactly from the upwind direction.
Now, it is clear that the excess in the backscattered Lyman-α intensity in models
1 and 2 as compared with model 3 is related to the higher number densities of H
atoms inside the heliosphere in models 1 and 2 arising from the lack of θ-dependent
boundary conditions. In model 3, the number density at 90 AU decreases with θ.
It also becomes clear why the intensities (and line-shifts) of the radiation scat-
tered by the primary interstellar atoms nearly coincide for all models as shown in
Fig. 2.19b. This is because on the one hand, the angle-dependence of the hydro-
gen parameters at 90 AU for the primary interstellar atoms is weaker than for the
secondary interstellar atoms. On the other hand, the distribution function of the
primary interstellar atoms is closer to a Maxwellian than the distribution function
of the secondary interstellar atoms.
From the middle column of Fig. 2.19a, c it can be seen that there are noticeable
differences in the line-shifts of models 2 and 3 in the downwind direction. The
differences are seen for the total radiation as well as for the radiation scattered by
the secondary interstellar atoms. At the same time, there are almost no discrepan-
cies in the line-shifts of model 1 and model 3, despite the fact that model 1 is the
simplest model without any of the effects of the heliospheric interface. Therefore
one would expect some differences.
In order to understand these results we calculated the contribution to the total
radial velocity of hydrogen at point 2 (7 AU in downwind) from the particles that
reach this point from different directions. Similar to n(ri , θb ), the contribution to
the radial velocity can be calculated as follows:

1
Vr (ri , θb ) = +∞ · f (ri , wi ) wi,r dwi
−∞
f (ri , wi )dwi Ωi
Figure 2.20b shows Vr (ri , θb ), calculated at point 2 for different models. It is

seen that the maximum of Vr (ri , θb ) is located approximately at θ=45◦ for all mod-
els. However, model 2 (dashed curve in Fig. 3b) has a non-negligible contribution of
negative values of Vr (ri , θb ) for θ > 100◦ . The negative values of Vr (ri , θb ) are due
to relatively hotter secondary interstellar atoms in the downwind hemisphere (due
to high temperature of the secondary interstellar atoms and absence of a decrease
in the number density from upwind to downwind in model 2), which can reach
point 2 from large values of angle θ. This makes the total line-shift of model 2 in
the downwind region smaller as compared with models 1 and 3. The contributions
of negative Vr (ri , θb ) in model 1 and model 3 are smaller than in model 2, but
due to different reasons. In model 3, the contribution of the particles with nega-
tive Vr (ri , θb ) is significantly reduced because of the low number density of such
particles at the outer boundary that comes from the self-consistent model results.
As for model 1, we do not see this effect due to the relatively smaller temperature
of the mixture of primary and secondary interstellar atoms as compared with the
temperature of the secondary interstellar atoms, which exists in model 2.
Now let us consider the line-widths (right column in Fig. 2.19). Plot a demon-
strates that there are essential qualitative differences in the line-widths calculated
on the basis of the three models. Model 1 shows a monotonic increase in the line-
width with θ. Model 2 shows a minimum of the line-width at θ ∼ 60◦ . Model 3
shows a small local maximum at θ ∼ 150◦ . These results demonstrate that the
line-width of the backscattered Lyman-α profiles is very sensitive to the properties
of the hydrogen distribution at the termination shock.
In the next section we will explain these qualitative differences and will show
that the main cause of the difference is the kinetic non-Maxwellian nature of the
hydrogen velocity distribution function at 90 AU. Note also that for the line-widths
the simplified model 2a results are close to model 3 in the upwind hemisphere.
However, large discrepancies in the downwind region still exist.
Now we will consider differences in the line-widths of the backscattered spectra
calculated separately for the primary and secondary populations of H atoms (right
columns of Fig. 2.19b, c). For both populations we observe large maxima in the
line-widths at θ ∼ 90◦ in the results of model 3. Such strong maxima do not exist for
model 2 and for model 2a. The latter means that the maxima are not due to the θ-
dependence of the hydrogen parameters at 90 AU but rather to the non-Maxwellian
velocity distribution function at the outer boundary. As was shown by Baranov
et al. (1998) (see also, Izmodenov et al. 2001 and Fig. 3e, f of Katushkina and
Izmodenov 2010) the components Tz and Tρ of the kinetic temperatures calculated
from the velocity distribution function of the interstellar hydrogen, i.e.

Tz (r) ∼ f (r, w) · (Vz (r) − wz )2 dw
and

Tρ (r) ∼ f (r, w) · (Vρ (r) − wρ )2 dw
are fundamentally different from each other inside the heliosphere.

Table 2.3: Comparison of intensities, line-shifts and line-widths of the backscat-

tered Lyman-α radiation for three models of the hydrogen distribution inside the
heliosphere
|Ilos −Ilos,3 | |Vlos −Vlos,3 | |Tlos −Tlos,3 |
θ = 0◦ Ilos (R) Vlos (km/s) Tlos (K) Ilos,3 Vlos,3 Tlos,3
Model 1 677 −24.2 14,710 2.3 1.2 14.5
Model 2 720 −24.5 14,158 8.7 0.0 10.3
Model 3 662 −24.5 12,841 0.0 0.0 0.0
Model 1 398 −0.9 17,505 4.7 200.0 18.0
Model 2 421 −0.6 15,561 10.8 100.0 4.9
Model 3 380 −0.3 14,834 0.0 0.0 0.0
Model 1 127 20.3 18,984 24.5 1.4 11.3
Model 2 129 17.4 23,946 26.5 15.5 40.4
Model 3 102 20.6 17,060 0.0 0.0 0.0
See text for a description of the models. Fractional differences in intensity (columns 5–7) are
shown as percentages
Recall that z-axis is the central axis of the axisymmetric heliosphere and ρ is
the distance from that axis in a cylindrical coordinate system. In other words,
the mean thermal velocities of the H atoms are different in different directions.
Moreover, this difference between Tz and Tρ temperatures increases approaching
to the Sun due to local effects. The large maxima of the line-widths at θ = 90◦
shown in Fig. 2.19b, c are related to the changes in the radial kinetic temperature
Tr of the gas at 90 AU as a function of θ. For example, when an observer looks
toward the upwind direction (i.e. θ = 0◦ ) Tr = Tz , for a line-of-sight with θ = 90◦
Tr = Tρ and for a line-of-sight of θ = 180◦ Tr = Tz again. In model 3 at the
outer boundary, Tρ is higher than Tz for each of the interstellar populations of H
atoms based on the Baranov–Malama model results. This results in the maxima of
Tr at θ = 90◦ and it is reflected in the maxima of the line-widths for the photons
scattered by each population of interstellar H atoms separately.
The differences in the Lyman-α intensities, line-shifts, and line-widths calcu-
lated in models 1–3 are summarized in Table 2.3. It shows the differences (both
absolute and relative) between the model results for the upwind (θ = 0◦ ), crosswind
(θ = 90◦ ), and downwind (θ = 180◦ ) directions. It shows that the one-component
hot model 1 gives differences from 2 % (in the upwind direction) to 24 % (in the
downwind direction) as compared with model 3 in intensities and about 10–18 %
difference in the line-widths. The two-component hot model 2 leads to 8–27 %
discrepancies with model 3 in the intensities, and from 10 % (upwind) to 40 %
(downwind) discrepancies in the line-widths. The differences in the line-shifts of
models 1 and 2 relative to model 3 are also large, especially for the crosswind di-
rection. However, they are related to very small values of the line-shifts and most
probably can not be detected experimentally.
Line-Width of the Backscattered Lyman-α Profile as a

Diagnostic of the Nature of the Heliospheric Interface
Costa et al. (1999) and Quémerais et al. (2006) have analyzed the spectral prop-
erties of the backscattered solar Lyman-α radiation measured by SOHO/SWAN in
1996–2003. The Lyman-α line-width as a function of line-of-sight direction was
studied. It was shown that there is a noticeable minimum in the line-width at
θ = 50◦ − 60◦ . This minimum was interpreted as an indication of the two (primary
and secondary) components of the interstellar H atoms in the heliosphere and,
therefore, as evidence of the heliospheric interface. Indeed, the results obtained
in the frame of the one-component classical hot model shows (dashed-dot curve in
Fig. 2.19a) a monotonic increase of the line-width from upwind to downwind. The
existence of the two components, which have the bulk velocities shifted with respect
to each other, can help to produce the minimum of the line-width in directions close
to the crosswind direction.
Models 2 and 3 described in this paper have two populations of H atoms at
the outer boundary at 90 AU. Therefore, one can expect that a minimum of the
line-widths should be obtained in these models. However, as Fig. 2.19a shows, the
minimum is seen only for model 2 and not for model 3. Instead, for model 3 we see
a small maximum in temperature at θ = 150◦ . In order to explain these features
we have performed a series of additional test calculations.
Our goal is to understand and distinguish the roles of different effects influenc-
ing the line-width of Lyman-α profiles. To do that we studied the effects of all
possible model parameters and established that the following factors contribute to
the dependence of the line-width (Tlos ) on the angle θ:
1. The ionization processes that change the parameters of H atoms near the Sun.
From our test calculations we have found that solar gravitation and radiation
pressure have a smaller influence on the line-width than ionization.
2. The existence of two populations of the interstellar hydrogen atoms that are
shifted in velocity space. This effect leads to the appearance of the minimum
of Tlos at θ = 90◦ , as discussed before.
3. The non-Maxwellian behavior of the velocity distribution function of the two

populations of hydrogen at 90 AU, or more precisely, the large difference
between the Tz and Tρ “temperature” components (z and ρ are cylindrical
coordinates). This effect exists for model 3.
4. The spectral shape of the solar Lyman-α line. It will be shown below that
the shape of the solar spectrum has some small but interesting effect on the
line-width as a function of θ.
Figure 2.22 summarizes the results of the test calculations and explores the
effects listed above. Table 2.2 spells out the model numbering scheme used in
Fig. 2.22. For all of the calculations in this section, we assume that the solar
gravitational force is in balance with solar radiation pressure, i.e. μ = 1 and the
trajectories of the H atoms are straight lines. To explore the effects of ionization
a b
Figure 2.22: Results of test calculations of the Lyman-α line-width as a function of

angle θ. Model numbers are described in Table 2.2. Results are based on the one-
component hot model (model 1 curve 1) and the two-component hot model (model
2 curves 2 and 2b) are shown in plot (a) and results based on model 3 (curves 3,
3b, 3c) are shown in plot (b). Curves 2 and 2b correspond to calculations based on
model 2 with and without ionization; curve 3 corresponds to model 3, and curve
3b and 3c are the results of model 3, where ionization was switched off. In all
calculations except model 3c, the solar spectrum from Lemaire et al. (1998) was
assumed. For model 3c, a flat solar spectrum was employed [From Katushkina and
Izmodenov (2011)]
in the vicinity of the Sun, we performed test calculations with a typical ionization
rate of βE = 5.9 × 10−7 s−1 as well as calculations with βE = 0 (models 2b, 3b, 3c).
The effects of the local ionization on the function of Tlos (θ) are clearly illustrated
in the frame of the one-component hot model 1 (curve 1 in Fig. 2.22a) since neither
the two-component nor the non-Maxwellian effects are taken into account in this
model. The line-width of the backscattered Lyman-α profile increases monotoni-
cally from the upwind direction to downwind. This behavior reflects the angular
variation in the effective radial temperature of the H atoms that increases from
upwind to downwind (see Fig. 3.5 in Izmodenov 2006). Such a behavior is due to a
combination of the so-called selection effect that leads to a reduction in the atom
temperature in the upwind direction and in the effect of broadening of the velocity
distribution function toward the downwind direction. The latter effect is due to
the fact that atoms penetrate into the downwind mainly from the sides that makes
the distribution function broader (see Fig. 3.4 in Izmodenov 2006). The selection
effect is the effect when the slower atoms are more ionized as compared with the
fast atoms, because they spend more time before approaching the vicinity of the
Sun (and Earth) and have more time to be ionized. For a more detailed descrip-
tion of these effects and evolution of the H atom velocity distribution inside the
heliosphere, see Izmodenov (2006).
Results of the two-component hot model, 2b, that do not account for the ioniza-
tion (curve 2b in Fig. 2.22) illustrate the effect of the two populations of the inter-
stellar H atoms. These two populations have different bulk velocities Vz (at 90 AU
in the upwind direction: Vz,primary ≈ −27 km/s and Vz,secondary ≈ −16 km/s) and
rather small thermal velocities. Therefore, in the upwind and downwind directions
(where Vr = ±Vz ), the line-of-sight projections of the velocity distribution func-
tions of the primary and secondary components overlap only partially in velocity
space. In the crosswind direction Vr = Vρ ≈ 0 for both primary and secondary
interstellar atoms. This means that in this direction the projections of the distri-
bution functions on the radial line-of-sight overlap completely. That is why the
radial temperature of the mixture of primary and secondary interstellar atoms is
smaller in the crosswind direction than in the upwind and downwind directions.
This minimum of the radial temperature of H atoms in the crosswind direction is
reflected in the Lyman-α line-widths as seen from model 2b.
Results of the two-component hot model, curve 2 of Fig. 2.22, where the ion-
ization is taken into account combine both the increase in Tlos from upwind to
downwind due to the ionization effect and the local minimum in the crosswind
direction due to the two populations of H atoms. Hence, a small minimum of Tlos
at θ = 50–60◦ is seen in curve 2.
The line-width obtained from model 3 is presented in Fig. 2.22b. Note that this
model takes into account all known effects of the heliospheric interface, namely: (1)
two populations of interstellar H atoms, (2) θ-dependence of the hydrogen parame-
ters at 90 AU, and (3) non-Maxwellian features of the hydrogen velocity distribution
function at the outer boundary of the model. In particular, the differences between
kinetic temperature’s components, Tz and Tρ , play an important role.
Curves 3b and 3c correspond to models that do not take ionization into account.
Additionally, model 3c assumes that the solar Lyman-α flux does not depend on
frequency. This case is referred to as the “flat” solar spectrum. Note that in all the
other models we use the shape of the solar spectrum from Lemaire et al. (1998).
It can be seen that the effect of minimum in Tlos at θ = 90◦ (due to the two
populations) almost disappears in model 3c, but is still visible. This disappearance
of the minimum is related to the growth of the radial kinetic temperature Tr with
a increase of θ at 90 AU. The effect of the Tr increase with θ compensates the effect
of the minimum in the radial hydrogen temperature in crosswind due to the two
populations of H atoms. That is why the value of the minimum of the line-widths
at θ = 90◦ is much smaller for model 3c as compared with model 2b.
It is interesting to note that for model 3b (which is more realistic than model
3c) the local minimum of Tlos at θ = 90◦ is replaced with a small maximum. This
effect is due to the shape of the solar spectrum. Remember that in model 3c we
use a flat solar spectrum while in model 3b we use the nonuniform solar spectra
from Lemaire et al. (1998).
Let us return to model 3 without any additional assumptions. It is seen (curve 3)
that in this case there is no minimum in Tlos at 50–60◦ at all, but instead a small
maximum at θ = 150◦ . This maximum can be explained by the θ-dependence of
the hydrogen kinetic temperature Tr . The radial temperature of the sum of the
primary and secondary interstellar atoms is shown in Fig. 2.23 (plot a corresponds
to model 2, plot b corresponds to model 3). It is seen that in the case of the two-
Figure 2.23: Kinetic radial temperature (Tr ) for the sum of the primary and
secondary interstellar atoms in the heliosphere; Plot a corresponds to the two-
component hot model 2, plot b corresponds to the results of the model 3. In these
calculations μ = 1 and βE = 5.9 × 10−7 s−1 [From Katushkina and Izmodenov
(2011)]
component hot model 2 ionization leads to the maximum of Tr in downwind. For

model 3 the maximum of Tr is located at about θ = 150◦ . This effect is reflected
in the line-widths of the Lyman-α radiation that is seen in curve 3 of Fig. 2.22b.
Summary of the Section

In this section we employed various models of hydrogen distribution in the
heliosphere to compute the spectral properties of the backscattered solar Lyman-α
radiation as it would be measured at 1 AU in the anti-solar directions. We have
found that imprints of the heliospheric interface in the H atom distribution inside
the heliosphere have a strong influence on the Lyman-α parameters.
We considered the minimum of the line-width of the backscattered Lyman-α
radiation at 50–60◦ from upwind that was observed by SWAN (Costa et al. 1999;
Quémerais et al. 2006). In the experimental data the line-width in the directions
of θ = 50–60◦ is smaller than in the upwind direction by 1,500–2,000 K. This
minimum is seen for 1996 and 1997 and practically not seen for 2002–2003 years
although data points for small angles are not available for this period (see Fig. 7
in Quémerais et al. 2006). This minimum was explained in Costa et al. (1999) and
Quémerais et al. (2006) by the existence of two different populations of interstellar
hydrogen atoms that are shifted in velocity space. However, we noticed that the
line-width calculated with the 2D stationary Baranov–Malama model (Quémerais
and Izmodenov 2002) has no minimum of Tlos . The non-stationary 2D Baranov–
Malama model (Quémerais et al. 2008) predicts a small minimum in 2003, but no
minimum in 1997.
In this work we explored theoretically the nature of the observed minimum in
Tlos on the basis of three cases of hydrogen distribution inside the heliosphere. It
Summary 57
was shown that the minimum of the line-widths appears only for the two-component
hot model, and there is no minimum at all for model 3 that takes into account all
effects of the heliospheric interface. It was found that the absence of the minimum
in model 3 is due to the effect of the two components being compensated by the
non-Maxwellian features of the velocity distribution of H atoms at 90 AU after
they passed the heliospheric interface region, namely, by a strong anisotropy of the
kinetic temperatures of H atoms (Tz < Tρ ).
Therefore, the question of why the minimum of the Lyman-α line-width exists
in the experimental data remains unanswered. Possibilities for reproducing the
minimum still exist in the context of model 3. Firstly, the models considered here
do not take into account the effects of latitudinal and solar cycle variations of
the photoionization and charge exchange rates as well as solar radiation pressure.
These local effects may potentially change the result of this paper.
Another option to explain the minimum in line-width would be to change the
boundary conditions at 90 AU, i.e. to make a change in the model of the heliospheric
interface. For example, interstellar magnetic field may play a key role (Izmodenov
et al. 2005). Another possibility is a change in the multi-component nature of both
the heliospheric and interstellar plasmas (Malama et al. 2006; Izmodenov et al.
2009; Chalov et al. 2010). In this non-equilibrium plasma model the interstellar
pickup ions would be treated as a separate kinetic component. The plasma tem-
perature in the vicinity of the heliopause is smaller in that model as compared
with the Baranov–Malama model. Therefore, we could expect a decrease in the
kinetic temperatures of the secondary interstellar atoms. This might result in a
larger velocity space separation of the primary and secondary interstellar atoms at
90 AU. The separation may enhance the effect of the two populations. As a result
one may hope that the observed minimum will appear in the model.
Summary
The interaction between the supersonic flow of partially ionized plasma of the
local interstellar medium and the solar wind produces a complicated flow pattern
consisting of one or two shock waves (the heliospheric termination shock and, pos-
sibly, the bow shock in the LISM), and a contact discontinuity, the heliopause. Due
to the charge exchange process, the region between the two shocks (heliospheric
interface) separating these flows plays the role of a filter for the penetration of the
interstellar hydrogen atoms (and also for O, N, and other species) into the Solar
System.
From a theoretical point of view, the interaction should be considered in the
frame of kinetic-continuum models where the interstellar H atom component is
described in the framework of kinetic theory, since for hydrogen atoms the Knudsen
number with respect to charge exchange is Kn ∼ 1. The first self-consistent model
of the SW/LIC interaction was developed by Baranov and Malama (1993). Since
that time the set of kinetic-continuum models was developed. The modern kinetic-
continuum models take into account the following physical components/effects:
• Ionized interstellar helium component and solar wind alpha particles;
• Anomalous and galactic cosmic rays;
• Interstellar magnetic field;

• Solar cycle variations of the solar wind parameters;
• The heliotail;
• Filtration of interstellar oxygen and nitrogen;
• The multi-component nature of the heliospheric plasma.
Extensive efforts by other groups have also been put into modelling the he-
liospheric interface (Opher et al. 2009; Heerikhuisen and Pogorelov 2011; and
references therein). Nevertheless, the complete time-dependent multi-component
kinetic-continuum model that includes all effects above (plus the interplanetary
magnetic field) simultaneously has not yet been developed. This leaves a challenge
for future theoretical studies.
The numerical kinetic-continuum models of the heliospheric interface in the
frame of the Baranov–Malama model led, first, to the prediction of the many
physical phenomena discovered later onboard spacecraft and, secondly, to the in-
terpretation of previously obtained experimental data. In December 2004, an event
expected for more than 30 years took place: the Voyager 1 spacecraft finally crossed
the heliospheric termination shock at a distance of 94 AU. This was predicted (with
a 10 % accuracy) more than 25 years ago (Baranov et al. 1981; Baranov 1990, 2002).
In spite of significant progress in recent years, the global self-consistent 3D mod-
els are too computationally expensive to routinely perform detailed calculations of
the velocity distribution function in the vicinity of the Sun that is needed for anal-
ysis of the backscattered Lyman-α radiation. Katushkina and Izmodenov (2010,
2011) have developed a model which combines the simplicity of the hot-type models
of the H atom distribution inside the heliosphere with the results of a global model
of the SW/LIC interaction. It has been shown that this newly developed model can
be used for modeling the velocity distribution of the interstellar atoms inside the
heliosphere with sufficient accuracy. This model has been used to analyze how the
imprint of the heliospheric interface in the velocity distribution of the interstellar
hydrogen can be seen at one or several AU. We have also demonstrated that the im-
print of the heliospheric interface in the H atom distribution inside the heliosphere
can have a strong influence on the observed backscattered Lyman-α parameters. It
has been shown that the theoretically calculated moments of the Lyman-α spectra
have in general very good agreement with the SOHO/SWAN data. However, some
measured spectral features can not be reproduced by the current model.
One such feature that is not reproduced by the model is the minimum of the
line-width of the backscattered Lyman-α radiation at 50–60◦ from the upwind
direction that was observed by SWAN (Costa et al. 1999; Quémerais et al. 2006).
This minimum is supposed to be a natural consequence of there being two (primary
and secondary) populations of H atoms in the heliosphere. However, the non-
Maxwellian features of the H atom velocity distribution at 90 AU makes the effect
of the minimum non-visible in any model that assumes the velocity distribution
function at 90 AU to be in accordance with the Baranov–Malama (1993) model.
Overall, from the results reviewed and presented in the paper we can conclude
the following:
Bibliography 59
• Backscattered solar Lyman-α and its spectral characteristics like line-widths

are an excellent tool to explore kinetic properties of the interstellar atoms,
and, therefore, a tool to study the region of the SW/LIC interaction.
• Our state-of-the-art model (Katushkina and Izmodenov 2011) has been devel-
oped specifically to model the distribution of H atoms inside the heliosphere
properly for the purpose of understanding the spectral parameters of the
backscattered solar Lyman-α. This model allows us to combine the simplic-
ity of the hot-type models of the H atom distribution inside the heliosphere
with the results of the global models of the SW/LIC interaction.
• The comparison of the theoretically calculated Lyman-α intensities, line-

shifts and line-widths with SOHO/SWAN measurements have shown quite
good agreement.
• Solar cycle variations of the solar wind parameters influences the H atom
distribution and should be taken into account in the analysis of backscattered
Lyman-α.
• The qualitative difference in the line-width minimum measured at ∼ 60◦

discussed in this paper could be explained by the effects of the interstel-
lar/interplanetary magnetic fields or by the effects of the non-thermal behav-
ior of the interstellar plasma component. The exploration of these effects is
the subject for further studies.
Acknowledgements
The authors would like to thank ISSI for their support of the working group.
The calculations were performed by using the supercomputers of the Russian
Academy of Sciences and Lomonosov Moscow State University (“Lomonosov” and
“Chebishev”). The work was performed under the Presidium RAS Programm 22.
V.I. and O.K were partially supported by RFBR grants 10-02-93113-CNRS-a, 10-
02-01316-a, 11-02-92605-RS-a, and the Regional Public Fund for Russian science.
M.B. acknowledges support from the Polish ministry for science and higher educa-
tion, grants NS-1260-11-09 and N-N203-513-038.
Bibliography
J.M. Ajello, A.I. Stewart, G.E. Thomas, A. Graps, Solar cycle study of interplanetary
Lyman-alpha variations: Pioneer Venus Orbiter sky background results. Astrophys. J.
317, 964–986 (1987)
D.B. Aleksashov, V. Baranov, E. Barsky, A. Myasnikov, An axisymmetric magnetohydro-
dynamic model for the interaction of the solar wind with the local interstellar medium.
Astron. Lett. 26, 743–749 (2000)
I.I. Alekseev, A.P. Kropotkin, Passage of energetic particles through a magnetohydrody-
namic discontinuity surface. Geomagn. Aeron. 10, 755 (1971)
D.B. Alexashov, V.V. Izmodenov, M. Opher, Effects of the helisopheric and interstellar
magnetic field on the heliospheric interface. 37th COSPAR Scientific Assembly D13-0015-
08, P025-TueWed (poster) (2008)
W.I. Axford, The interaction of the solar wind with the interstellar medium. In Solar wind,
ed. by C.P. Sonett, P.J. Coleman, J.M. Wilcox. (Scientific and Technical Information
Office, National Aeronautics and Space Administration, Washington), p. 609 format?
(1972)
W.I. Axford, A.J. Dessler, B. Gottlieb, Termination of solar wind and solar magnetic field.
Astrophys. J. 137, 1268–1278 (1963)
V.B. Baranov, Gasdynamics of the solar wind interaction with the interstellar medium.
Space Sci. Rev. 52, 89–120 (1990)
V.B. Baranov, The heliosheath as a special case of stellarsheaths and the hydrogen wall as
a signature of the heliosheath. Planet. Space Sci. 50, 535–539 (2002)
V.B. Baranov, M.K. Ermakov, M.G. Lebedev, A three-component model of solar wind-
interstellar medium interaction: some numerical results. Sov. Astron. Lett. 7, 206–209
(1981)
V.B. Baranov, V.V. Izmodenov, Y.G. Malama, On the distribution function of H atoms in
the problem of the solar wind interaction with the local interstellar medium. J. Geophys.
Res. 103, 9575–9586 (1998)
V.B. Baranov, K.V. Krasnobaev, A.G. Kulikovksy, Sov. Phys. Dokl. 15, 791 (1971)
V.B. Baranov, M.G. Lebedev, Y.G. Malama, The influence of the interface between helio-
sphere and the local interstellar medium on the penetration of the H-atoms to the solar
system. Astrophys. J. 375, 347–351 (1991)
V.B. Baranov, Y.G. Malama, Model of the solar wind interaction with the local interstellar
medium: numerical solution of self-consistent problem. J. Geophys. Res. 98, 15157–15163
(1993)
V.B. Baranov, Y.G. Malama, Axisymmetric self-consistent model of the solar wind interac-
tion with the LISM: basic results and possible ways of development. Space Sci. Rev. 78,
305–316 (1996)
V.B. Baranov, N.A. Zaitsev, On the problem of the solar wind interaction with magnetized
interstellar plasma. Astron. Astrophys. 304, 631–637 (1995)
C.A. Barth, Mariner 6 measurements of the Lyman α sky background. Astrophys. J.Lett.
161, L181–L184 (1970)
J.-L. Bertaux, J.E. Blamont, Evidence for a source of an extraterrestrial hydrogen Lyman
α emission: the interstellar wind. Astron. Astrophys. 11, 200 (1971)
J.-L. Bertaux, R. Lallement, Analysis of interplanetary Lyman α line profile with a hydro-
gen absorption cell: theory of the Doppler angular spectral scanning method. Astron.
Astrophys. 140, 230–242 (1984)
J.-L. Bertaux, A. Ammar, J.E. Blamont, OGO-5 determination of the local interstellar wind
parameters. Space Res. 12, 1559–1567 (1972)
J.-L. Bertaux, J.E. Blamont, N. Tabarie, W.G. Kurt, M.C. Bourgin, A.S. Smirnov, N.N.
Dementeva, Interstellar medium in the vicinity of the sun: a temperature measurement
obtained with the Mars-7 interplanetary probe. Astron. Astrophys. 46, 19–29 (1976)
J.L. Bertaux, J.E. Blamont, E.N. Mironova, V.G. Kurt, M.C. Bourgin, Temperature mea-
surement of interplanetary interstellar hydrogen. Nature 270, 156–158 (1977)
J.-L. Bertaux, E. Quémerais, R. Lallement, E. Kyrölä, W. Schmidt, T. Summanen, J.P.
Goutail, M. Berthé, J. Costa, T. Holzer, First results from the SWAN Lyman α solar
wind mapper on SOHO. Sol. Phys. 175, 737–770 (1997)
J.-L. Bertaux, E. Kyrölä, E. Quémerais, R. Lallement, W. Schmidt, T. Summanen, J. Costa,
T. Makinen, SWAN observations of the solar wind latitude distribution and its evolution
since launch. Space Sci. Rev. 87, 129–132 (1999)
P.W. Blum, H.-J. Fahr, Interaction between interstellar hydrogen and the solar wind. As-
tron. Astrophys. 4, 280–290 (1970)
P.W. Blum, J. Pfleiderer, C. Wulf-Mathies, Neutral gases of interstellar origin in interplan-
etary space. Planet. Space Sci. 23, 93–105 (1975)
P. Blum, P. Gangopadhyay, H.S. Ogawa, D.L. Judge, Solar-driven neutral density waves.
Astron. Astrophys. 272, 549–554 (1993)
J.C. Brandt, J.W. Chamberlain, Interplanetary gas. I. hydrogen radiation in the night sky.
Astrophys. J. 130, 670–682 (1959)
Bibliography 61
P.C. Brandt, E.C. Roelof, P. Wurz, S. Barabash, D. Bazell, R. DeMajistre, T. Sotirelis, R.

Decker, Energetic neutral atom (ENA) imaging of the heliosheath: spectral characteristics
and implications for shock acceleration from observations by the neutral particle detector
(NPD) on board Venus Express (VEX). American Geophysical Union, Spring Meeting,
abstract SH24A-01 (2009)
L.F. Burlaga, N.F. Ness, M.H. Acuña, Crossing the termination shock into the heliosheath:
magnetic fields. Science 309, 2027–2029 (2005). doi:10.1126/science.1117542
M. Bzowski, Response of the groove in heliospheric Lyman-alpha glow to latitude-dependent
ionization rate. Astron. Astrophys. 408, 1155–1164 (2003)
M. Bzowski, Survival probability and energy modification of hydrogen energetic neutral
atoms on their way from the termination shock to Earth orbit. Astron. Astrophys. 488,
1057–1068 (2008)
M. Bzowski, D. Ruciński, Solar cycle modulation of the interstellar hydrogen density distri-
bution in the heliosphere. Space Sci. Rev. 72, 467–470 (1995)
M. Bzowski, H.-J. Fahr, D. Ruciński, H. Scherer, Variation of bulk velocity and temperature
anisotropy of neutral heliospheric hydrogen during the solar cycle. Astron. Astrophys.
326, 396–411 (1997)
M. Bzowski, T. Summanen, D. Ruciński, E. Kyrölä, Response of interplanetary glow to
global variations of hydrogen ionization rate and solar Lyman α flux. J. Geophys. Res.
107, CiteID 1101 (2002). doi:10.1029/2001JA000141
M. Bzowski, T. Makinen, E. Kyrölä et al., Latitudinal structure and north-south asymmetry
of the solar wind from Lyman-alpha remote sensing by SWAN. Astron. Astrophys. 408,
1165–1177 (2003)
M. Bzowski, E. Möbius, S. Tarnopolski et al., Density of neutral interstellar hydrogen at the
termination shock from Ulysses pickup ion observations. Astron. Astrophys. 491, 7–19
(2008)
S.V. Chalov, H.-J. Fahr, Energetic particles from the outer heliosphere appearing as a
secondary pick-up ion component. Astron. Astrophys. 401, L1–L4 (2003)
S.V. Chalov, D.B. Alexashov, D. McComas et al., Scatter-free pickup ions beyond the
heliopause as a model for the interstellar boundary explorer ribbon. Astrophys. J.Lett.
716, L99–L102 (2010)
J. Costa, R. Lallement, E. Quémerais et al., Heliospheric interstellar H temperature from
SOHO/SWAN H cell data. Astron. Astrophys. 349, 660–672 (1999)
J.M.A. Danby, G.L. Camm, Stat. dynam. accretion Monthly Not. Royal Astron. Soc. 117,
50–71 (1957)
L. Davis Jr., Interplanetary magnetic fields and cosmic rays. Phys. Rev. 100, 1440–1444
(1955). doi:10.1103/PhysRev.100.1440
R.B. Decker, S.M. Krimigis, E.C. Roelof et al., Voyager 1 in the foreshock, termination
shock, and heliosheath. Science 309, 2020–2024 (2005)
H.-J. Fahr, On the influence of the neutral interstellar matter on the upper atmosphere.
Astrophys. Space Sci. 2, 474–495 (1968a)
H.-J. Fahr, Neutral corpuscular energy flux by charge transfer collisions in the vicinity of
the Sun. Astrophys. Space Sci. 2, 496–503 (1968b)
H.-J. Fahr, The interplanetary hydrogen cone and its solar cycle variations. Astron. Astro-
phys. 14, 263–274 (1971)
H.-J. Fahr, The extraterrestrial UV-background and the nearby interstellar medium. Space
Sci. Rev. 15, 483–540 (1974)
H.-J. Fahr, Change of interstellar gas parameters in stellar-wind-dominated astrospheres:
the solar case. Astron. Astrophys. 66, 103–117 (1978)
W.C. Feldman, J.J. Lange, F. Scherb, Interstellar helium in interplanetary space. In Solar
Wind, ed. by C.P. Sonett, P.J. Coleman, J.M. Wilcox. (Scientific and Technical Infor-
mation Office, National Aeronautics and Space Administration, Washington, 1972), p.
668
H.O. Funsten, F. Allegrini, G.B. Crew et al., Structures and spectral variations of the outer
heliosphere in IBEX energetic neutral atom maps. Science 326, 964 (2009)
S.A. Fuselier, F. Allegrini, H.O. Funsten et al., Width and variation of the ENA flux ribbon
observed by the interstellar boundary explorer. Science 326, 962 (2009)
K. Gringauz, V. Bezrukih, V. Ozerov, R. Ribchinsky, A study of the interplanetary ionized
gas, high-energy electrons and corpuscular radiation from the Sun by means of the three-
electrode trap for charged particles on the second soviet cosmic rocket. Sov. Phys. Doklady
5, 361 (1960)
M. Gruntman, V.V. Izmodenov, Mass transport in the heliosphere by energetic neutral
atoms. J. Geophys. Res. 109, A12108 (2004). doi:10.1029/2004JA010727
M. Gruntman, E.C. Roelof, D.G. Mitchell, H.-J. Fahr, H.O. Funsten, D.J. McComas, En-
ergetic neutral atom imaging of the heliospheric boundary region. J. Geophys. Res. 106,
15767–15782 (2001)
J. Heerikhuisen, N.V. Pogorelov, An estimate of the nearby interstellar magnetic field using
neutral atoms. Astrophys. J. 738, 29 (2011). doi:10.1088/0004-637X/738/1/29
M. Hilchenbach, K.C. Hsieh, D. Hovestadt et al., Detection of 55–80 keV hydrogen atoms
of heliospheric origin by CELIAS/HSTOF on SOHO. Astrophys. J. 503, 916–922 (1998)
M. Hilchenbach, K.C. Hsieh, D. Hovestadt, R. Kallenbach, A. Czechowski, E. Möbius, P.
Bochsler, Energetic neutral hydrogen of heliospheric origin observed with SOHO/CELIAS
at 1 AU. In The Outer Heliosphere: The Next Frontiers, ed. by K. Scherer, H. Fichtner,
H.J. Fahr, E. Marsch. (Pergamon, Elmsford, 2000), pp. 273–276
A.J. Hundhausen, Interplanetary neutral hydrogen and the radius of the heliosphere. Planet.
Space Sci. 16, 783–793 (1968)
P.A. Isenberg, Evolution of interstellar pickup ions in the solar wind. J. Geophys. Res. 92,
1067–1073 (1987)
V.V. Izmodenov, Physics and gasdynamics of the heliospheric interface. Astrophys. Space
Sci. 274, 55–69 (2000)
V.V. Izmodenov, Velocity distribution of interstellar H atoms in the heliospheric interface.
Space Sci. Rev. 97, 385–388 (2001)
V.V. Izmodenov, Early concepts of the heliospheric interface: H atoms. In The Physics
of the Heliospheric Boundaries, ed. by V.V. Izmodenov, R. Kallenbach. ISSI Scientific
Report No. 5 (ESA-ESTEC, Paris, 2006), pp. 45–65
V.V. Izmodenov, Local interstellar parameters as they are inferred from analysis of obser-
vations inside the heliosphere. Space Sci. Rev. 143, 139–150 (2009)
V.V. Izmodenov, D. Alexashov, A model for the tail region of the heliospheric interface.
Astronomy Lett. 29, 58–63 (2003)
V.V. Izmodenov, D. Alexashov, Kinetic vs multi-fluid models of H atoms in the heliospheric
interface. In Solar Wind 11 / SOHO 16 Connecting Sun and Heliosphere conference, ed.
by B. Fleck, T.H. Zurbuchen, H. Lacoste. (ESA SP-592, 2005), p. 355
V.V. Izmodenov, D. Alexashov, Multi-component 3d modeling of the heliospheric inter-
face: effects of interstellar magnetic field. In Physics of the inner heliosheath: Voyager
observations, theory, and future prospects. 5th annual IGPP international astrophysics
conference. AIP Conference Proceedings, vol. 858 (2006), pp. 14–19
V.V. Izmodenov, Y.G. Malama, A.P. Kalinin et al., Hot neutral H in the heliosphere: elastic
H-H, H-p collisions. Astrophys. Space Sci. 274, 71–76 (2000)
V.V. Izmodenov, M. Gruntman, Y.G. Malama, Interstellar hydrogen atom distribution
function in the outer heliosphere. J. Geophys. Res. 106, 10681 (2001)
V.V. Izmodenov, G. Gloeckler, Y.G. Malama, When will Voyager 1 and 2 cross the termi-
nation shock? Geophys. Res. Lett. 30, 3–1 (2003). doi:10.1029/2002GL016127
V.V. Izmodenov, D. Alexashov, A. Myasnikov, Direction of the interstellar H atom inflow in
the heliosphere: role of the interstellar magnetic field. Astron. Astrophys. 437, L35–L38
(2005)
V.V. Izmodenov, Y.G. Malama, M.S. Ruderman, Modeling of the outer heliosphere with
the realistic solar cycle. Adv. Space Res. 41, 318–324 (2008)
V.V. Izmodenov, Y.G. Malama, M.S. Ruderman et al., Kinetic-gasdynamic modeling of the
heliospheric interface. Space Sci. Rev. 146, 329–351 (2009)
J.A. Joselyn, T.E. Holzer, The effect of asymmetric solar wind on the Lyman α sky back-
ground. J. Geophys. Res. 80, 903–907 (1975)
Bibliography 63
O.A. Katushkina, V.V. Izmodenov, Effect of the heliospheric interface on the distribution
of interstellar hydrogen atom inside the heliosphere. Astronomy Lett. 36, 297–300 (2010)
O.A. Katushkina, V.V. Izmodenov, The influence of effects on the heliospheric interface on
parameters of backscattered solar Lα radiation measured at the Earth’s orbit. Cosmic
Res. 50, 141–151 (2012)
S.M. Krimigis, D.G. Mitchell, E.C. Roelof, P.C. Brandt, Energetic neutral atoms (ENA)
from the termination shock/heliosheath? the view from 10AU. Paper presented at Voy-
agers in the heliosheath: observations, models, and plasmas physics Kauai, Hawaii, 9–14
Jan 2009
S.M. Krimigis, E.C. Roelof, R.B. Decker, M.E. Hill, Zero outward flow velocity for plasma
in a heliosheath transition layer. Nature 474, 359–361 (2011)
S. Kumar, A.L. Broadfoot, Evidence from Mariner 10 of solar wind flux depletion at high
ecliptic latitudes. Astron. Astrophys. 69, L5–L8 (1978)
S. Kumar, A.L. Broadfoot, Signatures of solar wind latitudinal structure in interplanetary
Lyman α emissions: Mariner 10 observations. Astrophys. J. 228, 302–311 (1979)
J.E. Kupperian, E.T. Byram, T.A. Chubb, H. Friedman, Far ultraviolet radiation in the
night sky. Planet. Space Sci. 1, 3–6 (1959)
V.G. Kurt, Measurement of scattered Lyman α radiation in the vicinity of the Earth and
in interplanetary space. In Space Research: Transactions of the All-union Conference on
Space Physics, ed. by G.A. Skuridin et al. NASA Technical Translation: NASA TT F-
389, Science Publishing House, Moscow, 10–16 June 1965, p. 769 (Translation published
by NASA, Washington DC, USA, May 1966)
V.G. Kurt, Kosmicheskie Issledovania (in russian) 5(6), 769–775 (1967)
V.G. Kurt, T.A. Germogenova, Scattering of solar Lyman α radiation by galactic hydrogen.
Sov. Astron. 11, 278–282 (1967)
V.G. Kurt, R.A. Syunyaev, Observations and interpretation of the ultraviolet radiation of
the Galaxy. Sov. Astron. 11, 928–931 (1967)
E. Kyrölä, T. Summanen, P. Raback, Solar cycle and interplanetary hydrogen. Astron.
Astrophys. 288, 299–314 (1994)
R. Lallement, The interaction of the heliosphere with interstellar medium. In The Century
of Space Science, ed. by A.M. Bleeker, J. Geiss, M.C.E. Huber. (Kluwer, New York,
2001), pp. 1191–1216
R. Lallement, J.-L. Bertaux, Deceleration of interstellar hydrogen at heliopause crossing
suggested by Lyman-alpha spectral observations. Astron. Astrophys. 231, L3–L6 (1990)
R. Lallement, A.I.F.S. Stewart, Out-of-ecliptic Lyman α observations with Pioneer-Venus:
solar wind anisotropy degree in 1986. Astron. Astrophys. 227, 600–608 (1990)
R. Lallement, J.-L. Bertaux, V.G. Kurt, E.N. Mironova, Observed perturbations of the
velocity distribution of interstellar H atoms in the solar system with Prognoz Lyman α
measurements. Astron. Astrophys. 140, 243–250 (1984)
R. Lallement, J.-L. Bertaux, F. Dalaudier, Interplanetary Lyman α spectral profiles and
intensities for both repulsive and attractive solar force fields predicted absorption pattern
by a hydrogen cell. Astron. Astrophys. 150, 21–32 (1985a)
R. Lallement , J.-L. Bertaux, V.G. Kurt, Solar wind decrease at high heliographic latitudes
detected from Prognoz interplanetary Lyman α mapping. J. Geophys. Res. 90, 1413–1423
(1985b)
R. Lallement, E. Quémerais, J.-L. Bertaux, S. Ferron, D. Koutroumpa, R. Pellinen, De-
flection of the interstellar neutral hydrogen flow across the heliospheric interface. Science
307, 1447–1449 (2005)
R. Lallement, E. Queḿerais, D. Koutroumpa, J.-L. Bertaux, S. Ferron, W. Schmidt, P.
Lamy, The interstellar H flow: updated analysis of SOHO/SWAN Data. AIP Conf. Proc.
1216, 555–558 (2010)
M. Lee, H. Kucharek, E. Möbius et al., An analytical model of interstellar gas in the
heliosphere tailored to interstellar boundary explorer observations. Astrophys. J.Suppl.
198, article id 10 (2012)
P. Lemaire, C. Emerich, W. Curdt et al., Solar H I Lyman alpha full disk profile obtained
with the SUMER/SOHO spectrometer. Astron. Astrophys. 334, 1095–1098 (1998)
B.G. Lindsay, R.F. Stebbings, Charge transfer cross sections for energetic neutral atom data
analysis. J. Geophys. Res. 110, A12213 (2005)
J. Linsky, B. Wood, The alpha Centauri line of sight: D/H ratio, physical properties of
local insterstellar gas, and measurement of heated hydrogen (the “hydrogen wall”) near
the heliopause. Astrophys. J. 463, 254 (1996)
L.J. Maher, B.A. Tinsley, Atomic hydrogen escape rate due to charge exchange with hot
plasmaspheric ions. J. Geophys. Res. 82, 689–695 (1977)
Y.G. Malama, Monte-Carlo simulation of neutral atoms trajectories in the solar system.
Astrophys. Space Sci. 176, 21–46 (1991)
Y.G. Malama, V.V. Izmodenov, S.V. Chalov, Modeling of the heliospheric interface: multi-
component nature of the heliospheric plasma. Astron. Astrophys. 445, 693–701 (2006)
D.J. McComas, R.W. Ebert, H.A. Elliott et al., Weaker solar wind from the polar coronal
holes and the whole Sun. Geophys. Res. Lett. 35, L18103 (2008)
D.J. McComas, F. Allegrini, P. Bochsler et al., Global observations of the interstellar inter-
action from the interstellar boundary explorer (IBEX). Science v326, 959 (2009)
R.R. Meier, Some optical and kinetic properties of the nearby interstellar gas. Astron.
Astrophys. 55, 211–219 (1977)
D.C. Morton, J.D. Purcell, Observations of the extreme ultraviolet radiation in the night
sky using an atomic hydrogen filter. Planet. Space Sci. 9, 455–458 (1962)
H.-R. Mueller, V. Florinski, J. Heerikhuisen, V.V. Izmodenov, K. Scherer, D. Alexashov,
H.-J. Fahr, Comparing various multi-component global heliospheric models. Astron. As-
trophys. 491, 43–51 (2008)
M. Neugebauer, C.W. Snyder, Solar plasma experiment. Science 138, 1095 (1962)
M. Opher, F.A. Bibi, G. Toth, J.D. Richardson, V.V. Izmodenov, T.I. Gombosi, A strong,
highly-tilted interstellar magnetic field near the Solar System. Nature 462, 1036–1038
(2009)
E.N. Parker, Dynamics of the interplanetary gas and magnetic fields. Astrophys. J.Lett.
128, 664 (1958)
E.N. Parker, The stellar-wind regions. Astrophys. J.Lett. 134, 20–27 (1961)
T.N.L. Patterson, F.S. Johnson, W.B. Hanson, The distribution of interplanetary hydrogen.
Planet. Space Sci. 11, 767–778 (1963)
W.R. Pryor, J.M. Ajello, C.A. Barth et al., The Galileo and Pioneer Venus ultraviolet
spectrometer experiments: solar Lyman-alpha latitude variation at solar maximum from
interplanetary Lyman-alpha observations. Astrophys. J. 394, 363–377 (1992)
W.R. Pryor, J.L. Scott, I.F. Stewart et al., Interplanetary Lyman α observations from
Pioneer Venus over a solar cycle from 1978 to 1992. J. Geophys. Res. 103, 26833–26849
(1998)
W.R. Pryor, J.M. Ajello, D.J. McComas, M. Witte, W.K. Tobiska, Hydrogen atom lifetimes
in the three-dimensional heliosphere over the solar cycle. J. Geophys. Res. 108, 8034
(2003). doi:10.1029/2003JA009878
E. Quémerais, Angle dependent partial frequency redistribution in the interplanetary
medium at Lyman α. Astron. Astrophys. 358, 353–367 (2000)
E. Quémerais, V.V. Izmodenov, Effects of the heliospheric interface on the interplanetary
Lyman α glow seen at 1 AU from the Sun. Astron. Astrophys. 396, 269–281 (2002)
E. Quémerais, R. Lallement, J.-L. Bertaux et al., Interplanetary Lyman α line profiles:
variations with solar activity cycle. Astron. Astrophys. 445, 1135–1142 (2006)
E. Quémerais, V.V. Izmodenov, D. Koutroumpa, Y.G. Malama, Time dependent model of
the interplanetary Lyman α glow: applications to the SWAN data. Astron. Astrophys.
448, 351–359 (2008)
D. Ruciński, M. Bzowski, Modulation of interplanetary hydrogen density distribution during
the solar cycle. Astron. Astrophys. 296, 248–263 (1995)
N.A. Schwadron, M. Bzowski, G.B. Crew et al., Comparison of interstellar boundary ex-
plorer observations with 3D global heliospheric models. Science 326, 966 (2009)
H. Sherer, M. Bzowski, H.-J. Fahr, D. Ruciński, Improved analysis of interplanetary HST
Ly-a spectra using time-dependent modelings. Astron. Astrophys. 342, 601 (1999)
Bibliography 65
I.S. Shklovsky, On hydrogen emission in the night glow. Planet. Space Sci. 1, 63–65 (1959)
E.C. Stone, A.C. Cummings, F.B. McDonald et al., Science 309, 2017–2020 (2005)
B. Strömgren, The physical state of interstellar hydrogen. Astrophys. J. 89, 526–547 (1939)
T. Summanen, R. Lallement, J.-L. Bertaux, E. Kyrölä, Latitudinal distribution of solar
wind as deduced from Lyman α measurements: an improved method. J. Geophys. Res.
98, 13215–13224 (1993)
T. Summanen, The effect of the time and latitude-dependent solar ionisation rate on the
measured Lyman α-intensity. Astron. Astrophys. 314, 663–671 (1996)
T. Terasawa, Energy spectrum and pitch angle distribution of particles reflected by MHD
shock waves fast mode. Planet. Space Sci. 27, 193–201 (1979)
G.E. Thomas, Properties of nearby interstellar hydrogen deduced from Lyman α sky back-
ground measurements. In C.P. Sonett, P.J. Coleman, J.M. Wilcox (eds.), Solar Wind,
Scientific and Technical Information Office, National Aeronautics and Space Administra-
tion., Washington, p. 661 (1972)
G.E. Thomas, R.F. Krassa, OGO-5 measurements of the Lyman α sky background. Astron.
Astrophys. 11, 218 (1971)
M.K. Wallis, Local interstellar medium. Nature 254, 202–203 (1975)
B.Y. Welsh, Warm and hot gas in the local ISM. Space Sci. Rev. 143, 241–252 (2009)
L. Williams, D.T. Hall, H.L. Pauls, G.P. Zank, The heliospheric hydrogen distribution: a
multifluid model. Astrophys. J. 476, 366–384 (1997)
N. Witt, P.W. Blum, J.M. Ajello, Solar wind latitudinal variations deduced from Mariner
10 interplanetary H 1216 A observations. Astron. Astrophys. 73, 272–281 (1979)
N. Witt, P.W. Blum, J.M. Ajello, Polar solar wind and interstellar wind properties from in-
terplanetary Lyman-alpha radiation measurements. Astron. Astrophys. 95, 80–85 (1981)
F.M. Wu, D.L. Judge, Temperature and flow velocity of the interplanetary gases along solar
radii. Astrophys. J. 231, 594–605 (1979)
—3—
Solar Parameters for Modeling the

Interplanetary Background
Maciej Bzowski∗ and Justyna M. Sokól

Space Research Center,
Polish Academy of Sciences, Warsaw, Poland
Munetoshi Tokumaru and Kenichi Fujiki

Solar-Terrestrial Environment Laboratory,
Nagoya University, Nagoya, Japan
Eric Quémerais and Rosine Lallement

LATMOS-IPSL,
Stéphane Ferron
ACRI-ST,
Guyancourt, France
Peter Bochsler
Space Science Center & Department of Physics,
University of New Hampshire, Durham, NH, USA
Physikalisches Institut, University of Bern, Bern, Switzerland
David J. McComas
Southwest Research Institute, San Antonio, TX, USA
University of Texas at San Antonio,
San Antonio, TX, USA
Abstract
The goal of the working group on cross-calibration of past and present ultraviolet
(UV) datasets of the International Space Science Institute (ISSI) in Bern,
Switzerland was to establish a photometric cross-calibration of various UV and

68 3. Solar Parameters
extreme ultraviolet (EUV) heliospheric observations. Realization of this goal re-

quired a credible and up-to-date model of the spatial distribution of neutral in-
terstellar hydrogen in the heliosphere, and to that end, a credible model of the
radiation pressure and ionization processes was needed. This chapter describes
the latter part of the project: the solar factors responsible for shaping the distri-
bution of neutral interstellar H in the heliosphere. In this paper we present the
solar Lyman-α flux and the topics of solar Lyman-α resonant radiation pressure
force acting on neutral H atoms in the heliosphere. We will also discuss solar EUV
radiation and resulting photoionization of heliospheric hydrogen along with their
evolution in time and the still hypothetical variation with heliolatitude. Further-
more, solar wind and its evolution with solar activity is presented, mostly in the
context of charge exchange ionization of heliospheric neutral hydrogen, and dy-
namic pressure variations. Also electron-impact ionization of neutral heliospheric
hydrogen and its variation with time, heliolatitude, and solar distance is discussed.
After a review of the state of the art in all of those topics, we proceed to present
an interim model of the solar wind and the other solar factors based on up-to-date
in situ and remote sensing observations. This model was used by Izmodenov et al.
(2013, this volume) to calculate the distribution of heliospheric hydrogen, which in
turn was the basis for intercalibrating the heliospheric UV and EUV measurements
discussed in Quémerais et al. (2013, this volume). Results of this joint effort will
also be used to improve the model of the solar wind evolution, which will be an in-
valuable asset in interpretation of all heliospheric measurements, including, among
others, the observations of Energetic Neutral Atoms by the Interstellar Boundary
Explorer (IBEX).
Brief Description of the Physics of the Neutral

Interstellar Gas in the Inner Heliosphere
The distribution of neutral interstellar hydrogen and the ultraviolet radiation in
the inner heliosphere are closely interrelated. Absolute calibration of observations
of the heliospheric backscattered Lyman-α glow requires knowledge of the well-
calibrated solar EUV output and of other solar forcing factors, mainly the solar
wind. The role of those factors and their variabilities in shaping the distribution
of neutral interstellar hydrogen can be derived from modeling papers cited in the
remaining portion of this section.
Assuming inflow of a fully neutral gas with a finite velocity, v∞ , and temper-
ature, T∞ , far away from the Sun, as well as a spherically symmetric and time-
independent ionization rate, β (r), plus an effective force, F (r), acting on the
atoms. The distribution function of the gas at a distance r from the Sun will be
axially symmetric around the inflow direction and can be given by the equation:
F (r)
v · ∇r f (v, r) + · ∇v f (v, r) = −β f (v, r) , (3.1)
mH
where ∇x is the gradient operation in the x-direction, r and v are position and
velocity vectors of the gas cell element, and mH is the hydrogen atom mass.
Physics of neutral gas in the inner heliosphere 69
Together with the assumption that the gas “at infinity” (in practice: a few
hundreds of AU from the Sun or at solar distances relevant for heliospheric models)
is homogeneous, collisionless, and Maxwellian (Izmodenov et al. 2000), this is the
basis for the classical hot model of the distribution of neutral interstellar gas in
the heliosphere (Thomas 1978; Fahr 1978, 1979; Wu and Judge 1979; Lallement
et al. 1985b). Based on Liouville’s theorem, the solution of this equation (Danby
and Camm 1957) for the distribution function f (v, r, t) for time t, location r and
velocity v can be expressed as:
f (v, r, t) = f∞ (v ∞ (v, r) , r∞ (v, r)) W (v, r, t) , (3.2)
where W (v, r, t) is the survival probability of an atom that arrives at the time t at
location r with velocity v from a distant location r∞ where its velocity was v∞ .
For now, t is only a formal parameter here. The probability of existence of such an
atom in the distant region of the heliosphere is given by the distribution function
f∞ (v∞ , r∞ ), and the link between the local velocity and position vectors v (t),
r (t) and the corresponding vectors in the so-called source region of the atoms can
be obtained from the solution of the equation of motion of hydrogen atoms in the
heliosphere:
G mH M (1 − μ (vr , t)) r
F (r, t, vr ) = − . (3.3)
r2 r
In this expression, F is the total force acting on the atom with mass mH , G is
the gravitational constant, M the solar mass, vr the radial velocity of the atom at
time t, and μ is the ratio of solar resonant radiation pressure force to solar gravity.
Radiation pressure will be more fully discussed in the section “Radiation Pressure
and Its Variations”.
As can be seen from this description, the distribution of neutral interstellar
hydrogen in the inner heliosphere is determined on one hand by the dynamical
influence of the Sun through the counteracting gravity and radiation pressure forces,
and on the other hand by the ionization losses, collectively denoted β in Eq. 3.1.
Both will be extensively discussed later in this paper. Here we only note that the
ionization processes include charge exchange between the incoming neutral atoms
and solar wind protons, ionization by impact of solar wind electrons, and ionization
by the solar EUV radiation.
There are no important sources of neutral gas in the region of velocity phase
space occupied by neutral interstellar gas, hence the lack of source terms in Eq. 3.1.
Recombination could potentially be considered as such a source, but is not impor-
tant for two reasons: (1) its rate is small in comparison with the ionization rate
(Wachowicz 2006), and (2) the recombined solar wind particles maintain their pre-
reaction velocities, which are equal to solar wind velocity. They do not contribute
to the population of heliospheric atoms capable of scattering the solar FUV radia-
tion which are responsible for the helioglow. Recombination is one of the secondary
sources of the so-called Neutral Solar Wind (Bleszyński et al. 1992; Gruntman 1994;
Bzowski and Ruciński 1996; Ruciński et al. 1998), which is beyond the scope of
this chapter.
The classical hot model is almost analytical (in fact, numerical calculations are
needed only when integrating the local distribution function to yield its moments,
such as density and mean velocity) and thus convenient to use, but it is far from
being perfect because many of its assumptions are not valid.
First, the interstellar gas in the Local Cloud is not fully neutral. The interaction
of its ionized component with the plasma of the solar wind creates a boundary
region of the heliosphere: the heliospheric interface. This interface begins at the
termination shock of the solar wind, where the solar wind becomes subsonic and
eventually turns back at the heliopause. The heliopause can be approximated as a
thin layer separating the solar wind plasma from the interstellar plasma. Beyond
the heliopause there is the outer heliosheath, where the pristine neutral interstellar
gas is altered due to charge exchange interactions with protons from the piled-up
and heated interstellar plasma. The history of the development of modeling of this
region of the heliosphere can be found in Baranov (2006b) and modern views on
this topic have been recently reviewed by Fahr (2004), Baranov (2006a), Izmodenov
and Baranov (2006), and by Izmodenov et al. (2013, this volume). Second, the
solar factors are not stationary or spherically symmetric, as will be demonstrated
in the remaining portion of this chapter.
With these two observations in perspective it can be easily understood that
quantitative interpretation of measuremeints of the helioglow require improvements
in the classical hot model, which were realized quite early in the history of helio-
spheric research.
Lallement et al. (1985b) allowed for latitudinal modulation of the charge ex-
change rate, approximating it with a one-parameter formula: 1 − A sin2 φ. This
enabled them to vary the equator-to-pole ratio of the ionization rates, but required
keeping the width and range of the equatorial region of enhanced ionization fixed.
A different extension of the hot model was proposed by Ruciński and Fahr (1989,
1991) who pointed out that the rate of electron-impact ionization does not scale
as r −2 , even though its effects on the distribution of neutral interstellar hydrogen
in the heliosphere are noticeable only within a few AU from the Sun, where its
density is already strongly reduced. This aspect of the heliospheric physics has
been neglected until Bzowski (2008) and Bzowski et al. (2008) reintroduced it in a
refined, latitude-dependent manner.
The next generation of heliospheric models abandoned the assumption of in-
variability of solar radiation pressure and ionization rate. The first, though very
simplified, model was proposed by Kyrölä et al. (1994), followed by Ruciński and
Bzowski (1995); Bzowski and Ruciński (1995a,b); Bzowski et al. (1997). They
studied variations in density, bulk velocity, and temperature of neutral interstellar
hydrogen near the Sun, as well as variations in the helioglow intensity. Because
of the lack of sufficient observational data at that time, they adopted an analytic
model of the evolution of radiation pressure and ionization rate over the solar cycle.
Another modification to models of the heliosphere was introduced by Scherer
et al. (1999), who addressed the prediction by Baranov et al. (1991); Osterbart and
Fahr (1992); Baranov and Malama (1993) [see also (Malama et al. 2006)] that the
charge exchange processes in the boundary layer of the heliosphere create a new,
so-called secondary, collisionless population of neutral H atoms. They modified
Physics of neutral gas in the inner heliosphere 71
the time-dependent hot model by approximating the distribution function, f∞

in Eq. 3.2, by a sum of two Maxwellian functions with densities, bulk velocities
and anisotropic temperatures being functions of the offset angle from the upwind
direction. One of the Maxwellians represented the so-called primary population of
neutral interstellar gas, which enters the supersonic solar wind after a “filtration”
process in the outer heliosheath. The other, the so-called secondary population of
neutral gas, is created via charge exchange with interstellar plasma in the outer
heliosheath.
The values of temperature, density, and bulk velocity in these Maxwellian func-
tions are parametrized by the angular separation of the point r ∞ in Eq. 3.2 from
the upwind direction. The values of parameters of the distribution function f∞ for
a given set of interstellar parameters (interstellar neutral and plasma densities, flow
speed and temperature) in this model must be obtained from an external model,
such as the Moscow Monte Carlo model of the heliosphere [see, e.g., Izmodenov
et al. (2009)]. Such an approach was later expanded and improved by Katushkina
and Izmodenov (2010).
Along with the two-population non-Gaussian model, an approximation of ra-
diation pressure and ionization rate by a spherically symmetric series of sines and
cosines was added. The coefficients of those model functions were obtained from
fits to measurements of the ionization rate and radiation pressure in the ecliptic
plane. These approximations were described by Scherer et al. (1999) and Bzowski
(2001a,b). Further extensions of the hot model to better account for latitudinal
variations of solar wind speed and density are presented in the section “Histori-
cal Perspective: Insight from Heliospheric Backscatter Glow”. The most recent
development in modeling was an addition by Tarnopolski (2007), Tarnopolski and
Bzowski (2008a) of radiation pressure force as a function of the radial velocity of
an atom relative to the Sun. This effect will be discussed later in this chapter.
The list of modifications to the classical hot model presented above is also a
list of effects that need to be taken into account at the solar side to facilitate
inter-calibration of measurements of the helioglow with other UV observations in
space. Apart from the heliospheric side, there is also the physics of the heliospheric
interface and the conditions in the Local Interstellar Cloud [see, e.g., Frisch et al.
2009, 2011 for review] that must be taken into account, which, however, are beyond
the scope of this chapter. From the above description it is clear that accurate
modeling of neutral interstellar hydrogen in the inner heliosphere requires accurate
knowledge of the factors contributing to the ionization and radiation pressure which
are the main topic of this chapter.
In the following section, we sketch a global picture of the solar factors influ-
encing neutral interstellar gas in the heliosphere. The new contributions to the
picture, accomplished as a result of the ISSI working group’s activities, are pre-
sented separately in two research papers: by Sokól et al. (2012), who elucidate the
solar wind evolution in time and heliolatitude, and by Bochsler et al. (2012), who
develops new ionization rates of heliospheric species. This chapter is intended as a
review of the topic, even though the results of the working group are discussed in
greater detail than insights from other sources.
Radiation Pressure and Its Variations

Temporal Evolution of the Total Solar Lyman-α
Flux in the Ecliptic
The radiation pressure force that acts on neutral interstellar H atoms in the
heliosphere is proportional to the total flux in the solar Lyman-α spectral line,
which is defined as the spectral flux integrated over a 1 nm interval from 120
to 121 nm and is referred to as the composite solar Lyman-α flux. It has been
measured since the middle of the 1970s [for the history of measurements, see Woods
et al. 2000]. Despite all efforts, while precision of the measurements has been good,
the problem of absolute calibration, prone to changes with time, has affected the
accuracy from the very beginning. It is a measure of progress in this field that
the discrepancies have been reduced from a factor of 4 in the 1970s to ∼ 15 %
nowadays.
The composite Lyman-α time series, available from the Laboratory for At-
mospheric and Space Physics (LASP) at the University of Colorado in Boulder,
CO (http://lasp.colorado.edu/lisird/lya/) is scaled to the absolute calibration of
UARS/SOLSTICE (Woods et al. 1996, 2000).
The time series of Lyman-α irradiance from the Sun shown in Fig. 3.1 is dom-
inated by an 11 year period which matches the sunspot cycle. The irradiance at
this wavelength at solar maximum is nearly double the value at solar minimum. In
addition to the solar cycle signal, the time series also shows a strong 27-day period
due to the rotation of the Sun. Active regions are much brighter in Lyman-α than
the surrounding quiet Sun, so the irradiance rises as these features rotate into view
on the solar disk.
The cadence of deduced flux values is presently 1 day and the inevitable gaps
are usually filled using a hierarchy of proxies, as illustrated in Fig. 3.1. The most
widely used is the proxy based on the solar radio flux at the 10.7 cm wavelength, the
so-called F10.7 flux (Covington 1947; Tapping 1987). Another frequently used proxy
is the Mg II core-to-wing (MgIIc/w ) ratio (Heath and Schlesinger 1986; Viereck and
Puga 1999).
The use of proxies raises the question of credibility of the results (Floyd et al.
2002, 2005). The solar FUV radiation varies on many time scales, from hours to
more than solar cycle length. Proxies generally follow the variability of the quantity
being modeled, but not precisely and not on all time scales. In particular, even
though the correlation coefficients, calculated from a long time series of daily values
may exceed 0.9, the agreement between the corresponding elements of the two time
series can sometimes be in disagreement on short time scales. One cause of this
disagreement is due the differing center-to-limb behavior of the proxy. Depending
on the solar latitude of the active region producing the emission, the timing of the
peak emission for the FUV and for the proxy may be significant (Floyd et al. 2005).
An illustration can be found in Fig. 3.2. The upper panel presents the daily
change rate of the composite flux (see figure caption for definition of this quantity)
and the lower panel the change rate of the monthly-averaged composite flux. The
amplitude of the change rate depends on primarily on which proxy is used rather
than on the strength of the solar cycle. While the monthly rate does not seem
Radiation Pressure and Its Variations 73
Lyman−α LASP daily composite flux

8 × 1011
F10.7 AE-E SME MgIIc/w UARS TIMED SORCE

7 × 1011
flux [cm−2 s−1]
6 × 1011
5 × 1011
4 × 1011
3 × 1011
1950 1960 1970 1980 1990 2000 2010
time [y]
Figure 3.1: Wavelength- and disk-integrated solar Lyman-α flux from the Labora-
tory for Atmospheric and Space Physics (LASP), referred to as the total Lyman-α
flux Itot . The daily time series is a composite of actual measurements from vari-
ous experiments re-scaled to the common calibration of UARS/SOLSTICE, with
the gaps filled by proxies. Color codes: gray: F10.7 proxy, orange: AE-E, pur-
ple: SME, green: MgIIc/w proxy, aquamarine: UARS/SOLSTICE version 18, red:
TIMED/SEE, blue: SORCE/SOLSTICE (based on Woods et al. 2000)
to depend on the source proxy, the difference between the change rates derived
from the F10.7 proxy and the proxies based on FUV measurements of the Sun is
especially pronounced.
As a consequence, the quality of the approximation at these short scales is
reduced even though it may be quite satisfactory at longer time scales, e.g., for
solar rotation period averages, as can be judged from the behavior of the monthly
change rates shown in the lower panel of Fig. 3.2 and as suggested by Dudok de
Wit et al. (2009). For modeling the distribution of neutral interstellar gas in the
heliosphere, the Carrington period of solar rotation is the finest time scale presently
in use,1 so the short time scale proxy imperfections are not a big problem for this
purpose.
Variation of the Lyman-α Flux with Heliolatitude

The disk-averaged solar Lyman-α flux is made of at least three components
(Amblard et al. 2008): a quiet Sun contribution and two components that vary
with solar activity, i.e., from the coolest regions of the chromosphere and from the
hot lower corona. The inhomogeneous heliolatitude distribution of active regions
was pointed out by Cook et al. (1980), who constructed a two-component latitude-
dependent model of disk-averaged solar UV irradiance.
Cook et al. (1981) considered the solar Lyman-α emission and suggested that
the ratio of the disk-integrated solar flux at the pole to the flux at the equator
1A H atom traveling at 30 km s−1 covers ∼ 0.5 AU during one Carrington period.

Daily change rate in the composite Lyman−α flux

3 × 1010
ΔF/Δt [cm−2 s−1 day−1] F10.7 AE−E SME MgIIc/w UARS TIMED SORCE
2 × 1010
1 × 1010
0
−1 × 1010
−2 × 1010
−3 × 1010
1950 1960 1970 1980 1990 2000 2010
time [y]
Monthly change rate in the composite Lyman−α flux
3 × 1010
ΔF/Δt [cm−2 s−1 month−1]
2 × 1010
1 × 1010
0
−1 × 1010
−2 × 1010
−3 × 1010
1950 1960 1970 1980 1990 2000 2010
time [y]
Figure 3.2: Rates of change of the total Lyman-α flux Itot . Shown are the rates
of change per unit time ΔI tot
Δt = Itot (ti+1 )−Itot (ti )
tt+1 −ti of the daily (upper panel) and
monthly (lower panel) composite Lyman-α flux presented in Fig. 3.1, with identical
color coding
should be about aLya = 0.8 during solar minimum, when the active regions are
distributed in latitudinal bands. These suggestions were supported by direct solar
minimum observations of the solar corona by Auchère (2005). Such a ratio was
suggested to be also valid for solar maximum by Pryor et al. (1992), based on
indirect evidence from observations of the Lyman-α helioglow. Thus it seems that
the latitudinal anisotropy of the Lyman-α flux does not change substantially during
the solar cycle, although this conclusion certainly needs further verification.
Bzowski (2008) suggested that the latitude dependence of the disk-integrated
solar Lyman-α flux may be approximated by the formula:

Itot (φ) = Itot (0) aLya sin2 (φ) + cos2 (φ), (3.4)
where φ is heliolatitude, aLya the “flattening” factor and Itot (0) is the equatorial
Lyman-α flux.
Heliolatitude variation of the disk-integrated Lyman-α line profile is, to our
knowledge, unexplored. On one hand, some form of variation should be expected
because, as shown by Tian et al. (2009a,b,c), the line profile depends on the features
on the solar disk that are being observed and the latitude distribution of these
features is inhomogeneous and varies during solar cycle. On the other hand, the
apparent lack of strong variability of the disk-integrated flux might suggest that the
spectral variation with heliolatitude is mild. Should the disk-integrated spectral
flux indeed vary with heliolatitude, this would potentially have consequences both
for the photoionization rate and the radiation pressure force. However, recent
investigation of the variation of the solar spectrum with heliolatitude by Kiselman
et al. (2011) seems to have brought a negative result (i.e., no variation).
Mechanism of Radiation Pressure

The mechanism of resonant interaction of an H atom with solar radiation, which
leads to a repelling force of resonant radiation pressure acting on neutral H atoms
in the heliosphere, was extensively discussed by Brasken and Kyrölä (1998). In
brief, the probability fabs (λ) that a hydrogen atom in the ground state, whose
base wavelength is λ0 , absorbs an incoming photon at wavelength λ is equal to:
ΓR
fabs (λ) = 2 , (3.5)
Γ2R
2πc 1
λ
− 1
λ0
+ 4
where ΓR is the energetic width of the second orbital of the atom, corresponding
to a Doppler width of about ±25 m s−1 around the rest wavelength of the Lyman-
α transition, and c is speed of light. An atom whose radial velocity relative to
the Sun is vr = 0 will absorb photons from the very center of the solar line at
λ0 , but if its radial velocity vr is non-zero, then due to the Doppler effect it will
be tuned to a different portion of the solar line profile, namely to the wavelength
λ = λ0 (1 − vr /c). Within about 10−7 s after it absorbs a Lyman-α photon, the
atom will re-emit the photon at an angle ω relative to the impact direction with
the scatter-angle probability p (ω) described by:
cos (ω) 11
p (ω) = + . (3.6)
4 12
Hence, a resonant interaction of the atom with a suitable photon results in a change
of atomic momentum at the moment of absorption by Δp = hv = c h/λ in the
antisolar direction, followed after a time of 10−7 s by another momentum change in
the direction described by Eq. 3.6. However, the typical frequency of interactions
between solar photons and H atoms at 1 AU, which is proportional to the solar
spectral flux, is on the order of 1/500 Hz (Quémerais 2006). Statistically, on time
scales shorter than the time scales of a change in atomic velocity relative to the Sun,
the only dynamical net effect of the interaction of the atom with solar radiation is
the antisolar momentum change.2
Since the interplanetary medium is optically thin within a few AU from the
Sun (Quémerais 2006), the solar spectral flux scales as the inverse square of solar
distance. Consequently, the solar radiation pressure also scales with solar distance
2 One can expect another statistical effect: an increase in the local velocity spread in the
population of neutral H gas in the heliosphere, but, to our knowledge, this effect has not been
studied in the available literature.
as 1/r 2 , thus leading to a partial compensation of solar gravity. Hence the effective
solar force acting on an atom is conveniently expressed by the fraction μ (Itot (t) , vr )
of the solar gravity force. It is proportional to the spectral flux Fλ corresponding
to the Doppler-shifted wavelength λ = λ0 (1 + vr /c), which results from the instan-
taneous radial speed vr of the atom relative to the Sun. Since the spectral flux
varies with time, effectively the μ factor is a function of radial velocity and time,
as expressed in Eq. 3.3.
Solar Lyman-α Line and Resulting Radiation Pressure

Measurements of the solar Lyman-α line profile, although scarce, date back to
the 1970s (Vidal-Madjar 1975; Artzner et al. 1978; Bonnet et al. 1978; Lemaire et al.
1978; Woods et al. 1995). However, they were performed from within the Earth’s
exosphere and hence suffered from the absorption by geocoronal Hydrogen in the
spectral region most relevant for the helioglow. Only after the launch of SOHO,
which orbits at the L1 Lagrange point, was it possible to obtain an unobstructed
view of the full spectral range of the disk-integrated solar line (Warren et al. 1998c;
Lemaire et al. 1998, 2002, 2005).
The solar Lyman-α line features a self-reversed shape that previously was ap-
proximated by two Gaussian functions (Fahr 1979; Chabrillat and Kockarts 1997;
Scherer et al. 2000). Recently, Tarnopolski (2007) and Tarnopolski and Bzowski
(2008b) showed that the measurements by Lemaire et al. (2002) can all be fit by a
three-Gaussian model parameterized by the disk-integrated flux:
μ (vr , Itot (t)) =

A [1+BItot (t)] exp −Cvr2 1+Dexp F vr −Gvr2 +Hexp −P vr −Qvr2 (3.7)
with the following parameters:
A = 2.4543 × 10−9 B = 4.5694 × 10−4 C = 3.8312 × 10−5

D = 0.73879 F = 4.0396 × 10−2 G = 3.5135 × 10−4
H = 0.47817 P = 4.6841 × 10−2 Q = 3.3373 × 10−4
and vr expressed in km s−1 . The accuracy of the fit is similar to the accuracy of

the measurements, estimated to be ∼ 10 %.
With this formula, one can calculate the μ factor for an arbitrary radial velocity,
vr , providing that the total solar Lyman-α flux, Itot , is known. The dependence
of the μ factor on radial velocity for the total flux values representative for solar
minimum and maximum conditions is shown in Fig. 3.3, adapted from Tarnopolski
and Bzowski (2008b).
The spectral region of the solar Lyman-α line most relevant for modeling the
heliospheric Lyman-α glow is the wavelength band straddling the central wave-
length by approximately ±30 km s−1 . The spectral flux at line center is closely
correlated with the line-integrated flux Itot (Vidal-Madjar and Phissamay 1980).
This is beneficial to the modeling of the helioglow because it permits us to easily
calculate the μ factor based on measurements of the solar line-integrated Lyman-α
flux.
Compensation factor m as function of radial velocity
1.5
gravity compensation factor m
1.0
0.5
0.0
–200 –100 0 100 200
radial velocity [km s–1]
Figure 3.3: Ratio μ of solar radiation pressure force to solar gravity based on
the model specified in Eq. 3.7, shown as a function of radial velocity of a H atom
relative to the Sun for the total solar flux values corresponding to the minimum
(red) and maximum (blue) of solar activity. Thick lines indicate the spectral region
±30 km s−1 around 0 Doppler shift, relevant for the neutral interstellar hydrogen
gas in the heliosphere
Approximate Values of the μ Factor

Since calculating the properties of neutral interstellar hydrogen inside the he-
liosphere using a model that takes the full solar line profile is computationally
demanding, an approach where the μ factor does not depend on radial velocity is
widely used. To obtain an appropriate μ value, a formula to translate the line-
integrated flux into the μ factor is needed. In the past, this issue was addressed by
taking
μ (Itot ) = 3.0303 × 1010 cm2 s a Itot , (3.8)
where Itot is the disk- and line-integrated solar Lyman-α flux and a is a constant
usually assumed to be between 0.85 and 1. With increasing accuracy of the mea-
surements, more sophisticated formulae have become available. Emerich et al.
(2005) fit the following relation between the spectral flux at line center Fλ and the
total flux:
1.21
Fλ Itot
= 0.64 ± 0.08. (3.9)
1012 cm−2 s−1 nm−1 1011 s−1 cm−2
Bzowski et al. (2008) found a linear relation between the spectral flux averaged
over the range ±30 km s−1 about the line center and the line- and disk-integrated
flux:
μ (Itot ) = 3.473 10−12 Itot − 0.287. (3.10)
Solar gravity compensation factor m

as function of total flux
2.0
1.8
1.6
m factor
1.4
1.2
1.0
0.8
3 × 1011 4 × 1011 5 × 1011 6 × 1011 7 × 1011

total flux [cm−2 s −1]
Figure 3.4: Solar radiation pressure factor μ as a function of the total flux in the
Lyman-α line Itot . The blue line represents the relation from Eq. 3.9 by Emerich
et al. (2005), which connects the total flux Itot with the spectral flux precisely at
line center Fλ0 . The conversion to the μ factor is then performed using Eq. 3.8,
with a = 0.9. The red line shows the relation defined in Eq. 3.10 by Bzowski et al.
(2008). Here, the relation between the total solar flux and spectral flux is adopted
from averaged spectral flux over ±30 km s−1 around the line center
Figure 3.4 shows a comparison of the predicted μ values as a function of the solar
total Lyman-α flux obtained from Eqs. 3.8–3.10. It suggests that if one decides not
to use a model radiation pressure force which is dependent on the radial velocity of
the atom, then the calculation of a good effective μ factor is not a straightforward
task. In fact, it may be appropriate to use different formulae for the upwind and
downwind regions in the heliosphere, as can be inferred from the asymmetry of the
solar Lyman-α line profile seen in Fig. 3.3. Such an approach, to our knowledge,
has never been implemented. Calculating μ as a function of vr is required in order
to model interstellar deuterium in the inner heliosphere (Tarnopolski and Bzowski
2008b).
The μ values obtained from the relations defined in Eqs. 3.9 (Emerich et al.
2005) and 3.10 (Tarnopolski and Bzowski 2008b), calculated from the monthly
values3 of the LASP composite Lyman-α flux (Fig. 3.1) are presented in Fig. 3.5.
Differences between the values obtained from these equations are on the order of
10 %, i.e. on the order of the uncertainty of the total flux.
3 Throughout the text, “monthly” is used as synonymous with “averaged over one Carrington
rotation period”.
Ionization Processes 79
Monthly averages of m factor, models comparison
2.0
1.8
1.6
m factor
1.4
1.2
1.0
1990 1995 2000 2005 2010

time [y]
Figure 3.5: Comparison of the solar radiation pressure factors μ that approximate
the compensation of solar gravity by the resonant radiation pressure force acting on
neutral H atoms. The green line represents monthly values obtained from Eq. 3.9
(Emerich et al. 2005) and the red line shows the monthly values obtained from
Eq. 3.10 (Bzowski et al. 2008). The blue line represents the μ values calculated
from Eq. 3.8 with a = 0.9. The gray dots represent daily values of the μ factor,
calculated from daily values of the composite Lyman-α flux using Fig. 3.8
Ionization Processes
The three main ionization processes of neutral interstellar hydrogen atoms in
the heliosphere are the following:
charge exchange with solar wind charged particles (mostly protons) resulting in a
pickup proton (pPUI ) and an energetic neutral atom (ENA):
H + p → pPUI + HENA ,
photoionization by photons ν of solar EUV radiation:
H + ν → pPUI + e,
and ionization by impact of solar wind electrons:
H + e → pPUI + 2e.
As a result of the charge exchange reaction, a solar wind proton captures an

electron from a neutral interstellar H atom and becomes an energetic neutral atom
(ENA) However, it maintains its momentum and thus does not enter the interstellar
population. In this respect, even though one neutral H atom is replaced by another,
such a reaction is still a loss process for the neutral interstellar gas despite the fact
that the total number of H atoms in the system does not change.
The newly-created protons are picked up by the solar wind flow (Fahr 1973;
Vasyliunas and Siscoe 1976) regardless of the reaction they originate from, creating
a distinct population that can be measured (Möbius et al. 1985; Gloeckler et al.
1993) and analyzed (e.g. Gloeckler and Geiss 2001; Gloeckler et al. 2004; Bzowski
et al. 2008), but this topic is outside the scope of this text.
The charge exchange reaction does not deplete solar wind protons. A proton
from the core of the solar wind distribution function is replaced with a pickup
proton. The concentration of protons per unit volume is not changed, although the
distribution function of solar wind protons is modified.
The two remaining ionization reactions do cause genuine losses for the entire H
population: a H atom enters the reaction and is not simply replaced with another
one at a different velocity. Eventually, a proton–electron pair is created and the
proton is picked up by solar wind.
Charge Exchange
General Formula
The process of resonant charge exchange between H atoms and protons is of
crucial importance for the physics of the heliosphere. It contributes to the pressure
balance between the solar wind and the interstellar gas and enables the energy
and momentum transport across the heliopause. Charge exchange losses of neutral
interstellar gas in the supersonic solar wind, which is discussed in this section,
are only a small piece of a larger picture of the role of charge exchange in the
heliosphere.
The rate of charge exchange between neutral H atoms and solar wind protons
can be regarded as the probability of a charge exchange act within unit time in
a given location in space. For an H atom traveling with velocity v H and a local
proton distribution function fp (vp ), where v p is the velocity vector of an individual
proton, the rate of charge exchange can be calculated from the formula:4

βCX = σCX (|v H − vp |) |vH − v p | fp (v p ) dvp , (3.11)
where v H − v p ≡ v rel is the relative velocity between the H atom and an individual
proton and σCX (|v H − v p |) is the reaction cross section. The integration covers
the entire proton velocity space. This formula can be put into an equivalent form:

βCX = σCX (|v rel |) |vrel | fp (vH − v rel ) dv rel . (3.12)
Depending on the underlying plasma regime and on the velocity of the H atom,
various simplifications can be made. When the kinetic spread of the plasma uT,p
is small compared with the plasma flow velocity vSW :
uT,p |v SW |

2kTp /mp v p f (v p ) dv p /np ,

4 We adopt a convention where bold-italic characters mean vector quantities, while italics
symbolize scalars.
we can approximate the proton distribution function by a delta-function centered

at the solar wind speed. The formula for charge exchange rate then simplifies to:
βCX = σCX (vrel ) np vrel , (3.13)
where np is the local proton density and vrel becomes vrel ≡ |v H − vSW |. This is
the case for ENAs that travel in the supersonic solar wind at vH ∼ 50 km s−1 or
faster.
For vH vSW , i.e., for atoms from the thermal interstellar H populations in
the supersonic solar wind, vrel vp . Then the rate of charge exchange between H
atoms and solar wind protons is given by:
βCX = σCX (vSW ) np vSW . (3.14)
This is the baseline formula for charge exchange rate between protons from the
supersonic solar wind and neutral interstellar H atoms. It has been widely used
in the heliospheric physics and will be used in the later part of this chapter. The
accuracy of this approximation is subject of one of the following subsections.
Charge Exchange Cross Section

The collision speed range most relevant for heliospheric physics is from ∼
1 km s−1 to ∼ 1, 000 km s−1 , which is equivalent to the energies of 0.005 eV and
5.2 keV, respectively. Relative velocities between interstellar neutral H atoms and
protons in the supersonic solar wind range from ∼ 300 km s−1 to ∼ 1, 000 km s−1 .
A detailed discussion of the charge exchange process and of the cross section
for this reaction can be found in Fahr et al. (2007) and will not be repeated here.
For the purpose of this work it is important to point out that there were four cross
section formulae used in heliospheric physics in the past: from Fite et al. (1962),
Maher and Tinsley (1977), Barnett et al. (1990), and Lindsay and Stebbings (2005).
Fite et al. (1962) and Maher and Tinsley (1977) both proposed to approximate
the charge exchange cross section as a function of relative velocity vrel between the
colliding partners by the formula:
2
σCX (vrel ) = (a + b ln vrel ) . (3.15)
The valid range for the Fite et al. (1962) relationship was claimed to be between 20
and 2,000 eV. The domain of the Maher and Tinsley (1977) expression was taken
to be from 0.005 to 1 keV.
Barnett et al. (1990) fit a form of Chebyshev polynomials which were valid in
a broad energy range. Bzowski (2001b) approximated the data used by Barnett
et al. (1990) but restricted to vrel < 800 km s−1 by the following expression:

3
i
σCX (vrel ) = a0 + ai (ln vrel ) . (3.16)
i=1
The most recent and authoritative compilation of measurements and calculations

was presented by Lindsay and Stebbings (2005), who suggested the following formula
Charge exchange cross sections for

H0 − H+ → H+ + H0
8. × 10−15
Fite et al.
Maher & Tinsley
6. × 10−15 Barnett et al.
cross section [cm2]
Lindsay & Stebbings
4. × 10−15
2. × 10−15
0
10 20 50 100 200 500 1000 2000
collision velocity [km s−1]
Figure 3.6: Cross sections for charge exchange reaction between protons and H
atoms in the energy range most important in the heliospheric physics. The recom-
mended relation from Lindsay and Stebbings (2005) is compared with the formulae
used in the past by Fite et al. (1962), Maher and Tinsley (1977) and Barnett et al.
(1990)
for the cross section expressed in cm2 , valid for collision energies E between 0.005
and 600 keV:
σCX (E) = 10−16 (1 − exp [−67.3/E])

4.5 2
(4.15 − 0.531 ln E) . (3.17)
A comparison of the cross sections from the four formulae is presented in Fig. 3.6.
It is important to note that while all four formulae return similar results for
the collision speeds relevant to the supersonic solar wind, the one from Lindsay
and Stebbings (2005) returns a significantly larger cross section for lower energies,
which are relevant for the outer heliosheath. Thus, adoption of the older formulae
may result in a significant underestimation of the coupling strength between the
neutral interstellar gas and the plasma in the outer heliosheath where the secondary
population of interstellar H atoms is formed. This would have marked consequences
for the results of heliospheric modeling, as described by Izmodenov et al. (2012,
this volume).
Averaging, Approximating and Estimating Errors

in the Calculation of Charge Exchange Rate
Models of neutral heliospheric gas usually need charge exchange rates averaged
over specific time intervals, typically the Carrington rotation period or a year.
Carrington period averages of this quantity will be extensively discussed in later
sections of this paper. But what is the correct way of calculating these averages?
Since the solar wind speed and density vary with time on time scales from
minutes to centuries, the instantaneous values of charge exchange rate vary on the
same time scales. Theory immediately suggests that to calculate the losses of the
neutral atom population due to charge exchange over a time interval ΔT , one has
to integrate the instantaneous rate given by Eq. 3.14 over this interval. The mean
charge exchange rate βΔT over interval ΔT is calculated from:

1
βΔT = σCX (vSW (t)) np (t) vSW (t) dt. (3.18)
ΔT ΔT
In practice, however, this strict approach is usually not possible to follow be-
cause high-resolution data on solar wind density and speed throughout the helio-
sphere are not available. Hence, a simplified version of the Carrington averaging is
adopted:
βCX Carr = np Carr vSW Carr σCX (vSW Carr ) , (3.19)
where .Carr marks averaging over the Carrington rotation period.

We have verified that calculations performed on the hourly data from the
OMNI-2 web page5 into Carrington rotation periods using Eq. 3.19 instead of
Eq. 3.18 introduces a bias of ∼ 3.5 % in the computed monthly charge exchange
rate. The bias fluctuates in time from 0 % to ∼ 8 %. The magnitude of this bias
is on the order of half the typical electron impact ionization rate, as can seen
in Fig. 3.7.
One source of error in the charge exchange rate is incomplete data coverage. In-
evitably, some fragments of time series measured in space happen to be unavailable.
For some time intervals, one quantity (density or speed) may be present while the
other one is missing. For some instruments, the data gaps are correlated with the
values of solar wind speed. The gaps typically occur in series and are not randomly
distributed over a Carrington rotation. This may bias the Carrington averages and
induce errors in the calculated averaged charge exchange rates. To estimate the
magnitude of the resulting errors, we calculated Carrington period averages of so-
lar wind density and speed from all available OMNI-2 data and then computed
the monthly charge exchange rates using Eq. 3.19. Subsequently, we changed the
data selection criterion: we demanded both density and speed to be available in
the qualifying hourly records and repeated the calculation of the charge exchange
monthly averages. Comparison of the resulting two time series suggests that an
error in the charge exchange rate due to data availability is about 2 %, but no bias
is introduced. The errors are likely to be largest during the Carrington rotations
with lowest data coverage. A presentation of the data coverage in the OMNI-2
collection can be found in Veselovsky et al. (2010).
Another source of error is the approximation of stationary H atoms. In this
approximation, used in Eqs. 3.14–3.19, it is assumed that the atoms subjected to
charge exchange losses do not move relative to the Sun, i.e., that vrel = vSW in
Eq. 3.13. This is not the case in the heliosphere, even for the atoms of neutral
interstellar gas. In the inner heliosphere, they typically travel at ∼ 30 km s−1
5 The OMNI-2 dataset is described in the section “Evolution of Solar Wind in the Ecliptic
Plane”.
Monthly−averaged rates of ionization processes in the ecliptic
btot bCX bHph bel
1 × 10−6
ionization rate [s−1]
5 × 10−7
1 × 10−7
5 × 10−8
1970 1980 1990 2000 2010

time [y]
Figure 3.7: Monthly ionization rates in the ecliptic plane from all relevant processes.
Blue marks the charge exchange rate, calculated from Eq. 3.14 using the cross
section from Eq. 3.17 (Lindsay and Stebbings 2005) and monthly averages of solar
wind speed and density shown in Fig. 3.8. Green represents the photoionization
rate (Bochsler et al. 2012), and orange the electron-impact ionization rate (Bzowski
2008). The total ionization rate, being a sum of the three aforementioned rates, is
shown in black. Note that the vertical axis is logarithmic. To better highlight the
secular change in the charge exchange rate after the turn of the past century, an
exploded view of the charge exchange rate for the time interval from 1985 through
2011 is shown in Fig. 3.15
(Bzowski et al. 1997). The modification of the charge exchange rate due to the
proper motion of H atoms can be assessed as follows. The atoms approaching the
Sun, i.e., mostly in the upwind hemisphere, run against solar wind and thus the rela-
tive speed is the sum of the proper velocity of the atoms and of the solar wind speed.
For a typical solar wind speed in the ecliptic plane of ∼ 440 km s−1 , the change in
the charge exchange rate is by (440 + 30) σCX (440 + 30) / (440 σCX (440)) 1.035.
Similarly, for the atoms in the downwind hemisphere which recede from the Sun,
the change in charge exchange rate is 0.965 (for σCX defined in Eq. 3.17). Thus,
the error induced by the approximation of stationary atom is about ±3.5 %. It
systematically varies with the offset angle from the upwind direction, transitioning
from an underestimation in the upwind hemisphere to an overestimation in the
downwind hemisphere.
All in all, the errors in monthly charge exchange rates due to the approxima-
tions presented are of similar magnitude. Since they are independent sources of
error, they can be added in quadrature and total about ∼ 5.5 %. This should be
contrasted with the uncertainty of ∼ 10 % related to the uncertainty of the cross
section alone.
Another approximation frequently made in the heliospheric modeling is that the
charge exchange rate decreases with the square of solar distance. This assumption
originates from the inverse-quadratical reduction of solar wind density with dis-
tance, with approximately constant velocity. We will assess now how good this
approximation is.
The solar wind expands almost radially (i.e., its non-radial velocity components
are very small in comparison with the radial component). It expands with basically
constant speed between the outer boundary of the acceleration region near the Sun
(located inside a few solar radii) and approximately 10 AU, where the slowdown
effects of mass loading due to the ionization of neutral interstellar gas become
measurable (Fahr and Ruciński 1999, 2001, 2002; Lee et al. 2009; Richardson et al.
1995, 2008b). The overall slowdown continues up to a few AU upstream from
the termination shock. At that point, the flow speed has already been reduced
by about 67 km s−1 relative to the speed at 1 AU (Richardson et al. 2008b). It is
additionally slowed down by the component of protons reflected at the termination
shock (Liewer et al. 1993; Richardson et al. 2008a). The effect of pickup ions
on distant solar wind varies with the angle from the inflow direction (Fahr and
Ruciński 1999), but the strongest effect is expected in the upwind hemisphere,
where Voyager measurements were made.
Outside the acceleration region, the solar wind flux initially falls off with solar
distance as 1/r 2 . This relation, stemming directly from the continuity equation, is
not significantly altered by the interaction of the solar wind plasma with neutral
interstellar gas. The main reaction is charge exchange—a reaction that does not
result in a change in the local proton density, only in a shift of the reaction product
into the pickup ion region of the total distribution function. Only the two secondary
reactions, photoionization and ionization by electron impact, actually inject new
protons into the pickup ion region in phase space.
As a result, the adjusted6 solar wind density, treated as a sum of the core so-
lar wind protons and pickup protons, increases very slowly with solar distance.
This was approximately assessed by Lee et al. (2009), who give the following for-
mula for the absolute density at a distance r of pickup ions nPUI,ph (r) created by
photoionization
2
rE nH,TS βph,E
nPUI,ph (r) = , (3.20)
r vSW,E
where βph,E is the photoionization rate at rE = 1 AU, vSW,E the solar wind speed
at rE , and nH,TS the neutral interstellar H density at the termination shock.
For the H density at the termination shock located at 90 AU equal to 0.087 cm−3
(Bzowski et al. 2008), the photoionization rate is 10−7 s−1 (see below), and the
typical solar wind speed is 440 km s−1 . The density of pickup ions at 90 AU from the
Sun is then equal to 3.3 10−5 cm−3 , which scales to 0.27 cm−3 when quadratically
adjusted to 1 AU. For a typical solar wind density at 1 AU equal to 5 cm−3 , this
yields a ∼ 5 % excess of the total solar wind density with respect to the pure 1/r 2
drop.
Inserting this modified adjusted density into Eq. 3.14, along with the reduced
solar wind speed of 440 − 67 = 373 km s−1 , one obtains an adjusted charge ex-
change rate of 3.7 10−7 s−1 , which is ∼ 3.5 % less than the rate calculated for the
6 Throughout this chapter, we refer to various quantities as “adjusted” meaning that we take
their magnitudes scaled by r2 , i.e., multiplied by the square of solar distance expressed in AU.
unenhanced and undecelerated solar wind. Hence, the modification of the charge
exchange rate in the outer supersonic solar wind from the 1/r 2 decrease is on the
order of the uncertainty related to the averaging and approximating the charge
exchange rate or less.
The remaining issue is whether or not using Eq. 3.14 to calculate the charge ex-
change rate of neutral H with pickup ions is justified. Pickup ions in the supersonic
solar wind make a special case in the calculation of the charge exchange rate. Even
though their bulk speed is equal to the solar wind expansion speed, the width of
their distribution function is comparable to the magnitude of the expansion speed
(Vasyliunas and Siscoe 1976; Möbius et al. 1988) and consequently the full version
of the formula for the charge exchange rate, defined in Eq. 3.12, should in principle
be used.
The simplification of using Eq. 3.14 for pickup ions seems acceptable because
it is not expected to introduce major errors. Close to the Sun (within ∼ 10 AU),
their abundance in the solar wind is small and consequently their contribution to
the total charge exchange rate is negligible. Outside of this region, their content
in the solar wind increases, but the total charge exchange rate becomes small.
Consequently, the contribution of charge exchange operating in this region of space
to the total losses of neutral interstellar gas observed from the inner heliosphere, es-
pecially via helioglow measurements, is also small. Another issue in the calculation
of charge exchange rate may be departures of the solar wind flow from purely ra-
dial expansion at larger heliocentric distances, which could result in a change of the
latitudinal structure with solar distance, as suggested by Fahr and Scherer (2004).
Detailed studies to support this qualitative discussion are missing in the literature,
as far as we know, except for some insight provided by Bzowski and Tarnopol-
ski (2006) on the radial profiles of survival probabilities of ENAs approaching
the Sun.
In our opinion, consistently taking all these effects into account would require
using a comprehensive, multi-fluid, three-dimensional and time-dependent model
of solar wind. Such a model should include the core solar wind, electrons, and
pickup ions as well as the solar wind magnetic field and be able to address both
the large-scale behavior of the solar wind flow and small-scale disturbances such
as turbulence. As boundary conditions, it should take actually and continuously
measured solar wind parameters as close to the corona as practical. To our knowl-
edge, such a model is now in development (see Usmanov et al. 2011 and references
therein).
Photoionization
From Solar Spectrum to Photoionization Rate
Photoionization is a secondary ionization factor of neutral interstellar H, but
its significance has recently increased because of the decrease in the average solar
wind flux observed since the last solar maximum (see upper panel in Fig. 3.8).
This has resulted in a decrease in the intensity of the dominant charge exchange
ionization rate.
Solar wind density from OMNI dataset

14
12
density [cm−3]
10
4
1985 1990 1995 2000 2005 2010
time [y]
Solar wind speed from OMNI dataset

650
600
speed [km s−1]
550
500
450
400
350
1985 1990 1995 2000 2005 2010
time [y]
Solar wind flux from OMNI dataset

6× 108
6 × 108
flux [cm−2 s−1]
6 × 108
6 × 108
6 × 108
1985 1990 1995 2000 2005 2010
time [y]
Figure 3.8: Carrington rotation averages of solar wind density adjusted to 1 AU

(upper panel), speed (middle panel), and adjusted flux (lower panel) calculated from
hourly averages from the OMNI-2 database (King and Papitashvili 2005). The thin
vertical lines mark the time interval which is expanded to daily time cadence in
Fig. 3.14
The rate of photoionization βph (t) at a time t can be calculated from the
formula:

λion
βph (t) = σph (λ) Fλ (λ, t) dλ, (3.21)

0
Cross section for photoionization of H
1
cross section [Mb]
0.01
10−4
10−6
0 20 40 60 80
wavelength [nm]
Figure 3.9: Cross section for hydrogen photoionization based on Eq. 3.22, adapted
from Verner et al. (1996)
where σph (λ) is the cross section for photoionization for wavelength λ and Fλ (λ, t)
is the solar spectral flux at the time t and wavelength λ. λion is the wavelength for
the threshold ionization energy. In the case of hydrogen, the spectral range of the
radiation capable of knocking out electrons from H atoms is entirely in the EUV
range. The cross section for photoionization of H can be expressed by the following
formula (Verner et al. 1996):
−2.963
−10 9.36664
σHph (λ) = 6.82297 × 10 √ +1 (λ − 2, 884.69) 2 λ2.0185 , (3.22)
λ
where the cross section is expressed in megabarns (Mb), wavelength in nm, and
λ ≤ 91.18 nm. The cross section is shown in Fig. 3.9, and the importance of various
portions of the spectrum for the photoionization is illustrated in Fig. 3.10. It shows
the integrand function from Eq. 3.21 for day of year 122 in 2001 (Bochsler et al.
2012).
It is clear from Fig. 3.10 that the most important portion of the spectrum for
the photoionization of hydrogen is the one immediately above the photoioniza-
tion threshold. The photoionization rate varies because the solar EUV spectral
flux varies. Direct measurements of the solar EUV spectrum in the entire rele-
vant energy range have been available on a 2-hourly basis since 2002, when the
TIMED/SEE experiment (Woods et al. 2005) was launched. Before then, the
spectral coverage was intermittent and one had to resort to indirect methods to
estimate the solar EUV flux.
Basically, these methods can be grouped into two classes: (1) direct integration
using Eq. 3.21 with the spectrum Fλ (λ, t) calculated from proxies, and (2) using
correlation formulae between selected proxies and photoionization rates obtained
for the times when the spectrum measurements are available. Since measurements
covering only a portion of the spectrum with a relatively low spectral resolution are
also available (CELIAS/SEM onboard SOHO; Hovestadt et al. 1995; Judge et al.
1998), a variant of method (2) would require finding a correlation between these
partial direct measurements and the photoionization rate. Neither of the methods
Integrand to calculate hydrogen photoionization rate

for 2010 DOY 122
0.00020
sHph(l) F(l) 0.00015
0.00010
0.00005
0.00000
0 20 40 60 80
wavelength [nm]
Figure 3.10: Integrand function from Eq. 3.21 for the flux obtained from
TIMED/SEE for 2010 DOY 122. The photoionization cross section used is de-
fined in Eq. 3.22 and shown in Fig. 3.9
gives perfect results. Furthermore, a reliable application of method (2) became

possible only after a sufficiently rich database of precise measurements of the solar
spectrum became available.
Inspection of Fig. 3.10 shows that the band from 50 to 30 nm, which is the most
relevant for helium and has been observed by CELIAS/SEM, is of secondary sig-
nificance for hydrogen. Since, however, variations in various portions of the solar
EUV spectrum are correlated to some extent, the evolution of the EUV flux in the
CELIAS/SEM bands is a reasonable indicator of the evolution of the photoioniza-
tion rates of H as well (Bochsler et al. 2012).
Using method (1) requires a proxy model of the solar spectrum in the relevant
photon energy range. A number of such models were developed in the past, in-
cluding SERF1 (also known as HFG or EUV81; Hinteregger et al. 1981), EUVAC
(Richards et al. 1994), SOLAR2000 (Tobiska et al. 2000), and NRLEUV (Warren
et al. 1998a,b; Lean et al. 2003; Warren 2006). These solar proxy models are based
on empirical correlations found between various portions of the solar spectrum and
selected time series of available measurements, including typically the F10.7 radio
flux, the MgIIc/w index, the solar Ca II K index, and even the sunspot time series.
The methodology and problems of creation of such models have been recently re-
viewed by Floyd et al. (2002, 2003, 2005) and Lean et al. (2011). The question of
finding suitable proxies has been discussed by Kretzschmar et al. (2006); Dudok de
Wit et al. (2005, 2008, 2009). Typically, linear correlations have been sought, but
they do not seem to be optimum representations of the true correlations in some
cases (Bochsler et al. 2012) as is shown in the following subsection.
Temporal Variation in the Photoionization Rate of Hydrogen

in the Ecliptic Plane
The process of photoionization of heliospheric hydrogen was extensively dis-
cussed by Ogawa et al. (1995) and we will not repeat such a discussion in this
chapter. We briefly present current views on the rate of photoionization of
heliospheric hydrogen based on both measurements and models. Bochsler et al.

(2012) developed a model of evolution of the Carrington period averages of the hy-
drogen photoionization rates that is based on directly measured solar spectra from
TIMED/SEE and a hierarchy of proxies. From the spectra at full time resolution,
available from 2002 until present, the photoionization rates were calculated using
Eq. 3.21. Since the time series obtained showed clear signatures of flares and local
particle-background contamination, it was filtered against the outliers beyond two
sigmas that show up in the time series of change rates (β (ti+1 ) − β (ti )) / (ti+1 − ti ).
Monthly averages were computed from the filtered time series, which are shown in
red in the lower panel of Fig. 3.11.
The TIMED/SEE coverage is limited in time, but intercalibrating or comparing
various measurements in the heliosphere taken at different times requires knowl-
edge of homogeneously-derived time series of ionization rates. Thus a hierarchy of
proxies was used to extend backwards in time the directly-obtained photoioniza-
tion rates. The proxy-based time series is available until the end of 1947, when the
measurements of the F10.7 flux became available.
Bochsler et al. (2012) started from directly integrated photoionization rates
calculated from the filtered TIMED spectra, which cover the full interval from solar
maximum to solar minimum. They calculated a time series of monthly averages
and found a correlation formula between these values on one hand and Carrington
rotation averaged measurements in Channel 1 SEMCh1 and Channel 2 SEMCh2
from CELIAS/SEM and the time series of Lyman-α flux Itot from LASP on the
other hand. This formula is the following:
βHph = 5.39758 × 10−20 Itot + 2.36415 × 10−16 SEM0.765549

Ch1 +
+ 3.98461 × 10−16 SEM0.765549
Ch2 + 2.05152 × 10−8 . (3.23)
Using this formula, Bochsler et al. (2012) calculated the Carrington averages of
photoionization rates for the entire interval for which the SEM data were available.
For the times when SEM data was unavailable, but the MgIIc/w index from LASP
was, they used another correlation formula:
βHph = 3.56348 × 10−6 MgIIc/w − 8.5947 × 10−7 . (3.24)
For epochs where the MgIIc/w index was unavailable, Bochsler et al. (2012) used
the following correlation formula with the F10.7 flux expressed in sfu units (i.e.,
10−22 W m−2 Hz−1 ):
βHph = 1.31864 × 10−8 F0.474344

10.7 − 1.7745 × 10−8 . (3.25)
It is worth pointing out that the exponent at the F10.7 flux is close to 1/2, not 1,
as is frequently adopted. The results of the model with color coding of the sources
used is shown in the lower panel of Fig. 3.11 for the time interval since 1990 until
present.
The photoionization rate obtained in the way described above can be compared
with the rate from the SOLAR2000 model, used extensively in previous studies
(e.g. Bzowski et al. 2009). The comparison is shown in the upper panel of Fig. 3.11.
The two models agree to about 10–15 %, with the direct integration model giving
Monthly photoionization rates of hydrogen:

SOLAR2000 vs present model
1.8 × 10−7 SOLAR2000 F10.7 MgIIc/w SEM/Lyman−α TIMED

photoionization rate [Hz]
1.6 × 10−7
1.4 × 10−7
1.2 × 10−7
1.× 10−7
8.× 10−7
1950 1960 1970 1980 1990 2000 2010
time [y]
SOL2000 vs direct integration relative difference
0.10
0.05
0.00
−0.05
−0.10
−0.15
−0.20
−0.25
1950 1960 1970 1980 1990 2000 2010
time [y]
Monthly photoionization rates of hydrogen:
SOLAR2000 vs present model
1.6 × 10−7
photoionization rate [Hz]
1.4 × 10−7
1.2 × 10−7
1.× 10−7
8.× 10−7
1990 1995 2000 2005 2010

time [y]
Figure 3.11: Comparison of the monthly averages of the hydrogen photoionization

rate obtained from the SOLAR2000 (Tobiska et al. 2000) and direct-integration
model, extended using the SEM/Lyman-α, Mg IIc/w , and F10.7 proxies (Bochsler
et al. 2012) in the upper panel. Relative difference between the SOLAR2000 and
the Bochsler et al. models are shown in the middle panel. Exploded view of
the photoionization rate from the Bochsler et al. (2012) multi-proxy model dur-
ing the two past solar cycles is shown in the lower panel. In this panel one can
assess the agreement of the multi-proxy model by Bochsler et al. (2012) with their
model solely based on the F10.7 proxy and with the ionization rate obtained from
SOLAR2000. The color code, common for the upper and lower panels, is shown in
the upper panel
almost always higher values. Such an accuracy is basically equal to the present
accuracy of the EUV measurements, especially in the low-energy portion of the
spectrum which contributes most to the ionization.
It seems that the cause of these small discrepancies is the difference in the way
the photoionization rates are calculated by SOLAR2000 and by Bochsler et al.
(2012). SOLAR2000 first calculates the solar spectrum from their sophisticated
system of proxies and then integrates the spectra to yield the ionization rate. The
quality of the derived spectrum seems to worsen when going backwards in time.
This is understandable because it is gradually based on fewer available measure-
ments of solar proxies. Bochsler et al. (2012) used measured solar spectra for the
time intervals when they were available and the absolute calibration was most cred-
ible. For other times they used proxies and correlation formulae specially developed
to connect a proxy measurement with the photoionization rate of a given species.
In this sense, technically, it is not a hierarchy of proxies, where one proxy model
is based on another proxy model. Rather, it is a collection of proxy models, which
all are based on one reliable series of direct measurements.
The lower panel of Fig. 3.11 presents the best estimate of the hydrogen pho-
toionization rate obtained by Bochsler et al. (2012) from their collection of proxies
(the colored line) with the rate calculated solely from the F10.7 proxy (black line)
and with the rate presented by SOLAR2000 (gray line). Particularly interesting
is the comparison of the model from the F10.7 proxy with the results of direct
integration of the solar spectrum (red vs black line), and with the SOLAR2000
results. While the F10.7 proxy model exactly tracks the direct rate, showing only
some departures above or below the red line, SOLAR2000 is consistently below.
Also outside the interval when the direct TIMED spectra could be used, the
F10.7 proxy model tracks quite well the best proxy model of Bochsler et al. (2012),
which leads us to believe that the system developed by these authors is self-
consistent and reliable, unless the relation of the F10.7 radio flux to the solar EUV
spectrum changed between 1948 and 1979, when the Mg IIc/w proxy became avail-
able. We are not aware of any such change described in the literature.
On the other hand one has to remember that the model by Bochsler et al.
(2012) is only able to calculate Carrington period averages of the ionization rate
of hydrogen and a few other selected species (He, O, Ne), while SOLAR2000’s
ambition is to provide an estimate for an arbitrarily selected day within its validity
interval. In fact, it allows one to calculate the ionization rate of any species since
it produces an approximate solar spectrum at a resolution of 1 nm. Given all the
challenges it has to face, it seems to be doing it remarkably well.
Latitude Variation of the Photoionization Rate

Just as for the solar Lyman-α radiation, radiation in the spectral region rele-
vant for photoionization is also expected to vary with heliolatitude. Auchère et al.
(2005b) constructed a model of equatorial and polar flux in the solar 30.4 nm line,
mostly responsible for ionization of helium, and demonstrated that a ∼ 0.8 pole-
to-equator ratio (fluctuating) can be expected. A similar modeling for the spectral
range relevant for photoionization of hydrogen is not available, but coronal ob-
servations by Auchère et al. (2005a) suggest that such an anisotropy can indeed
be expected and that some north-south asymmetry cannot be ruled out. In the
absence of a complete model and sufficient data we surmise that a latitude varia-
tion of hydrogen photoionization rate approximated by a formula similar to Eq. 3.4
may be tentatively adopted. The subject certainly needs further studies.
Electron Ionization
The significance of the electron-impact ionization reaction for the distribution
of neutral interstellar gas in the inner heliosphere was pointed out by Ruciński
and Fahr (1989), who developed a model of the electron ionization rate based on
the local electron temperature and density. Further insight into the problem of
electron-impact ionization of neutral interstellar hydrogen inside the heliosphere
can be found in Bzowski et al. (2008) and Bzowski (2008).
The ionization rate in the electron-impact reaction at a location described by
the radius-vector r can be calculated from the following formula (Owocki et al.
1983):
∞
8π
βel (r) = 2 σel (E) fe (E, r) E dE, (3.26)
me
Eion
where σel is the energy-dependent reaction cross section, E the collision energy
and Eion the ionization threshold energy. For hydrogen, Eion equals ∼ 13.6 eV.
Ionization occurs for the H atom—electron collisions with energies exceeding the
limiting energy Eion . Practically, almost all of the energy of the electrons in the
solar wind at a few AU from the Sun is in thermal motions.
The kinetic energy of an electron moving at a typical solar wind expansion speed
of 440 km s−1 is about 0.5 eV, which is much less than the ionization threshold.
Since the temperature of the electron fluid at 1 AU is on the order of ∼ 105 –
∼ 106 K, the thermal speeds of the solar wind electrons are on the order of a
few thousand of km s−1 , which strongly exceeds the expansion speed. Thus, the
expansion speed of the electron fluid can be neglected in the calculation of the
electron-impact ionization rate.
The formula for the cross section for electron ionization was proposed by Lotz
(1967b) and simplified for H by Lotz (1967a):

NA a q ln E
i i Pi 1 − bi exp −ci PEi − 1
σel (E) = , (3.27)
i=1
EPi
where NA is the number of electrons per ion and the summation goes over the
partial cross sections for knocking out all individual electrons from the ion. Pi
is the ionization potential for a given charge state of the ion, E is the impacting
electron energy, ai , bi , ci are parameters specific to a given ion and its charge state,
and qi is the statistical weight. For hydrogen, there is only one electron to be
knocked out and Eq. 3.27 takes the form:

E ln (E/13.6)
σel (E) = 4.0 × 10−14 1 − 0.60 exp −0.56 −1 , (3.28)
13.6 13.6E
Cross section for H electron−impact ionization
6. × 10−17
5. × 10−17
cross section [cm2]
4. × 10−17
3. × 10−17
2. × 10−17
1. × 10−17
0
50 100 500 1000 5000
collision energy [eV]
Figure 3.12: Cross section for electron-impact ionization of hydrogen, defined in

Eq. 3.26 from Lotz (1967a), as a function of H atom–electron collision energy in eV
where E ≥ 13.6 is expressed in eV. It is claimed by Lotz (1967a) to be accurate to

∼ 10 % and shown in Fig. 3.12.
The electron density can be estimated from quasi-neutrality and continuity
conditions in the solar wind and calculated from the formula:
ne = np (1 + 2ξα ) , (3.29)
where ξα is the local alpha-particle abundance relative to solar-wind protons. The
temperature behavior is much less simple. The distribution function of electrons in
the solar wind and its evolution with solar distance is fairly complex and requires
further studies. Note that measurements performed on Ulysses using two different
techniques—Quasi Thermal Noise (Issautier et al. 2001) and particle-measurements
(Salem et al. 2001)—return somewhat discrepant results (Issautier 2009). Basically,
the electron distribution function can be decomposed into three components: a
warm core of ∼ 105 K; a hot halo of ∼ 106 K; and a fluctuating strahl, stretched
along the local magnetic-field direction (Pilipp et al. 1987a,b).7
The core and halo can both be approximated by a Maxwellian function:
3/2
me E
fe (r, Te , E) = ne exp − , (3.30)
2π kB Te (r) kB Te (r)
where kB is the Boltzmann constant, me the electron mass, E the kinetic energy
of electron, Te the temperature of the electron fluid, and r the radius-vector of the
7 Note that the electron distribution function can be approximated by the kappa function—
originally by Maksimovic et al. (1997) and recently by Le Chat et al. (2010, 2011)—which naturally
covers both the core and halo components.
location in space. The contribution of the halo population to the net rate is on
the level of a few percent and is an increasing function of the heliocentric distance
(Maksimovic et al. 2005; Štveràk et al. 2009). Estimates by Bzowski et al. (2008)
show that at 1 AU, the ionization rate due to the core population of the solar
wind electrons is equal to about 0.4 × 10−7 s−1 and to the halo population less
than 0.04 × 10−7 s−1 , respectively. The amplitude of fluctuations in the electron
ionization rate may reach one order of magnitude, which is much more than the
long-time variations related to variations in solar activity. On the other hand, the
electron data from WIND (Salem et al. 2003) imply an in-ecliptic solar minimum
(1995) electron ionization rate of ∼ 0.68 × 10−7 s−1 and a solar maximum (2000)
rate of ∼ 0.73 × 10−7 s−1 . Thus, assuming a constant electron ionization rate over
the solar cycle is a reasonable approximation.
Observations done with Ulysses (Phillips et al. 1995a; Issautier et al. 1998; Le
Chat et al. 2011) suggest that the electron ionization rate is a 3D, time dependent
function of the solar cycle phase. Both the temperature magnitude and the cooling
rate differ between the fast and slow solar wind. Bzowski (2008) adopted the
following radial profiles of the core Tc and halo Th temperatures and the halo-to-
core density ratios ξhc = nh /nc for the slow solar wind (after Scime et al. 1994):
Tc = 1.3 · 105 r −0.85
Th = 9.2 · 105 r −0.38 (3.31)
ξhc = 0.06 r −0.25
and for the fast solar wind (after Issautier et al. 1998; Maksimovic et al. 2000):
Tc = 7.5 · 104 r −0.64
Th /Tc = 13.57 (3.32)
ξhc = 0.03.
In both solar wind regimes, the core nc and halo nh densities are calculated
from the equations:
1 + 2ξα
nc = np
1 + ξhc
nh = ξhc nc , (3.33)
with the solar wind alpha abundance ξα = 0.04, adopted to be identical in both
fast and slow wind regimes. np in this expression is the proton density. We adopt it
as velocity-independent because it introduces a relatively small modification to the
electron density, even though Kasper et al. (2012) show variations of this quantity
with solar activity as a function of solar wind speed.
Ruciński and Fahr (1989) inserted the formulae from Eqs. 3.27, 3.29, and 3.30
to the integrand function in Eq. 3.26 and obtained a formula for each of the terms
i contributing to the total electron-impact cross section, which we present here in
a slightly modified form:
⎛ ⎞
Γ 0, c + Pi exp [c ] b P
ai qi 8 me ⎝ Pi i kB Te i i i
βel,i (Te ) = ne Γ 0, − ⎠,
me Pi π kB Te kB Te Pi + ci kB Te
(3.34)
Equatorial and polar electron ionization rate

normalized to 1 AU
5 × 10−8
normalized ionization rate [s−1]
2 × 10−8
1 × 10−8
5 × 10−9
2 × 10−9
1 × 10−9
0.5 1.0 2.0 5.0 10.0 20.0
heliocentric distance [AU]
Figure 3.13: Normalized radial profiles r 2 βel (r) of the equatorial (red) and polar
(blue) rates of electron-impact ionization of neutral interstellar H atoms in the
supersonic solar wind, defined by Eqs. 3.35 and 3.36, respectively (adapted from
Bzowski 2008). Since those rates are calculated based on measurements performed
between 0.3 and 5.5 AU from the Sun, their validity is limited to this distance range
where Γ(a, z) is the incomplete gamma function and for hydrogen i = 1, qi = 1,

Pi = Eion = 13.6 eV, ai = 4.0 × 10−14 , bi = 0.60, ci = 0.56.
Using these relations, Bzowski (2008) employed the approach proposed by
Ruciński and Fahr (1989), assuming that the core and halo temperatures are
isotropic, which is not exactly the case, as shown by Štverák et al. (2008), and
calculated radial profiles of the electron-impact ionization rates separately for the
fast and slow solar wind. Subsequently, they approximated the results by the
following phenomenological formulae, which are valid between ∼ 0.3 and ∼ 5 to
10 AU:

np ln r (541.69 ln r − 1, 061.32) + 1, 584.32
βel,s (r, np ) = exp (3.35)
r2 (ln r − 29.17) ((ln r − 2.02) ln r + 2.91)

np ln r (348.73 ln r − 917.39) + 2, 138.05
βel,f (r, np ) = exp . (3.36)
r2 (ln r − 18.97) ((ln r − 2.53) ln r + 5.74)
As is evident in these formulae, the electron-impact ionization rates are parametrized
by local proton densities np normalized to 1 AU and are fixed functions of helio-
centric distance, which differ appreciably from the 1/r 2 profiles that are typical of
the solar wind flux and photoionization rate. An illustration of the departure of
the electron-impact ionization rate from 1/r 2 behavior is shown in Fig. 3.13.
The behavior of the electron temperature in the compressed solar wind resulting
from stream-stream interactions is not easily modeled and thus the behavior of the
Evolution of Solar Wind in the Ecliptic Plane 97
electron ionization rate in this solar wind regime is poorly known. Bzowski (2008)
approximated the electron ionization in the solar wind of intermediate velocity as
a weighted mean value between the rates specific for the fast and slow wind.
The differences in the electron fluid parameters and their variation with solar
distance, as well as the differences between the electron fluids in the fast and slow
wind result in a pronounced latitude anisotropy of the electron ionization rate
throughout the solar cycle. An exception is a brief interval at solar maximum,
when the solar wind becomes almost spherically symmetric.
The significance of the electron impact ionization in the overall balance of the
contributing ionization reactions increases towards the Sun. Eventually, the role
of electron ionization and its anisotropy is greater in the downwind hemisphere,
where neutral interstellar gas has already passed the Sun (i.e., the streamlines of
the H gas flow have passed their perihelia).
Because of the fast cooling of the electron fluid with solar distance, electron-
impact ionization almost negligibly affects the distribution of neutral interstellar
hydrogen in the upwind hemisphere at the distances beyond ∼ 2 AU (Tarnopolski
and Bzowski 2008b). This distance seems to be close to the distance to the Max-
imum Emissivity Region of the heliospheric Lyman-α backscatter glow.8 Hence,
the limited validity range of the formulae specified in Eqs. 3.35 and 3.36 seems to
not hinder modeling of the effective ionization rate of heliospheric neutral atoms
at larger distances from the Sun within the supersonic solar wind.
Summary of the Variations in the In-ecliptic

Ionization Rates
A summary of the variations in the in-ecliptic ionization rates of hydrogen,
scaled to 1 AU, is presented in Fig. 3.7 for a time interval from 1970 through the
end of 2011. Charge exchange with solar wind protons dominates, photoionization
is the next largest rate, and electron ionization is the third. The total ionization
rate does not seem to be a periodic function with a period close to the solar cycle
period, even though one of the components, the photoionization rate, does feature
a strong solar cycle periodicity. Rather, a much longer periodicity may be guessed.
On the time scale of spaceborne solar EUV measurements, a secular decrease in
the total ionization rate by about 30 % between 1980s and 1990s and the present
decade is observed.
Evolution of Solar Wind in the Ecliptic Plane

The existence of the solar wind was theoretically predicted by Parker (1958)
and experimentally discovered at the very beginning of the space age, when the
spacecraft Lunnik II and Mariner 2 had left the magnetosphere (Gringauz et al.
1960; Neugebauer and Snyder 1962). Regular measurements of solar wind param-
eters started in the beginning of the 1960s and up to now, data from more than
8 The Maximum Emissivity Region is by definition the region where the maximum of the
source function for the Lyman-α backscatter glow is located.

20 spacecraft have been available, obtained using various observations and data
processing techniques.
The highly supersonic solar wind in the ecliptic plane consists of a sequence
of various interleaved components: a “genuine” slow solar wind, a fast solar wind,
solar wind plasma from stream-stream interaction regions, and (intermittently)
interplanetary coronal mass ejections (CME). The Mach number of the flow at
1 AU varies from ∼ 3 to ∼ 10. The balance between the populations varies with
solar activity.
Historically, measurements of the solar wind speed obtained from various ex-
periments generally agreed among themselves with an accuracy of ∼ 5 %, but
systematic differences between density values from different experiments up to
∼ 30 % existed (for extensive discussion, see the OMNIWeb documentation at
http://omniweb.gsfc.nasa.gov/html/omni2 doc.html). Hence any study of a long-
term behavior of the solar wind required intercalibration of the results from dif-
ferent experiments. Such an initiative brought the OMNI data collection [King
and Papitashvili available at at http://omniweb.gsfc.nasa.gov/]; see also King and
Papitashvili (2005) where historical and present measurements of the solar wind
density, velocity, temperature, alpha abundance, and magnetic field vector were
brought to a common calibration.
Originally, the OMNI collection was created in the 1970s by the National Space
Science Data Center (NSSDC) at the NASA Goddard Space Flight Center. In 2003,
a successor database OMNI-2 was made available, which has been maintained until
present. The OMNI-2 data collection was the basis for the present study of solar
wind parameters in the ecliptic and is used in this chapter to construct the charge
exchange ionization rates.
The early period of the OMNI databases (1963–1971) includes data from mul-
tiple spacecraft (Bonetti et al. 1969; Neugebauer 1970), the middle-period data
are mostly from IMP-8 and span from 1971 to 1994, and the later periods, from
1994 until present, include mostly data from IMP-8, or the WIND Solar Wind
Experiment (SWE) (Kasper 2002), and ACE/SWEPAM (McComas et al. 1998a).
Since there was no overlap between the early and middle period, the data from
the early period in the OMNI-2 database are adopted unchanged from the original
OMNI collection. The data obtained from various spacecraft during the early
period were extensively intercalibrated, but no intercalibration with the middle
and recent periods was possible. Still, owing to the overlap of the data between the
middle and recent periods, it was possible to perform an intercalibration between
the data from these two periods.
It is important to mention the significance of the correlation work that the
OMNI team performed on the data from the IMP-8 spacecraft and early WIND and
ACE measurements. IMP-8 operated for 28 years from 1973 to 2001 and provided
a bridge between the early and present observations, enabling the presentation of
a more or less homogeneous series of solar wind parameters shown further down in
this chapter.
Most of the solar wind plasma data used in the OMNI collection were ob-
tained from the MIT Faraday Cups (Bridge et al. 1965; Lyon et al. 1967, 1968;
Lazarus and Paularena 1998) and LANL electrostatic analyzers (Bame et al. 1971,
1978a,b; Hundhausen et al. 1967; Ogilvie et al. 1968; Feldman et al. 1973; Asbridge
et al. 1976; McComas et al. 1998a). In the OMNI-2 series, King and Papitashvili
(2005) adopted measurements of solar wind parameters performed by the WIND

spacecraft as basis for the common calibration. They followed in this respect an
analysis performed by Kasper (2002) and Kasper et al. (2006). The latter paper
includes a comprehensive, physics-based analysis of the accuracy of solar wind mea-
surements, especially of the measurements performed using the WIND spacecraft
Faraday cups.
The preparation of the data published in the OMNI collection involves removing
potential Earth bow shock contamination and incomplete records from the original
high-resolution data supplied by the Principal Investigators of the experiments,
and subsequently time-shifting the data from the spacecraft location to the Earth.
The primary source of the solar wind data are currently WIND measurements,
but they are being superceded by measurements from the Advanced Composition
Explorer (ACE). These, unlike the data from WIND, are free from possible bow
shock contamination because the ACE spacecraft operates near the Earth at a
Lissajous orbit close to the L1 Lagrange point about a million kilometers upstream
of the Earth’s bow shock.
Because of different time scales of the processing of the data from various space-
craft, typically an interim data product becomes available once the first data are
obtained, which is superseded with the final product when all the data needed be-
come available or are declared as unavailable. This results in some updates to the
published records over time. Our experience shows, however, that the changes are
seldom significant for the Carrington averages.
The data from different experiments are scaled to a common calibration using
linear fits based on results of linear regression analysis. The result of the intercali-
bration process is a time series of hourly-averaged solar wind parameters. Because
of the varying quality of individual records, the time coverage of the parameters
is not uniform and gaps may exist in some parameters, while correct data for the
same time interval may be available for others.
Since the distribution function of solar wind is inhomogeneous and varies rapidly
in time, the values of solar wind parameters retrieved from observations depend
on the method used to process the data. Typically, the LANL team take mo-
ments of the observed distribution function to calculate the density, speed and
temperature of the solar wind, while the MIT team fit the measurements to an
anisotropic Maxwellian or bi-Maxwellian function using a nonlinear fit method. To
assess differences resulting from the two aforementioned approaches, the MIT team
calculated the density, speed and temperature from the WIND/SWE distributions
using both methods.
King and Papitashvili at http://omniweb.gsfc.nasa.gov/html/omni2 doc.html
extensively discuss the differences and correlations between data from various
sources. They show that the velocities are very well linearly correlated, with the
coefficients of the relation vSWE = a + b vACE equal to a = −2.135 ± 0.387 and
b = 1.010 ± 0.001.
In the case of densities, it is the logarithms of density which are linearly corre-
lated and the coefficients of the formula log nSWE = a+b log nACE slowly vary with
speed, a changing from 0.006 for v < 350 km s−1 to 0.091 for 350 < v < 450 km s−1
to 0.082 for v > 450 km s−1 and b changing from 1 for v < 350 km s−1 to 1.036 for
350 < v < 450 km s−1 to 1 for v > 450 km s−1 .
This leads to differences on the order up to 20 % between ACE and WIND,

which are comparable to the uncertainty in density coming up from the application
of various methods of parameter derivation discussed earlier. In a nutshell, while a
very good correlation of speeds is obtained, the correlation between the logarithm of
the densities is close to linear, but with a scatter of approximately 30 % around the
fit line. This is probably a good measure of the inherent uncertainty of the densities
even without taking into account the uncertainties in the absolute calibrations.
Based on the OMNI database, we constructed a time series of Carrington
period-averaged parameters of the solar wind normalized to 1 AU, with the grid
points set precisely at halves of the Carrington rotation intervals. Small deviations
of the times from the halves of the rotation periods were linearly interpolated.
Averaging over the Carrington rotation enables the construction of a consistent,
axially symmetric model of the ionization rate. The time series of density, veloc-
ity, and charge exchange rate in the approximation of neutral H atoms stationary
relative to the Sun is presented in Fig. 3.8. The daily averaged values for density
and speed of the solar wind for a much shorter time interval is shown in Fig. 3.14.
The time interval shown in Fig. 3.15 starts before the solar activity minimum
in 1986 and includes the solar minima of 1995 and 2007, as well as the two recent
maxima of 1990 and 2001. One observes a striking difference in the appearance
of the solar wind equatorial parameters in comparison with the behavior of solar
EUV radiation (compare Fig. 3.8 with Fig. 3.1; see also the behavior of the charge
exchange rate contrasted with the photoionization rate in Fig. 3.7). Neither density
nor speed seem to be correlated with the level of solar activity. There is no clear
minimum–maximum–minimum variation, which is clearly seen in the EUV-related
time series. Speed shows multi-timescale variations, but its mean value is basically
constant over time.
By contrast, the density features a secular change, which begun just before the
last solar maximum and leveled off shortly before the present minimum. The overall
drop in the average solar wind density is on the order of 30 % between 1998 and
2005. Thus the solar wind density features a “plateau” until 1998, then a “cliff”
and a “foot” starting from 2002. Within the “foot,” density fluctuations seem to
exist that are anticorrelated with solar wind speed. These can be associated with
the persistence of coronal holes at equatorial latitudes, as convincingly illustrated
by de Toma (2011).
The present rate of charge exchange (oscillating about 4 10−7 s−1 , cf Fig. 3.15)
is at a level similar to the charge exchange rate observed by Ulysses at the poles
during its first fast latitude scan (Phillips et al. 1995a; McComas et al. 1999). The
reduction relative to the pre-drop values of ∼ 6.5 10−7 s−1 is by ∼ 40 %.
The overall long-standing drop in the ionization rate must result in an overall
enhancement in the neutral interstellar gas density in the inner heliosphere (see,
e.g., Bzowski et al. 2009 for the effects of various parameters on the behavior of this
quantity). The secular variation of the solar wind, on time scales significantly longer
than the solar cycle, suggest that the heliosphere does not evolve strictly period-
ically and that monitoring of these variations is an essential element of any effort
aimed at a quantitative analysis of all kinds of heliospheric observations.
Selected interval of daily solar wind density averages

20
density [cm−3] 15
10
0
2055 2056 2057 2058 2059 2060 2061 2062
Carrington rotation number
Selected interval of daily solar wind speed averages

700
600
speed [km s−1]
500
400
300
200
2055 2056 2057 2058 2059 2060 2061 2062
Carrington rotation number
Figure 3.14: Daily averages of solar wind density adjusted to 1 AU (upper panel)
and speed (lower panel) calculated from hourly averages obtained from the OMNI-
2 database (King and Papitashvili 2005) for an example time interval covering
7 full Carrington rotation periods, presented to illustrate the complex structure
of solar wind parameter evolution at different time scales and the approximate
anticorrelation of density and speed. The interval shown covers 7 full Carrington
rotations in 2008 and is placed in the context of long-term solar wind evolution
shown in Fig. 3.8
CX rate of H from OMNI database

1. × 10−6
9. × 10−7
8. × 10−7
rate [s−1]
7. × 10−7
6. × 10−7
5. × 10−7
4. × 10−7
3. × 10−7
1985 1990 1995 2000 2005 2010
time [y]
Figure 3.15: Carrington rotation averages of the rate of charge exchange between
solar wind protons and neutral interstellar H atoms, calculated from monthly av-
erages of solar wind speed and adjusted density using Eq. 3.19. The thin vertical
lines mark the time interval expanded to daily time resolution in Fig. 3.14
Latitudinal Structure of the Solar Wind

and Its Evolution During the Solar Cycle
Shortly after the discovery of the solar wind, the question of whether or not it
is spherically symmetric appeared. The solar wind is currently known to feature
a latitudinal structure which varies with the solar activity cycle. While direct
observations of the solar wind in the ecliptic plane have been conducted since the
early 1960s, information on its latitudinal structure was available only from indirect
and mixed observations of the cometary ion tails (Brandt et al. 1972, 1975). The
situation changed when radio-astronomy observations of interplanetary scintillation
(IPS) and appropriate interpretation of observations of the Lyman-α helioglow
became available. Apart from the in situ measurements obtained from Ulysses,
these two techniques remain the only source of global, time-resolved information
on the solar wind structure.
The launch of the Ulysses spacecraft (Wenzel et al. 1989), the first and up to now
the only interplanetary probe to leave the ecliptic plane and sample interplanetary
space in the polar regions, improved our understanding of the 3D behavior of the
solar wind. Its measurements provided a very high resolution in latitude but a
poor resolution in time. The same latitudes were visited only a few times during
the ∼ 20-year mission at uneven time intervals. Hence the studies of solar wind
parameters as a function of time and heliolatitude are still a work in progress and
therefore are discussed in a separate section in this chapter.
Historical Perspective: Insight from Interplanetary

Scintillation
Interplanetary scintillation measurements involve radiotelescope observations of
remote compact radio sources (like quasars), searching for fluctuations of the signal.
The variations are caused by the diffraction of radio waves on electron density
fluctuations ∼ 200 km in diameter, occurring along the line of sight. Specifically,
the observable quantity is called the scintillation index, defined as a quotient of
the r.m.s. of the observed intensity fluctuation to the mean intensity of a source
(Manoharan 1993).
The scintillation signal is the sum of waves scattered along the line of sight to
the observed radio source. Most of the scattering occurs at the closest distances
to the Sun along the line of sight because the strength of the electron density fluc-
tuations rapidly decreases with solar distance. If the solar wind were spherically
symmetric, it would be possible to define a weighting function and the IPS “mid-
point” speed would be the spatial average of the solar wind speed centered at the
closest point along the line of sight (Coles and Maagoe 1972). But the solar wind
is not spherically symmetric and it features streams, which result in discrepancies
of the measured speed from the actual one (Houminer 1971). Owing to a corre-
lation between the fluctuations amplitude and the solar wind speed, it is possible
to approximately deduce the solar wind speed from careful analysis of registered
diffraction patterns in the observed radio source signal (Hewish et al. 1964).
Latitudinal structure and evolution of the solar wind 103
An improvement in the accuracy of measurements of the solar wind speed was

achieved from a better identification of the scintillation patterns along the lines of
sight when they are simultaneously observed by multiple stations displaced longi-
tudinally on the Earth surface, and by using the solar wind tomography technique
(Jackson et al. 1997, 1998; Kojima et al. 1998; Asai et al. 1998). Despite the high
sophistication of the tomography technique, its accuracy inevitably depends on
geometrical considerations (the telescopes are located in the northern hemisphere
of the Earth, whose orbit is tilted at an angle of 7.25◦ to the solar equator), on
the number of available observations, and on the fidelity of the adopted correlation
relationship between the solar wind speed and the density turbulence level.
While inferring the correlation between the solar wind speed and density fluc-
tuations was achieved early on for the equatorial solar wind (Harmon 1975), the
out-of-ecliptic IPS measurements could only be calibrated once in situ data from
Ulysses became available (Kojima et al. 2001). Ideally, such a calibration should be
repeated separately for each solar cycle because, as discussed earlier in this chapter,
the solar wind features secular changes.
Early measurements of the solar wind speed using the IPS technique brought
mixed conclusions: while Dennison and Hewish (1967) discovered an increase in
solar wind speed outside the ecliptic plane, Hewish and Symonds (1969) did not
find such an increase. Further observations, however, reported by Coles and Rickett
(1976), clearly showed that the solar wind is structured, with a band of slow speed
around the solar equator and a much faster wind near the poles.
An extensive program of IPS observations of the solar wind, initiated in the
1980s in the Solar-Terrestrial Environmental Laboratory at Nagoya University,
Japan (Kojima and Kakinuma 1990), resulted in a homogeneous dataset that spans
almost three solar cycles and enables studies of the evolution of the solar wind speed
profile with changes in solar activity (Kojima and Kakinuma 1987). Even before
the introduction of the Computer Assisted Tomography technique, they suggested
that the solar wind structure varies with solar activity, with the slow wind reaching
polar regions when the activity is high.
Supported and interpreted by the tomography technique (Hayashi et al. 2003),
IPS observations enable detailed studies of the structure of the solar wind with
varying solar activity conditions (Kojima et al. 1999, 2001, 2007; Ohmi et al. 2001,
2003; Fujiki et al. 2003a,b,c; Tokumaru et al. 2009, 2010). The solar wind speed
data obtained from these observations are discussed later in this chapter.
Historical Perspective: Insight from Heliospheric

Backscatter Glow
Observations of the Lyman-α heliospheric backscatter glow, carried out since
the beginning of the 1970s (Bertaux and Blamont 1971) have also been used as a
tool to discover the 3D structure of the solar wind and its evolution with the solar
cycle. The bi-modal structure of the solar wind, with the slow and dense wind in
an equatorial band and a rarefied, fast wind at the polar caps, results in a distinctly
structured charge exchange ionization rate.
The direction of inflow of neutral interstellar gas is very close to the solar equa-
tor. Thus, the inflowing atoms that approach the Sun close enough to significantly
contribute to the backscattered signal and whose orbits happen to be in a plane

close to the solar equator spend their entire travel through the heliosphere in the
region of increased probability of ionization. In contrast, those traveling in planes
inclined at greater angles to solar equator spend relatively little time in the equa-
torial band of the increased ionization rate or even do not get close to this region
at all. Consequently, more of them are able to survive the travel towards the Sun
(see Fig. 1 in Lallement et al. 1985a for an illustrative sketch). As a result, a region
of reduced density of neutral interstellar hydrogen gas is created close to the solar
equator. At higher heliolatitudes, the density of the gas at similar solar distances
is higher than that within the equatorial band.
This gas is illuminated by an approximately spherically symmetric solar Lyman-
α radiation, which is backscattered by resonance fluorescence. Since at equatorial
latitudes the density of the gas is reduced and the illuminating flux is almost
homogeneous in heliolatitude, the intensity of the backscattered radiation is lower
at equatorial latitudes than at the polar caps. This equatorial dimming of the
helioglow, referred to as the heliospheric groove, was observed in the 1970s and
1980s (Kumar and Broadfoot 1978, 1979; Lallement et al. 1985b, 1986; Lallement
and Stewart 1990) and correctly interpreted as due to the enhanced ionization level
at equatorial latitudes because of the anisotropy of the solar wind.
It is important to stress, however, that analysis of the heliospheric Lyman-α
backscatter glow is only able to yield the latitude structure of the total ionization
rate of neutral interstellar hydrogen. From the view point of density structure of
neutral interstellar hydrogen near the Sun, the nature of the ionization processes is
not important, only the results they produce, i.e. a decrease in the total density in
an equatorial latitude band. Hence, no differentiation between the charge exchange,
photoionization, and electron-impact ionization can be made based solely on the
heliospheric glow analysis.
Since, however, charge exchange is the dominant process and photoionization is
only slightly anisotropic in latitude, to first order the latitude variation in the total
rate can be regarded as a latitude variation in the charge exchange rate, which
is proportional to the latitude variation in the total solar wind flux, modulated
by the dependence of the charge exchange rate on solar wind speed (see Eqs. 3.14
and 3.17).
Lallement et al. (1985a) proposed to describe the rate of charge exchange be-
tween the solar wind protons and neutral interstellar H atoms as a function of
heliolatitude, φ, defined by the formula:

βCX (φ) = β0 1 − A sin2 φ . (3.37)

This is a two-parameter relation, where β0 corresponds to the rate at the equator
and A is a pole-to-equator amplitude, which can be fit from observations of the
helioglow. This formula approximately reproduced the limited observations of the
groove obtained in the early stages of the research. It is able to reproduce various
pole-to-equator contrasts in the ionization rate, as well as the situation when the
contrast virtually disappears and the solar wind becomes almost spherically sym-
metric. However, the profile of the ionization rate obtained from this formula has
a full width at half maximum ∼ 45◦ and is perfectly symmetric about the solar
equator. Consequently, it is not able to reproduce the north-south asymmetries in
the solar wind and the situations when the range of the slow solar wind significantly
differs from ∼ 45◦ .
Despite these deficiencies, Eq. 3.37 was successfully used by a number of authors
(e.g. Lallement et al. 1985a; Pryor et al. 1998, 2003) to qualitatively infer the
solar wind structure. The conclusions were similar to those obtained from the IPS
analysis: during solar minimum the solar wind is latitudinally structured, with a
band of enhanced flux at the equator and two polar caps of a rarefied and fast
wind. During solar maximum the ordered structure changes and the polar caps
become almost fully covered with the slow wind.
However, it became clear that the simple model given by Eq. 3.37 is not fully
adequate to describe reality and a need for more observations became evident.
Bertaux et al. (1995) proposed an experiment to study solar wind anisotropies
using the technique of analysis of the heliospheric Lyman-α backscatter glow, which
was implemented in the French/Finnish project SWAN9 onboard the Solar and
Heliospheric Observatory mission (SOHO).
Already shortly after the launch of SOHO it became evident that the early
conclusions on the evolution of the heliospheric groove, and thus the solar wind,
with the solar activity cycle were confirmed (Bertaux et al. 1996, 1997, 1999; Kyrölä
et al. 1998), but the formula used to describe the latitude profile of the ionization
rate needed modification. Thus in the latter work, Eq. 3.37 was modified to describe
separately the northern and southern hemispheres:

βCX (φ) = β0 Θ (φ) 1 − AN sin2 φ + Θ (−φ) 1 − AS sin2 φ + B (φ) , (3.38)
where Θ is the Heavyside step function, AN , AS are the separate anisotropy param-
eters for the northern and southern hemispheres, and B (φ) is used to narrow the
width of the equatorial band of enhanced solar wind flux. An even more sophisti-
cated approach was proposed by Summanen (1996), who suggested to approximate
the equatorial band in the latitudinal profile of the total ionization rate by:

βCX (φ, t) = β0 1 − A (t) sin2 (c φ) (3.39)
for heliolatitudes −40◦ ≤ φ ≤ 40◦ , where c = 9/4 limits the equatorial band to
±40◦ , and by:
2
t − P/2
βCX (φ, t) = β0 exp − (3.40)
0.2P
outside the ±40◦ equatorial band, where P is the solar cycle length and t, time.
In this formula, there was no north-south asymmetry allowed, but it was possible
to homogeneously reproduce the variations of the anisotropy parameters with the
solar activity cycle.
The north-south anisotropies in the ionization rate and their evolution with
solar activity were evident on one hand, and on the other hand the first fast latitude
scan by Ulysses (Phillips et al. 1995a) suggested that the profile of solar wind
parameters can be approximated by an equatorial plateau (with a “rough surface”
of the gusty slow wind) standing out from a flat “foot” of the fast polar wind.
9 For: Solar Wind ANisotropies.

Bzowski (2001a, 2003) and Bzowski et al. (2002) suggested to approximate it by

the formula:
βCX (φ, t) = (βCX,pol + δCX φ) + (βCX,eqtr (t) − βCX,pol ) (3.41)

N
2φ − φN (t) − φS (t)
× exp − ln 2 ,
φN (t) − φS (t)
where φ is heliographic latitude and N is a shape factor; βCX,pol is the average

ionization rate at the poles and the term (βCX,pol + δCX φ) describes the north-
south asymmetry of the polar ionization rates. The term (βSW,pol + δCX φ) +
(βSW,eqtr (t) − βCX,pol) for φ = 0 corresponds to the charge exchange rate at solar
N
N −φS
equator; and the term exp − ln 2 2φ−φ φN −φS describes the latitudinal depen-
dence of the ionization rate (see also Bzowski 2008). The shape of the central bulge
is controlled by the exponent N ; for N = 2 the shape is Gaussian, for N = 8 it is
close to rectangular.
The parameters in the model by Bzowski (2003) are the north and south bound-
aries of the equatorial slow wind band φN and φS , the equator/north pole and equa-
tor/south pole ratios of the ionization rate, and the polar north/south asymmetry
parameters. Thus, to obtain absolute values of the ionization rate, an independent
assessment of the ionization rate at the equator is needed.
Bzowski et al. (2003) used the theory developed by Bzowski (2003) and inter-
preted a carefully selected subset of SWAN observations. To maintain in the data
as much symmetry around the inflow axis as possible and simultaneously include
the full span of heliolatitudes, they chose observations taken within a week of the
passage of SWAN through the projection of inflow axis in the ecliptic plane, i.e. at
the beginning of June and December of each year, and they restricted the field of
view to a narrow strip going through the projections on the celestial sphere of the
solar equator and poles.
They eliminated the “searchlights” (i.e., reflections on sky of the point-like
active regions on the solar disk, traveling across the sky with the angular velocity
of the solar rotation) discovered by Bertaux et al. (2000). They also cleaned the
data from contamination by extraheliospheric “chaff” (Milky Way, stars etc.). To
eliminate possible bias from an imperfect absolute calibration, they normalized the
lightcurves to equatorial values of the helioglow intensity instead of attempting to
fit the absolute values.
In agreement with other studies, Bzowski et al. (2003) found that the ordered
structure of the solar wind, present during the solar minimum phase, disappears
with the increasing solar activity. The boundaries between the fast and slow wind
regions move polewards and ultimately at solar maximum the slow wind encom-
passes the entire space. The motion of the fast/slow wind boundaries in the north
and south hemispheres were found to be shifted in phase by approximately a year.
Bzowski et al. (2008) discovered that the areas of the polar fast wind regions
are linearly correlated with the areas of the polar holes observed by Harvey and
Recely (2002), which enabled them to calculate the variation in the boundaries
between the fast and slow solar wind φN , φS for the time span of the polar coronal
holes observations, i.e. from 1990 until 2002. Bzowski et al. (2008) also suggested
that the evolution of φN , φS can be approximated by:

φN,S (t) = φ0 + φ1 exp − cos3 (ωφ t) , (3.42)
where ωp was obtained as 2π/main period of the composite Lyman-α flux (cf
Fig. 3.1) and the free parameters φ0 and φ1 are fit separately for the northern
and southern hemispheres. The validity range of this approximation is limited to
the time interval from 1990 to 2002. Based on this model of ionization, it was
possible to infer the evolution of solar wind speed and density as a function of time
and latitude. This topic is covered later in this chapter.
The model by Bzowski et al. (2003) and Bzowski et al. (2008) was able to more
correctly reproduce the latitudinal span of the slow wind region and its evolu-
tion with solar activity, but was inadequate to correctly reproduce the boundaries
themselves. Observations from SWAN showed that especially during the transi-
tion phases of solar activity, the photometric latitudinal profiles of the groove are
complex and variable in time.
To address this problem, it turned out to be necessary to use an approach orig-
inally proposed by Summanen et al. (1993) and to model the ionization rate as
a multi-step function, with different levels in fixed—though arbitrarily selected—
latitudinal bands. Such an approach to the modeling of the ionization rate was
numerically implemented already by Lallement et al. (1985a), but these authors
filled the latitudinal “slots” with values obtained from the analytical models dis-
cussed earlier in this chapter.
In the refined approach, presented by Quémerais et al. (2006) and Lallement
et al. (2010), the ionization rates in the latitudinal bands are free parameters fit
to the maps of heliospheric backscatter glow, without any assumptions on the
relations between the neighboring bands. In contrast to the approach exercised by
Bzowski et al. (2003), Quémerais et al. (2006) and Lallement et al. (2010) used
all the data available, cleaned only by appropriate masking to cut off the known
extraheliospheric “chaff”.
As a result, a time series of the ionization rate profiles at a resolution of ∼ 10◦
in heliolatitude and ∼ 2 days in time was obtained from fitting the model to
the filtered full-sky maps and subsequent scaling to the in-ecliptic ionization rates
obtained from in situ measurements. An illustration of results of this analysis is
shown by Lallement et al. (2010). They support the general picture of the evolution
of the solar wind structure with solar activity, but they point out that the width
of the equatorial band of enhanced ionization rate was wider during the extended
solar minimum in 2005 through 2009 than during the former minimum.
On the other hand, inspection of the heliolatitude vs time map of the ionization
rate in Lallement et al. (2010) reveals some morphological features that are not
supported by Ulysses or IPS observations. These discrepancies were one of the
reasons to take a closer look at the calibration of the heliospheric FUV observations.
Historical Perspective: Ulysses Measurements

Ulysses, launched in October 1990, was the first spacecraft to traverse the polar
regions of the heliosphere and provide a unique view of the solar wind (Wenzel et al.
1989; Smith et al. 1991). After a cruise to Jupiter, carried out close to the ecliptic
Ulysses and Earth heliolatitude vs solar activity
80
60
heliolatitude [deg]
40
20
0
−20
−40
−60
−80
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
time [y]
Figure 3.16: Illustration of the heliolatitude track of Ulysses (red) and the Earth
(green) during the time span of Ulysses mission. The pale blue line is the F10.7
solar radio flux, superimposed to correlate variations in solar activity with Ulysses
heliolatitudes during its more than three orbits in a polar plane almost perpendic-
ular to the inflow direction of neutral interstellar gas. Adapted from Sokól et al.
(2012)
plane, it performed a Jupiter gravity-assist maneuver that cast it away from the
ecliptic plane on an elliptical nearly polar orbit. The orbit has aphelion ∼ 5.5 AU
and perihelion ∼ 1.4 AU and is nearly perpendicular both to the ecliptic plane and
solar equator. It is also nearly perpendicular to the inflow direction of interstellar
gas. The period of the orbit is about 6 years. The heliolatitude track of Ulysses is
shown in Fig. 3.16.
The spacecraft was launched at solar maximum and its radial cruise was com-
pleted just when solar activity was beginning to decrease. The first dive towards
the south solar pole was carried out during an interval of decreasing solar activity,
followed by the first so-called fast latitude scan, when the spacecraft coasted from
the south to the north solar pole, almost covering the full span of heliolatitudes
during just about a year at solar minimum activity conditions.
This fast latitude scan preceded the first observations of the helioglow by SWAN
by about a year. This provided an opportunity to calibrate the model of ionization
rates obtained from the analysis of the heliospheric Lyman-α glow. Ulysses contin-
ued on its polar orbit, performing its first full slow latitude scan during an interval
of increasing solar activity. The second fast latitude scan occurred in a totally
different solar cycle phase, namely during solar maximum. This scan also took
about a year and was performed during an interval of dynamically variable solar
wind structure. In contrast to the first slow scan, the subsequent slow latitude scan
occurred during decreasing solar activity. Afterwards, the spacecraft performed its
last fast latitude scan, again during minimum solar activity.
The geometry and timing of Ulysses trajectory produced a unique dataset of
direct in situ measurements of the solar wind plasma parameters, obtained from
SWOOPS (Bame et al. 1992) and SWICS (Gloeckler et al. 1992) experiments. The
discoveries and findings from the plasma measurements were presented in dozens
of papers (e.g. Phillips et al. 1995a,b; Marsden and Smith 1997, McComas et al.
1998a, McComas et al. 1999, 2000a,b, 2002a,b, 2003, 2006, 2008). An additional
benefit from this unique mission is the use of the Ulysses flight spare plasma instru-
ment with only minor modifications on the ACE mission (McComas et al. 1998a),
which facilitates intercalibrating the Ulysses measurements with the OMNI time
series.
The Ulysses solar wind dataset is unique and invaluable because it is the first
and only direct in situ measurement of the solar wind parameters outside the eclip-
tic plane. The evolution of solar wind speed, adjusted density, and adjusted flux
during the previous and current solar minima and during the previous maximum
are compiled in Fig. 3.17, where the parameter values are averaged over 10-degree
bins in heliolatitude. “Adjustment” throughout this text means scaling to 1 AU
assuming an average dropoff with heliocentric distance as 1/r 2 .
It can be seen that the heliolatitude structure during the two minima is basically
similar, featuring an equatorial enhancement in density with the associated reduc-
tion in velocity (the slow wind region), and that during solar maximum the slow
wind expands to all heliolatitudes (see also Fig. 3.18 and discussion of time scales in
the variation of the solar wind structure). However, the region of slow wind seems
to reach farther in heliolatitude during the last solar minimum than during the min-
imum of 1995, which, interestingly, is much less conspicuous in density. Thus the
findings from the Ulysses in situ measurements and SWAN remote-sensing obser-
vations by Lallement et al. (2010), discussed in the section “Historical Perspective:
Insight from Heliospheric Backscatter Glow”, qualitatively agree in this respect.
A striking feature seen in Fig. 3.17 is a strong reduction in flux with heliolati-
tude, reported by McComas et al. (2008) and Ebert et al. (2009). The reduction is
visible as a continuous trend from the 1995 minimum through the 2002 maximum
until present. Another interesting trend is a variation of ∼ 1 km s−1 deg−1 in the
fast polar solar wind, discovered by McComas et al. (2000b) and expanded upon
by Ebert et al. (2009).
Retrieval of Solar Wind Evolution: Introduction

In this section, we describe the history of our knowledge of the evolution of the
solar wind density and speed during the past solar cycle. Our goal is to provide
modelers a tool to develop a model of the neutral interstellar hydrogen distribution
(see Izmodenov et al. 2012, this volume) that could be used as background for
the inter-calibration of various sets of UV observations. This effort was recently
described in a research paper by Sokól et al. (2012).
Sokól et al. (2012) attempted to use all relevant datasets, paying special at-
tention to their absolute calibrations and possible biases. As shown later in this
section, this is still a work in progress. Since an essential part of information must
be drawn from a careful interpretation of the Lyman-α helioglow observations from
SWAN, which are subject to improved absolute calibration, the effort of construct-
ing a homogeneous set of solar wind parameters must be an iterative one.
In the following, we will present a construction of the procedure to retrieve
the variability of solar wind speed and density in time and heliographic latitude
using the available datasets, developed by Sokól et al. (2012). That construction
SW speed for Ulysses fast scans
800
speed [km s−1]
700
600
500
400
−80 −60 −40 −20 0 20 40 60 80
heliolatitude [deg]
SW density for Ulysses fast scans
10
8
[cm−3]
6
density
−80 −60 −40 −20 0 20 40 60 80

heliolatitude [deg]
SW flux for Ulysses fast scans
4.0 × 108
3.5 × 108
flux [cm−2 s−1 ]
3.0 × 108
2.5 × 108
2.0 × 108
1.5 × 108
1.0 × 108
−80 −60 −40 −20 0 20 40 60 80
heliolatitude [deg]
Figure 3.17: Solar wind speed (upper panel), adjusted density (middle panel), and
adjusted flux (lower panel) as a function of heliolatitude for the first (blue), sec-
ond (red) and third (green) Ulysses fast latitude scans, obtained from SWOOPS
(Bame et al. 1992). The parameters are averaged over 10-degree heliolatitude bins.
Adapted from Sokól et al. (2012)
SW speed during 1st fast scan / 1994.7 - 1995.58

800
700
speed [km s-1]
600
500
400
300
-80 -60 -40 -20 0 20 40 60 80
heliolatitude [deg]
SW speed during 2nd fast scan / 2000.9 - 2001.78

800
700
speed [km s-1]
600
500
400
300
-80 -60 -40 -20 0 20 40 60 80
heliolatitude [deg]
SW speed during 3rd fast scan / 2007.1 - 2008.03

800
700
speed [km s-1]
600
500
400
300
-80 -60 -40 -20 0 20 40 60 80
heliolatitude [deg]
Figure 3.18: Solar wind speed profiles from Ulysses measurements and IPS obser-
vations for the three fast Ulysses scans. Red: Ulysses, blue: IPS during the year of
beginning of a Ulysses fast scan, gray: IPS during the year of the end of Ulysses
fast scan. Top panel: the first fast scan during solar minimum, middle panel: the
second fast scan during solar maximum, bottom panel: the third fast scan during
minimum. Adapted from Sokól et al. (2012)
relies on the absolute calibration of the OMNI dataset both in speed and density
in the ecliptic plane. Out of ecliptic, the baseline is the absolute calibration of
Ulysses SWOOPS measurements and IPS. The prime source of information on
the heliolatitude structure of the solar wind are IPS observations, interpreted by
tomography modeling, that generally agree quite well with the Ulysses SWOOPS
in situ measurements out of the ecliptic and with the OMNI measurements in the
ecliptic. Up to now, no continuous measurements of solar wind density as a function
of ecliptic latitude have been available.
Tests revealed that an appropriate balance between the latitudinal resolution
of the coverage and the fidelity of the results is obtained at a subdivision of the
heliolatitudes’ range into 10-degree bins. Concerning the time resolution, the most
welcome would be Carrington rotation averages, identical with the resolution of
the photoionization rate and Lyman-α flux. Regrettably, such a high resolution
seems to be hard to achieve because (1) the time coverage in the data from IPS has
gaps that typically occur during ∼ 4 months at the beginning of each year, which
would induce an artificial 1-year periodicity in the data, and (2) the fast latitude
scans by Ulysses were about 12 months long, hence differentiating between time
and latitude effects in its measurements is challenging. Thus a reliable latitude
structure of the solar wind can only be obtained on a time scale of 1 year and this
is the time resolution of the model that Sokól et al. (2012) developed.
Concerning the global mapping of solar wind parameters from the ecliptic, it
has to be pointed out that the accuracy of measurements of solar wind parameters
decreases with latitude because of geometry. The polar values are the most un-
certain (and possibly biased) because the signal in the polar lines of sight is only
partly formed in the polar region of space, which is illustrated in Fig. 3.19.
In the following section, we will present a procedure to retrieve the solar wind
speed evolution in time and heliolatitude, plus two procedures for retrieval of solar
wind density. One, by Sokól et al. (2012), is based on a correlation between the solar
wind speed and density that was established from the three Ulysses fast latitude
scans and must be regarded as an interim solution, to be used until the other one,
based on the SWAN Lyman-α helioglow observations, will be available.
Latitude Profiles of Solar Wind Velocity from Interplanetary

Scintillation Observations
IPS observations carried out by the Solar-Terrestrial Environment Laboratory
(STEL) of Nagoya University (Japan) enable Sokól et al. (2012) to derive the lati-
tude structure of solar wind speed and its variations in time. They used data from
1990 to 2011—with a one year gap in 2010—obtained from 3 antennas (Toyokawa,
Fuji and Sugadaira) and from another antenna (Kiso) since 1994. The 4-antenna
system was operated until 2005, when the Toyokawa antenna was closed (Tokumaru
et al. 2010). Since then, the system has operated in a 3-antenna setup.
The IPS data from STEL are typically collected on a daily basis during 11
Carrington rotations per year: there is a break in winter because the antennas get
covered with snow. The IPS observations are line of sight integration of the solar
wind speed weighted by density turbulence and a Fresnel filter. Each day, 30–40
lines of sight for selected scintillating radio sources are observed.
signal from pole
LOS
Earth
Sun
Figure 3.19: Illustration of the geometry of line of sight (LOS) in a remote sensing
observation of the solar polar region. The observer is close to the Earth in the
ecliptic plane and aims its instrument (e.g., a radio telescope antenna or a Lyman-α
photometer) at a target so that the line of sight crosses a cone with a small opening
angle centered at the pole. The signal is collected from the full length of the line
of sight, but the contributions from various parts are different and depend on the
observations technique. In the case of IPS observations, the strongest contribution
to the signal is from the point nearest to the Sun along the line of sight, marked with
P, because the source function of the scintillation signal decreases with the square
of solar distance. In the case of helioglow observations, the maximum of the signal
comes from the so-called Maximum Emissivity Region, which is located within 1.5–
5 AU from the Sun, so it is important to carefully select the solar elongation of the
line of sight. Note that the angular area of the polar region is quite small, so with
an observations program that maps the entire sky only a small region in the maps
indeed includes the signal from polar regions. Adapted from Sokól et al. (2012)
The line of sight integration effect is deconvolved using the Computer Assisted
Tomography (CAT) method developed by the STEL group (Kojima et al. 1998,
2007). The LOS’s are projected on the source surface at 2.5 solar radii (R ), which
is used as a reference surface in time sequence tomography.
The heliolatitude coverage by IPS observations is not uniform and is strongly
correlated with the Sun’s position in the sky, which changes during the year. The
coverage is also constrained by the target distribution in the sky, with relatively
few of them near the solar poles. Additionally, the observations of the south pole
are of lower quality than those of the north pole because of the low elevation of
the Sun during winter in Japan. The original latitude coverage was improved by
adding the new antenna in 1994 and by optimization of the choice of the targets.
In the following, we briefly present the approach and results from Sokól et al.
(2012). For the analysis, they took the solar wind speed from 1990 to 2011, without
year 2010, when the number of observations was not sufficient to retrieve reliable
solar wind speeds. The solar wind speed observations were mapped at the source
surface on a grid of 11×360×180 records per year, which corresponds to a series of
Carrington rotations. The data were organized in heliolatitude from 89.5◦ North
to 89.5◦ South.
A comparison of the tomography-derived solar wind speed with the in situ
measurements by Ulysses performed by Sokól et al. (2012) showed that the accuracy
of the tomographic results depends on the number of IPS observations available
for a given rotation. The intervals with a small number of data points clearly tend
to underestimate the speed. Consequently, the Carrington rotations with the total
number of points less than 30 000 were removed from the data. Small numbers of
available observations typically happen at the beginning and at the end of the year
and at the edges of data gaps. The selection of data by the total number of points
per rotation constrained the coverage mainly to the summer and autumn months,
when all latitudes are well sampled.
The selected subset of data was split into years, and within each yearly subset
into 19 heliolatitude bins, equally spaced from −90◦ to 90◦ . The speeds averaged
over bins and over year for the latitudinal bins yield the yearly latitudinal profiles
of solar wind speed, shown in Figs. 3.20 and 3.21. They cover half of solar cycle 22
and the full solar cycle 23. In the analysis a two-step calculation was used.
First, Carrington rotation averaged values per bin were computed. Next, the
yearly averages were calculated from the monthly ones. It is worth noting that the
solar wind speed profiles for individual Carrington rotations during a given year
typically have very similar shapes to the related yearly profile, which suggests that
the latitude structure is stable during a year and changes only on a time scale
comparable with solar activity variations.
The results confirm that the solar wind speed is bimodal during solar minimum,
slow at latitudes close to the solar equator (and thus the ecliptic plane) and fast at
the poles. The latitude structure evolves with the solar activity cycle and becomes
flatter when the activity is increasing. The structure is approximately homogeneous
in latitude only during a short time interval during the peak of solar maximum,
when the solar wind is slow at all latitudes (see the panel for 2000 in Fig. 3.20).
Shortly after the activity maximum, the bimodal structure reappears and the fast
wind at the poles is observed again, but switchovers from the slow to fast wind at
the poles may still occur during the high activity period, as shown in the panels
for 2001 and 2003 in Figs. 3.20 and 3.21.
During the descending and ascending phases of solar activity there is a wide
band of slow solar wind that straddles the equator and extends to midlatitudes; the
fast wind is restricted to the polar caps and upper midlatitudes. At solar minimum,
the structure is sharp and stable during a few years straddling the turn of solar
cycles, with high speed at the poles and at midlatitudes and a rapid decrease at
the equator.
Thus, apart from short time intervals at solar maximum, the solar wind struc-
ture close the poles is almost flat, with a steady fast speed value typical of wide
polar coronal holes, in perfect agreement with the measurements from Ulysses
(Phillips et al. 1995b; McComas et al. 2000b, 2006).
1990.51 1991.53 1992.53

800 800 800
]
]
−1
−1
−1
700 700 700
speed [km s
speed [km s
speed [km s
600 600 600
500 500 500
400 400 400
300 300 300
−90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90
heliolatitude [deg] heliolatitude [deg] heliolatitude [deg]
1993.53 1994.67 1995.52
800 800 800
]
]
−1
−1
−1
700 700 700
speed [km s
speed [km s
speed [km s
600 600 600
500 500 500
400 400 400
300 300 300
−90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90
1996.56 1997.59 1998.57
800 800 800
]
]
−1
−1
−1
700 700 700
speed [km s
speed [km s
speed [km s
600 600 600
500 500 500
400 400 400
300 300 300
−90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90
1999.53 2000.55 2001.52
800 800 800
]
]
−1
−1
−1
700 700 700
speed [km s
speed [km s
speed [km s
600 600 600
500 500 500
400 400 400
300 300 300
−90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90
Figure 3.20: Heliolatitude profiles of yearly averaged solar wind speed for 1990–2001
obtained by Sokól et al. (2012) from the Computer Assisted Tomography analysis
of interplanetary scintillation observations (Tokumaru et al. 2010). Adapted from
Sokól et al. (2012)
To further verify the results obtained from the IPS analysis, we compared them
with the data from the three Ulysses fast latitude scans and the OMNI measure-
ments in the ecliptic plane. The Ulysses velocity profiles used for this comparison
were constructed from subsets of hourly averages available from the NSSDC, split
into identical heliolatitude bins as those used for the IPS data analysis and aver-
aged. They are shown in Fig. 3.18 as red lines. Since the acquisition of the Ulysses
profiles took one year each and the fast scans straddle the turn of the years, we
show the IPS results for the years straddling the fast latitude scans. They are
presented in blue and gray in Fig. 3.18.
The fast scans were performed at the perihelion half of the Ulysses elliptical
orbit, with the perihelion close to the solar equator plane. Hence, the angular
speed of its motion was highest close to the equator and traversing the 10-degree
bin took it less than one solar rotation period. The apparent bin-to-bin fluctuation
is an effect of incomplete Carrington longitude coverage of the bimodal solar wind,
with slow wind interleaved with fast wind streams.
Near the poles, the angular speed was slower and it took more than 1 Carrington
rotation to scan the 10-degree bin. Still, the longitudinal coverage was uneven.
2002.53 2003.61 2004.55

800 800 800
]
]
−1
−1
−1
700 700 700
speed [km s
speed [km s
speed [km s
600 600 600
500 500 500
400 400 400
300 300 300
−90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90
2005.53 2006.64 2007.51
800 800 800
]
]
−1
−1
−1
700 700 700
speed [km s
speed [km s
speed [km s
600 600 600
500 500 500
400 400 400
300 300 300
−90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90
2008.55 2009.48 2011.48
800 800 800
]
]
−1
−1
−1
700 700 700
speed [km s
speed [km s
speed [km s
600 600 600
500 500 500
400 400 400
300 300 300
−90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90
Figure 3.21: Heliolatitude profiles of yearly averaged solar wind speed for 2002–
2011 obtained by Sokól et al. (2012) from the Computer Assisted Tomography
analysis of interplanetary scintillation observations (Tokumaru et al. 2010). Note
that data for 2010 are missing because of a non-sufficient number of observations
available. Adapted from Sokól et al. (2012)
Thus, during solar maximum, when the gusty slow wind engulfed the whole space,
the “sawtooth effect” expanded into the full latitude span. By contrast, during the
low-activity scans the solar wind speed at high latitude was stable, which resulted
in the lack of the small-scale latitude variations in the Carrington rotation averages
at high latitudes, even though the issue of uneven longitudinal coverage persisted.
The solar wind speed profiles obtained from IPS and Ulysses observations are
very similar, but some systematic differences do exist. On one hand, it seems that
∼ 50 km s−1 is a typical difference between Ulysses and IPS values in the polar
regions, with the northern region usually in better agreement than the southern.
On the other hand, sometimes the agreement is almost perfect.
The difference between the blue and gray lines in the top and middle panels of
Fig. 3.18 is a measure of true variation of the latitudinal profile of solar wind speed
during one year. Ulysses was moving from south to north during the fast latitude
scans, so the south limb of the profile from Ulysses ought to be closer to the south
limb of the blue profile obtained from the IPS analysis, while the north limb of the
Ulysses profile should agree better with the north limb of the gray IPS profile.
Such behavior is observed in the second panel of Fig. 3.18, which corresponds
to the solar maximum in 2001. In our opinion, it is a very interesting observation
because (1) it shows how rapidly the latitude structure of the solar wind varies
during solar maximum, and (2) it confirms both Ulysses and IPS results regarding
solar wind structure.
Solar wind speed in the ecliptic plane

650
IPS
600 OMNI
Ulysses fast
speed [km s −1]
550 Ulysses slow
500
450
400
350
1990 1995 2000 2005 2010
time [y]
Figure 3.22: Yearly averages of solar wind speed from the IPS analysis (blue) and
in-situ measurements collected in the OMNI-2 database (green), compared with
the average solar wind speed measured by Ulysses during its passage within 5◦
from the solar equator plane during the three fast latitude scans (red diamonds)
and three slow scans (red triangles). Adapted from Sokól et al. (2012)
Originally, the interpretation of the speed profiles obtained from Ulysses was
not clear. It was pondered whether the north-hemisphere increase in the solar
wind speed was a long-standing feature of the solar wind or was just due to a time-
variability of the wind at the north pole. Similarly, it was pondered whether the
IPS analysis is able to reliably reproduce the solar wind profiles given the fact that
some of the profiles obtained approximately at the time of the fast scan seemed to
disagree with the in situ data.
The analysis by Sokól et al. (2012) suggests that the yearly-averaged velocity
profiles obtained from the tomography analysis of IPS observations agree with the
in-situ observations from Ulysses even at times when the solar wind is restructuring
rapidly at the peak of solar activity. The IPS data are in a very good agreement
with the OMNI data collected in the ecliptic plane (see Fig. 3.22). Up to 2004, the
agreement is almost perfect, afterwards small differences appear. The agreement is
better than with the in-ecliptic Ulysses measurements from the fast latitude scans.
This, in our opinion, is because the measurements of solar wind parameters in the
ecliptic plane obtained during the fast latitude scans from Ulysses are challenging
to directly compare with the OMNI and IPS measurements. The reason is that
Ulysses was passing through the ecliptic 10-degree latitude bin in a time equal to
about half of a Carrington rotation and thus a reliable longitude averaging of the
solar wind parameters could not be obtained. A detailed discussion of the non-
trivial issue of comparing the OMNI, Ulysses, and IPS solar wind speed close to
the ecliptic plane is provided by Sokól et al. (2012).
The IPS solar wind speed profiles provide a solar wind latitude structure that
can be adopted as an interim solution. They agree well both with the OMNI time
series in the ecliptic and with the Ulysses measurements out of the ecliptic for the
time intervals when they can be directly compared.
1990.51 1991.53 1992.53

800 800 800
speed [km s ]
speed [km s ]
speed [km s−1]

−1
−1
700 700 700
600 600 600
500 500 500
400 400 400
300 300 300
−90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90
1993.53 1994.67 1995.52
800 800 800
speed [km s ]
speed [km s ]
speed [km s−1]

−1
−1
700 700 700
600 600 600
500 500 500
400 400 400
300 300 300
−90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90
1996.56 1997.59 1998.57
800 800 800
speed [km s ]
speed [km s ]
speed [km s−1]

−1
−1
700 700 700

600 600 600
500 500 500
400 400 400
300 300 300
−90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90
1999.53 2000.55 2001.52
800 800 800
speed [km s ]
speed [km s−1]
speed [km s−1]

−1
700 700 700

600 600 600
500 500 500
400 400 400
300 300 300
−90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90
Figure 3.23: Latitude profiles of yearly-averaged solar wind speed for 1990–2001
obtained from the interim procedure described in the text. The red lines show the
speed profiles obtained from the Ulysses fast scans. The profile for 2010 is a linear
interpolation between the profiles for 2009 and 2011 (Sokól et al. 2012)
For further analysis, Sokól et al. (2012) smoothed the yearly speed profiles as-
suming that the variation of the solar wind speed at high latitude close to the poles
is linear and the variation outside the polar caps can be approximated by a series
of smoothly-transitioning parabolae. The resulting smoothed yearly heliolatitude
profiles of solar wind speeds are shown in Figs. 3.23 and 3.24. The gap in the ob-
servations in 2010 was filled by linear interpolation between the profiles from 2009
and 2011. The formula used to smooth the profiles along with the numerical values
of their parameters are listed in Sokól et al. (2012). These smoothed profiles will be
used in the remainder of this paper to obtain the solar wind density and flux profiles.
Latitude Structure of Solar Wind Density and Flux

The time- and latitude-dependence of the solar wind flux and density can be
obtained using two different methods. The first one (Sokól et al. 2012), which
we regard as an interim solution, relies on an approximate correlation between
the solar wind speed and density inferred from from fast latitude scan data taken
by Ulysses. The other, expected to be the ultimate one, will be based on future
analysis of the SWAN Lyman-α helioglow measurements, once completed. Here we
will only introduce the first method.
2002.53 2003.61 2004.55

800 800 800
speed km s 1
speed km s 1
speed km s 1
700 700 700
600 600 600
500 500 500
400 400 400
300 300 300
90 60 30 0 30 60 90 90 60 30 0 30 60 90 90 60 30 0 30 60 90
heliolatitude deg heliolatitude [deg] heliolatitude deg
2005.53 2006.64 2007.51
800 800 800

speed km s 1
speed km s 1
speed km s 1
700 700 700
600 600 600
500 500 500
400 400 400
300 300 300
90 60 30 0 30 60 90 90 60 30 0 30 60 90 90 60 30 0 30 60 90
heliolatitude deg heliolatitude deg heliolatitude deg
2008.55 2009.48 2010.5

800 800 800
speed km s 1
speed km s 1
speed km s 1
700 700 700
600 600 600
500 500 500
400 400 400
300 300 300
90 60 30 0 30 60 90 90 60 30 0 30 60 90 90 60 30 0 30 60 90
heliolatitude deg heliolatitude deg heliolatitude deg
2011.48
800
speed km s 1
700
600
500
400
300
90 60 30 0 30 60 90
heliolatitude deg
Figure 3.24: Latitude profiles of yearly-averaged solar wind speed for 2002–2011
obtained from the interim procedure described in the text. The red lines show the
speed profiles obtained from the Ulysses fast scans. The profile for 2010 is a linear
interpolation between the profiles for 2009 and 2011 (Sokól et al. 2012)
Solar Wind Density and Flux from Density-Speed Correlation

As discussed by Sokól et al. (2012), the solar wind speed and density seem to
be related to heliolatitude and thus correlated with each other, at least during
the solar minimum conditions and for the observations collected during the fast
latitude scans. The correlations seem to be slightly different between the first and
third latitude scans, as illustrated in Fig. 3.25:
nUlysses (v) = ascan + bscan v, (3.43)
where ascan and bscan are fit separately for the speed and density values averaged
over 10-degree bins using the ordinary least squares bisector method (Isobe et al.
1990) which allows for uncertainty in both ordinate and abscissa. The fitting is
done separately for the first and third latitude scans. For the first scan (blue line
and points in Fig. 3.25) we obtain afirst = 12.69 and bfirst = −0.01332. For the
third scan (green line and dots in Fig. 3.25), the correlation formula parameters
are athird = 10.01 and bthird = −0.01107. Thus, the slopes are almost identical and
the main difference between the two relations is in the y-intercept, which reflects
the overall secular decrease in solar wind density.
Density−speed correlation for Ulysses fast scans
8
−3]
6
density [cm
400 500 600 700 800

speed [km s −1]
Figure 3.25: Anticorrelation between the solar wind density and speed obtained
from Ulysses fast latitude scans. Blue corresponds to the first scan at solar min-
imum (see Fig. 3.16), red to the second scan (performed during solar maximum
conditions), and green to the third scan excecuted during solar minimum. The
dots represent speeds and densities averaged over the 10-degree heliolatitude bins,
the blue and green lines are the linear correlations specified in Eq. 3.43. The dotted
red line is the density—speed relation proposed for the transition interval close to
the solar maximum of 2002, calculated as the mean of the correlation relations
obtained from the first and third latitude scans. Adapted from Sokól et al. (2012)
The relation between density and speed for the second scan, which occurred
during solar maximum, does not seem to be linear, but in this case the spatial and
temporal effects seem to be convolved (as discussed earlier in this section). There-
fore, Sokól et al. (2012) proposed to use an arithmetic mean of the relations for the
first and third scans: asecond = (afirst + athird) /2 and bsecond = (bfirst + bthird) /2.
This relation is shown in Fig. 3.25 as the red broken line. A comparison of the
density values actually measured during the second latitude scan and calculated
from the correlation formula is shown in Fig. 3.26.
The interval of applicability of the latter formula is from ∼1998 until 2002. The
formula from the first scan is applicable to the interval before 1998 and the formula
from the third scan for the interval after 2002.
Sokól et al. (2012) calculated the interim yearly profiles of solar wind density as
a function of heliolatitude by applying Eq. 3.43 to the speed profiles presented in
the preceding subsection. Since this is an interim and very approximate solution,
which by its nature is not very accurate, the authors did not attempt to further
adjust it at the equator to the corresponding OMNI densities. The results are
shown in Figs. 3.27 and 3.28. The accuracy of the results in the polar regions is in
the range of 20–40 %.
With the density and speed profiles on hand, one can easily calculate the flux:
F (φj , ti ) = v (φj , ti ) n (φj , ti ) , (3.44)

Solar wind density for 2. fast scan

measured vs calculated
10
8
−3]
density [cm
−80 −60 −40 −20 0 20 40 60 80

heliolatitude [deg]
Figure 3.26: Comparison of solar wind density averaged over 10-degree heliolatitude
bins actually measured by Ulysses (red) and calculated from the correlation formula
in Eq. 3.43 (blue) (Sokol et al., Solar Physics, 2012)
dynamic pressure:
1 2
pdyn (φj , ti ) = mp n (φj , ti ) (v (φj , ti )) , (3.45)
2
and charge exchange rate (in the stationary H atom approximation; Eq. 3.14). They
are collectively shown in Fig. 3.29.
To obtain the solar wind parameters (speed, density, flux, dynamic pressure,
charge exchange rate and total ionization rate) at monthly resolution, Sokól et al.
(2012) replaced the equatorial bin directly with the Carrington rotation averaged
series from OMNI, linearly interpolated to halves of Carrington rotations. The
±10◦ bins were replaced with values linearly interpolated between the ±20 deg
bins and the equatorial bin.
The pole values were calculated from the the parabolic interpolation between
the ±70 and ±80◦ bins, because due to the problems discussed earlier in this chapter
direct measurements over the poles are not available. The remaining latitudinal
bins were linearly interpolated in time between the yearly profiles. As a result
of such a treatment, all available information on the equatorial bin of the solar
wind was utilized. Away from the equatorial bin, where such an information is
not available, there is a smooth transition into the latitude region where the low
time-resolution model is used.
Figure 3.29 shows contour maps of solar wind flux (upper), charge exchange
rate (middle), and dynamic pressure (lower) as a function of time (horizontal axis)
and heliolatitude (vertical axis). The magnitude of the quantity being shown is
indicated by color code and isocontours.
The latitudinal profile of a given quantity at a desired moment in time can be
retrieved by taking a vertical strip of the map from a location corresponding to the
1990.51 1991.53 1992.53

10 10 10
density [cm ]
density [cm−3]
density [cm−3]
−3
8 8 8
6 6 6
4 4 4
2 2 2
−90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90
1993.53 1994.67 1995.52
10 10 10
density [cm ]
density [cm−3]
density [cm−3]
−3
8 8 8
6 6 6
4 4 4
2 2 2
−90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90
1996.56 1997.59 1998.57
10 10 10
density [cm ]
density [cm−3]
density [cm−3]
−3
8 8 8
6 6 6
4 4 4
2 2 2
−90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90
1999.53 2000.55 2001.52
10 10 10
density [cm−3]
density [cm ]
density [cm ]
−3
−3
8 8 8
6 6 6
4 4 4
2 2 2
−90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90
Figure 3.27: Latitude profiles of yearly averaged solar wind density for 1990–2001
obtained from the interim procedure. The red lines show the Ulysses fast scan
profiles (Sokól et al. 2012)
desired time on the horizontal axis. The evolution of a given quantity in time at
a selected heliolatitude can be retrieved taking a horizontal strip for the latitude
indicated at the vertical axis. The magnitude of this quantity is color-coded and
the color code is given in the color bar next to the panels. The isocontours indicate
regions in (time vs heliolatitude) space where the quantity has a specific value.
As discussed by Sokól et al. (2012), the structure of solar wind flux and charge
exchange painted in Fig. 3.29 shows a clear variation with solar activity level. Dur-
ing the epochs of two solar minima presented in the figure, the flux and charge
exchange rate were almost spherically symmetric, as indicated by the light bands
spanning the whole latitude range in 1991 and in 2001.
By contrast, the flux and charge exchange structure during prolonged intervals
of lower activity is quite different. The flux and charge exchange rate feature clear
maxima at the equatorial latitude. One can observe some north-south asymmetry
in the latitudinal range of this equatorial enhancement, as well as a small (< 1 year)
shift in phase between the northern and southern hemispheres.
The absolute magnitude of the flux and charge exchange rate dropped after
the solar maximum of activity in 2001, which can be seen as a general dimming
of the maps in their right-hand halves. This weakening exists at all heliolatitudes.
2002.53 2003.61 2004.55

10 10 10
density [cm−3]
density [cm−3]
density [cm−3]
8 8 8
6 6 6
4 4 4
2 2 2
−90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90
2005.53 2006.64 2007.51
10 10 10
density [cm−3]
density [cm−3]
density [cm−3]
8 8 8
6 6 6
4 4 4
2 2 2
−90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90
2008.55 2009.48 2010.5
10 10 10
density [cm−3]
density [cm−3]
density [cm−3]
8 8 8
6 6 6
4 4 4
2 2 2
−90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90 −90 −60 −30 0 30 60 90
2011.48
10
density [cm−3]
8
6
4
2
−90 −60 −30 0 30 60 90
heliolatitude [deg]
Figure 3.28: Latitude profiles of yearly averaged solar wind density for 2002–2011
obtained from the interim procedure. The red lines show the Ulysses fast scan
profiles (Sokól et al. 2012)
Interestingly, the behavior of the solar wind ram pressure is different. This quantity
is much less structured in heliolatitude than radiation pressure and charge exchange
are. This is true at all times, practically regardless of the solar activity level. The
solar wind ram pressure also weakened, similarl to the flux and charge exchange
rate. However, the drop in pressure began earlier than the drop in the other
quantities, namely just before the last solar maximum, i.e., about 1998.
Summing up this section, we have the structure of the solar wind speed from
the smoothed IPS profiles (Figs. 3.23 and 3.24). The density structure is obtained
(Figs. 3.27 and 3.28) from the density-speed correlation from Ulysses (Eq. 3.43,
Fig. 3.25). From these, one calculates the solar wind flux, dynamic pressure, and—
from Eq. 3.19—charge exchange rate between solar wind protons and neutral H
atoms (Fig. 3.29).
Results: solar wind flux [108 cm−2 s−1]

4.5
80 4.25
60 4.
3.75
40
heliolatitude [deg]
3.5
20 3.25
3.
0
2.75
−20 2.5
−40 2.25
2.
−60 1.75
−80 1.5
1.25
1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 1.
time [y]
Results: charge exchange rate [10−7s−1]
8.
80
7.5
60 7.
40 6.5
heliolatitude [deg]
6.
20 5.5
0 5.
4.5
−20
4.
−40 3.5
−60 3.
2.5
−80
2.
1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 1.5
time [y]
Results: dynamic pressure [nPa]
2.
80 1.9
1.8
60
1.7
40
heliolatitude [deg]
1.6
1.5
20
1.4
0 1.3
1.2
−20 1.1
−40 1.
0.9
−60 0.8
−80 0.7
0.6
1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 0.5
time [y]
Figure 3.29: Contour maps of solar wind flux (in 108 cm−2 s−1 ), charge exchange
(in 10−7 s−1 ) and dynamic pressure (in nPa) shown as a function of time (horizontal
axis) and heliolatitude (vertical axis). The magnitudes are color-coded, with the
color code shown in the color bars accompanying the panels (Sokól et al. 2012)
Outlook: Latitudinal Structure of Solar Wind Flux and Density

from IPS and Lyman-α Helioglow Observations
(A Sketch of the Method)
In this section we propose a method to obtain the solar wind density as a
function of heliolatitude by combining information from IPS measurements and
from analysis of photometric observations of the Lyman-α backscatter glow from
SWAN. We assume that the total ionization rate from SWAN is the sum of the
charge exchange and photoionization rates of H, with electron impact ionization
neglected (because βel βCX for the maximum emissivity region (MER) distance
and beyond). Thus the heliolatitude- and time-dependent charge exchange rate
can be calculated as
βCX (φj , ti ) = βtot (φj , ti ) − βHph (0, ti ) , (3.46)
where βtot (φj , ti ) is obtained from the inversion of a SWAN map and the photoion-
ization rate βHph (0, ti ), here assumed to be spherically symmetric, is from one of
the formulae specified in Eqs. 3.23, 3.24, or 3.25.
Since the total ionization rate data from SWAN for the equator does not always
agree with the rate derived from direct in situ measurements, we calculate the
contrasts κβ,SWAN (φj , ti ) of the SWAN-derived ionization rates as a function of
heliolatitude using:
βSWAN (φj , ti )
κβ,SWAN (φj , ti ) = , (3.47)
βSWAN (0, ti )
where βSWAN (φj , ti ) is an ionization rate at φj th heliolatitude and βSWAN (0, ti )
is an ionization rate at the equator. In this way we can build yearly contrasts for
the 19 heliolatitude bins and for the years since 1996. Multiplying the contrasts
κβ,SWAN (φj , ti ) with the monthly averages of the total equatorial ionization rates
(see Fig. 3.7), we obtain latitudinal profiles of the total ionization rate that agree
with the baseline values for the equator.
We can calculate the absolute charge exchange ionization rate for all heliolat-
itudes by subtracting the photoionization rate from the total rate from Eq. 3.46.
Considering the formula for charge exchange rate given in Eq. 3.14, with the solar
wind velocity profile obtained from IPS observations v (φj , ti ) we can now calculate
the total solar wind flux as a function of heliolatitude in the following way:
F (φj , ti ) = βCX (φj , ti ) /σCX (v (φj , ti )) . (3.48)
With the flux and solar wind speed profile on hand, it is straightforward to calculate
the profile of density:
n (φj , ti ) = F (φj , ti ) /v (φj , ti ) . (3.49)
In that way we can obtain a model of the evolution of the solar wind speed and
density as a function of time and heliolatitude. This model can be used to calculate
the evolution of the ionization rate for H atoms traveling at arbitrary speeds and
can be used as an input to global models of the heliosphere to calculate the flux
and dynamic pressure of the solar wind. This calculation will be performed once
the intercalibration of the heliospheric EUV measurements has been completed and
the final inversion of the SWAN photometric observations for the ionization rate
profiles is performed.
It is worthwhile to mention here that the proposed method involves a time-
latency because the inflowing hydrogen gas needs some time to adapt its density
distribution to the changing ionization rate and radiation pressure. This effect was
discussed by Bzowski et al. (2002), who showed that the latency is a function of the
offset angle of the line of sight from the upwind direction. Its magnitude depends
on the velocity and temperature of the gas, but it is almost constant on a level
between 0.5 and 1.1 years for the offset angles between 0◦ and 90◦ . For higher offset
angles, between 90◦ and 140◦ , it increases by a factor of 1.5–2. This suggests that
a time dependent model of the distribution of neutral interstellar hydrogen in the
inner heliosphere should be used for modeling.
Summary
Intercalibration of heliospheric UV and EUV measurements requires a common
basis of heliospheric ionization processes. The ionization processes affect the dis-
tribution of neutral interstellar gas in the heliosphere and thus influence both the
spectrum and intensity distribution of the heliospheric backscatter glow.
In this chapter we presented a review of the solar factors affecting the distri-
bution of neutral interstellar gas in the heliosphere. We discussed the radiation
pressure, solar EUV ionizing radiation, and the solar wind parameters. We re-
viewed the history of measurements of these factors and developed a model of
time and heliolatitude evolution of solar wind speed and density based on data
available from in situ measurements of the solar wind parameters, from remote
sensing interplanetary scintillation observations of the solar wind speed structure,
and from correlation between the solar wind density and speed, inferred from in
situ measurements.
The results of this model are used as input in the global heliospheric models,
discussed by Izmodenov et al. (2012, this volume). The results of the global he-
liospheric modeling can in turn be used to fine tune the absolute calibration of
heliospheric EUV measurements and, in a new iteration, to further refine the solar
wind evolution model.
The refined model will cover at least the two recent solar cycles and will thus
provide a common homogeneous basis for interpretation of the present and past
heliospheric experiments. It will also be used to interpret the observations of En-
ergetic Neutral Atoms by the Interstellar Boundary Explorer (IBEX) (McComas
et al. 2009a,b) and hopefully other past and future heliospheric measurements.
Acknowledgements
M.B. and J.S. are obliged to Marty Snow for sharing his insight into the realm of EUV mea-
surements. The authors acknowledge the use of NASA/GSFC’s Space Physics Data Facility’s
ftp service for Ulysses/SWOOPS and TIMED/SEE data, SOHO/CELIAS/SEM (http://
www.usc.edu/dept/space science/semdatafolder/long/daily avg/), and OMNI2 data col-
lection (ftp://nssdcftp.gsfc.nasa.gov/spacecraft data/omni/). The F10.7 solar radio flux
Bibliography 127
was provided by the NOAA and Pentincton Solar Radio Monitoring Programme operated
jointly by the National Research Council and the Canadian Space Agency (ftp://ftp.ngdc.
noaa.gov/STP/SOLAR DATA/SOLAR RADIO/FLUX/Penticton Adjusted/ and ftp://
ftp.geolab.nrcan.gc.ca/data/solar flux/daily flux values/). The composite Lyman-α flux
and MgIIc/w were obtained from LASP, accessed through the LISIRD Web page at
(http://lasp.colorado.edu/lisird/lya/). The SOLAR2000 Research Grade historical ir-
radiances are provided courtesy of W. Kent Tobiska and SpaceWx.com. These historical
irradiances have been developed with funding from the NASA UARS, TIMED, and SOHO
missions.
M.B. and J.S. were supported by the Polish Ministry for Science and Higher Education
grants NS-1260-11-09 and N-N203-513-038. P.B. acknowledges support from NASA SR&T
Grant NX09AW32G. Contributions from D.M. were supported by NASA’s IBEX Explorer
mission. Support from the International Space Science Institute in Bern, Switzerland, is
gratefully acknowledged.
Bibliography
P.O. Amblard, S. Moussaoui, T. Dudok de Wit, J. Aboudarham, M. Kretzschmar,
J. Lilensten, F. Auchère, The euv sun as the superposition of elementary suns. Astron.
Astrophys. 487, L13–L16 (2008). doi:10.1051/0004-6361:200809588
G. Artzner, J.C. Vial, P. Lemaire, P. Gouttebroze, J. Leibacher, Simultaneous time-resolved
observations of the H L-alpha Mg K 2795 A, and Ca K solar lines. Astrophys. J. Lett.
224, L83–L85 (1978). doi:10.1086/182765
K. Asai, M. Kojima, M. Tokumaru, A. Yokobe, B.V. Jackson, P.L. Hick, P.K. Manoharan,
Heliospheric tomography using interplanetary scintillation observations. iii - correlation
between speed and electron density fluctuations in the solar wind. J. Geophys. Res. 103,
1991–2001 (1998). doi:10.1029/97JA02750
J.R. Asbridge, S.J. Bame, W.C. Feldman, M.D. Montgomery, Helium and hydro-
gen velocity differences in the solar wind. J. Geophys. Res. 81, 2719–2727 (1976).
doi:10.1029/JA08li016p02719
F. Auchère, Effect of the H I Ly α chromospheric flux anisotropy on the total inten-
sity of the resonantly scattered coronal radiation. Astrophys. J. 622, 737–743 (2005).
doi:10.1086/427903
F. Auchère, J.W. Cook, J.S. Newmark, D.R. McMullin, R. von Steiger, M. Witte,
Model of the all-sky He II 30.4 nm solar flux. Adv. Space Res. 35, 388–392 (2005a).
doi:10.1016/j.asr.2005.02.036
F. Auchère, D.R. McMullin, J.W. Cook et al., A model for solar euv flux helium photoion-
ization throughout the 3-dimensional heliosphere, in Proceedings of the Solar Wind 11
/ SOHO 16 “Connecting the Sun and Heliosphere” Conference, Whistler, Canada, June
2005. S.P. ESA-592, ed. by B. Fleck, T.H. Zurbuchen, H. Lacoste (2005b), pp. 327–329
S.J. Bame, E.W. Hones Jr., S.-L. Akasofu, M.D. Montgomery, J.R. Asbridge, Geomagnetic
storm particles in the high-latitude magnetotail. J. Geophys. Res. 76, 7566–7583 (1971).
doi: 10.1029/JA076i031p07566
S.J. Bame, J.R. Asbridge, H.E. Felthauser, J.P. Glore, H.L. Hawk, J. Chavez, ISEE-C solar
wind plasma experiment. IEEE Trans. Geosci. Electron. 16, 160–162 (1978a)
S.J. Bame, J.R. Asbridge, H.E. Felthauser, J.P. Glore, G. Paschmann, P. Hemmerich,
K. Lehmann, H. Rosenbauer, ISEE-1 and ISEE-2 fast plasma experiment and the ISEE-1
solar wind experiment. IEEE Trans. Geosci. Electron. 16, 216–220 (1978b)
S.J. Bame, D.J. McComas, B.L. Barraclough, J.L. Phillips, K.J. Sofaly, J.C. Chavez,
B.E. Goldstein, R.K. Sakurai, The Ulysses solar wind plasma experiment. Astron. Astro-
phys. Supp. 92, 237–265 (1992)
V.B. Baranov, Kinetic and hydrodynamic approaches in space plasma, in The Physics of
the Heliospheric Boundaries, ed. by V.V. Ismodenov, R. Kallenbach, Vol. SR-005 of ISSI
Science Report, pp. 1–26, ESA Publications Division, EXTEC ISBN 1608-280X (2006a)
V.B. Baranov, Early concepts of the heliospheric interface: plasma, in The Physics of the
Heliospheric Boundaries, ed. by V.V. Ismodenov, R. Kallenbach, vol. SR-005 of ISSI
Science Report, pp. 27–44 (2006b)
medium—numerical solution of self-consistent problem. J. Geophys. Res. 98, 15157–15163
(1993). doi:10.1029/93JA01171
V.B. Baranov, M.G. Lebedev, Y.G. Malama, The influence of the interface between the
heliosphere and local interstellar medium on the penetration of H atoms to the solar
system. Astrophys. J. 375, 347–351 (1991). doi:10.1086/170194
C.F. Barnett, H.T. Hunter, M.I. Kirkpatrick, I. Alvarez, C. Cisneros, R.A. Phaneuf, Atomic
data for fusion, in Volume 1: Collisions of H, H2 , He, and Li Atoms and Ions with Atoms
and Moleucles. vol. ORNL-6086/V1 (Oak Ridge National Laboratories, Oak Ridge, Ten-
nessee, 1990)
J.-L. Bertaux, J.E. Blamont, Evidence for a source of an extraterrestrial hydrogen Lyman
alpha emission. Astron. Astrophys. 11, 200–217 (1971)
J.-L. Bertaux, E. Kyrölä, E. Quémerais, R. Pellinen, R. Lallement, W. Schmidt, M. Berthé,
E. Dimarellis, J.P. Goutail, C. Taulemasse, C. Bernard, G. Leppelmeier, T. Summanen,
H. Hannula, H. Huomo, V. Kehlä, S. Korpela, K. Leppälä, E. Strömmer, J. Torsti, K.
Viherkanto, J.-F. Hochedez, G. Chretiennot, R. Peyroux, T. Holzer, SWAN: a study
of solar wind anisotropies on SOHO with Lyman alpha sky mapping. Sol. Phys. 162,
403–439 (1995). doi:10.1007/BF00733435
J.-L. Bertaux, E. Quémerais, R. Lallement, Observations of a sky Lyman α groove related
to enhanced solar wind mass flux in the neutral sheet. Geophys. Res. Lett. 23, 3675–3678
(1996). doi:10.1029/96GL03475
J.-L. Bertaux, E. Quémerais, R. Lallement, E. Kyrölä, W. Schmidt, T. Summanen,
T. Mäkinen, The first 1.5 years of observation from SWAN Lyman-alpha solar wind map-
per on SOHO, in Fifth SOHO Workshop: The Corona and Solar Wind near Minimum
Activity, Oslo, Norway, June 1997, ed. by A. Wilson. S.P. ESA-404, p. 29 (1997)
J.-L. Bertaux, E. Kyrölä, E. Quémerais, R. Lallement, W. Schmidt, J. Costa, T. Mäkinen,
SWAN observations of the solar wind latitude distribution and its evolution since launch.
Space Sci. Rev. 87, 129–132 (1999). doi:10.1023/A:1005178402842
J.-L. Bertaux, E. Quémerais, R. Lallement, E. Lamassoure, W. Schmidt, E. Kyrölä, Mon-
itoring solar activity on the far side of the sun from sky reflected Lyman α radiation.
Geophys. Res. Lett. 27, 1331–1334 (2000). doi:10.1029/1999GL003722
S. Bleszyński, S. Grzȩdzielski, D. Ruciński, J. Jakimiec, Expected fluxes of about 1 keV neu-
tral H atoms in interplanetary space—comparison with the uv background and possibility
of detection. Planet. Space Sci. 40, 1525–1533 (1992). doi:10.1016/0032-0633(92)90049-T
P. Bochsler, M. Bzowski, L. Didkovsky, H. Kucharek, J.M. Sokól, T.N. Woods, Ionization
rates (preliminary), (2012) in preparation
A. Bonetti, G. Moreno, S. Cantarano, A. Egidi, R. Marconero, F. Palutan, G. Pizella, Solar
wind observations with satellite ESRO HEOS-1 in December 1969. Nuovo Cimento B
Series 46, 307–323 (1969). doi:10.1007/BF02711013
R.M. Bonnet, P. Lemaire, J.C. Vial, G. Artzner, P. Gouttebroze, A. Jouchoux,
A. Vidal-Madjar, J.W. Leibacher, A. Skumanich, The LPSP instrument on OSO 8.
ii—in-flight performance and preliminary results. Astrophys. J. 221, 1032–1053 (1978).
doi:10.1086/156109
J.C. Brandt, R.G. Roosen, R.S. Harrington, Interplanetary gas. xvii. an astrometric deter-
mination of solar wind velocities from orientations of ionic comet tails. Astrophys. J. 177,
277–284 (1972). doi:10.1086/151706
J.C. Brandt, R.S. Harrington, R.G. Roosen, Interplanetary gas. xx. does the radial solar
wind speed increase with latitude. Astrophys. J. 196, 877–878 (1975). doi:10.1086/153478
M. Brasken, E. Kyrölä, Resonance scattering of Lyman alpha from interstellar hydrogen.
J.S. Bridge, A. Egidi, A.J. Lazarus, E. Lyon, L. Jacobson, Preliminary results of plasma mea-
surements on IMP-A, in Space Research. V:969–978, ed. by D.G. King-Hele, P. Muller,
G. Righini (North Holland, Amsterdam, 1965)
Bibliography 129
M. Bzowksi, Time dependent radiation pressure and time dependent 2d ionisation rate for
heliospheric modelling, in The Outer Heliosphere: The Next Frontiers, Cospar Colloquia
Series, vol. 11, ed. by K. Scherer, H. Fichtner, H.-J. Fahr, E. Marsch (Pergamon Press,
Amsterdam, 2001a), pp. 69–72
M. Bzowski, A model of charge exchange of interstellar hydrogen on a time-dependent, 2d
solar wind. Space Sci. Rev. 97, 379–383 (2001b). doi:10.1023/A:1011814125384
M. Bzowski, Response of the groove in heliospheric Lyman-α glow to latitude-dependent ion-
ization rate. Astron. Astrophys. 408, 1155–1164 (2003). doi:10.1051/0004-6361:20031023
M. Bzowski, Survival probability and energy modification of hydrogen energetic neutral
atoms on their way from the termination shock to earth orbit. Astron. Astrophys. 488,
1057–1068 (2008). doi:10.1051/0004-6361:200809393
M. Bzowski, D. Ruciński, Solar cycle modulation of the interstellar hydrogen density distri-
bution in the heliosphere. Space Sci. Rev. 72, 467–470 (1995a). doi:10.1007/BF00768821
M. Bzowski, D. Ruciński, Variability of the neutral hydrogen density distribution due
to solar cycle related effects. Adv. Space Res. 16, 131–134 (1995b). doi:10.1016/0273-
1177(95)00325-9
M. Bzowski, D. Ruciński, Neutral solar wind evolution during solar cycle, in Solar Wind
Eight, ed. by D. Winterhalter, J.T. Gosling, S.R. Habbal, W.S. Kurth, M. Neugebauer.
AIP Conference Proceedings, vol. 382 (American Institute of Physics, Woodbury, New
York, 1996), pp. 650–654. doi:10.1063/1.51452
M. Bzowski, S. Tarnopolski, Neutral atom transport from the termination shock to
1 au, in Physics of the Inner Heliosheath, ed. by J. Heerikhuisen, V. Florinski,
G.P. Zank, N.V. Pogorelov. AIP Conference Series, vol. 858, pp. 251–256 (2006).
doi:10.1063/1.2359335
M. Bzowski, H.-J. Fahr, D. Ruciński, H. Scherer, Variation of bulk velocity and temperature
anisotropy of neutral heliospheric hydrogen during the solar cycle. Astron. Astrophys.
326, 396–411 (1997)
M. Bzwoski, T. Summanen, D. Ruciński, E. Kyrölä, Response of interplanetary glow to
global variations of hydrogen ionization rate and solar Lyman-α flux. J. Geophys. Res.
107, ssh2-1 (2002). doi:10.1029/2001JA000141
M. Bzowski, T. Mäkinen, E. Kyrölä, T. Summanen, E. Quémerais, Latitudinal structure
and north-south asymmetry of the solar wind from Lyman-α remote sensing by SWAN.
Astron. Astrophys. 408, 1165–1177 (2003). doi:10.1051/0004-6361:20031022
M. Bzowski, E. Möbius, S. Tarnopolski, V. Izmodenov, G. Gloeckler, Density of neutral
interstellar hydrogen at the termination shock from Ulysses pickup ion observations.
Astron. Astrophys. 491, 7–19 (2008). doi:10.1051/0004-6361:20078810
M. Bzowski, E. Möbius, S. Tarnopolski, V. Izomdenov, G. Gloeckler, Neutral H density at
the termination shock: a consolidation of recent results. Space Sci. Rev. 143, 177–190
(2009)
S. Chabrillat, G. Kockarts, Simple parameterization of the absorption of the solar Lyman-
alpha line. Geophys. Res. Lett. 24, 2659–2662 (1997)
W.A. Coles, S. Maagoe, Solar-wind velocity from IPS observations. J. Geophys. Res. 77,
5622–5624 (1972). doi: 10.1029/JA077i028p05622
W.A. Coles, B.J. Rickett, IPS observations of the solar wind speed out of the ecliptic. J.
Geophys. Res. 81, 4797–4799 (1976)
J.W. Cook, G.E. Brueckner, M.E. van Hoosier, Variability of the solar flux in the far ultra-
violet 1175–2100 Å. J. Geophys. Res. 85, 2257–2268 (1980)
J.W. Cook, R.R. Meier, G.E. Brueckner, M.E. van Hoosier, Latitudinal anisotropy of the
solar far ultraviolet flux—effect on the Lyman alpha sky background. Astron. Astrophys.
97, 394–397 (1981)
A.E. Covington, Micro-wave solar noise observations during the partial eclipse of November
23, 1946. Nature 159, 405–406 (1947). doi:10.1038/159405a0
J.M.A. Danby, J.L. Camm, Statistical dynamics and accretion. Monthly Not. Royal Astron.
Soc. 117, 150 (1957)
G. de Toma, Evolution of coronal holes and implications for high-speed solar wind during the
minimum between cycles 23 and 24. Solar Phys. 274, 195–217 (2011). doi:10.1007/s11207-
010-9677-2
P.A. Dennison, A. Hewish, The solar wind outside the plane of the ecliptic. Nature 213,
343–346 (1967). doi:10.1038/213343a0
T. Dudok de Wit, J. Lilensten, J. Aboudarham, P.-O. Amblard, M. Kretzschmar, Retrieving
the solar euv spectrum from a reduced set of spectral lines. Ann. Geophys. 23, 3055 (2005)
T. Dudok de Wit, M. Kretzschmar, J. Aboudarham, P.-O. Amblard, F. Auchère, J. Lilen-
sten, Which solar euv indices are best for reconstructing the solar euv irradiance? Adv.
Space Res. 42, 903–911 (2008). doi:10.1016/j.asr.2007.04.019
T. Dudok de Wit, M. Kretzschmar, J. Lilensten, T. Woods, Finding the best proxies for the
solar uv irradiance. Geophys. Res. Lett. 36, L10 (2009). doi:10.1029/2009/GL037825
R.W. Ebert, D.J. McComas, H.A. Elliott, R.J. Forsyth, J.T. Gosling, Bulk properites
of the slow and fast solar wind and interplanetary coronal mass ejections measured
by Ulysses: three polar orbits of observations. J. Geophys. Res. 114, A1109 (2009).
doi:10.1029/2008JA013631
C. Emerich, P. Lemaire, J.-C. Vial, W. Curdt, U. Schüle, K. Wilhelm, A new relation be-
tween the central spectral solar H I Lyman α irradiance and the line irradiance measured
by sumer/SOHO during cycle 23. Icarus 178, 429–433 (2005)
H.-J. Fahr, Non-thermal solar wind heating by supra-thermal ions. Solar Phys. 30, 193–206
(1973)
H.-J. Fahr, Change of interstellar gas parameters in stellar wind dominated atmospheres:
solar case. Astron. Astrophys. 66, 103–117 (1978)
H.-J. Fahr, Interstellar hydrogen subject to a net repulsive solar force field. Astron.
Astrophys. 77, 101–109 (1979)
H.-J. Fahr, The 3d heliosphere: three decades of growing knowledge. Adv. Space Res. 32,
3–13 (2004)
H.-J. Fahr, D. Ruciński, Neutral interestellar gas atoms reducing the solar wind number
and fractionally neutralizing the solar wind. Astron. Astrophys. 350, 1071–1078 (1999)
H.-J. Fahr, D. Ruciński, Modification of properties and dynamics of distant solar wind
due to its interaction with neutral interstellar gas. Space Sci. Rev. 97, 407–412 (2001).
doi:10.1023/A:1011874311272
H.-J. Fahr, D. Ruciński, Heliospheric pick-up ions influencing thermodynamics and dynam-
ics of the distant solar wind. Nonlinear Proc. Geophys. 9, 377–386 (2002)
H.-J. Fahr, K. Scherer Perturbations of the solar wind flow by radial and latitudinal pick-up
ion pressure gradients. Ann. Geophys. 22, 2229–2238 (2004)
H.-J. Fahr, H. Fichtner, K. Scherer, Theoretical aspects of energetic neutral atoms as mes-
sengers from distant plasma sites with emphasis on the heliosphere. Rev. Geophys. 45,
RG4003 (2007). doi:10.1029/2006RG000214
W.C. Feldman, J.R. Asbridge, S.J. Bame, M.D. Montgomery, Double ion streams in the
solar wind. J. Geophys. Res. 78, 2017–2027 (1973). doi:10.1029/JA078i013p02017
W.L. Fite, A.C.S. Smith, R.F. Stebbins, Charge transfer in collisions involving symmetric
and asymmetric resonance. Proc. R. Soc. London Ser. A 268, 527 (1962)
L. Floyd, D.K. Prinz, P.C. Crane, L.C. Herring, Solar uv irradiance variation during cycles
22 and 23. Adv. Space Res. 29, 1957–1962 (2002)
L. Floyd, G. Rottman, M. Deland, J. Pap, 11 years of solar uv irradiance measurements from
UARS, in Solar Variability as an Input to the Earth’s Environment, ed. by A. Wilson.
ESA SP-535, pp. 195–203 (2003)
L. Floyd, J. Newmark, J. Cook, L. Herring, D. McMullin, Solar euv and uv spec-
tral irradiances and solar indices. J. Atmos. Sol. Terr. Phys. 67, 3–15 (2005).
doi:10.1016/j.jastp.2004.07.013
P.C. Frisch, M. Bzowski, E. Grün, V. Izmodenov, H. Krüger, J.L. Linsky, D.J. McComas,
E. Möbius, S. Redfield, N. Schwadron, R. Shelton, J.D. Slavin, B.E. Wood, The galactic
environment of the sun: interstellar material inside and outside of the heliosphere. Space
Sci. Rev. 146, 235–273 (2009). doi:10.1007/s11214-009-9502-0
Bibliography 131
P.C. Frisch, S. Redfield, J.D. Slavin, The interstellar medium surrounding the sun. Ann.
Rev. Astron. Astrophys. 49, 237–279 (2011). doi:10.1146/annurev-astro-081710-102613
K. Fujiki, M. Kojima, M. Tokumaru, T. Ohmi, A. Yokobe, K. Hayashi, Solar cycle de-
pendence of high-latitude solar wind, in Solar Wind Ten, ed. by M. Velli, R. Bruno,
F. Malara, B. Bucci. American Institute of Physics Conference Series, vol. 679 (American
Institute of Physics, Woodbury, New York, 2003a), pp. 141–143. doi:10.1063/1.1618561
K. Fujiki, M. Kojima, M. Tokumaru, T. Ohmi, A. Yokobe, K. Hayashi, D.J. McComas,
H.A. Elliott, Solar wind velocity structure around the solar maximum observed by in-
terplanetary scintillation, in Solar Wind Ten, ed. by M. Velli, R. Bruno, F. Malara, B.
Bucci. American Institute of Physics Conference Series, vol 679 (American Institute of
Physics, Woodbury, New York, 2003b) pp. 226–229. doi:10.1063/1.1618583
K. Fujiki, M. Kojima, M. Tokumaru, T. Ohmi, A. Yokobe, K. Hayashi, D.J. McComas,
H.A. Elliott, How did the solar wind structure change around the solar maximum?
from interplanetary scintillation observation. Ann. Geophys. 21, 1257–1261 (2003c).
doi:10.5194/angeo-21-1257-2003
G. Gloeckler, J. Geiss, Heliospheric and interstellar phenomena deduced from pickup ion
observations. Space Sci. Rev. 97, 169–181 (2001)
G. Gloeckler, J. Geiss, H. Balsiger, P. Bedini, J.C. Cain, J. Fisher, L.A. Fisk, A.B. Galvin, F.
Gliem, D.C. Hamilton, The solar wind ion composition spectrometer. Astron. Astrophys.
Supp. 92, 267–289 (1992)
G. Gloeckler, J. Geiss, H. Balsinger, L.A. Fisk, A.B. Galvin, F.M. Ipavich, K.W. Ogilvie,
R. von Steiger, B. Wilken, Detection of interstellar pickup hydrogen in the solar system.
Science 261, 70–73 (1993)
G. Gloeckler, E. Möbius, J. Geiss, M. Bzowksi, S. Chalov, H.-J. Fahr, D.R. McMullin,
H. Noda, M. Oka, D. Ruciński, R. Skoug, T. Terasawa, R. von Steiger, A. Yamazaki, T.
Zurbuchen, Observations of the helium focusing cone with pickup ions. Astron. Astrophys.
426, 845–854 (2004)
K. Gringauz, V. Bezrukih, V. Ozerov, R. Ribchinsky, A study of the interplanetary ionized
gas, high-energy electrons and corpuscular radiation from the sun by means of hte three
electrode trap for charged particles on the second soviet cosmic rocket. Sov. Phys. Doklady
5, 361 (1960)
M.A. Gruntman, Neutral solar wind properties: advance warning of major geomagnetic
storms. J. Geophys. Res. 99, 19213–19227 (1994)
J.K. Harmon, Scintillation studies of density microstructure in the solar wind plasma. Dis-
sertation, University of California, San Diego, 1975
K.L. Harvey, F. Recely, Polar coronal holes during cycles 22 and 23. Solar Phys. 211, 31–52
(2002)
K. Hayashi, M. Kojima, M. Tokumaru, K. Fujiki, MHD tomography using interplanetary
scintillation measurement. J. Geophys. Res. 108, 1102 (2003). doi:10.1029/2002JA009567
D.F. Heath, B.M. Schlesinger, The Mg 280-nm doublet as a monitor of changes in solar ultra-
violet irradiance. J. Geophys. Res. 91, 8672–8682 (1986). doi:10.1029/JD091iD08p08672
A. Hewish, M.D. Symonds, Radio investigation of the solar plasma. Planet. Space Sci. 17,
313 (1969). doi:10.1016/0032-0633(69)90064-6
A. Hewish, P.F. Scott, D. Wills, Interplanetary scintillation of small diameter radio sources.
Nature 203, 1214–1217 (1964). doi:10.1038/2031214a0
H.E. Hinteregger, K. Fukui, B.R. Gilson, Observational, reference and model data on
solar euv, from measurements on AE-E. Geophys. Res. Lett. 8, 1147–1150 (1981).
doi:10.1029/GL008i011p01147
Z. Houminer, Radio source scintillation—evidence of plasma streams corotating about the
sun. Nature 231, 165 (1971)
D. Hovestadt, M. Hilchenbach, A. Bürgi, B. Klecker, P. Laeverenz, M. Scholer,
H. Grünwaldt, W.I. Axford, S. Livi, E. Marsch, B. Wilken, H.P. Winterhoff, F.M. Ipavich,
P. Bedini, M.A. Coplan, A.B. Galvin, G. Gloeckler, P. Bochsler, H. Balsiger, J. Fischer,
J. Geiss, R. Kallenbach, P. Wurz, K.-U. Reiche, F. Gliem, D.L. Judge, H.S. Ogawa, K.C.
Hsieh, E. Möbius, M.A. Lee, G.G. Managadze, M.I. Verigin, M. Neugebauer, CELIAS—
charge, element and isotope analysis system for SOHO. Solar Phys. 162, 441–481 (1995).
doi:10.1007/BF00733436
A.J. Hundhausen, J.R. Asbridge, S.J.B.H.E. Gilbert, I.B. Strong, Vela 3 satellite observa-
tions of solar wind ions. J. Geophys. Res. 72, 1979 (1967). doi:10.1029/JZ072i007p01979
T. Isobe, E.D. Feigelson, M.G. Akritas, G.J. Babu, Linear regression in astronomy. Astro-
phys. J. 364, 104–113 (1990). doi:10.1086/169390
K. Issautier, Diagnostics of the solar wind plasma, in Turbulence in Space Plasmas, ed.
by P. Cargill, L. Vlahos. Lecture Notes in Physics, vol. 778 (Springer, Berlin, 2009), pp.
223–246
K. Issautier, N. Meyer-Vernet, M. Moncuquet, S. Hoang, Solar wind radial and latitudi-
nal structure—electron density and core temperature from Ulysses thermal noise spec-
troscopy. J. Geophys. Res. 103, 1969–1979 (1998)
K. Issautier, R.M. Skoug, J.T. Gosling, S.P. Gary, D.J. McComas, Solar wind plasma pa-
rameters on Ulysses: detailed comparison between the urap and swoops experiments. J.
Geophys. Res. 106, 15665–15676 (2001). doi:10.1029/2000JA000412
V.V. Izmodenov, V.B. Baranov, Modern multi-component models of the heliospheric inter-
face, in The Physics of the Heliospheric Boundaries, ed. by V.V. Izmodenov, R. Kallen-
bach. ISSI Scientific Report Series, SR-005, pp. 67–136 (2006)
V.V. Izmodenov, Y.G. Malama, A.P. Kalinin, M. Gruntman, R. Lallement, I.P. Rodionova,
Hot neutral H in the heliosphere: elastic H-H, H-p collisions. Astrophys. Space Sci. 274,
71–76 (2000). doi:10.1023/A:1026531519864
V.V. Izmodenov, D.B. Alexashov, S.V. Chalov, O.A. Katushkina, Y.G. Malama,
E.A. Provornikova, Kinetic-gasdynamic modeling of the heliospheric interface: global
structure, interstellar atoms and heliospheric enas. Space Sci. Rev. 146, 329–351 (2009).
doi:10.1007/s11214-009-9528-3
V.V. Izmodenov, O.A. Katushkina, E. Quémerais, M. Bzowski, Distribution of interstellar
H atoms in the heliosphere and backscattered solar Lyman-α, in Cross-Calibration of Far
uv Spectra of Solar System Objects and the Heliosphere, ed. by E. Quémerais, M. Snow,
R.M. Bonnet. ISSI Scientific Report Series, SR-013 (2013) (this volume)
B.V. Jackson, P.L. Hick, M. Kojima, A. Yokobe, Heliospheric tomography using interplan-
etary scintillation observations. Adv. Space Res. 20, 23–26 (1997). doi:10.1016/S0273-
1177(97)00474-2
B.V. Jackson, P.L. Hick, M. Kojima, A. Yokobe, Heliospheric tomography using interplan-
etary scintillation observations. i. combined nagoya and cambridge data. J. Geophys. Res.
103, 12049–12067 (1998)
D.L. Judge, D.R. McMullin, H.S. Ogawa, D. Hovestadt, B. Klecker, M. Hilchenbach,
E. Möbius, L.R. Canfield, R.E. Vest, R. Watts, C. Tarrio, M. Kuehne, P. Wurz, First solar
euv irradiances obtained from SOHO by the CELIAS/SEM. Solar Phys. 177, 161–173
(1998)
J.C. Kasper, Solar wind plasma: kinetic properties and micro-instabilities. Dissertation,
Massachusetts Institute of Technology, Cambridge, 2002
J.C. Kasper, A.J. Lazarus, J.T. Steinberg, K.W. Ogilvie, A. Szabo, Physics-based tests to
identify the accuracy of solar wind ion measurements: a case study with the wind faraday
cups. J. Geophys. Res. 111, A03105 (2006). doi:10.1029/2005JA011442
J.C. Kasper, M.L. Stevens, K.E. Korreck, B.A. Maruca, K.K. Kiefer, N.A. Schwadron, S.T.
Lepri, Evolution of the relationships between helium abundance, minor ion charge state,
and solar wind speed over the solar cycle. Astrophys. J. 745, 162 (2012). doi:10.1088/0004-
637X/745/2/162
O.A. Katushkina, V.V. Izmodenov, Effect of the heliospheric interface on the distribution
of interstellar hydrogen atom inside the heliosphere. Astron. Lett. 36, 297–306 (2010).
doi:10.1134/S1063773710040080
J.H. King, N.E. Papitashvili, Solar wind spatial scales in and comparisons of hourly wind
and acd plasma and magnetic field data. J. Geophys. Res. 110, 2104–2111 (2005).
doi:10.1029/2004JA010649
D. Kiselman, T. Pereira, B. Gustafsson, M. Asplund, J. Meléndez, K. Langhans, Is the solar
spectrum latitude dependent? an investigation with SST/TRIPPEL. Astron. Astrophys.
535, A18 (2011)
Bibliography 133
M. Kojima, T. Kakinuma, Solar cycle evolution of solar wind speed structure between 1973
and 1985 observed with the interplanetary scintillation method. J. Geophys. Res. 92,
7269–7279 (1987). doi:10.1029/JA092iA07p07269
M. Kojima, T. Kakinuma, Solar cycle dependence of global distribution of solar wind speed.
Space Sci. Rev. 53, 173–222 (1990). doi:10.1007/BF00212754
M. Kojima, M. Tokumaru, H. Watanabe, A. Yokobe, K. Asai, B.V. Jackson, P.L. Hick,
Heliospheric tomography using interplanetary scintillation observations. 2. latitude and
heliocentric distance dependence of solar wind structure at 0.1-1 au. J. Geophys. Res.
103, 1981–1989 (1998)
M. Kojima, K. Fujiki, T. Ohmi, M. Tokumaru, A. Yokobe, K. Hakamada, The highest solar
wind velocity in a polar region estimated from IPS tomography analysis. Space Sci. Rev.
87, 237–239 (1999). doi:10.1023/A:1005108820106
M. Kojima, K. Fujiki, T. Ohmi, M. Tokumaru, A. Yokobe, K. Hakamada, Latitudinal veloc-
ity structures up to the solar poles estimated from interplanetary scintillation tomography
analysis. J. Geophys. Res. 106, 15677–15686 (2001)
M. Kojima, M. Tokumaru, K. Fujiki, K. Hayashi, B.V. Jackson, IPS tomographic observa-
tions of 3d solar wind structure. Astron. Astrophys. Trans. 26, 467–476 (2007)
M. Kretzschmar, J. Lilensten, J. Aboudarham, Retrieving the solar euv spectral
irradiance from the observation of 6 lines. Adv. Space Res. 37, 341–346 (2006).
doi:10.1016/j.asr.2005.02.029
S. Kumar, A.L. Broadfoot, Evidence from mariner 10 of solar wind flux depletion at high
ecliptic latitudes. Astron. Astrophys. 69, L5-L8 (1978).
S. Kumar, A.L. Broadfoot, Signatures of solar wind latitudinal structure in interplanetary
Lyman-α emissions: mariner 10 observations. Astrophys. J. 228, 302–311 (1979)
E. Kyrölä, T. Summanen, P. Råback, Solar cycle and interplanetary hydrogen. Astron.
Astrophys. 288, 299–314 (1994)
E. Kyrölä, T. Summanen, T. Mäkinen, E. Quémerais, J.-L. Bertaux, R. Lallement, J. Costa,
Preliminary retrieval of solar wind anisotropies / SOHO observations. J. Geophys. Res.
103, 14523–14538 (1998)
R. Lallement, A.I. Stewart, Out-of-ecliptic lyman-alpha observations with Pioneer-Venus:
solar wind anisotropy degree in 1986. Astron. Astrophys. 227, 600–608 (1990)
R. Lallement, J.-L. Bertaux, F. Dalaudier, Interplanetary lyman α spectral profiles and
intensities for both repulsive and attractive solar force fields: predicted absorption pattern
by a hydrogen cell. Astron. Astrophys. 150, 21–32 (1985a)
R. Lallement, J.-L. Bertaux, V.G. Kurt, Solar wind decrease at high heliographic lati-
tudes detected from prognoz interplanetary lyman alpha mapping. J. Geophys. Res. 90,
1413–1420 (1985b)
R. Lallement, T.E. Holzer, R.H. Munro, Solar wind expansion in a polar coronal hole: infer-
ences from coronal white light and interplanetary lyman alpha observations. J. Geophys.
Res. 91, 6751–6759 (1986)
R. Lallement, E. Quémerais, P. Lamy, J.L. Bertaux, S. Ferron, W. Schmidt, The solar
wind as seen by SOHO/SWAN since 1996: comparison with SOHO/LASCO C2 coronal
densities. In Proceedings of SOHO 23 Workshop, ed. by S. Cranmer, T. Hoeksma, J.
Kohl. ASP Conference Series, vol. 428 (2010), pp. 253–258
A.J. Lazarus, K. Paularena, A comparison of solar wind parameters from experiments on
the IMP 8 and Wind spacecraft. In Measurement Techniques in Space Plasmas, ed. by
E. Borovsky, F. Pfaff, T. Young. AGU Geophysical Monograph Series, vol. 102 (1998),
pp. 85–90
G. Le Chat, K. Issautier, N. Meyer-Vernet, I. Zouganelis, M. Moncuquet, S. Hoang, Quasi-
thermal noise spectroscopy: preliminary comparison between kappa and sum of two
Maxwellian distributions, in Twelfth International Solar Wind Conference, vol. 1216,
pp. 316–319 (2010). doi:10.1063/1.3395864
G. Le Chat, K. Issautier, N. Meyer-Vernet, S. Hoang, Large-scale variation of solar wind
electron properties from quasi-thermal noise spectroscopy: Ulysses measurements. Solar
Phys. 271, 141–148 (2011). doi:10.1007/s11207-011-9797-3
J.L. Lean, H.P. Warren, J.T. Mariska, J. Bishop, A new model of solar euv irradiance
variability 2. comparisons with empirical models and observations and implications for
space weather. J. Geophys. Res. 108, 1059 (2003). doi:10.1029/2001JA009238
J.L. Lean, T.N. Woods, F.G. Eparvier, R.R. Meier, D.J. Strickland, J.T. Correira,
J.S. Evans, Solar extreme ultraviolet irradiance: present, past, and future. J. Geophys.
Res. 116, A01102 (2011). doi:10.1029/2010JA015901
M.A. Lee, H.J. Fahr, H. Kucharek, E. Möbius, C. Prested, N.A. Schwadron, P. Wu,
Physical processes in the outer heliosphere. Space Sci. Rev. 146, 275–294 (2009).
doi:10.1007/s11214-009-9522-9
P. Lemaire, J. Charra, A. Jouchoux, A. Vidal-Madjar, G.E. Artzner, J.C. Vial, R.M. Bonnet,
A. Skumanich, Calibrated full disk solar H I Lyman-alpha and Lyman-beta profiles.
Astrophys. J. Lett. 223, L55–L58 (1978). doi:10.1086/182727
P. Lemaire, C. Emerich, W. Curdt, U. Schühle, K. Wilhelm, Solar HI Lyman α full disk
profile obtained with the SUMER/SOHO spectrometer. Astron. Astrophys. 334, 1095–
1098 (1998)
P.L. Lemaire, C. Emerich, J.-C. Vial, W. Curdt, U. Schühle, K. Wilhelm, Variation of the
full sun hydrogen lyman α and β profiles with the activity cycle, in ESSP A-508: From
solar min to max: half a solar cycle with SOHO, 2002, pp. 219–222
P. Lemaire, C. Emerich, J.-C. Vial, W. Curdt, U. Schühle, K. Wilhelm, Variation of the full
sun hydrogen Lyman profiles through solar cycle 23. Adv. Space Res. 35, 384–387 (2005)
P.C. Liewer, B.E. Goldstein, N. Omidi, Hybrid simulations of the effects of interstellar pickup
hydrogen on the solar wind termination shock. J. Geophys. Res. 981, 15211–15220 (1993).
doi:10.1029/93JA01172
B.G. Lindsay, R.F. Stebbings, Charge transfer cross sections for energetic neutral atom data
analysis. J. Geophys. Res. 110, A12213 (2005). doi:10.1029/2005JA011298
W. Lotz, An empirical formula for the electron-impact ionization cross-section. Zeitschrift
f. Phys. 206, 205–211 (1967a)
W. Lotz, Electron-impact ionization cross-sections and ionization rate coefficients for atoms
and ions. Ap. J. Suppl. 14, 207–238 (1967b)
E.F. Lyon, H.S. Bridge, J.H. Binsack, Explorer 35 plasma measurements in the vicinity of
the moon. J. Geophys. Res. 72, 6113–6117 (1967). doi:10.1029/JZ072i023p06113
E.F. Lyon, A. Egidi, G. Pizella, H.S. Bridge, J.S. Binsack, R. Baker, R. Butler, Plasma
measurements on Explorer 33 (I) interplanetary region. Space Research, VIII, 99 (1968)
L.J. Maher, B.A. Tinsley, Atomic hydrogen escape rate due to charge exchange with hot
plasmaspheric ions. J. Geophys. Res. 82, 689–695 (1977)
M. Maksimovic, V. Pierrard, P. Riley, Ulysses distributions fitted with Kappa functions.
Geophys. Res. Lett. 24, 1151–1154 (1997). doi:10.1029/97GL00992
M. Maksimovic, S.P. Gary, R.M. Skoug, Solar wind electron suprathermal strength and
temperature gradients: Ulysses observations. J. Geophys. Res. 105, 18337–18350 (2000)
M. Maksimovic, I. Zouganelis, J.-Y. Chaufray, K. Issautier, E.E. Scime, J.E. Littleton, E.
Marsch, D.J. McComas, C. Salem, R.P. Lin, H. Elliott, Radial evolution of the electron
distribution functions in the fast solar wind between 0.3 and 1.5 AU. J. Geophys. Res.
110, A9104 (2005). doi:10.1029/2005JA011119
Y. Malama, V.V. Izmodenov, S.V. Chalov, Modeling of the heliospheric interface: multi-
component nature of the heliospheric plasma. Astron. Astrophys. 445, 693–701 (2006)
P.K. Manoharan, Three-dimensional structure of the solar wind: Variation of density with
the solar cycle. Sol. Phys. 148, 153–167 (1993). doi:10.1007/BF00675541
R.G. Marsden, E.J. Smith, Ulysses: a summary of the first high-latitude survey. Adv. Space
Res. 19, (6)825–(6)834 (1997)
D.J. McComas, S. J. Bame, P. Barker, W. C. Feldman, J. L. Phillips, P. Riley, J.W. Griffee,
Solar wind electron proton alpha monitor (SWEPAM) for the Advanced Composition
Explorer. Space Sci. Rev. 86, 563–612 (1998a)
D.J. McComas, S.J. Bame, B.L. Barraclough, W.C. Feldman, W.C. Funsten, J.T. Gosling,
P. Riley, R. Skoug, Ulysses’ return to the slow solar wind. Geophys. Res. Lett. 25(1), 1-4
(1998)
Bibliography 135
D.J. McComas, H.O. Funsten, J.T. Gosling, W.R. Pryor, Ulysses measurements of variations
in the solar wind – interstellar hydrogen charge exchange rate. Geophys. Res. Lett. 26,
2701–2704 (1999)
D.J. McComas, B.L. Barraclough, H.O. Funsten, J.T. Gosling, E. Santiago-Muñoz, B.E.
Goldstein, M. Neugebauer, P. Riley, A. Balogh, Solar wind observations over Ulysses first
full polar orbit. J. Geophys. Res. 105, 10419–10433 (2000a)
D.J. McComas, J.T. Gosling, R.M. Skoug, Ulysses observations of the irregularly structured
mid-latitude solar wind during the approach to solar maximum. Geophys. Res. Lett. 27,
2437–2440 (2000b)
D.J. McComas, H.A. Elliot, R. von Steiger, Solar wind from high-latitude coronal holes at
solar maximum. Geophys. Res. Lett. 29, 1314 (2002a). doi:10.1029/2001GL013940
D.J. McComas, H.A. Elliott, J.T. Gosling, D.B. Reisenfeld, R.M. Skoug, B.E. Goldstein, M.
Neugebauer, A. Balogh, Ulysses second fast-latitude scan: Complexity near solar maxi-
mum and the reformation of polar coronal holes. Geophys. Res. Lett. 29, 1290 (2002b).
doi:10.1029/2001GL014164
D.J. McComas, H.A. Elliot, N.A. Schwadron, J.T. Gosling, R.M. Skoug, B.E. Goldstein,
The three-dimensional solar wind around solar maximum, Geophys. Res. Lett. 30, 24–1,
(2003). doi:10.1029/2003GL017136
D.J. McComas, F. Allegrini, L. Bartolone, P. Bochsler, M. Bzowski, M. Collier, H. Fahr, H.
Fichtner, P. Frisch, H. Funsten, S. Fuselier, G. Gloeckler, M. Gruntman, V. Izmodenov, P.
Knappenberger, M. Lee, S. Livi, D. Mitchell, E. Möbius, T. Moore, S. Pope, D. Reisenfeld,
E. Roelof, H. Runge, J. Scherrer, N. Schwadron, R. Tyler, M. Wieser, M. Witte, P. Wurz,
G. Zank, The Interstellar Boundary Explorer (IBEX): Update at the end of phase B, in
Physics of the Inner Heliosheath, ed. by J. Heerikhuisen, V. Florinski, G.P. Zank, N.V.
Pogorelov. American Institute of Physics Conference Series, vol. 858 (American Institute
of Physics, Woodbury, New York, 2006), pp. 241–250
D.J. McComas, R.W. Ebert, H.A. Elliot, B.E. Goldstein, J.T. Gosling, N.A. Schwadron,
R.M. Skoug, Weaker solar wind from the polar coronal holes and the whole sun, Geophys.
Res. Lett. 35, L18103 (2008). doi:10.1029/2008GL034896
D.J. McComas, F. Allegrini, P. Bochsler, M. Bzowski, E.R. Christian, G.B. Crew, R. De-
Majistre, H. Fahr, H. Fichtner, P.C. Frisch, H.O. Funsten, S.A. Fuselier, G. Gloeckler,
M. Gruntman, J. Heerikhuisen, V. Izmodenov, P. Janzen, P. Knappenberger, S. Krimigis,
H. Kucharek, M. Lee, G. Livadiotis, S. Livi, R.J. MacDowall, D. Mitchell, E. Möbius,
T. Moore, N.V. Pogorelov, D. Reisenfeld, E. Roelof, L. Saul, N.A. Schwadron, P.W.
Valek, R. Vanderspek, P. Wurz, G.P. Zank, Global observations of the interstellar inter-
action from the Interstellar Boundary Explorer (IBEX). Science 326, 959–962 (2009a).
doi:10.1126/science.1180906
D.J. McComas, F. Allegrini, P. Bochsler, M. Bzowski, M. Collier, H. Fahr, H. Fichtner,
P. Frisch, H.O. Funsten, S.A. Fuselier, G. Gloeckler, M. Gruntman, V. Izmodenov, P.
Knappenberger, M. Lee, S. Livi, D. Mitchell, E. Möbius, T. Moore, S. Pope, D. Reisen-
feld, E. Roelof, J. Scherrer, N. Schwadron, R. Tyler, M. Wieser, M. Witte, P. Wurz,
G. Zank, IBEX - Interstellar Boundary Explorer, Space Sci. Rev. 146, 11–33 (2009b).
doi:10.1007/s11214-009-9499-4
E. Möbius, D. Hovestadt, B. Klecker, M. Scholer, G. Gloeckler, Direct observation of He+
pick-up ions of interstellar origin in the solar wind. Nature 318, 426–429 (1985)
E. Möbius, B. Klecker, D. Hovestadt, M. Scholer, Interaction of interstellar pick-up ions
with the solar wind. Astrophys. Space Sci. 144, 487–505 (1988)
M. Neugebauer, Initial deceleration of solar wind positive ions in the earth’s bow shock. J.
Geophys. Res. 75, 717–733 (1970)
M. Neugebauer, C.W. Snyder, Solar plasma experiment. Science 138, 1095–1097 (1962).
doi:10.1029/JA075i004p00717
H.S. Ogawa, C.Y.R. Wu, P. Gangopadhyay, D.L. Judge, Solar photoionization as a loss
mechanism of neutral interstellar hydrogen in interplanetary space. J. Geophys. Res.
100, 3455–3462 (1995)
K.W. Ogilvie, L.F. Burlaga, T.D. Wilkerson, Plasma observations on Explorer 34.
J. Geophys. Res. 73, 6809–6824 (1968). doi:10.1029/JA073i021p06809
T. Ohmi, M. Kojima, A. Yokobe, M. Tokumaru, K. Fujiki, K. Hakamada, Polar low-speed

solar wind at the solar activity maximum. J. Geophys. Res. 106, 24923–24936 (2001).
doi:10.1029/2001JA900094
T. Ohmi, M. Kojima, K. Fujiki, M. Tokumaru, K. Hayashi, K. Hakamada, Polar low-speed
solar wind reappeared at the solar activity maximum of cycle 23. Geophys. Res. Lett. 30,
1409 (2003). doi:10.1029/2002GL016347
R. Osterbart, H.-J. Fahr, A Boltzmann-kinetic approach to describe entrance of neutral
interstellar hydrogen into the heliosphere. Astron. Astrophys. 264, 260–269 (1992)
S.P. Owocki, T.E. Holzer, A.J. Hundhausen, The solar wind ionization state as a coronal
temperature diagnostic. Astrophys. J. 275, 354–366 (1983)
E.N. Parker, Dynamics of the interplanetary gas and magnetic fields. Astrophys. J. 128,
664–676 (1958)
J.L. Phillips, S.J. Bame, A. Barnes, B.L. Barrcalough, W.C. Feldman, B.E. Goldstein, J.T.
Gosling, G.W. Hoogveen, D.J. McComas, M. Neugebauer, S.T. Suess, Ulysses solar wind
plasma observations from pole to pole. Geophys. Res. Lett. 22, 3301–3304 (1995a)
J.L. Phillips, S.J. Bame, W.C. Feldman, J.T. Gosling, C.M. Hammond, D.J. McComas, B.E.
Goldstein, M. Neugebauer, Ulysses solar wind plasma observations during the declining
phase of solar cycle 22. Adv. Space Res. 16, (9)85–(9)94 (1995b)
W.G. Pilipp, K.-H. Muehlhaeuser, H. Miggenrieder, M.D. Montgomery, H. Rosenbauer,
Unusual electron distribution functions in the solar wind derived from the HELIOS plasma
experiment—double-strahl distributions and distributions with an extremely anisotropic
core. J. Geophys. Res. 92, 1093–1101 (1987a)
W.G. Pilipp, K.-H. Muehlhaeuser, H. Miggenrieder, M.D. Montgomery, H. Rosenbauer,
Characteristics of electron velocity distribution functions in the solar wind derived from
the HELIOS plasma experiment. J. Geophys. Res. 92, 1075–1092 (1987b)
W.R. Pryor, J.M. Ajello, C.A. Barth, C.W. Hord, A.I.F. Stewart, K.E. Simmons, W.E.
McClintock, B.R. Sandel, D.E. Shemansky, The Galileo and Pioneer Venus ultraviolet
spectrometer experiments: solar Lyman-α latitude variation at solar maximum from
interplanetary Lyman-α observations. Astrophys. J. 394, 363–377 (1992)
W.R. Pryor, M. Witte, J.M. Ajello, Interplanetary Lyman α remote sensing with the Ulysses
interstellar neutral gas experiment. J. Geophys. Res. 103, 26813–26831 (1998)
in the three-dimensional heliosphere over the solar cycle. J. Geophys. Res. 108, 8034
(2003). doi:10.1029/2003JA009878
E. Quémerais, The interplanetary Lyman-α background, in The Physics of the Heliospheric
Boundaries, ed. by V.V. Izmodenov, R. Kallenbach, ISSI Scientific Report Series, SR-005,
pp. 283–310 (2006)
E. Quémerais, R. Lallement, S. Ferron, D. Koutroumpa, J.-L. Bertaux, E. Kyrölä, W.
Schmidt, Interplanetary hydrogen absolute ionization rates: retrieving the solar wind
mass flux latitude and cycle dependence with SWAN/SOHO maps. J. Geophys. Res.
111, 9114–9131 (2006). doi:10.1029/2006JA011711
P.G. Richards, J.A. Fennelly, D.G. Torr, EUVAC: a solar euv flux model for aeronomic
calculations. J. Geophys. Res. 99, 8981–8992 (1994). doi:10.1029/94JA00518
J.D. Richardson, K.I. Paularena, A.J. Lazarus, J.W. Belcher, Radial evolution of the solar
wind from IMP-8 to Voyager 2. Geophys. Res. Lett. 22, 325–328 (1995)
J.D. Richardson, J.C. Kasper, C. Wang, J.W. Belcher, A.J. Lazarus, Cool heliosheath
plasma and deceleration of the upstream solar wind at the termination shock. Nature
454, 63–66 (2008a). doi:10.1038/nature07024
J.D. Richardson, Y. Liu, C. Wang, D.J. McComas, Determining the LIC H density from
the solar wind slow down. Astron. Astrophys. 491, 1–5 (2008b). doi:10.1051/0004-
6361:20078565
D. Ruciński, M. Bzowski, Solar cycle dependence of the production of H+ pick-up ions in
the inner heliosphere. Adv. Space Res. 16, 121–124 (1995)
D. Ruciński, H.-J. Fahr, The influence of electron impact ionization on the distribution of
interstellar helium in the inner heliosphere: possible consequences for determination of
interstellar helium parameters. Astron. Astrophys. 224, 290–298 (1989)
Bibliography 137
D. Ruciński, H.-J. Fahr, Nonthermal ions of interstellar origin at different solar wind con-
ditions. Ann. Geophys. 9, 102–110 (1991)
D. Ruciński, M. Bzowski, H.-J. Fahr, Minor helium components co-moving with the solar
wind. Astron. Astrophys. 334, 337–354 (1998)
C. Salem, J.-M. Bosqued, D.E. Larson, A. Mangeney, M. Maksimovic, C. Perche, R.P.
Lin, J.-L. Bougeret, Determination of accurate solar wind electron parameters using
particle detectors and radio wave receivers. J. Geophys. Res. 106, 21701–21717 (2001).
doi:10.1029/2001JA900031
C. Salem, S. Hoang, K. Issautier, M. Maksimovic, C. Perche, WIND-Ulysses in-situ thermal
noise measurements of solar wind electron density and core temperature at solar maximum
and minimum. Adv. Space Res. 32, 491–496 (2003). doi:10.1016/S0273-1177(03)00354-5
H. Scherer, M. Bzowski, H.-J. Fahr, D. Ruciński, Improved analysis of interplanetary HST-
HLyα spectra using time-dependent modelings. Astron. Astrophys. 342, 601–609 (1999)
H. Scherer, H.-J. Fahr, M. Bzowski, D. Ruciński, The influence of fluctuations of the solar
emission line profile on the Doppler shift of interplanetary H Lyα lines observed by the
Hubble-Space-Telescope. Astrophys. Space Sci. 274, 133–141 (2000)
E.E. Scime, S.J. Bame, W.C. Feldman, S.P. Gary, J.L. Phillips, Regulation of the solar wind
electron heat flux from 1 to 5 au JGR 99, 23401–23410 (1994)
G.J. Smith, L.K. Johnson, R.S. Gao, K.A. Smith, R.F. Stebbings, Absolute differential cross
sections for electron capture and loss by kilo-electron-volt hydrogen atoms. Phys. Rev. A
44, 5647–5652 (1991)
J.M. Sokól, M. Bzowski, M. Tokumaru, K. Fujiki, D.J. McComas, Heliolatitude and time
variations of solar wind structure from in-situ measurements and interplanetary scintil-
lation observations. Solar Phys. (2012). doi: 10.1007/s11207-012-9993-9.
Š. Štverák, P.M. Trávnı́ček, M. Maksimovic, E. Marsch, A.N. Fazakerley, E.E. Scime, Elec-
tron temperature anisotropy constraints in the solar wind. J. Geophys. Res. 113, A03103
(2008). doi:10.1029/2007JA012733
Š. Štveràk, M. Maksimovic, P.M. Trávnı́ček, E. Marsch, A.N. Fazakerley, E.E. Scime,
Radial evolution of nonthermal electron populations in the low-latitude solar wind:
Helios, Cluster, and Ulysses observations. J. Geophys. Res. 114, A05104 (2009).
doi:10.1029/2008JA013883
T. Summanen, The effect of the time and latitude-dependent solar ionisation rate on the
measured Lyman-α-intensity. Astron. Astrophys. 314, 663–671 (1996)
T. Summanen, R. Lallement, J.-L. Bertaux, E. Kyrölä, Latitudinal distribution of solar
wind as deduced from Lyman alpha measurements: an improved method. J. Geophys.
Res. 98, 13215–13224 (1993)
K.F. Tapping, Recent solar radio astronomy at centimeter wavelengths - the tem-
poral variability of the 10.7-cm flux. J. Geophys. Res. 92, 829–838 (1987).
doi:10.1029/JD092iD01p00829
S.T. Tarnopolski, Expected distribution of interstellar deuterium in the heliosphere.
Dissertation. Space Research Centre PAS, 2007
S. Tarnopolski, M. Bzowski, Detectability of neutral interstellar deuterium by a forthcoming
SMEX mission IBEX. Astron. Astrophys. 483, L35-L38 (2008a). doi:10.1051/0004-
6361:200809593
S. Tarnopolski, M. Bzowski, Neutral interstellar hydrogen in the inner heliosphere under
the influence of wavelength-dependent solar radiation pressure. Astron. Astrophys. 493,
207–216 (2008b). doi:10.1051/0004-6361:20077058
G.E. Thomas, The interstellar wind and its influence on the interplanetary environment.
Ann. Rev. Earth Planet. Sci. 6, 173–204 (1978)
H. Tian, W. Curdt, E. Marsch, U. Schühle, Hydrogen Lyman-α and Lyman-β spectral radi-
ance profiles in the quiet sun. Astron. Astrophys. 504, 239–248 (2009a). doi:10.1051/0004-
6361/200811445
H. Tian, W. Curdt, L. Teriaca, E. Landi, E. Marsch, Solar transition region above sunspots.
Astron. Astrophys. 505, 307–318 (2009b). doi:10.1051/0004-6361/200912114
H. Tian, L. Teriaca, W. Curdt, J.-C. Vial, Hydrogen Ly α and Ly β radiances and profiles
in polar coronal holes. Astrophys. J. Lett. 703, L152–L156 (2009c). doi:10.1088/0004-
637X/703/2/L152
W.K. Tobiska, T. Woods, F. Eparvier, R. Viereck, L.E. Floyd, D. Bouwer, G. Rottman, O.R.
White, The SOLAR2000 empirical solar irradiance model and forecast tool. J. Atmos.
Sol. Terr. Phys. 62, 1233–1250 (2000)
M. Tokumaru, M. Kojima, K. Fujiki, K. Hayashi, Non-dipolar solar wind structure
observed in the cycle 23/24 minimum. Geophys. Res. Lett. 360, L09101 (2009).
doi:10.1029/2009GL037461
M. Tokumaru, M. Kojima, K. Fujiki, Solar cycle evolution of the solar wind speed distribu-
tion from 1985 to 2008. J. Geophys. Res. 115, A04102 (2010). doi:10.1029/2009JA014628
A.V. Usmanov, W.H. Matthaeus, B.A. Breech, M.L. Goldstein, Solar wind modeling
with turbulence transport and heating. Astrophys. J. 727, 84 (2011). doi:10.1088/0004-
637X/727/2/84
V. Vasyliunas, G. Siscoe, On the flux and the energy spectrum of interstellar ions in the
solar wind. J. Geophys. Res. 81, 1247–1252 (1976)
D.A. Verner, G.J. Ferland, T.K. Korista, D.G. Yakovlev, Atomic data for astrophysics. ii.
new fits for photoionization cross-sections of atoms and ions. Astrophys. J. 465, 487–498
(1996)
I.S. Veselovsky, A.V. Dmitriev, A.V. Suvorova, Algebra and statistics of the solar wind.
Cosmic Res. 48, 113–128 (2010). doi:10.1134/S0010952510020012
A. Vidal-Madjar, Evolution of the solar Lyman alpha flux during four consecutive years.
Solar Phys. 40, 69–86 (1975)
A. Vidal-Madjar, B. Phissamay, The solar L α flux near solar minimum. Solar Phys. 66,
259–271 (1980)
R.A. Viereck, L.C. Puga, The NOAA Mg II core-to-wing solar index: construction of a
20-year time series of chromospheric variability from multiple satellites. J. Geophys. Res.
104, 9995–10006 (1999). doi:10.1029/1998JA900163
M.E. Wachowicz, Global model of distribution of ionization states of heavy ions from solar
plasma in the heliosphere (in Polish). Dissertation, Space Research Centre PAS, 2006
H.P. Warren, NRLEUV 2, A new model of solar euv irradiance variability. Adv. Space Res.
37, 359–365 (2006). doi:10.1016/j.asr.2005.10.028
H.P. Warren, J.T. Mariska, J.L. Lean, A new reference spectrum for the euv irradiance
of the quiet sun 1. emission measure formulation. J. Geophys. Res. 103, 12077–12090
(1998a). doi:10.1029/98JA00810
H.P. Warren, J.T. Mariska, J.L. Lean, A new reference spectrum for the euv irradiance of
the quiet sun 2. comparisons with observations and previous models. J. Geophys. Res.
103, 12091–12102 (1998b). doi:10.1029/98JA00811
H.P. Warren, J.T. Mariska, K. Wilhelm, High-resolution observations of the solar hydrogen
Lyman lines in the quiet sun with the SUMER instrument on SOHO. Astrophys. J. Suppl
119, 105–120 (1998c). doi:10.1086/313151
K.-P. Wenzel, R.G. Marsden, D.E. Page, E.J. Smith, Ulysses: the first high-latitude helio-
spheric mission. Adv. Space Res. 9, 25–29 (1989). doi:10.1016/0273-1177(89) 90089-6
T.N. Woods, G.J. Rottman, O.R. White, J. Fontenla, E.H. Avrett, The solar Ly-alpha line
profile. Astrophys. J. 442, 898–906 (1995). doi:10.1086/175492
T.N. Woods, D.K. Prinz, G.J. Rottman, J. London, P.C. Crane, R.P. Cebula, E. Hilsenrath,
G.E. Brueckner, M.D. Andrews, O.R. White, M.E. VanHoosier, L.E. Floyd, L.C. Her-
ring, B.G. Knapp, C.K. Pankratz, P.A. Reiser, Validation of the UARS solar ultraviolet
irradiances: comparison with the ATLAS 1 and 2 measurements. J. Geophys. Res. 101,
9541–9570 (1996). doi:10.1029/96JD00225
T.N. Woods, W.K. Tobiska, G.J. Rottman, J.R. Worden, Improved solar Lyman irradiance
modeling from 1979 through 1999 based on UARS observations. J. Geophys. Res. 105,
27195–27215 (2000)
T.N. Woods, F.G. Eparvier, S.M. Bailey, P.C. Chamberlin, J. Lean, G.J. Rottman, S.C.
Solomon, W.K. Tobiska, D.L. Woodraska, Solar euv experiment (SEE): mission overview
and first results. J. Geophys. Res. 110, A01312 (2005). doi:10.1029/2004JA010765
F.M. Wu, D.L. Judge, Temperature and velocity of the interplanetary gases along solar
radii. Astrophys. J. 231, 594–605 (1979)
Part II
Interplanetary Hydrogen
—4—
Thirty Years of Interplanetary

Background Data: A Global View
Eric Quémerais∗
LATMOS-IPSL, Université Versailles-Saint Quentin, Guyancourt, France
Bill R. Sandel
Lunar and Planetary Laboratory, University of Arizona, Tucson, AZ, USA
Vladislav V. Izmodenov
Lomonosov Moscow State University, School of Mechanics and Mathematics
Institute for Problems in Mechanics, Russian Academy of Sciences,
Moscow, Russia
Space Research Institute, Russian Academy of Sciences, Moscow, Russia
G. Randall Gladstone
Abstract
This chapter compares results of models of the interplanetary background, such
as the one presented in Chap. 1, to different datasets obtained in the outer helio-
sphere (Voyager-UVS, Alice New-Horizons) and in the inner heliosphere (SWAN-
SOHO, STIS-HST). The aim of this work is to combine these datasets and the
models and to derive calibration factors that give a coherent picture of the various
instruments and the interplanetary background. These datasets do not overlap and
the models are used to bridge the gaps in distance or in time. In the case of Voy-
ager 1 and 2 UVS instruments, the calibration factors derived here are significantly
different from the values published by Hall (Ultraviolet resonance radiation and
the structure of the heliosphere. Dissertation, University of Arizona, 1992).
Introduction
The aim of this section is to show that the interplanetary (IP) ultraviolet (UV)
background can be used to calibrate UV instruments at a given wavelength, i.e.

142 4. Thirty Years of Interplanetary Background Data
H Lyman-α at 121.566 nm. This emission can also be used to follow the temporal
evolution of the sensitivity of the instrument at the same wavelength. Most UV
instruments studying planetary atmospheres, exospheres or cometary atmospheres
are also able to measure the interplanetary background. To give an example of
how this can be done, we will use different datasets obtained at various positions
in the heliosphere and compare them with models. The models are described by
Izmodenov et al. (2013, this volume). Our aim is also to show how accurate the
determination of the absolute calibration can be. There are various uncertainties
that need to be taken into account in this work.
In the following chapter, we will briefly present the models of the interplanetary
background used here. They are described in greater length in Quémerais et al.
(2008). Following that, we will compare the models with the data obtained by the
two UVS instruments aboard Voyager 1 and Voyager 2 (Broadfoot et al. 1977). We
will show that from this model comparison we can derive a new absolute calibration
of these instruments at Lyman-α.
In the following section, we will show that this new calibration is compatible
with the calibration of two other UV instruments: the UV spectrometer ALICE
on the New Horizons mission and the Solar Wind ANisotropies (SWAN) photome-
ter on the Solar Heliospheric Observatory (SOHO). Both instruments have also
observed the interplanetary background but at times and regions of the helio-
sphere that are different from the Voyager observations. The SWAN calibration
has been recently revised and is significantly different from the values published
by Quémerais and Bertaux (2002). This revision is based on a comparison of the
SWAN measurements with Hubble Space Telescope measurements (Clarke et al.
1998; Quémerais et al. 2003).
The Interplanetary Background Models

The Interplanetary UV Background was first observed in the late 1960s by
two instruments on the Orbiting Geophysical Observatory-5 (OGO-5) spacecraft
(Bertaux and Blamont 1971; Thomas and Krassa 1971). This ubiquitous emission
is due to the backscatter of solar Lyman-α photons by hydrogen atoms present
in the interplanetary medium. The flow of neutral hydrogen atoms through the
heliosphere is called the interstellar wind. It is due to the relative motion between
the solar system and the local interstellar cloud.
Models of the hydrogen atom distribution in the inner heliosphere were de-
veloped in the early 1970s (Blum and Fahr 1970; Thomas 1978). In the 1990s,
following the work of Baranov (1990) and Baranov and Malama (1993), it was
realized that the distribution of hydrogen atoms in the heliosphere is influenced by
the interface between the solar wind and the interstellar plasma. Studies of the
interplanetary background have been used to infer parameters of the interstellar
wind flow and even the direction of the interstellar magnetic field (Bertaux et al.
1985; Lallement et al. 2005; Quémerais et al. 2007). Details of the distribution of
hydrogen atoms in the heliosphere are found Izmodenov et al. (2013, this volume).
Models of the interplanetary background must include radiative transfer effects,
even in the inner heliosphere. Close to the Sun, the hydrogen number density
The Interplanetary Background Models 143
is small and the single scattering approximation can be used to compute local
emissivities. However, this simplification does not work for calculating intensities.
To model background intensities, it is necessary to perform an integration of the
emissivity along the line of sight. The integration has to be performed until ex-
tinction on the line of sight is significant, which means that the optically thin
approximation is not valid anymore. Beyond five astronomical units (AU) from
the Sun, multiple scattering effects have to be included in the model. Details on
the computations can be found in Quémerais (2000), Quémerais and Izmodenov
(2002), and Quémerais et al. (2008).
Our method combines a standard iterative computation of the first- and second-
order scattering terms using Monte Carlo simulations to compute the higher or-
ders of scattering (i.e., photons that are scattered more than two times). The
second-order scattering term is computed from integration over the whole sky of
the first-order component. It is also computed independently by the Monte Carlo
simulation. This provides a way to check the validity of both methods (Quémerais
2000).
The iterative scheme, i.e. computing order ns from integration over the whole
sky of terms of scattering order ns + 1, could theoretically be applied to any order
(see Sect. 2.3 of Quémerais 2000). However, starting from ns = 3, it is much more
efficient to use a Monte Carlo approach. The general assumptions of our model are
the following:
• The photon source frequency profile is derived from the Solar Ultraviolet Mea-
surements of Emitted Radiation (SUMER) measurements on SOHO (Lemaire
et al. 1998). The source profile is shown in Quémerais (2000).
• The local distribution of hydrogen atoms is represented by the sum of three
components. The first component is the pristine interstellar component,
the second one is created by charge-exchange with protons in the outer he-
liosheath, and the third one is created by charge exchange with protons in
the inner heliosheath. These three components are described in Izmodenov
et al. (2013, this volume).
• The density distribution of the three populations is symmetrical around the
wind axis going through sun center.
• At each point in the heliosphere, the local distribution of each component is
described by six values: a number density, a mean velocity (two components
in the plane containing the wind axis), and three pseudo-temperatures that
are the widths of the velocity distribution projected on the three directions
of the local frame. These three pseudo-temperatures can be different, which
means that the local distribution is not necessarily a Maxwell–Boltzmann
distribution.
• The scattering process is computed following the Angle Dependent Partial
Frequency Redistribution (Mihalas 1970). This means that the frequency of
the outgoing photons depends on the frequency of the incoming photons, the
scattering angle, and the velocity distribution at the point of scattering. The
numerical expressions of the redistribution function is given in Quémerais
(2000).
• The scattering redistribution function includes the phase function given by

Brandt and Chamberlain (1959). The improved phase function by Brasken
and Kyrölä (1998) has not been included here because it leads to changes
which are much smaller than the expected accuracy of our Monte Carlo com-
putation.
Most previous publications have used the assumption of Complete Frequency

Redistribution (Keller et al. 1981; Hall 1992; Quémerais and Bertaux 1993), for
which the frequency of the outgoing photon only depends on the local density
distribution at the point of scattering. Unfortunately, this more simple assumption
does not allow for an exact computation of backscattered line profiles because it
neglects the incoming photon frequency profile. This applies to a medium with
large optical thickness where many scatterings occur, not in the inner heliosphere.
To be able to compute exact line profiles, we had to implement a scheme based on
the actual Angle Dependent Partial Redistribution Function described by Mihalas
(1970). Quémerais (2000) devoted a whole section (Sect. 4) to the comparison
between CFR and ADPFR results. Similarly, Gangopadhyay et al. (2005) have
gone beyond the use of Complete Frequency Redistribution. To improve our line
profile computations in the future, we will include a full description of the local
velocity distribution and not only the width of the local distribution projected on
the local frame axes. However, we don’t expect large differences with our current
results.
Outer Heliosphere Data

The Voyager 1 and 2 spacecraft were launched in late 1977. After the last
planetary encounter by Voyager 2 in 1989, the Voyager mission to the planets be-
came a mission dedicated to the study of the outer heliosphere and the heliospheric
interface. Each of the Voyager spacecraft carries a UV Spectrometer (UVS) de-
veloped by the Lunar and Planetary Laboratory in Tucson (Broadfoot et al. 1977,
1989). During their long cruise in the solar system, the UVS instruments have
made many observations of the interplanetary background. A study of the data
obtained between 1980 and 1995 can be found in de Toma et al. (1997).
After the last planetary encounter in 1989, the UVS were used to study the
outer heliosphere. A study of the data obtained during the roll maneuvers was
published by Hall et al. (1993). They were the first to show that the intensity
variation with solar distance in the upwind direction did not decrease as expected
from a standard hot model. The authors suggested that this was a signature of
the hydrogen pile-up near the nose of the heliopause which had been predicted by
Baranov (1990) and Baranov and Malama (1993). This result has later been con-
firmed by Quémerais et al. (2003) on a much longer database. However, Quémerais
et al. (2003) showed that the numerical values given by Hall et al. (1993) were in-
compatible with the measurements obtained after 1993. This is most likely due to
an erroneous correction for the solar Lyman-α flux variations (Pryor et al. 1992;
de Toma et al. 1997). This did not change the fact that the hydrogen distribution
in the outer heliosphere is strongly influenced by the heliospheric interface.
Outer Heliosphere Data 145
In 1993, the UVS spectrometers started a campaign of outer heliosphere scans.

The idea was to scan from the upwind to the downwind direction to search for
signs of emission from the hydrogen wall predicted by Baranov (1990). To alle-
viate the problem of the solar flux correction, special observations were designed.
They started in late 1993 for Voyager 1 and early 1994 for Voyager 2. They are
presented in Quémerais et al. (1995) and Quémerais et al. (2003). These scans
were made of about 20 points regularly distributed on a great circle going from
the upwind direction to the downwind direction. Each individual observation with
a fixed direction of the line of sight lasted 12 or 24 h. The idea was to use the
downwind direction data to scale the model of the background and then to look for
an excess in the upwind direction. On a scale of 10–20 days, it is only necessary
to take into account the rotational modulation of the solar flux, and because of
the large distance between the Sun and the spacecraft, the rotational modulation
amounts only to a few percent. The early results are shown in Figs. 9 and 10 of
Quémerais et al. (1995). Comparisons with hot models showed a systematic excess
intensity in the upwind direction. This results confirmed independently the results
of Hall et al. (1993) and was independent of corrections for long-term variations of
the solar flux. From 1993 to 2003, V1 UVS performed more than 100 scans. At the
time when the Voyager 2 UVS was shut down in 1998, it had performed 40 scans of
the outer heliosphere. Since late 2003, the Voyager 1 UVS observations have been
maintained but following the shutdown of the heaters of the platform actuators,
movements of the scan platform have been stopped. The Voyager 1 UVS line of
sight is pointed within a few degrees of the upwind direction. Spectra are obtained
continuously and averaged to obtain daily values. Voyager 1 UVS is still operating
as of the time of this publication.
Quémerais et al. (1996) tried to fit this excess with a model computation in-
cluding effects of the heliospheric interface on the distribution of hydrogen atoms
in the outer heliosphere. The authors failed to obtain an excess as large as the
one appearing in the data and they suggested that some emission may come from
recombination processes happening in HII regions in the milky way, i.e. at much
greater distance than the length scale of the local interstellar cloud. Lallement
et al. (2011) has recently shown that part of the excess emission in the upwind
direction is well correlated with H α emission from hot clouds in the galactic plane.
However, most of the excess is in fact due to the existence of the hydrogen wall in
the upwind direction.
Ben-Jaffel et al. (2000) suggested an alternate explanation. Based on one ob-
servation from the Goddard High Resolution Spectrograph (GHRS) on the Hub-
ble Space Telescope (HST) of the IP Lyman-α in the upwind direction made in
1995 (Clarke et al. 1995), the authors identified a small spectral feature at shorter
wavelength than the core interplanetary line. They interpreted that feature as a
signature of the Fermi acceleration that may accelerate photons scattering back
and forth between media of different velocity. However this interpretation didn’t
withstand a more detailed analysis. First, the spectral feature was exactly at the
position of the D Lyman-α line of the geocorona. This suggested that the feature
was simply a contamination of the observation by the atmosphere of the Earth.
Second, a new observation was performed in 2001 by the Space Telescope Imag-
ing Spectrograph (STIS) instrument on HST which replaced the GHRS. This new
observation was analysed by Quémerais et al. (2009) who showed that the feature
was absent from the data for a different line of sight which was less contaminated
by the Earth exosphere.
Figure 4.1 shows the variation of the interplanetary background data as a func-
tion of time for UVS-Voyager 1 and UVS-Voyager 2, respectively. We should note
that before 1993, observations were randomly done. Starting in 1993, observ-
ing campaigns scanning the outer heliosphere were conducted more systematically
(Quémerais et al. 1995). For both datasets, we see a very similar behaviour. The
general trend of the measurements show a decrease of intensity as a function of
solar distance. For a constant intensity in an optically thin medium, we would
expect a decrease as 1/r, however variations of hydrogen number density with ion-
ization effects and multiple scattering effects modify the dependence with solar
distance. The high frequency variations are due to the variations of the direction
of the line of sight, with higher count rates when line of sight is closer to the
inner heliosphere and the Maximum Emissivity Region. The data also show the
variations of the solar illuminating flux that varies with solar activity. There are
local maxima in 1991 and 2001 at times of solar activity maximum. The Voyager
2 UVS measurements stopped in 1998. On the other hand, Voyager 1 UVS is still
active as of July 2012. Starting in 2003, the outer heliosphere scans were stopped.
It appears clearly in Fig. 4.1 where high frequency variations disappear after 2003.
It should also be noted that the Voyager 1 data seem to flatten after 1997. Through
mid-2012, this flattening is stronger that the predictions of our model computation
including heliospheric and time-dependent effects.
Voyager UVS Calibration Factor

The initial radiative transfer computations used in Quémerais et al. (1995, 1996)
were based on the assumption of Complete Frequency Redistribution (CFR). The
numerical code is described in details in Quémerais and Bertaux (1993). CFR
is a simplifying approximation which assumes that the energy of the photon after
scattering only depends on the local velocity profile of the hydrogen atoms (velocity
and temperature). The work presented in Quémerais et al. (1995) also assumed
that the bulk velocity was constant and the hydrogen flow was isothermal. Those
simplifying assumptions were also adopted by Quémerais et al. (1996) although the
hydrogen number density values were obtained from the Baranov–Malama model of
the hydrogen wall. At that time, it was thought that these assumptions were good
enough to compute background intensities in the heliosphere, even if computing
the background line shape was impossible given the assumptions.
Quémerais (2000) modified the numerical code to be able to compute line
shapes. To do that, the CFR numerical scheme was replaced by Angle Dependent
Partial Frequency Redistribution (ADPFR). In that scheme, the energy distribu-
tion of the scattered photons depends on the local velocity distribution, the energy
of the incoming photons and the phase angle. Complete velocity distributions of
hydrogen atoms were also used including first moments (local velocity) and second
moments (projected temperature). The model also includes the scattering phase
function dependence. This model was later adapted to multi-population hydrogen
distribution models (Izmodenov et al. 2001; Quémerais and Izmodenov 2002).
Voyager UVS Calibration Factor 147
Figure 4.1: (Top) IP background measurements made by UVS-Voyager 1 between

1980 and 2011. (Bottom) IP background measurements from UVS-Voyager 2 be-
tween 1980 and 1998. The x-axis is the decimal year and the ordinate is in units
of counts per second (integrated over the nine channels of the Lyman-α line). The
line of sight is non-uniformly distributed in the sky. Changes in the line of sight
direction cause the high frequency spikes. Between 1980 and 1995, the data de-
crease more or less like 1/r. Since 1995, the background measurements show only
a small decrease with distance. These data also show the variation of the solar
illuminating flux with the solar cycle. The local bumps in 1990 and 2001 (top) and
1990–1991 (bottom) are caused by the larger values of the solar flux at Lyman-α
at solar maximum
The latest version of this model is presented in Quémerais et al. (2008).

Izmodenov et al. (2005, 2008) modified the hydrogen distributions to account for
time variations of the solar parameter (radiation pressure, total ionization rate,
see Izmodenov et al. 2013 or Bzowski et al. 2013 this volume). It was found that
time dependence has very little effect on the hydrogen distribution in the outer
heliosphere but is very strong close to the Sun.
Using this model from 2008, we have computed the interplanetary background
model values for each of the UVS observations presented in Fig. 4.1. The model is
based on the following parameters for the pristine interstellar medium(i.e. before
interaction with the heliosphere): VLIC = 26.4 km s−1 and TLIC = 6527 K (Witte
2004; Möbius et al. 2004; Gloeckler et al. 2004; Lallement et al. 2004), NH,LIC =
0.18 cm−3 , np,LIC = 0.06 cm−3 .
The solar flux variations are taken into account by multiplying the model re-
sult by scaling factor proportional to the solar flux measurements performed by
the SOLar-STellar Irradiance Comparison Experiment (SOLSTICE) instrument on
the SOlar Radiation and Climate Experiment (SORCE) spacecraft (Woods et al.
2000). The values are available on line from http://lasp.colorado.edu/lisird/lya/.
However, it should be noted that the solar flux measured by SORCE is the total
line intensity and is expressed in photons cm−2 s−1 . The hydrogen atoms in the
interplanetary medium are only affected by photons with a frequency near line
center (within 30 km/s of the rest frame) and the solar illuminating flux is often
expressed in units cm−2 s−1 per Angströms. Following Emerich et al. (2005), we
used an equivalent width for the solar Lyman-α line center of 0.9 per Angström.
This means that SORCE measurements of the solar flux values are multiplied by
0.9 to get the illuminating flux at line center. This ratio varies with the solar cycle
because it depends on the activity on the solar disk.
Figure 4.2 shows the relation between the observed intensities and the model.
For these plots we have selected all data obtained within 50 AU from the Sun. In
that way, we avoid the influence of the strong flattening of the data as seen after
1997 in the Voyager 1 UVS data shown in the top panel of Fig. 4.1. For both
datasets we get a very good agreement between data and model. Assuming that
our model is correct, we will get back to that point later on, it is then possible to
derive an absolute calibration for the two Voyager UVS instruments based on the
relations shown in Fig. 4.2. For Voyager 1 UVS, we obtain an absolute calibration
of 89.7 ± 4.9 Rayleigh per counts per sec. Similarly for Voyager 2 UVS, we obtain
an absolute calibration at Lyman-α of 110.4 ± 8.7 Rayleigh per counts per sec.
Figure 4.3 presents the result of this analysis. The data are now expressed
in Rayleighs and the corresponding models are shown (in orange for Voyager 1
UVS and green for Voyager 2 UVS). The curves at the bottom show the difference
between data and models. From this figure, we see that a very good agreement
is obtained when the spacecraft is between 10 and 50 AU from the Sun. How-
ever, below 10 AU from the Sun, there are some systematic variations that need
to be considered. In the models used here, the time-dependence is averaged over a
1-year period. This means that shorter time-scale variations are not well repre-
sented by the model. This is fine beyond 10 AU from the Sun but leads to system-
atic discrepancies closer to the Sun. This also means that the variations of the solar
parameters need to be accounted for in a very accurate way if the we wish to use
Voyager UVS Calibration Factor 149
Figure 4.2: Determination of the linear relation between the Voyager IPH data
and the model. The top panel shows Voyager 1 data, the bottom panel shows
Voyager 2. The correlation is very good. The slopes with uncertainties are added.
The agreement between the data and the model is much better for Voyager 1 than
Voyager 2. The uncertainties for the Voyager 2 measurements are about twice as
large as for Voyager 1. This plot includes all data obtained between 10 and 50 AU
from the Sun. If we take into account the difference in solar flux values, our result
is in good agreement with the result of Hall (1992)
Figure 4.3: Comparison of the UVS Voyager data and model computations in
absolute values. Top panel shows Voyager 1 and bottom panel shows Voyager 2.
The abscissae are in AU and show the distance between the spacecraft and the Sun.
Closer to the Sun (within 10 AU) it is necessary to take into account accurately
the variations of the solar parameters
ALICE New Horizons Roll Data 151
models to calibrate UV instruments. The latest models as described by Izmodenov

et al. (2013, this volume) are able to improve the agreement between the data and
the model for distances up to 10 AU.
Finally, it should be noted that the absolute calibration derived here is sig-
nificantly different from the value used usually for UVS Voyager 1 and Voyager
2 (Hall, Ph.D. Thesis, University of Arizona, 1992). The values quoted by Hall
(1992) are C1 = 218 Rayleigh per counts per second for UVS Voyager 1 and C2 =
172 Rayleigh per counts per second for UVS Voyager 2. In our estimate both UVS
are more sensitive than what was previously found. It is likely that a good portion
of the discrepancy comes from the values of the solar illuminating. At the time of
the work of Hall (1992), the reference solar Lyman-α flux was based on the SME
measurements (Rottman et al. 1982). After the UARS SOLSTICE measurements
started in the 1990s the average value of the solar flux increased by a factor of 1.5
(Pryor et al. 1996). It is therefore not surprising that our estimates of the absolute
calibration of UVS V2 differ by a factor of 1.56 that compensates for the discrep-
ancy in mean solar illuminating flux. For V1 UVS, the ratio of calibration factors
is larger (2.4). Other factors must be considered to explain the discrepancy. This
point has not yet been resolved.
Gangopadhyay et al. (2005) tried to reappraise the calibration of Voyager 2 by
comparing some observations obtained between 1993 and 1998 with a model that
as many similarities with ours. They came to the conclusion that the intensities
recorded by UVS V2 need to be multiplied by a factor in the range 0.5–0.76.
In this work, we find a change of 89.7/172 = 0.52. This is in good agreement
with the conclusion of Gangopadhyay et al. (2005). In the same work, the authors
showed that the Pioneer 10 photometer data also need to be revised. This revision
will be performed in a future publication.
ALICE New Horizons Roll Data

Gladstone et al. (2013, this volume) present the interplanetary hydrogen (IPH)
data obtained by the ALICE UV Spectrometer on board the New Horizons mission.
The spacecraft is on its way to Pluto and during the cruise phase of the mission,
it has performed a few rolls with observations of the interplanetary background.
The instrument, data and observing conditions are detailed in Gladstone et al.
(2013, this volume). Our aim here is to compare the ALICE interplanetary data
to the same model we used in the previous section to calibrate the Voyager UVS
spacecraft.
The New Horizons spacecraft performed three rolls around a fixed rotation axis
on three different occasions. In Table 4.1 we provide the date, distance to the Sun,
and the solar illuminating flux for each roll maneuver. More details can be found
in Table 5.1 of Gladstone et al. (2013, this volume).
Figure 4.4 shows the geometry of the rolls. The plot shows a contour plot of
the interplanetary background as seen by an observer at 17 AU from the Sun. The
map is shown in west ecliptic longitude and ecliptic latitude. At that distance
from the Sun the maximum intensity is close to the direction of the Sun. The anti-
sunward direction has the minimum intensity. For the New Horizons spacecraft,
Table 4.1: ALICE IPH observations

Observation date Distance from the sun (AU) Solar illuminating flux
(cm−2 s−1 )
ACO-1 10/7/2007 7.6 3.54 × 1011 phot

ACO-2 10/18/2008 11.3 3.53 × 1011 phot
ACO-3 6/19/2010 17.0 3.74 × 1011 phot
the upwind direction is not far from the direction with minimum intensity. The
three New Horizons rolls have a similar geometry. The roll axis was chosen so that
the plane of the roll passed close to upwind and downwind but avoided most of the
stars along the plane of the milky way. The region close to the Sun is the most
difficult to model because it is more sensitive to the solar parameters of ionization
and radiation pressure. The emission from the rest of the sky is little affected by
variations of solar parameters.
The comparison with our model computed for the correct positions of the New
Horizons spacecraft at the times of the observations is shown in Fig. 4.5. The data
from the three rolls (ACO-1 is incomplete) are shown by the green, black and red
curves with high frequency noise. The spikes correspond to stars in the field of
view of the instrument. In the figure, the data have been scaled to a constant solar
flux by dividing by the SORCE SOLSTICE solar Lyman-α flux of the day and
multiplying by a reference flux of 0.9 × 3.32 × 1011 phot cm−2 s−2 Å−1 . The factor
0.9 Å−1 has been discussed before. It corresponds to the change from the total
line flux to the line center flux at Lyman-α. This is the exact same scheme that
we have applied in the previous section to determine the Voyager UVS calibration
factors.
The agreement between the data and the model in Fig. 4.5 is rather good except
in the region closest to the Sun. This is the region that is most affected by solar
parameters and is more difficult to model accurately. The IPH emission from rest
of the sky shows little dependence on solar parameters. The excellent agreement
between the data and the model suggests that the model correctly represents the
interplanetary background and the interplanetary hydrogen distribution from 10
to 20 AU in the region where these data were obtained.
We performed a linear regression between the data and a model where the solar
flux was not corrected by the 0.9 Å−1 factor (i.e. the SOLSTICE flux was used).
Using the three rolls, it was found that the best ratio between total solar Lyman-
α line and the solar line core value was not 0.9 but 0.86. Given the potential
residual bias in the model, we decided to keep the 0.9 factor with an uncertainty
of about 5 %.
In conclusion, the comparison between the ALICE New Horizon roll data and
our model of the interplanetary background in the range 7–17 AU, validates our
model within about 5 % accuracy. Based on the extrapolation of the model to larger
values of solar distance (up to 50 AU), we have linked the calibration factor of the
two UVS instrument to the calibration of the ALICE spectrometer at Lyman-α.
This means that the Voyager UVS calibration is not based on the model alone but
Inner Heliosphere Data 153
Figure 4.4: Map of the Interplanetary Background as seen by an observer at 17 AU

from the Sun. The coordinates are in west ecliptic longitude and ecliptic latitude.
The direction of the Sun is close to the downwind direction because the spacecraft
is roughly going towards the upwind direction. The geometry of the rolls performed
by New Horizons is shown by the white line. The line of sight is aligned along a
plane passing close to the upwind and downwind directions but avoids most of the
hot stars in the plane of the milky way
also on the comparison with the ALICE New Horizons roll datasets. This is an
important step in the creating a global view of the interplanetary background from
the inner to the outer heliosphere.
Inner Heliosphere Data

In this section, we present a comparison between our model and data obtained
in the inner heliosphere. Our aim is to show that this model is in good agreement
with the measurements of the SOHO SWAN instrument. In the first part, we
explain how the SWAN calibration was obtained, based on a comparison with HST
STIS and HST GHRS measurements. In the second part, we compare the SWAN
dataset and the model.
SWAN Calibration
The SWAN instrument is dedicated to the study of the solar wind flux aniso-
tropies (Bertaux et al. 1995). The SOHO mission is a cooperation between ESA
Figure 4.5: Plot of the ALICE IPH measurements for the three rolls. The data are
shown by the green curve (ACO-1), black curve (ACO-2), and red curve (ACO-4).
The spikes are due to stars crossing the line of sight. The models (see text) are
shown by the blue, green and yellow curves and are plotted over the data. The
data-model agreement is excellent except close to the direction of the Sun
and NASA and was launched in December 1995. The SWAN instrument started its
operation in early 1996 and has been operating since almost continuously, except
during the summer of 1998. It is a UV photometer with a bandwidth from 110
to 190 nm. It consists of two sensor units with identical designs, but sitting on
opposite sides of the SOHO spacecraft. The combination of the data from both
sensors allows to make daily observations of the whole sky.
An initial estimate of the SWAN calibration was published by Quémerais and
Bertaux (2002). The two sensor units of SWAN are compared on a regular basis
because they can both view the same area of the sky at the same time. Therefore,
It is necessary to calibrate only one sensor and the calibration of the second sensor
is derived from that knowledge. As shown by Quémerais and Bertaux (2002),
the two sensors have shown very distinct temporal variations in sensitivity, even
though they have been built following the same design. The sensor on the +Z side
of SOHO, facing the north ecliptic sky most of the time, has been fairly stable over
its 15 years of activity. On the other hand, the −Z sensor (facing south ecliptic
most of the time) has been degrading since the early days of the mission and has
lost more than half of its responsivity since 1995. Therefore, very early in the
mission, it was decided to use the more stable unit as a reference and to use the
almost daily sensor cross-calibration observations to scale the data to the values
of that sensor unit. In the spring of 1996, GHRS on the Hubble Space Telescope
Inner Heliosphere Data 155
Table 4.2: List of common observations between SOHO and HST
Date HST values SWAN values ratio

GHRS or STIS
March 1996 260 R 175 R 1.49

March 2000 890 R 578 R 1.54
March 2000 840 R 568 R 1.48
March 2001 1450 R 870 R 1.67
measured the IPH background close to the downwind direction (Clarke et al. 1998).
This observation was made at the same time as a regular SWAN observation and
served as a reference point for the uncalibrated SWAN measurements. However,
for unclear reasons, the value used by Quémerais and Bertaux (2002) was not the
final GHRS result and this led to an inexact estimate of the absolute calibration of
SWAN. It should be noted also that because the downwind direction has a lower
brightness by nearly a factor of three compared to the upwind direction as seen
from 1 AU, the derived calibration factor had a large uncertainty.
Following the replacement of GHRS by STIS on HST, new observations were
made (Quémerais et al. (2009), private communication). A comparison between the
four HST (one by GHRS and three by STIS) measurements and the corresponding
SWAN values is shown in Table 4.2. As shown in the table, the SWAN calibration
factor has been corrected to match the HST measurements. The current SWAN
calibration is equal to 4.1 Rayleigh per count per second for the +Z sensor unit.
Model of the Inner Heliosphere

Now that we have derived the SWAN calibration from comparison to HST
measurements, which means that this calibration is directly derived from the GHRS
and STIS calibrations, we are going to compare the results of our model with the
SWAN observations.
Modeling the interplanetary background in the inner heliosphere is more diffi-
cult than it is in the outer heliosphere. In the vicinity of the Sun, the hydrogen
distribution is strongly affected by changes in solar radiation pressure and ion-
ization rates. These parameters are not uniformly spatially distributed, and have
never been measured for all directions in space. However, we can use the SWAN
observations to derive the spatial distribution of the ionization rates and the solar
illuminating flux (Quémerais and Bertaux 2002; Quémerais et al. 2006).
To fit the full-sky measurements from SWAN, one needs a full three-dimensional
model of the interplanetary background. This work has not been completed yet and
here we will use the axi-cylindrical model used in the previous sections. However,
Quémerais et al. (2006) have shown that the solar ionization rate is very close to
being isotropic at times of solar maximum. This means that our axi-symmetric
model can be used to try to fit the SWAN data obtained near solar maximum.
Figure 4.6 shows the IPH intensity recorded by SOHO SWAN between 1996 and
2011 in the direction of the North ecliptic pole as a function of time. The values are
corrected for solar flux variations based on the SORCE SOLSTICE measurements.
The values are scaled to the reference solar flux of 0.9 × 3.32 × 1011 phot cm−2
s−2 Å−1 as was the case for the ALICE Lyman-α line center values described in
the previous section. The data show two main modulations, one with a period of 1
year corresponding to the rotation of SOHO around the Sun and one with a period
of about 10–11 years corresponding to the solar activity cycle. The 1-year period is
easy to understand. The distribution of hydrogen is not symmetric around the Sun.
In the upwind direction, hydrogen atoms get closer to the Sun and the downwind
region is almost void of hydrogen atoms. During the orbit of the spacecraft around
the Sun it crosses the denser upwind region and then downwind cavity which is
the almost void of hydrogen. The contrast in density is the same in emissivity
and therefore also appears in the intensity. The 1-year modulation disappears if
the hydrogen cavity surrounding the Sun gets much larger than the radius of the
spacecraft orbit.
The 11-year period is caused by the change in solar activity and distribution
between fast and slow solar wind. As mentioned by Quémerais et al. (2006), during
solar minimum, ionization rates at the solar poles are much lower than at the solar
equator. This is because the fast solar wind at the solar poles has a lower mass
flux than the slow solar wind at the equator (Quémerais et al. 2006). But this
difference disappears during solar maximum when the fast and slow solar wind
can be found at different latitudes. This means that the mean ionization rate is an
average between different conditions. As appears in Fig. 4.6, after correction for the
solar illumination, the scaled intensity at the poles is minimum at solar maximum
because this is when the ionizing fluxes are stronger. The ionizing fluxes are also
more or less isotropic. In Fig. 4.6, we have added the isotropic ionization models
computed for two solar conditions detailed in Quémerais et al. (2008). Because the
models are isotropic and independent of time, they give the same intensities each
year. The only variation is due to the position of the observer with respect to the
interstellar wind axis. The polar values at solar minimum (1996 and 2007) are not
well reproduced by the model shown in the figure. To reproduce the polar values,
we would need to use a full three-dimensional model.
Figure 4.7 shows the same curves as Fig. 4.6 but concentrates on the period
around the previous solar maximum in 2001. First, it is interesting to note that the
minimum of intensity, i.e. the lowest column density of hydrogen atoms is obtained
in 2003, almost 2 years after the solar maximum. This is the time necessary for
the changes in ionization rates to modify the hydrogen distribution. Second, the
two models are isotropic models with different solar wind fluxes. The parameters
for the orange model are given in Fig. 1 of Quémerais et al. (2008). The values
correspond to solar minimum conditions in-ecliptic (Quémerais et al. 1996). For the
blue curve, the parameter value correspond to in-ecliptic solar maximum conditions
(i.e. values for 2001 in Fig. 4.1 of the same paper). The two models are not
very different because they are isotropic models with in-ecliptic conditions. Once
anisotropy of the ionizing flux is included, models for solar minimum and solar
maximum conditions become very different at the poles. However, it is interesting
to note that the SWAN data for the year 2003 fall between the two models. The two
Conclusion 157
Figure 4.6: Plot of the IPH intensity recorded by SOHO SWAN between 1996 and
2011 in the direction of the North ecliptic pole as a function of time (diamonds).
The values are corrected for solar flux variations based on the SORCE SOLSTICE
measurements. The data show two main modulations, one with a period of 1 year
corresponding to the rotation of SOHO around the Sun and one with a period of
about 10–11 years corresponding to the solar activity cycle. The bottom curves
(orange and blue) show models for uniform ionization rates for different solar wind
mass flux values (see text)
models give the range of variation for an isotropic model given the uncertainty on
the solar ionizing fluxes, therefore we can conclude that the SWAN data obtained
for solar maximum conditions are in good agreement with the isotropic models
presented here. The blue model is the same one that has been used to compare to
the ALICE-NH rolls and to calibrate the two Voyager UVS.
In conclusion, based on our model and data comparison for solar maximum
conditions and for an observer at 1 AU, the SWAN calibration gives values that
agree with the computed values at 1 AU. Therefore we can conclude that the SWAN
and ALICE calibration are in good agreement assuming that the extrapolation
based on our model is correct.
Conclusion
In this chapter, we have compared four different sets of observations of the In-
terplanetary background covering the inner heliosphere to the outer heliosphere. In
one case, we were able to compare data that were obtained from similar positions
Figure 4.7: The figure is a close-up of the previous figure for the period 2000–2005,
around the solar maximum of 2001. As shown by Quémerais et al. (2006) the
total solar ionization rate is close to uniform for all latitudes at solar maximum.
The orange model corresponds to medium isotropic ionization and the blue model
corresponds to strong isotropic ionization. The data show that the minimum of
IPH intensity is found almost 2 years after the solar maximum of 2001. This delay
is caused by the latency in the reaction of the ionization processes
(close to 1 AU) and at the same time. In that case the comparison is straight-
forward. The result of this comparison is that the SWAN calibration is directly
derived from the calibration of the STIS and GHRS instruments from the Hubble
Space Telescope.
However, it is not always possible to observe at the same time and position.
In that case, we can use a model combining accurate distributions of hydrogen
atoms in the heliosphere with radiative transfer computations to correct for the
differences in location of the observer. We have also shown that close to the Sun, it
is necessary to use three-dimensional models except for solar maximum conditions
when the ionizing fluxes from the Sun become more or less isotropic. Based on
comparisons with models described in Izmodenov et al. (2013, this volume), we have
cross calibrated ALICE and UVS Voyager 1 and 2. We have also cross calibrated
SOHO SWAN with both instruments.
The calibration factors derived here for Voyager 1 and 2 UVS are different from
the ones published by Hall (1992). However, for Voyager 2, the difference is simply
explained by the change in values of the solar illuminating flux. In 1992, the mean
solar Lyman-α flux was assumed to be smaller than current estimates based on
SORCE SOLSTICE values (Woods et al. 2009).
Bibliography 159
In the models used here, the solar illuminating flux is derived from SORCE
SOLSTICE measurements. However, we use a correction factor of 0.9 Å−1 to com-
pute the line center flux. This was derived from measurements made by SOHO
SUMER (Emerich et al. 2005). This correction is assumed to be independent of
solar activity. This assumption is likely to be inexact and future works will have
to include an updated correction factor.
Future works will extend this analysis. In the last few years, we have made
many mutual observations between various UV instruments measuring the inter-
planetary background. For instance, various cross-calibrations have been performed
between SWAN and the Spectroscopy for Investigation of Characteristics of the At-
mosphere of Venus SPICAV instrument on the Venus-Express spacecraft (Bertaux
et al. 2007). Similarly, we have performed cross-calibrations measurements be-
tween the Mercury Atmospheric and Surface Composition Spectrometer (MASCS)
on the MErcury Surface, Space ENvironment, GEochemistry, and Ranging (MES-
SENGER) spacecraft (McClintock and Lankton 2007) and SPICAV. This will allow
us to link these two instruments to the others that we presented here. And finally,
we will link these results with the results shown in Snow et al. (2013, this volume)
by comparing the Spectroscopy for Investigation of the Characteristics of the At-
mosphere of Mars (SPICAM) and SPICAV observations of the same stars.
Bibliography
V.B. Baranov, Gasdynamics of the solar wind interaction with the interstellar medium.
Space Sci. Rev. 52, 89–120 (1990)
medium—numerical solution of self-consistent problem. J. Geophys. Res. 98, 15157 (1993)
L. Ben-Jaffel, O. Puyoo, R. Ratkiewicz, Far -ultraviolet echoes from the frontier between
the solar wind and the local interstellar cloud. Astrophys. J. 533, 924–930 (2000)
J-L. Bertaux, J.E. Blamont, Evidence for an extra-terrestrial Lyman-alpha emission, the
interstellar wind. Astron. Astrophys. 11, 200 (1971)
J-L. Bertaux, R. Lallement, V.G. Kurt, E.N. Mironova, Characteristics of the local in-
terstellar hydrogen determined from PROGNOZ 5 and 6 interplanetary Lyman-alpha
line profile measurements with a hydrogen absorption cell. Astron. Astrophys. 150, 1–20
(1985)
J-L. Bertaux et al., SWAN: A study of solar wind anisotropies on SOHO with Lyman alpha
sky mapping. Solar Phys. 162, 403–439 (1995)
J-L. Bertaux et al., SPICAV on Venus Express: three spectrometers to study the global
structure and composition of the Venus atmosphere. Planet. Space Sci. 55, 1673–1700
(2007)
P.W. Blum, H-J. Fahr, The distribution of interplanetary hydrogen. Astrophys. Lett. 6, 127
(1970)
J.C. Brandt, J.W. Chamberlain, Interplanetary gas. I. Hydrogen radiation in the night sky.
Astrophys. J. 130, 670 (1959)
M. Brasken, E. Kyrölä, Resonance scattering of Lyman alpha from interstellar hydrogen.
A.L. Broadfoot, B.R. Sandel et al., Ultraviolet spectrometer experiment for the Voyager
mission. Space Sci. Rev. 21, 183 (1977)
A.L. Broadfoot, S.K. Atreya, J-L. Bertaux, J.E. Blamont, A.J. Dessler, T.M. Donahue, W.T.
Forrester, D.T. Hall, F. Herbert, J.B. Holberg, Ultraviolet spectrometer observations of
Neptune and Triton. Science 246, 1459–1466 (1989)
M. Bzowski et al., in Cross-Calibration of Far UV Spectra of Solar System Objects and the
Heliosphere, ed. by E. Quémerais, M. Snow, R.M. Bonnet. ISSI Scientific Report Series,
SR-013 (2013)
J.T. Clarke, R. Lallement, J-L. Bertaux, E. Quémerais, HST/GHRS observations of the
interplanetary medium downwind and in the inner solar system. Astrophys. J. 448, 893
(1995)
J.T. Clarke, R. Lallement, J-L. Bertaux, H-J. Fahr, E. Quémerais, H. Scherer, HST/GHRS
observations of the velocity structure of interplanetary hydrogen. Astrophys. J. 499, 482
(1998)
G. de Toma, E. Quémerais, B.R. Sandel, Long-term variation of the interplanetary H Ly
alpha glow: Voyager measurements and implications for the solar H Ly alpha irradiance.
Astrophys. J. 491, 980 (1997)
C. Emerich, P. Lemaire, J-C. Vial, W. Curdt, U. Schühle, K. Wilhelm, A new relation be-
tween the central spectral solar H I Lyman α irradiance and the line irradiance measured
by SUMER/SOHO during the cycle 23. Icarus 178, 429–433 (2005)
P. Gangopadhyay, V.V. Izmodenov, D.E. Shemansky, M.A. Gruntman, D.L. Judge, Reap-
praisal of the Pioneer 10 and Voyager 2 Lyα intensity measurements. Astrophys. J. 628,
514–519 (2005)
G.R. Gladstone et al., New Horizons cruise observations of Lyα emissions from the inter-
planetary medium. In: Cross-Calibration of Far UV Spectra of Solar System Objects and
the Heliosphere, ed. by E. Quémerais, M. Snow, R.M. Bonnet. ISSI Scientific Report
Series, SR-013 (2013)
G. Gloeckler, E. Möbius, J. Geiss, M. Bzowski, S. Chalov, H. Fahr, D.R. McMullin, H.
Noda, M. Oka, D. Ruciński, Observations of the helium focusing cone with pickup ions.
D.T. Hall, Ultraviolet resonance radiation and the structure of the heliosphere. Dissertation,
University of Arizona, 1992
D.T. Hall, D.E. Shemansky, D.L. Judge, P. Gangopadhyay, M.A. Gruntman, Heliospheric
hydrogen beyond 15 AU - evidence for a termination shock. J. Geophys. Res. 98, 15185–
15192 (1993)
V.V. Izmodenov, M.A. Gruntman, Y.G. Malama, Interstellar hydrogen atom distribution
function in the outer heliosphere. J. Geophys. Res. 106, 10681–10690 (2001)
V.V. Izmodenov, Y.G. Malama, M.S. Ruderman, Solar cycle influence on the interaction of
the solar wind with local interstellar cloud. Astron. Astrophys. 429, 1069–1080 (2005)
V.V. Izmodenov, Y.G. Malama, M.S. Ruderman, Modeling of the outer heliosphere with
the realistic solar cycle. Adv. Space Res. 41, 318–324 (2008)
V.V. Izmodenov et al., in Cross-Calibration of Far UV Spectra of Solar System Objects
and the Heliosphere, ed. by E. Quémerais, M. Snow, R.M. Bonnet. ISSI Scientific Report
Series, SR-013 (2013)
H.U. Keller, K. Richter, G.E. Thomas, Multiple scattering of solar resonance radiation in
the nearby interstellar medium. ii. Astron. Astrophys. 102, 415–423 (1981)
R. Lallement, J.C. Reymond, J. Vallerga, M. Lemoine, F. Delaudier, J.-L. Bertaux, Model-
ing the interstellar-interplanetary helium 58.4 nm resonance glow: towards a reconcilia-
tion with particle measurements. Astron. Astrophys. 426, 875–884 (2004)
R. Lallement, E. Quémerais, J-L. Bertaux, S. Ferron, D. Koutroumpa, R. Pellinen, Deflec-
tion of the interstellar neutral hydrogen flow across the heliospheric interface. Science
307, 1447–1449 (2005)
R. Lallement, E. Quémerais, J-L. Bertaux, B.R. Sandel, V. Izmodenov, Voyager measure-
ments of hydrogen Lyman-α diffuse emission from the milky way. Science 334, 1665
(2011)
P. Lemaire, C. Emerich, W. Curdt, U. Schuehle, K. Wilhelm, Solar HI Lyman alpha full
disk profile obtained with the SUMER/SOHO spectrometer. Astron. Astrophys. 334,
1095–1098 (1998)
W.E. McClintock, M.R. Lankton, The mercury atmospheric and surface composition spec-
trometer for the MESSENGER mission. Space Sci. Rev. 131, 481–521 (2007)
D. Mihalas, Stellar Atmospheres (Freeman, San Francisco, 1970)
Bibliography 161
E. Möbius, M. Bzowski, S. Chalov, H-J. Fahr, G. Gloeckler, V.V. Izmodenov, R. Kallenbach,

R. Lallement, D. McMullin, H. Noda, Synopsis of the interstellar He parameters from
combined neutral gas, pickup ion and uv scattering observations and related consequences.
W.R. Pryor, J.M. Ajello, C.A. Barth, C.W. Hord, A.I.F. Stewart, K.E. Simmons, W.E.
McClintock, B.R. Sandel, D.E. Shemansky, The Galileo and Pioneer Venus ultraviolet
spectrometer experiments - solar Lyman-alpha latitude variation at solar maximum from
interplanetary Lyman-alpha observations. Astrophys. J. 394, 363–377 (1992)
W.R. Pryor, C.A. Barth, C.W. Hord, A.I.F. Stewart, K.E. Simmons, J.J. Gebben, W.E.
McClintock, S. Lineaweaver, J.M. Ajello, W.K. Tobiska, Latitude variations in interplan-
etary Lyman-α data from the Galileo EUVS modeled with solar He 1083 nm images.
Geophys. Res. Lett. 23, 1893–1896 (1996)
medium at Lyman alpha. Astron. Astrophys. 358, 353–367 (2000)
E. Quémerais, J.L. Bertaux, Radiative transfer in the heliosphere at Lyman α: comparison
of numerical and Monte Carlo simulations. Adv. Space Res. 13, 298–298 (1993)
E. Quémerais, J-L. Bertaux, 14-day forecast of solar indices using interplanetary Lyman α
background data. Geophys. Res. Lett. 29, 1018 (2002). doi:10.1029/2001GL013920
E. Quémerais, V.V. Izmodenov, Effects of the heliospheric interface on the interplanetary
Lyman alpha glow seen at 1 AU from the Sun. Astron. Astrophys. 396, 269–281 (2002)
E. Quémerais, B.R. Sandel, R. Lallement, J-L. Bertaux, A new source of Lyα emission
detected by Voyager UVS: heliospheric or galactic origin? Astron. Astrophys. 299, 249
(1995)
E. Quémerais, Y.G. Malama, B.R. Sandel, R. Lallement, J-L. Bertaux, V.B. Baranov, Outer
heliosphere Lyman α background derived from two-shock model hydrogen distributions:
application to the Voyager UVS data. Astron. Astrophys. 308, 279–289 (1996)
E. Quémerais, J-L. Bertaux, R. Lallement, B.R. Sandel, V.V. Izmodenov, Voyager 1/UVS
Lyman α glow data from 1993 to 2003: hydrogen distribution in the upwind outer helio-
sphere. J. Geophys. Res. 108, 8029 (2003). doi:10.1029/2003JA009871
E. Quémerais, R. Lallement, S. Ferron, D. Koutroumpa, J-L. Bertaux, E. Kyrölä, W.
Schmidt, Interplanetary hydrogen absolute ionization rates: retrieving the solar wind
mass flux latitude and cycle dependence with SWAN/SOHO maps. J. Geophys. Res.
111, A09114 (2006)
E. Quémerais, R. Lallement, D. Koutroumpa, P. Lamy, Velocity profiles in the solar corona
from multi-instrument observations. Astrophys. J. 667, 1229–1234 (2007)
E. Quémerais, V.V. Izmodenov, D. Koutroumpa, Y.G. Malama, Time dependent model of
the interplanetary Lyman α glow: applications to the SWAN data. Astron. Astrophys.
488, 351–359 (2008)
E. Quémerais, R. Lallement, B.R. Sandel, J.T. Clarke, Interplanetary Lyman α observations:
intensities from Voyagers and line profiles from HST/STIS. Space Sci. Rev. 143, 151–162
(2009)
G.J. Rottman, C.A. Barth, R.J. Thomas, G.H. Mount, G.M. Lawrence, D.W. Rusch, R.W.
Sanders, G.E. Thomas, J. London, Solar spectral irradiance, 120 to 190 nm, October 13,
1981 - January 3, 1982. Geophys. Res. Lett. 9, 587–590 (1982)
M. Snow et al., A new catalog of ultraviolet stellar spectra for calibration. In: Cross-
Calibration of Far UV Spectra of Solar System Objects and the Heliosphere, ed. by E.
Quémerais, M. Snow, R.M. Bonnet. ISSI Scientific Report Series, SR-013 (2013)
G.E. Thomas, The interstellar wind and its influence on the interplanet-
ary environment. Ann. Rev. Earth Planet. Sci. 6, A78-38764 16-42 (1978).
doi:10.1146/annurev.ea.06.050178.001133
G.E. Thomas, R.F. Krassa, OGO 5 measurements of the Lyman alpha sky background.
Astron. Astrophys. 11, 218 (1971)
M. Witte, Kinetic parameters of interstellar neutral helium. review of results obtained dur-
ing one solar cycle with the Ulysses/GAS-instrument. Astron. Astrophys. 426, 835–844
(2004)
T.N. Woods, W.K. Tobiska, G.J. Rottman, J.R. Worden, Improved solar Lyman α irradiance
modeling from 1947 through 1999 based on UARS observations. J. Geophys. Res. 105,
27195–27216 (2000)
T.N. Woods et al., Solar irradiance reference spectra (SIRS) for the 2008 whole heliosphere
interval (WHI). Geophys. Res. Lett. 36, L01101 (2009). doi:10.1029/2008GL036373
—5—
Lyman-α Models for LRO LAMP from

MESSENGER MASCS and SOHO
SWAN Data
Wayne R. Pryor∗
Central Arizona College, Coolidge, AZ, USA
(also at LASP, University of Colorado and Space Environment Technologies,
Palisades, CA, USA)
Gregory M. Holsclaw, William E. McClintock,

Martin Snow
Laboratory for Atmospheric and Space Physics, University of Colorado,
Boulder, CO, USA
Ronald J. Vervack, Jr.

Applied Physics Laboratory, The Johns Hopkins University, Laurel, MD, USA
G. Randall Gladstone
S. Alan Stern
Southwest Research Institute, Boulder, CO, USA
Kurt D. Retherford and Paul F. Miles

Abstract
From models of the interplanetary Lyman-α glow derived from observations by
the Mercury Atmospheric and Surface Composition Spectrometer (MASCS) inter-
planetary Lyman-α data obtained in 2009–2011 on the MErcury Surface, Space
ENvironment, GEochemistry, and Ranging (MESSENGER) spacecraft mission,
daily all-sky Lyman-α maps were generated for use by the Lunar Reconnaissance

164 5. LAMP Lyman-α Models
Orbiter (LRO) LAMP Lyman-Alpha Mapping Project (LAMP) experiment. These

models were then compared with Solar and Heliospheric Observatory (SOHO) Solar
Wind ANistropy (SWAN) Lyman-α maps when available. Although the empirical
agreement across the sky between the scaled model and the SWAN maps is ade-
quate for LAMP mapping purposes, the model brightness values best agree with
the SWAN values in 2008 and 2009. SWAN’s observations show a systematic de-
cline in 2010 and 2011 relative to the model. It is not clear if the decline represents
a failure of the model or a decline in sensitivity in SWAN in 2010 and 2011. MES-
SENGER MASCS and SOHO SWAN Lyman-α calibrations systematically differ
in comparison with the model, with MASCS reporting Lyman-α values some 30 %
lower than SWAN.
Introduction
The Lunar Reconnaissance Orbiter (LRO) LAMP Lyman-Alpha Mapping
Project (LAMP, Gladstone et al. 2010) experiment is designed to view lunar per-
manently shadowed regions (PSRs) that may contain ice and other volatiles using
the surface reflectance of faint ultraviolet (UV) light from interplanetary hydrogen
emissions and stars (Gladstone et al. 2010). We developed a time-dependent model
for the hydrogen Lyman-α illumination of the Moon that was merged with a stel-
lar illumination model for use in determining far-UV lunar albedos in LAMP data.
Preliminary LAMP results indicate that PSRs are spectrally distinct from neigh-
boring regions, perhaps due to ice deposits and porosity enhancement (Gladstone
et al. 2011). This paper describes the heliospheric Lyman-α model and how it
was tuned with MErcury Surface, Space ENvironment, GEochemistry, and Rang-
ing (MESSENGER) Mercury Atmospheric and Surface Composition Spectrometer
(MASCS) data (McClintock and Lankton 2007), then tested and linearly scaled
in brightness by comparisons with Solar and Heliospheric Observatory (SOHO)
SWAN (Solar Wind ANisotropy) instrument data (Bertaux et al. 1995).
Lyman-α Models
Hydrogen Lyman-α emissions in the inner solar system result from resonance
scattering of solar Lyman-α photons by slow neutral hydrogen. Hydrogen in the
inner heliosphere is dominated by interstellar wind hydrogen: neutral gas that
has penetrated the solar wind termination shock from the “upwind” direction. Hot
models consider processes affecting this hydrogen near the Sun but do not explicitly
consider outer heliospheric processes (Thomas 1978). We apply a hot model which
was previously used to model Cassini Ultraviolet Imaging Spectrograph (UVIS)
cruise Lyman-α data to describe the hydrogen distribution in the inner heliosphere
(Pryor et al. 2008). The following neutral hydrogen parameters are assumed for
large distances from the Sun (but inside the termination shock) in the model:
density n = 0.12 cm−3 to match Cassini UVIS brightness levels obtained with its lab
calibration soon after launch, effective velocity v = 20 km s−1 , and temperature T =
12000 K on the basis of SWAN hydrogen cell studies (Costa et al. 1999). For this
study the downwind flow direction is assumed to be toward 74.7◦ ecliptic longitude
Lyman-α Models 165
and −5.2◦ ecliptic latitude based on Ulysses Interstellar Gas Experiment (GAS)
measurements of neutral helium (Witte 2004). We have neglected the small offset
between hydrogen and helium flow directions (Lallement et al. 2005). The primary
loss processes affecting the slow neutral hydrogen distribution are charge-exchange
with solar wind protons and solar Extreme UltraViolet (EUV) photoionization.
Solar wind proton fluxes have been extracted from the NASA Goddard Space Flight
Center’s OMNI dataset (a multi-spacecraft compilation of near-Earth solar wind
magnetic field and plasma measurements) through their online website OMNIWeb
(King and Papitashvili 2005), and have been time-averaged to reflect the slow
(4.22 AU/year) motion of hydrogen atoms through the solar system. The solar
EUV photoionization rates for hydrogen are from Space Environment Technologies’
Solar Irradiance Platform (SIP) version 2.37, a solar database at http://www.
spacewx.com (Tobiska et al. 2000). They have been time-averaged as follows: solar
wind charge-exchange data, photoionization data, and radiation pressure data are
averaged over 1–8 year intervals, with 1 year used far upwind, 1–2 years used
4.22–8.44 AU upwind of the Sun, 2–3 years used 0–4.22 AU upwind, 3–4 years used
0–4.22 AU downwind of the Sun, 4–5 years used 4.22–8.44 AU downwind, 5–6 years
8.44–12.66 AU downwind, 6–7 years used 12.66–16.88 AU downwind, 7–8 years used
16.88–21.10 AU downwind, and 8 years used beyond 21.10 AU downwind. The idea
is that downwind volume elements contain hydrogen that has had a longer period
of interaction with the solar wind, solar EUV, and the solar Lyman-α radiation
pressure force (Pryor et al. 1998a).
Slowly moving interstellar hydrogen atoms scatter solar Lyman-α photons, with
solar flux values taken here from the solar Lyman-α composite available at the
Laboratory for Atmospheric and Space Physics (LASP) website for the Thermo-
sphere Ionosphere Mesosphere Energetics and Dynamics (TIMED) mission Solar
EUV Experiment (SEE) (Woods et al. 2000). The relationship between integrated
Lyman-α line flux and the line-center Lyman-α line flux is taken from Emerich
et al. (2005). Lyman-α radiative transfer is assumed to be dominated by single
scattering near 1 AU, with a significant multiple scattering correction (Ajello et al.
2004) indicated by full exponential integral multiple scattering calculations (Hall
1992). The brightness of a particular sky direction on a particular day is deter-
mined from an integral over volume elements along a line of sight. A solar Lyman-α
value is assigned to each volume element on the basis of its solar longitude and the
brightness value of that solar longitude when measured at the Earth. The solar
measurements used are those closest in time to the observation date, sometimes up
to 13 days earlier or later on the basis of a 27-day solar rotation as seen from the
Earth. The assumption is that solar active regions and their Lyman-α fluxes change
slowly compared with 13 days. The model uses a simple formalism (Cook et al.
1981) for the modest solar Lyman-α falloff with solar latitude (Pryor et al. 1992).
The quiet sun polar Lyman-α resonance scattering coefficient (g-factor value) at
1 AU is taken as gquiet = 1.7 × 10−3 s−1 .
From the equatorial g-factor value, Lyman-α g-values at other latitudes (glat )
are estimated as:
gpole = gquiet + 0.35 ∗ (gave − gquiet ) (5.1)
where gave is the 27-day equatorial average and gpole is the estimated polar g-value.
fpara = (gfact − gquiet )/(3.6 ∗ gquiet ) (5.2)
is the fraction of the equatorially-viewed disk covered by plage and gfact is the
equatorial g-value on a given day, and
fperp = (gpole − gquiet )/(3.6 ∗ gquiet ) (5.3)
is the fraction of the polar-viewed disk covered by plage. The fraction of the disk
covered by plage as viewed at a particular latitude is given by
flat = fperp + (fpara − fperp) ∗ (cos2 (latitude)), (5.4)
and finally,
glat = gquiet (1 + 3.6 ∗ flat ) (5.5)
gives the estimated g-value at that latitude (Cook et al. 1981; Pryor et al. 1992).
MESSENGER MASCS
The MASCS instrument on MESSENGER consists of an Ultraviolet-Visible
Spectrometer (UVVS) and a Visible-Infrared Spectrograph (VIRS). UVVS cov-
ers the wavelength ranges of the far ultraviolet (115–180 nm), middle ultraviolet
(160–320 nm), and visible (250–600 nm), with an average spectral resolution of
0.6 nm (McClintock and Lankton 2007). UVVS uses an entrance slit with a field of
view of 1◦ by 0.04◦ for its airglow studies. MASCS UVVS has been studying Mer-
cury’s exosphere and surface reflectance properties initially from flyby observations
(McClintock et al. 2009) and more recently from orbit.
The radiometric sensitivity of the far-ultraviolet (FUV) channel of MASCS was
determined prior to the launch of the MESSENGER spacecraft. Measurements
were conducted in vacuum by observing the output from a monochromator with
both MASCS and a photomultiplier detector, which itself was calibrated against a
National Institute of Science and Technology (NIST) photodiode. Flight observa-
tions of stellar sources provide an opportunity to validate the MASCS radiometric
calibration. An adjustment of approximately 20 % was applied to the MASCS
spectral sensitivity in order to provide agreement with the Solar Radiation and
Climate Experiment (SORCE) satellite SOLar STellar Irradiance Comparison Ex-
periment (SOLSTICE)-measured irradiance of the star Alpha Virginis (Spica) in
the wavelength range 130–190 nm (Snow et al. 2013, this volume) and near Lyman-
α [M. A. Snow, private communication, 2012].
After launch on August 3, 2004, MESSENGER had an extended cruise phase
of its mission in the inner heliosphere until its orbital insertion at Mercury on
March 18, 2011. During part of the cruise phase, in 2009–2011, Lyman-α observa-
tions were obtained by UVVS over great circles as the spacecraft rolled at roughly
right angles to the spacecraft-Sun line at distances from the Sun ranging from 0.30
to 0.57 Astronomical Units (AU). The locations of MESSENGER during these
observations are indicated in Fig. 5.1. Because UVVS is a spectrometer, it is pos-
sible to separate the spectrum into heliospheric Lyman-α emissions (excess counts
MESSENGER MASCS 167
0.6 0.6 AU
2011-045
0.5 AU
0.4 0.4 AU
0.3 AU
0.2
Y_EC (AU)
–0.0 Sun
LISM flow
–0.2
–0.4
2009-124
–0.6
–0.6 –0.4 –0.2 –0.0 0.2 0.4 0.6
X_EC (AU)
Figure 5.1: MESSENGER spacecraft locations in heliocentric ecliptic coordinates

during interplanetary Lyman-α observations. In this coordinate system the x-axis
points towards the location on the celestial sphere of the Sun as seen from the
Earth at northern vernal equinox. Symbols are colored by observation time. The
flow direction of the local interstellar medium (LISM) is indicated by an arrow
within 0.3 nm of the Lyman-α 121.6 nm emission) and longer-wavelength starlight.

By combining these observations, a Lyman-α map of the full sky was obtained.
MASCS Observations and Model Comparisons

We applied the Lyman-α model described above to the MESSENGER MASCS
data, assuming the solar flux and solar wind databases are accurate and the inter-
stellar gas flow parameters are well-known from other sources (see Izmodenov et al.
2013, this volume). Currently, the most unconstrained part of the problem is the
changing solar wind latitudinal structure (see Bzowski et al. 2013, this volume).
Lyman-α maps are strongly modified by the spatial distribution of solar wind mass
flux. This was shown from the study of Prognoz-5 data from 1977 (Bertaux et al.
1996). These data showed a narrow dip, called the “groove,” in the Lyman-α
brightness maps—mainly in the upwind direction—caused by an enhanced solar
wind flux of protons at low solar latitudes (Bertaux et al. 1996). When Ulysses
made a rapid south-north sweep of solar latitudes in 1994–1995, Summanen et al.
(1997) showed that the measurements of proton flux enhancements at low solar
latitudes made by Ulysses Solar Wind Observations Over the Poles of the Sun
(SWOOPS) were suitable for explaining the Prognoz results. SWAN observations
initially showed a Lyman-α groove, as Prognoz had (Bertaux et al. 1997), but later
Summanen et al. (2002) reported that the solar wind proton flux changed to a more
isotropic distribution as the Sun changed from solar minimum to solar maximum
Figure 5.2: MESSENGER interplanetary cruise Lyman-α data (top panel) and a
corresponding groove model (lower panel). The model was scaled by a factor of
0.66 to best agree with the data
conditions. Later, Pryor et al. (2003) assessed solar wind latitudinal changes in
Ulysses data from 1990 to 2001 and Quémerais et al. (2006) analyzed the changes
in SOHO SWAN data from 1996 to 2005. The SWAN data showed that the groove
was a narrow well-defined structure in the 1996 solar minimum observations, van-
ished in the 2002 solar maximum data, and reformed as a broader structure by the
next solar minimum in 2005.
We initially tried to fit MASCS data with a Ulysses-derived solar wind groove
model appropriate to the previous solar minimum and detailed by Pryor et al.
(1998b) and McComas et al. (1999). Figure 5.2 shows that this model has too
narrow a groove to describe the current epoch. The width of the groove may be
related to the varying tilt of the heliospheric neutral current sheet.
This Lyman-α model also has a built-in option (Witt et al. 1979) to vary the
solar wind with latitude. When this option is used, the solar latitude distribution
of the lifetime, t, of H atoms against charge-exchange with solar wind protons is
characterized by an asymmetry parameter, A, as follows:
t(latitude) = t(equator)/(1 − A sin2 (latitude)). (5.6)
A = 0 would characterize a spherically symmetric solar wind; typically somewhat

larger values are found reflecting increased hydrogen atom survival lifetimes away
from the solar equator. To model the MASCS data we performed a two-parameter
grid search with separate asymmetry parameter values AN and AS for the northern
(N ) and southern (S) hemispheres of the Sun as was used for Ulysses data by Pryor
et al. (2003). The best agreement (in a root-mean square (RMS) fit sense) of the
SOHO SWAN 169
linearly scaled model with the MASCS data is found for AN = 0.8 and AS = 0.5
(Fig. 5.3). This solution indicates that there is a north-south asymmetry with a
longer hydrogen lifetime in the north, explaining the brighter northern Lyman-α
measurements seen in the MASCS data of Fig. 5.3. A scaling factor was needed to
match our model to the MASCS data: the best-fitting model values for the IPH
brightness must be multiplied by 0.62 to match MESSENGER MASCS Lyman-α
data.
The latitudinal behavior of the solar wind lifetime derived from the MESSEN-
GER MASCS data, assuming AN = 0.8 and AS = 0.5, is shown in Fig. 5.4.
SOHO SWAN
SWAN, mounted on the SOHO spacecraft, provides regular Lyman-α maps of
the sky on a 1◦ by 1◦ grid (Bertaux et al. 1997). The SWAN instrument, described
by Bertaux et al. (1995) uses twin two-mirror periscopes mounted on opposite sides
of the spacecraft (+Z and −Z) to construct these maps. The detectors measure 5◦
by 5◦ on the sky at a time, using a Hamamatsu 5×5 multi-anode microchannel plate
array detector tube with a MgF2 window. The CsI photocathode of the detector
renders the instrument solar-blind at wavelengths above 200 nm, but leaves the
instrument with sensitivity to UV stars. Each unit has four unilluminated side
pixels suitable for instrumental dark current subtractions.
The initial published sensitivity of SWAN to Lyman-α radiation in photometer
mode was 0.75 counts per sec per Rayleigh per 1◦ by 1◦ pixel (Bertaux et al. 1995).
This ground calibration was based on comparison with an aluminum standard
photodiode from the National Institute of Standards and Technology (NIST) at a
synchrotron light source at Orsay. In-flight cross-calibration with Hubble Space
Telescope Goddard High Resolution Spectrometer (GHRS) data from March 9,
1996 led to slightly revised values that depended on which sensor was used and
its high-voltage setting (Bertaux et al. 1997). SWAN Lyman-α values cited in this
paper were the current values provided by the SWAN team in 2011.
SWAN Observations and Model Comparisons

The SWAN team provided calibrated SWAN maps from 2008 to 2011. A sam-
ple SWAN image with a band of stars visible along the galactic plane is shown
in Fig. 5.5. The blacked-out region in the figure is the region inaccessible to the
periscope because of spacecraft obscurations and the solar avoidance zone. SWAN’s
L1 orbital location places it in an ideal setting to study heliospheric Lyman-α with-
out geocoronal Lyman-α contamination due to the Earth’s hydrogen exosphere.
A SWAN image for a single day and the corresponding all-sky model are shown
in Fig. 5.6. The dark band snaking across the images is the galactic plane, which
has been filtered from these comparisons. The Bright Star Catalogue (Hoffleit and
Warren 1991) has also been used to filter 929 selected bright O and B stars brighter
than magnitude 6, although some stars remain in the comparisons.
The quality of the model fits to the SWAN maps can be assessed with a root
mean square (RMS) analysis of the effect of variations in the asymmetry parameter.
Figure 5.3: MESSENGER interplanetary cruise Lyman-α data (top panel) and a
corresponding scaled model with AN = 0.8 and AS = 0.5 (lower panel). The result
of the model was scaled by a factor of 0.62 to best agree with the data
Figure 5.4: Best-fitting latitudinal behavior of the lifetime of interplanetary hy-

drogen atoms against solar-wind charge-exchange (thin line). Also shown is the
groove profile used to construct the model in Fig. 5.3 (thick line)
Conclusions 171
Figure 5.5: Sample SWAN Lyman-α all-sky map from March 6, 2011
Before assessing the RMS fit, the sum of the model values is linearly scaled to match
the sum of the data values. Figure 5.7 shows the RMS fits versus time from 2008
to 2011. Although the RMS fits do not change markedly over the orbit in 2010
or 2011, there is some seasonal variation in the fit in 2008 and 2009, suggesting
the solar wind fitting parameters derived from MASCS (mostly from 2010 to 2011
data) may no longer be appropriate. The agreement is adequate for our objective
of providing illumination values for the lunar poles for the LRO LAMP work, since
LRO was launched on June 18, 2009 and mapping operations began on September
15, 2009. Of greater interest is the trend in the derived scale factors shown in
Fig. 5.7. While the scale factor values seem steady and near 0.9 in 2008 and 2009,
a systematic decline is seen in 2010 and 2011. If the model were perfect, this
decline would suggest some degradation in the SWAN sensitivity. Alternatively,
this could be an indicator of model problems or some combination of data and
model problems.
Conclusions
We have identified a substantial offset between the MESSENGER MASCS val-
ues and the model: the best model used here must be scaled by a factor of 0.62 to
match the MASCS data. This same model is a better absolute fit to SWAN values
for 2008 and 2009 than to SWAN values in 2010 and 2011. Other techniques may
be valuable for assessing the hydrogen densities and deciding which brightnesses
are best calibrated. Some examples of such techniques are as follows. Damping at
increasing heliospheric distances of solar Lyman-α 27-day waves due to multiple
Figure 5.6: Scaled model output with AN = 0.8 and AS = 0.5 (top panel) and
the corresponding SWAN map for January 1, 2010 (bottom panel). The model
was scaled by a factor of 0.90 to best agree with the data. The diagram shows
the location of the Earth (E) and Sun and indicates the interstellar wind upwind
(u) and downwind (d) directions. The small diagram below the panels shows the
relative locations of the Earth, spacecraft, and the interstellar wind directions
scattering effects is observed (Pryor et al. 2008), indicating a hydrogen density

near the termination shock 0.085–0.095 cm−3 with a poorly determined uncer-
tainty. This value is slightly lower than the 0.12 cm−3 density value used in
these model runs that generally fit the SWAN maps from 2008 to 2009. Bzowski
(2008) estimated the neutral hydrogen density at the termination shock at 0.087 ±
0.022 cm−3 based on pickup proton density fluxes measured at 5 A.U. by Ulysses So-
lar Wind Ion Composition Spectrometer (SWICS). The MASCS Lyman-α bright-
nesses are consistent with a lower density of 0.12 cm−3 times a scale factor of 0.62,
or 0.074 cm−3 .
Our tuned model for MASCS was developed largely from 2010 data, so we can
best address the SWAN/MASCS cross-calibration for that year. The average scale
factor to apply to the model to match the SWAN data in 2010 is 0.81. The derived
scale factor to apply to the model to match the MASCS data is 0.62. Using the
model to compare the Lyman-α brightnesses, we estimate that SWAN values for
2010 are a factor of 0.81/0.62 = 1.3 times brighter than MASCS values.
Conclusions 173
Figure 5.7: Top panel shows the derived RMS fits of the scaled model, with AN =
0.8 and AS = 0.5, to the filtered SWAN maps. The bottom panel shows the scale
factors applied to the model to agree with the data
In summary, MESSENGER MASCS provided unusually clean heliospheric

Lyman-α data on the solar wind latitudinal dependence during the cruise phase
of its mission to Mercury. This latitudinal information proved generally consistent
with SWAN observations that provide almost full-sky coverage on a regular basis,
but with a notable stellar background. These datasets were used to validate a
time-varying all-sky Lyman-α illumination model for LRO LAMP lunar studies.
We identified evidence for a possible decline in SWAN sensitivity (or a modeling
issue), and also identified a substantial offset between the Lyman-α calibrations of
SOHO SWAN and MESSENGER MASCS.
Acknowledgements
Solar Irradiance Platform historical irradiances were provided courtesy of W.
Kent Tobiska and Space Environment Technologies. The historical irradiances have
been developed with partial funding from the NASA Upper Atmosphere Research
Satellite (UARS), TIMED, and ESA-NASA SOHO missions. We also acknowl-
edge use of the NASA Goddard Space Flight Center Space Physics Data Facility’s
OMNIWeb service and OMNI data. Calibrated SWAN data were provided by
Eric Quémerais (Laboratoire Atmosphères, Milieux, Observations Spatiales (LAT-
MOS)/Centre National de la Recherche Scientifique (CNRS)) and Stephane Ferron
(ACRI-ST). The MESSENGER contributions to this work are supported by the
NASA Discovery Program under contracts NAS5-97271 to The Johns Hopkins Uni-
versity Applied Physics Laboratory and NASW-00002 to the Carnegie Institution
of Washington. R. J. Vervack is supported by the MESSENGER Participating Sci-

entist Program. We thank the Lunar Reconnaissance Orbiter project and project
team at NASA Goddard Space Flight Center for conducting the LAMP atmo-
spheric observations. The authors would also like to thank the International Space
Science Institute, Bern, Switzerland for their support of the working group. This
work was supported under a subcontract to contract NNG05EC87C from NASA.
Bibliography
J.M. Ajello et al., Observations of interplanetary Lyman-alpha with the Galileo ultraviolet
spectrometer: multiple scattering effects at solar maximum. Astron. Astrophys. 289,
283–303 (1994)
J-L. Bertaux et al., SWAN: a study of solar wind anisotropies on SOHO with Lyman alpha
sky mapping. Sol. Phys. 162, 403–439 (1995)
J-L. Bertaux, E. Quémerais, R. Lallement, Observations of a sky Lyman alpha groove
related to enhanced solar wind mass flux in the neutral sheet. Geophys. Res. Lett. 23,
3675–3678 (1996)
J-L. Bertaux et al., First results from SWAN Lyman α solar wind mapper on SOHO. Sol.
Phys. 175, 737–770 (1997)
M. Bzowski, Density of neutral interstellar hydrogen at the termination shock from Ulysses
pickup ion observations. Astron. Astrophys. 491, 7–19 (2008)
M. Bzowski et al., Solar parameters for modeling the interplanetary background. In: Cross-
J.W. Cook et al., Latitudinal anisotropy of the solar far ultraviolet flux: effect on the
Lyman-alpha sky background. Astron. Astrophys. 97, 394–397 (1981)
J. Costa et al., Heliospheric interstellar H temperature from SOHO/SWAN H cell data.
C. Emerich et al., A new relation between the central spectral solar H I Lyman alpha
irradiance and the line irradiance measured by SUMER/SOHO during the cycle 23. Icarus
178, 429–433 (2005)
G.R. Gladstone et al., LAMP: The Lyman alpha mapping project on NASA’s lunar recon-
naissance orbiter mission. Space Sci. Rev. 150, 161–181 (2010)
G.R. Gladstone et al., Far-ultraviolet reflectance properties of the Moon’s permanently
shadowed regions. J. Geophys. Res. 117, E00H04 (2011). doi:10.1029/2011JE003913
D.T. Hall, Ultraviolet resonance radiation and the structure of the heliosphere Dissertation,
University of Arizona, 1992
D. Hoffleit, W.H. Warren Jr., The Bright Star Catalogue, 5th Revised edn. (National Space
Science Data Center/Astronomical Data Center, 1991)
V.V. Izmodenov et al., Distribution of interstellar hydrogen atoms in the heliosphere and
backscattered solar Lyman-α. In: Cross-Calibration of Far UV Spectra of Solar System
Objects and the Heliosphere, ed. by E. Quémerais, M. Snow, R.M. Bonnet. ISSI Scientific
Report Series, SR-013 (2013)
J.H. King, N.E. Papitashvili, Solar wind spatial scales in and comparisons of hourly
Wind and ACE plasma and magnetic field data. J. Geophys. Res. 110, A02104 (2005).
doi:10.1029/2004JA010649
R. Lallement et al., Deflection of the interstellar neutral hydrogen flow across the heliospheric
interface. Science 307, 1447–1449 (2005)
W.E. McClintock, M.R. Lankton, The mercury atmospheric and surface composition spec-
trometer for the MESSENGER mission. Space Sci. Rev. 131, 481–521 (2007)
W.E. McClintock et al., MESSENGER observations of Mercury’s exosphere: detection of
magnesium and distribution of constituents. Science 324, 610–613 (2009)
D.J. McComas et al., Measurements of variations in the solar wind-interstellar hydrogen
charge exchange rate. Geophys. Res. Lett. 26, 2701–2704 (1999)
Bibliography 175
W.R. Pryor et al., The Galileo and Pioneer Venus ultraviolet spectrometer experiments:
solar Lyman-alpha latitude variation at solar maximum from interplanetary Lyman-alpha
observations. Astrophys. J. 394, 363–377 (1992)
W.R. Pryor, S.J. Lasica, A.I.F. Stewart, D.T. Hall, S. Lineaweaver, W.B. Colwell, J.M.
Ajello, O.R. White, W.K. Tobiska, Interplanetary Lyman alpha observations from Pioneer
Venus over a solar cycle from 1978 to 1992. J. Geophys. Res. 103, 26833–26849 (1998a)
W.R. Pryor, M. Witte, J.M. Ajello, Interplanetary Lyman alpha remote sensing with the
Ulysses interstellar neutral gas experiment. J. Geophys. Res. 103, 26813–26831 (1998b)
in the three-dimensional heliosphere over the solar cycle. J. Geophys. Res. 108, A108034
(2003). doi:10.1029/2003JA009878
W.R. Pryor et al., Radiation transport of heliospheric Lyman-alpha from combined Cassini
and Voyager data sets. Astron. Astrophys. 491, 21–28 (2008)
E. Quémerais et al., Interplanetary hydrogen absolute ionization rates: retrieving the solar
wind mass flux latitude and cycle dependence with SWAN/SOHO maps. J. Geophys.
Res. 111, A09114 (2006). doi:10.1029/2006JA011711
T.R. Summanen, R. Lallement, E. Quémerais, Solar wind proton flux latitudinal variations:
comparisons between Ulysses in situ data and indirect measurements from interstellar
Lyman alpha mapping. J. Geophys. Res. 102, 7051–7062 (1997)
T.R. Summanen et al., Interplanetary Lyman alpha observations of SWAN during the rising
phase of the 23rd solar cycle. Adv. Space Res. 29, 457–462 (2002)
W.K. Tobiska et al., The SOLAR2000 empirical solar irradiance model and forecast tool. J.
Atmos. Sol. Terr. Phys. 62, 1233–1250 (2000)
N. Witt, J.M. Ajello, P.W. Blum, Solar wind latitudinal variations deduced from Mariner
10 interplanetary H (1216 A) observations. Astron. Astrophys. 73, 272–281 (1979)
M. Witte, Kinetic parameters of interstellar neutral helium. review of results obtained dur-
ing one solar cycle with the Ulysses/GAS-instrument. Astron. Astrophys. 426, 835–844
(2004)
T.N. Woods, W.K. Tobiska, G.J. Rottman, J.R. Worden, Improved solar Lyman alpha
irradiance modeling from 1947 through 1999 based on UARS observations. J. Geophys.
Res. 105, 27195–27215 (2000)
—6—
New Horizons Cruise Observations

of Lyman-α Emissions from the
Interplanetary Medium
G. Randall Gladstone∗
Southwest Research Institute,
San Antonio, TX, USA
S. Alan Stern
Southwest Research Institute,
Boulder, CO, USA
Wayne R. Pryor
Central Arizona College,
Coolidge, AZ, USA
Abstract
Initial results are presented for observations of interplanetary Lyman-α and
Lyman-β emissions in the outer solar system obtained by the Alice ultraviolet
spectrograph on the New Horizons spacecraft (the first new such data from outside
the orbit of Saturn since the Voyager spacecraft). The observations consist of 6◦ ×
360◦ great-circle swaths on the sky, centered on the ecliptic direction λ = 51.3◦ , β =
44.8◦ , which passes within ∼ 33◦ of the upstream and downstream directions of the
interstellar wind. To date, three such scans have been acquired: on October 7, 2007,
October 18, 2008, and June 19, 2010 (at which times the New Horizons spacecraft
was 7.6, 11.3, and 17.0 AU from the Sun, respectively). The data compare fairly
well with model simulations, although the brightness of interplanetary Lyman-α
emissions falls off more slowly than expected with radial distance from the Sun.
The ratio of Lyman-α/Lyman-β brightnesses in the interplanetary medium agrees
well with previous measurements by the Voyager ultraviolet spectrometers.
Introduction
The interplanetary medium (IPM) and nearby interstellar medium (ISM) may be
fruitfully studied by measuring the sky background Lyman-α (121.6 nm) emission,

178 6. New Horizons IPM Observations
at nearly any location in the solar system (e.g. Bertaux and Blamont 1971; Thomas
and Krassa 1971; Fahr 1974; Adams and Frisch 1977; Holzer 1977; Thomas 1978;
Ajello et al. 1994; Murthy et al. 1999; Quémerais et al. 2009; Izmodenov 2009).
These emissions result from the scattering of solar Lyman-α photons by neutral
hydrogen atoms of the ISM as they pass by the Sun, and present an interesting
problem in resonance line radiative transfer (e.g. Brandt and Chamberlain 1959;
Meier 1977; Keller et al. 1981; Braskén and Kyrölä 1998; Quémerais and Bertaux
1993; Quémerais 2000).
The characteristic brightness of IPM Lyman-α in the vicinity of the Earth is
about 600 Rayleighs—i.e., about 6 × 108 photons cm−2 s−1 over the entire sky
(Ajello et al. 1987). For comparison, the direct Lyman-α flux from the Sun is
currently about 3–5 ×1011 photons cm−2 s−1 at the Earth. Although the direct
flux is a much more important source for, e.g., photochemistry in the inner solar
system, the interplanetary source falls off much more slowly than the direct flux
(Ajello et al. 1987), so that the two sources are of comparable strength at the orbit
of Neptune (Broadfoot et al. 1989), and the interplanetary source dominates from
there on out into the Kuiper Belt.
The general behavior of the scattered resonance line radiation may be approx-
imated by combining simple expressions for the optically thin and optically thick
limits. The backscattered IPM brightness BIPM may be approximated under opti-
cally thick and optically thin conditions as
1 F
BIPM ≈ (optically thick)
2 r2
1 F ΔλIPM
BIPM ≈ ksca (r) (optically thin),
2 r Δλ
where r is the distance from the Sun, F is the solar line flux, ΔλIPM and Δλ
are the line-widths of the appropriate resonance line in the interplanetary medium
and emitted by the Sun, respectively, and ksca is the line-center extinction due to
scattering. By simply combining these two limits, we obtain an expression useful
for approximating the backscattered mean intensity in both optically thick and
optically thin regimes:

1 F r ksca (r) ΔλIPM
Δλ
BIPM ≈ .
2 r 2 1 + r ksca (r) Δλ IPM
Δλ
The solar Lyman-α line is quite broad (FWHM ∼ 0.1 nm ≈ 250 km/s) com-
pared to the width (T ∼ 12, 000 K ≈ 14 km/s) and offsets (from ∼ −22 km/s to
∼ 22 km/s) of the scattered emissions from the IPM hydrogen, so that the illumi-
nation can be taken to be the same anywhere in the solar system (after accounting
for extinction on the way out from the Sun (Wu and Judge 1979). At a typ-
ical interplanetary H-atom number density of nH ∼ 0.1 cm−3 and temperature
of TH ∼ 12, 000 K (e.g. Izmodenov 2009), the line-center scattering cross section
σ0 is
πe2 f
σ0 = √
me c πΔνD
New Horizons observations 179
πe2 f λ0
= √
me c π 2kTH /mH
√
= (0.02654) (0.4162/ π) (1.216 × 10−5 /1.41 × 106 )
= 5.4 × 10−14 cm2 ,
so that the path length for unit optical depth at line center is
1
L =
nH σ0
1
= cm ∼ 12 AU.
(0.1) (5.4 × 10−14 )
In the above equations, e is the elementary charge, me is the electron mass, c is

the speed of light, f is the line oscillator strength, ΔνD is the IPM Doppler line
width, λ0 is the wavelength at line center, and k is the Boltzmann constant.
The ISM wind comes from a direction (in ecliptic coordinates) of α = 252.5◦
and β = 8.9◦ (e.g., Lallement et al. 2010), at a velocity of vH ∼ 22 km/s (e.g.
Quémerais et al. 2009). Interestingly, the radiation pressure of the solar Lyman-α
line on interstellar H atoms nearly balances the force of gravity, so that they pass
through the solar system under nearly force-free conditions. The lifetime of the
H atoms decreases near Sun, primarily due to charge exchange with much faster
solar wind protons, so that a density cavity forms, weighted toward the downwind
side. This density cavity imparts a brightness modification to the IPM Lyman-α
emissions in the inner solar system; near the Earth, the IPM Lyman-α brightness
is nearly twice as large looking upwind as looking downwind.
New Horizons observations

The New Horizons (NH) mission currently en route to the Pluto system and
the Kuiper Belt (Weaver and Stern 2008) provides an excellent platform for ob-
servations of IPM Lyman-α emissions in the outer solar system. During some of
its annual checkouts (ACOs) the NH-Alice ultraviolet spectrograph on the New
Horizons spacecraft is used to observe the IPM Lyman-α signal along a fixed great
circle on the sky. To date, IPM observations have been made during ACO-1, ACO-
2, and ACO-4. Figure 6.1 shows where these observations occurred in the plane
of the ecliptic, as the New Horizons spacecraft follows its trajectory toward the
Pluto system (the closest approach to Pluto is scheduled for 14 July 2015, at 11:50
UTC).
The NH-Alice ultraviolet spectrograph is comprised of a telescope, a Rowland-
circle spectrograph, a double-delay-line (DDL) microchannel plate (MCP) detector
at the focal plane, and associated electronics and mechanisms (Stern et al. 2008).
The entrance slit has two contiguous sections: a 2◦ × 2◦ “Box” and a 0.1◦ × 4◦
“Slot”. The bandpass is 52–187 nm with a filled-slit spectral resolution of 0.9 nm.
Figure 6.2 shows the wavelength dependence of the NH-Alice effective area, as
established through in-flight observations of UV-bright stars (whose fluxes have
been determined by IUE measurements—see also Snow et al. (2013) this volume).
Figure 6.1: During its annual checkouts (ACOs) the NH-Alice ultraviolet spectro-
graph is used to observe the IPM Lyman-α signal along a fixed great circle on the
sky. Such observations can be used to provide a measurement of the distribution
of interstellar hydrogen atoms flowing through the solar system. The trajectory
of New Horizons takes the spacecraft approximately radially upwind at a rate of
∼3 AU/year. Currently, we have IPM scans for ACOs 1, 2, and 4 on October 7,
2007, October 18, 2008, and June 19, 2010, at solar ranges of rNH = 7.6, 11.3, and
17.0 AU, respectively. These observations will continue at approximately yearly
intervals for as long as possible (at least out to 32 AU)
The filled-slit IPM Lyman-α emissions fall on bare, KBr-coated, and CsI-coated
sections of the MCP. The effective area for IPM Lyman-α in the Box is 0.056 cm2 ,
while in the Slot it is 0.007 cm2 . The overall sensitivity SIPM of NH-Alice to IPM
Lyman-α is given by
106
SIPM = [AEFF (Box) ω(Box) + AEFF (Slot) ω(Slot)]
4π
106
= [0.056 (2.0π/180)2 + 0.007 (π/180)2 (0.1)(4.0)]
4π
= 5.43 + 0.07 = 5.5 counts/s/R,
where AEFF is the NH-Alice effective area and ω represents solid angle.
Figures 6.3–6.5 show the great circle swath used during ACO-1, ACO-2, and
ACO-4, respectively, along with the locations of UV-bright stars and the Sun, in
ecliptic coordinates. The IPM swaths are identically positioned at 90◦ from the
ecliptic direction λ = 51.3◦ , β = 44.8◦ (in order to avoid overly bright UV stars and
the Sun). This great circle passes to within about 33◦ of the IPM upstream and
downstream directions. The circumstances during the three ACOs presented here
are given in Table 6.1. The observations made during each ACO were somewhat
different. During ACO-1, although it was planned that time-tagged data be would
Figure 6.2: NH-Alice nominal effective area, with the average effective areas shown
for IPM Lyman-α filling the 2◦ × 2◦ box and the 0.1◦ × 4◦ slot. Using these av-
erages, the expected sensitivity of NH-Alice to IPM Lyman-α is 5.5 counts/s/R.
The reduced sensitivity around Lyman-α (known as a “Lyman-α gap”) is accom-
plished by leaving the microchannel plate detector bare of any photocathode at
near-Lyman-α wavelengths (done to make the NH-Alice count rates due to bright
Lyman-α emissions more comparable to the count rates from other emissions)
Table 6.1: Circumstances of IPM observations

Parameter ACO-1 ACO-2 ACO-4
Date 10/7/2007 10/18/2008 6/19/2010
Start UTC 02:13:45 11:47:51 09:07:42
End UTC 03:10:44 13:51:02 11:21:46
NH-Sun-Earth angle (◦ ) +113.96 +115.96 −8.93
rNH (AU) 7.624 11.337 16.991
(λNH ,βNH ) (◦ ) (259.32,1.17) (269.33,1.50) (276.56,1.72)
Lyman-α πF (photons/cm2 /s @ 1 AU) 3.07 × 1011 2.86 × 1011 3.04 × 1011
Lyman-β πF (photons/cm2 /s @ 1 AU) 2.30 × 109 2.14 × 109 2.02 × 109
Solar Lyman-α/Lyman-β 133.5 133.6 150.5
IPM Lyman-α brightness (Rayleighs) 547.4 ± 1.9 404.3 ± 1.2 300.6 ± 1.3
IPM Lyman-β brightness (Rayleighs) 0.73 ± 0.09 0.67 ± 0.06 0.37 ± 0.11
IPM Lyman-α/Lyman-β 754 ± 90 602 ± 52 811 ± 233
acquired, a delay in voltage ramp-up left the NH-Alice instrument in histogram

mode, and histogram data (tEXP = 30 s each) were obtained over only about one-
half of the entire great circle. During ACO-2, time-tag data were obtained over
two complete spacecraft spins of 1 h each in duration. During ACO-4, histogram
data (tEXP = 80 s each) were obtained over one complete spacecraft spin. As the
New Horizons spacecraft gets further from Earth, data volume becomes a larger
burden, and time-tag and even histogram IPM observations will be less common.
However, since the total count rate is dominated by the IPM Lyman-α signal, only
housekeeping data is required in order to get useful IPM Lyman-α data.
Figure 6.3: Model IPM brightness contours appropriate for ACO-1 (rNH = 7.6 AU,
on October 7, 2007) are shown in blue at intervals of 50 Rayleighs. The great circle
swath of the IPM observed through the 6◦ -long NH-Alice slit is shown in green.
Bright UV stars are shown in orange (NH-Alice counts rates > 1, 000 counts/s) and
red (NH-Alice count rates > 5, 000 counts/s), and the Sun and the NH-Alice 20◦
solar keep-out zone are shown in orange. The upstream and downstream directions
for interstellar H atoms (Lallement et al. 2010) are indicated
on October 18, 2008) are shown in blue at intervals of 50 Rayleighs. The great circle
on June 19, 2010) are shown in blue at intervals of 50 Rayleighs. The great circle
The brightness data presented in Fig. 6.6 is entirely based on count rate in-
formation from housekeeping data, with a minor correction for nonlinearity, the
subtraction of ∼ 120 counts/s background (∼ 100 counts/s due to particles from
the radioisotope thermoelectric generator (RTG) plus ∼ 20 counts/s due to fidu-
cial “stim” counts), and conversion to brightness units assuming the filled-slit IPM
sensitivity of 5.5 counts/s/R estimated above. The data for the three ACOs are
shown in comparison with model predictions based on Pryor et al. (2008). The IPM
Lyman-α brightness is largest in the direction closest to the Sun, and the ratio of
the brightest/dimmest signal is about 2.2. The model fit is generally reasonable in
the upstream direction but tends to overestimate brightnesses near the Sun.
As mentioned above, one of the interesting differences between direct solar
Lyman-α and IPM Lyman-α is that the latter falls off much more slowly than
1/r 2 , so that it becomes relatively more important in the outer solar system (e.g.,
for methane photochemistry in the atmospheres of Neptune, Triton, and Pluto). As
noted by Gladstone (1993) and more recently by Throop (2011), photochemistry
by IPM Lyman-α was likely very significant in the early solar nebula. While more
NH-Alice data are needed, it appears that the first three ACOs show that the falloff
with rNH is even less than the ∼ 1/r dependence seen by Voyager (Hall 1992; Hall
et al. 1993). Figure 6.7 shows that, in the rNH ∼ 8–17 AU range at least, that
the current dependence is closer to ∼ 1/r 1/2 . It is interesting that the observed
dependence on r is much shallower than predicted by the simple expression for
Figure 6.6: IPM Lyman-α brightnesses observed by NH-Alice compared with the
model described by Pryor et al. (2008). Excursions due to UV-bright stars have not
been removed. The shifting longitude of the peak emissions is due to the changing
solar ecliptic longitude as seen from the NH spacecraft
BIPM presented in the introduction, even though the measurements encompass the
region where the turnover from an optically thin to an optically thick falloff should
occur. At this point we have no explanation for this behavior, and eagerly wait for
additional data with increased coverage in r to further investigate the problem.
Using the actual science data acquired during the three ACOs (i.e., spectral
images), we have searched for additional IPM emissions besides the dominant
Lyman-α emissions. Figure 6.8 shows the spectra integrated over the swaths ob-
tained during ACO-1, ACO-2, and ACO-4, with contamination from the brightest
stars removed. Although there is considerable structure in these spectra, most
of the behavior is explained by the broad wings of the Lyman-α line and edge
effects of the detector (at both short and long wavelength limits). An expected
few-Rayleigh feature due to resonant scattering of the solar 58.4 nm line by He
atoms in the interstellar wind is not clearly seen. This non-detection is likely due
to a combination of (1) detector edge effects, (2) the relatively large background
(∼ 100 counts/s over the entire detector, ∼ 2 count/s over the 58.4 nm line) from
the New Horizons RTG power source, and (3) the low instrumental sensitivity at
this wavelength (∼ 0.03 counts/s/R). However, a feature at 102.6 nm is identified
as due to resonant scattering of the solar Lyman-β line by IPM hydrogen atoms
(the Lyman-β line is far from the detector edge and the estimated instrumental
sensitivity is ∼ 0.10 counts/s/R). This line is very weak, at about the 0.5–1 R
Conclusions 185
Figure 6.7: IPM Lyman-α brightnesses observed by NH-Alice, scaled by

[10/rNH (AU)]1/2 and centered on the solar ecliptic longitude. The IPM Lyman-α
−1/2
brightness scales fairly well with rNH over the range rNH = 8–17 AU
level, as seen in Table 6.1 (the IPM brightnesses and brightness ratios provided
in Table 6.1 are derived from the swath-averaged spectra of Fig. 6.8). The IPM
Lyman-α/Lyman-β ratios are 754 ± 90 (ACO-1), 602 ± 52 (ACO-2), and 811 ± 233
(ACO-4), and are in good agreement with the Voyager result (using data from
inside 15 AU) of 700 ± 200 Murthy et al. 1999). The reason that the IPM Lyman-
α/Lyman-β ratio is so much larger than the solar Lyman-α/Lyman-β ratio (the
solar line center brightness ratio would be even smaller than the integrated line
ratio, since the Lyman-α line is broader than the Lyman-β line) is likely due to
the lack of substantial multiple scattering of Lyman-β in the IPM, due mostly to
its smaller resonance scattering cross section and to the substantial likelihood of a
branching to the Hα transition during each scattering event.
Conclusions
The IPM measurements made by the NH-Alice instrument on the New Hori-
zons mission to the Pluto system presented here represent the first new data from
outside of Saturn’s orbit since Voyager. It is clear that even with total count rate
information only that useful results are obtained with regard to the distribution
and brightness of IPM Lyman-α emissions in the outer solar system. We plan on
continuing such observations on a more or less yearly basis, until they are no longer
possible. Current extrapolations of the RTG performance suggest that NH-Alice
Figure 6.8: IPM spectra (±1σ) observed by NH-Alice on ACO-1, ACO-2 and ACO-
4, averaged over the great-circle swaths on the sky shown in Figs. 6.3–6.5, using
only detector rows corresponding to the 0.1◦ part of the slit (for improved spectral
resolution) and with star contributions removed. Besides Lyman-α, a feature due
to IPM resonant scattering of the solar line H 102.5 nm (Lyman-β) is detected,
although an expected signal at He 58.4 nm is not clearly seen
could be operable for ∼ 40 or so more years, although other factors are also im-
portant (e.g., how much fuel remains for IPM scans after the planned Kuiper Belt
Object flyby and whether an extended mission is approved). In addition, more
extensive IPM data will be available as a side effect of various airglow observations
that are planned for NH-Alice during the Pluto flyby (and its rehearsal). As New
Horizons moves further from the Sun the effects of multiple scattering in the ter-
mination shock should become more apparent, and the NH-Alice data should allow
for useful remote sensing studies of the termination shock region.
Acknowledgements
We thank the New Horizons mission team and the New Horizons science team. New
Horizons is funded by NASA, whose financial support we gratefully acknowledge. We also
thank ISSI for support of the working group. The authors would also like to thank the
reviewers and Priscilla Frisch for their useful comments.
Bibliography 187
Bibliography
T.F. Adams, P.C. Frisch, High-resolution observations of the Lyman alpha sky background.
Astrophys. J. 212, 300–308 (1977)
J.M. Ajello, A.I.F. Stewart, G.E. Thomas, A. Graps, Solar cycle study of interplanetary
Lyman-alpha variations: pioneer venus orbiter sky background results. Astrophys. J.
317, 964–986 (1987)
J.M. Ajello et al., Observations of interplanetary Lyman-α with the Galileo ultraviolet
spectrometer: multiple scattering effects at solar maximum. Astron. Astrophys. 289,
283–303 (1994)
J-L. Bertaux, J.E. Blamont, Evidence for a source of extraterrestrial hydrogen Lyman-alpha
emission: the interstellar wind. Astron. Astrophys. 11, 200–217 (1971)
J.C. Brandt, J.W. Chamberlain, Interplanetary gas. i. hydrogen radiation in the night sky.
Astrophys. J. 130, 670–682 (1959)
M. Braskén, E. Kyrölä, Resonance scattering of Lyman alpha from interstellar hydrogen.
A.L. Broadfoot et al., Ultraviolet spectrometer observations of Neptune and Triton. Science
246, 1459–1466 (1989)
H-J. Fahr, The extraterrestrial uv-background and the nearby interstellar medium. Space
Sci. Rev. 15, 483–540 (1974)
G.R. Gladstone, Photochemistry in the primitive solar nebula. Science 261, 1058 (1993)
D.T. Hall, Ultraviolet resonance line radiation and the structure of the heliosphere. Disser-
tation, University of Arizona, Tucson, 1992
D.T. Hall et al., Heliospheric hydrogen beyond 15 au: evidence for a termination shock. J.
Geophys. Res. 98, 15185–15192 (1993)
T.E. Holzer, Neutral hydrogen in interplanetary space. Rev. Geophys. Space Phys. 15,
467–490 (1977)
V.V. Izmodenov, Local interstellar parameters as they are inferred from analysis of obser-
vations inside the heliosphere. Space Sci. Rev. 143, 139–150 (2009)
H.U. Keller, K. Richter, G.E. Thomas, Multiple scattering of solar resonance radiation in
the nearby interstellar medium. Astron. Astrophys. 102, 415–423 (1981)
R. Lallement, E. Quémerais, D. Koutroumpa, J-L. Bertaux, S. Ferron, W. Schmidt, P. Lamy.
The interstellar H flow: updated analysis of SOHO/SWAN data. In: 12th International
Solar Wind Conference, vol 1216. AIP Conference Proceedings, pp. 555–558 (2010)
R.R. Meier, Some optical and kinetic properties of the nearby interstellar gas. Astron.
Astrophys. 55, 211–219 (1977)
J. Murthy, D. Hall, M. Earl, R.C. Henry, J.B. Holberg, An analysis of 17 years of Voy-
ager observations of the diffuse far-ultraviolet radiation field. Astrophys. J. 522, 904–914
(1999)
W.R. Pryor et al., Radiation transport of heliospheric Lyman-α from combined Cassini and
Voyager data sets. Astron. Astrophys. 491, 21–28 (2008)
medium at Lyman α. Astron. Astrophys. 358, 353–367 (2000)
E. Quémerais, J-L. Bertaux, Radiative transfer in the interplanetary medium at Lyman
alpha. Astron. Astrophys. 277, 283–301 (1993)
E. Quémerais, R. Lallement, B.R. Sandel, J.T. Clarke, Interplanetary Lyman α observations:
intensities from Voyagers and line profiles from HST/STIS. Space Sci. Rev. 143, 151–162
(2009)
S.A. Stern et al., Alice: the ultraviolet imaging spectrograph aboard the new horizons
Pluto-Kuiper belt mission. Space Sci. Rev. 140, 155–187 (2008)
G.E. Thomas, R.F. Krassa, OGO 5 measurements of the Lyman alpha sky background.
H.B. Throop, UV photolysis, organic molecules in young disks, and the origin of meteoritic
amino acids. Icarus 212, 885–895 (2011)
H.A. Weaver, S.A. Stern, New Horizons: NASA’s Pluto-Kuiper belt mission, in: The so-
lar system beyond neptune, ed. by M.A. Barrucci, H. Boenhardt, D.P. Cruikshank, A.
Morbidelli (University of Arizona Press, Tucson, 2008), p. 592
F.M. Wu, D.L. Judge, Modification of solar lines propagating through the interplanetary
medium. J. Geophys. Res. 84, 979–982 (1979)
Part III
Instrument
Cross-Calibration
—7—
A New Catalog of Ultraviolet Stellar

Spectra for Calibration
Martin Snow∗
Laboratory for Atmospheric and Space Physics,
University of Colorado,
Boulder, CO, USA
Aurélie Reberac and Eric Quémerais

LATMOS-IPSL,
John Clarke
Boston University,
Boston, MA, USA
W. E. McClintock and T. N. Woods

Boulder, CO, USA
Abstract
The SOLar-STellar Irradiance Comparison Experiment (SOLSTICE) on the
SOlar Radiation and Climate Experiment (SORCE) observes both the Sun and
stars in the ultraviolet (115–300 nm). Prior to launch, it was calibrated at the
SURF-III synchrotron. Spectra from the International Ultraviolet Explorer (IUE)
corrected to the white dwarf flux scale are in good agreement with SOLSTICE
observations, validating the two completely independent methods of calibration.
Measurements of stars in the SOLSTICE catalog are then used to transfer this
calibration to the SPectroscopy for the Investigation of the Characteristics of the
Atmosphere of Mars (SPICAM) instrument. We describe the steps used to calculate
the effective area for SPICAM to calibrate its stellar observations. Since only a
handful of stars in the IUE archive have been converted to the white dwarf scale
and many of them are relatively faint, the SOLSTICE catalog of bright stars can
be an extremely useful resource for inflight calibration of ultraviolet spectrographs.

192 7. Catalog of UV Stellar Spectra
Introduction
Early-type stars shine brightly in the ultraviolet (UV), and many can be easily
detected by astronomical space-based instruments. Additionally, the UV spectrum
of these main sequence stars is very stable over time (Mihalas and Binney 1981).
These two properties make them extremely attractive as calibration targets. Ob-
servations of a set of stars by one instrument can be compared to measurements
taken by a different instrument even if the two datasets were collected many years
apart.
A and B main sequence stars such as those included in the SOLar-STellar
Irradiance Comparison Experiment (SOLSTICE; McClintock et al. 2005a) and
SPectroscopy for the Investigation of the Characteristics of the Atmosphere of Mars
(SPICAM; Bertaux et al. 2006) target lists have higher effective temperatures than
a star like the Sun. Consequently, the peak wavelength of the blackbody emission
from their photospheres is in the ultraviolet rather than in the visible range. On
the Sun, the ultraviolet spectrum is formed in the chromosphere and is far more
variable than the visible spectrum which is formed at the photosphere. For A and B
spectral type stars, the stable photosphere produces a bright ultraviolet spectrum
over long timescales (cf Mihalas and Binney 1981). These stars can therefore be
an important transfer standard in cross-calibrating a large number of instruments.
All that is required is a set of absolutely calibrated spectra, ideally at a wide range
of apparent magnitude and widely distributed across the celestial sphere.
Instruments in orbit around the Earth or other celestial body often have sea-
sonal observing constraints much like ground-based observatories. A star in one
part of the sky may only be within viewing constraints for part of the year. An
ensemble of stars distributed around the sky will increase the probability that on
any given day, at least one of them will be available for observation. A wide range
of stellar brightnesses is also desired to make the catalog useful to the greatest
number of instruments. Depending on the design of the spectrometer, a given star
may be either too dim to produce a reasonable signal or be too bright and saturate
the detector system. Additionally, well-calibrated spectra of different levels may
provide diagnostic information about an instrument’s linearity.
In this paper, we will focus on two main topics. The first is a description of
the catalog of stars observed by the absolutely-calibrated SOLSTICE, including
comparisons to widely used archived spectra. These SOLSTICE stellar ultraviolet
spectra cover the wavelength range of 120–300 nm and are only observable from
space. The Earth’s atmosphere completely absorbs radiation at these shorter ul-
traviolet wavelengths. The second goal of this paper is to show how the catalog of
stellar spectra can be used to calibrate an instrument inflight that was launched
without a highly accurate ground calibration. In particular, we will apply this
technique to the UV channel of the SPICAM instrument onboard Mars Express
(Bertaux et al. 2006). After describing the observations by SOLSTICE and
SPICAM, we will compare their measurements to spectra of IUE standard stars
corrected to the white dwarf scale, and then finally discuss the results.
SOLSTICE 193
Grating Drive
Entrance Slit Assembly Sphere-Sphere
Assembly Kinematic Mount
Purge/
Fill Assy
Sunshade Fold Mirror
Assembly Assembly
Door Mech &

EMI Filter Assy Sphere-Rigid
Kinematic Mount
Sphere-Translate
Kinematic Mount
Camera Mirror Exit Slit Detector Head

Assembly Assembly Assembly
Figure 7.1: Layout of SORCE SOLSTICE (From McClintock et al. 2005a)
SOLSTICE
SOLSTICE (McClintock et al. 2005a) onboard the SOlar Radiation and Climate
Experiment (SORCE; Rottman 2005) is a grating spectrometer whose primary
mission is to measure the solar spectral irradiance on a daily basis. SORCE was
launched in January 2003, and began taking measurements of both solar and stellar
spectral irradiance soon thereafter. The SOLSTICE instrument has two channels:
one observing the far ultraviolet (FUV) from 115 nm to 180 nm, and the other
observing the middle ultraviolet (MUV) from 180 nm to 300 nm. Figure 7.1 shows
the optical layout of the SOLSTICE instrument.
The instrument is designed to use the spectral irradiance from a set of early-
type stars as a calibration reference in order to track long-term changes in the
instrument. It accomplishes this goal by observing both the Sun and the stars
with the same optics and detector, changing only the size of the entrance aperture
and exit slit, in a highly controlled manner (Snow et al. 2005). Since the instru-
ment observes the solar spectral irradiance every orbit, it is expected to degrade
significantly over its lifetime. Repeated observations of the stellar calibration stars
allow this degradation to be corrected in the published data.
As described in McClintock et al. (2005b), the instrument was calibrated
before launch at the NIST SURF III facility in Gaithersburg, MD (Arp et al. 2000)
with an accuracy of about 3 %. During an early on-orbit observing campaign, this
calibration was transfered to the stars. All the stars in the SOLSTICE catalog were
observed in the first few days of the mission before the instrument was exposed to
the Sun. Therefore, SOLSTICE stellar spectra are absolutely calibrated and can
be useful for cross-calibration of other instruments that observe the same stars in
the ultraviolet.
Table 7.1: SOLSTICE calibration stars
Star name Bright star RA (2000) Dec (2000) V mag Spectral type
Per 1,220 3 h 57.8 min 40◦ 0 2.90 B0.5 III

β CMA 2,294 6 h 22.7 min −17◦ 57 1.98 B1 II-III
α CMa 2,491 6 h 45.1 min −16◦ 43 −1.46 A1 V
κ Vel 3,734 9 h 22.1 min −55◦ 1 2.50 B2 IV-V
α Leo 3,982 10 h 8.4 min 11◦ 58 1.35 B7 V
δ Cen 4,621 12 h 8.4 min −50◦ 43 2.60 B2 IVne
α Cru 4,730 12 h 26.6 min −63◦ 7 1.35 B0.5 IV + B1 V
α Vir 5,056 13 h 25.1 min −11◦ 10 0.97 B1 IV + B2 V
η UMa 5,191 13 h 47.5 min 49◦ 15 1.86 B3 V
ζ Cen 5,231 13 h 55.5 min −47◦ 17 2.55 B2.5 IV
β Cen 5,267 14 h 3.8 min −60◦ 22 0.61 B1 III
δ Sco 5,953 16 h 0.3 min −22◦ 37 2.32 B0.5 IV
τ Sco 6,165 16 h 35.9 min −28◦ 13 2.82 B0 V
α Lyr 7,001 18 h 36.9 min 38◦ 47 0.03 A0 Va
σ Sgr 7,121 18 h 55.3 min −26◦ 18 2.02 B2.5 V
α Pav 7,790 20 h 25.6 min −56◦ 44 1.94 B2.5 V
α Gru 8,425 22 h 8.2 min −46◦ 58 1.74 B7 IV
α PsA 8,728 22 h 57.7 min −29◦ 37 1.16 A3 V
The stars observed by SOLSTICE are listed in Table 7.1. As described in

McClintock et al. (2005a) and Snow et al. (2005), they were chosen to be bright,
non-variable in the ultraviolet, and free of other bright UV sources in the field-of-
view. The original list of 31 SOLSTICE targets published in Rottman et al. (1993)
for the SOLSTICE on the Upper Atmosphere Research Satellite (UARS) mission
was reduced to the 18 stars shown in Table 7.1 for SORCE. Some of the original
31 stars were eliminated because they were either not bright enough or else they
were in the same general part of the sky as a brighter target. The remaining 18
stars have been observed repeatedly by SORCE SOLSTICE since early 2003. As of
the end of 2010, SOLSTICE has taken more than 700 stellar spectra in the MUV
and over 1,000 in the FUV. The reason that more FUV spectra were acquired than
MUV spectra is simply that the FUV wavelength range is smaller than the MUV
range (115–180 nm is only 65 nm while 180–300 nm is 120 nm). Therefore the FUV
spectral scan experiment was easier to schedule and consequently happened slightly
more often than the MUV spectral scan.
Unfortunately, due to degradation of the spacecraft battery system, stellar ob-
servations on SORCE have now been greatly curtailed. The only stellar observations
that continue to the current time are fixed-wavelength calibration measurements
(Snow et al. 2005) rather than full spectra, and only during eclipse periods shorter
than 30 min. But the existing database of full spectra give excellent signal-to-noise
ratios for all the stars and can serve as an archive of well-calibrated observations
of these reference stars.
SOLSTICE 195
Table 7.2: SORCE SOLSTICE spectroscopic parameters
Parameter FUV channel MUV channel
Wavelength range 115–180 nm 180–320 nm

Grating ruling density 3,600 grooves mm−1 3,600 grooves mm−1
Stellar entrance slit 16 mm diam. 16 mm diam.
Stellar exit slit 0.75 mm × 6 mm 1.5 mm × 6 mm
Stellar bandpass 1.1 nm 2.2 nm
Detector photocathode Cesium iodide (CsI) Cesium telluride (CsTe)
The SOLSTICE on SORCE is the follow-on to the highly successful SOLSTICE

instrument on the UARS (Rottman et al. 1993). The basic design for the two
instruments is similar, but SORCE SOLSTICE has one key improvement in the
stellar observing mode. UARS SOLSTICE had an interference filter just before the
exit slit designed to reduce the solar signal at the long-wavelength end of the MUV
channel. Unfortunately, this filter also reduced the stellar signal. The SORCE SOL-
STICE instrument replaced this interference filter with a pair of neutral density
filters which can be moved in or out of the optical path during a solar measure-
ment. By removing these filters during stellar observations, the amount of starlight
reaching the detector is greatly increased for SORCE SOLSTICE.
Fully-calibrated spectra for all the stars in Table 7.1 are shown in the appendix
section of this paper. FITS files of all the spectral data including uncertainties are
available for download from the FONDUE web page, http://bdap.ipsl.fr/fondue/.
SOLSTICE Instrument Description

The full description of the SOLSTICE instrument can be found in McClintock
et al. (2005a), but we will include a synopsis of the instrument design as it per-
tains to stellar mode in this document for completeness. The SOLSTICE package
on SORCE consists of two identical scanning grating monochromators. These two
redundant instruments are known as SOLSTICE A and SOLSTICE B. Figure 4
of McClintock et al. (2005a) shows the opto-mechanical layout of each instrument.
In stellar mode, a 16 mm diameter circular aperture collects the stellar irradiance.
The light is dispersed by a plane grating and ultimately collected with a photomul-
tiplier tube detector. Table 7.2 contains the spectroscopic parameters for stellar
mode.
The spectral resolution in stellar mode is about 1.1 nm in the FUV channel
and varies between 2.2 nm and 1.8 nm in the MUV channel. During a typical
observation, the spectrum is sampled every 0.3 nm with an integration time of 1 s,
thus the spectrum is sampled adequately in the FUV channel and oversampled in
the MUV channel.
SOLSTICE Data Reduction

The data reduction steps for SOLSTICE stellar observations are fairly straight-
forward. Before each spectral scan, the grating is rotated so that zeroth order
falls on the detector. This allows us to confirm that the star is well-centered in
the field-of-view and to measure the current value of the offset, , in the grating
equation (Eq. 1 of Snow et al. 2005):
λ = 2d sin(θ + /2)cos((φ − )/2). (7.1)

In this equation, d is the grating groove spacing, θ is the grating rotation
angle, φ is the half-angle between the diffracted and incident beams, and is the
pointing offset. By measuring through the location of zero order, we have excellent
knowledge of the wavelength scale for each spectrum.
The detector counts are converted to a count rate and a small correction is
made for detector dead time and dark current (McClintock et al. 2005b). The
corrected count rate is then converted to an irradiance by applying the responsivity
determined by the preflight calibration (McClintock et al. 2005b) and the inflight
degradation correction described in Snow et al. (2005). It should be noted that the
stellar degradation analysis uses the ensemble of all stars observed by SOLSTICE,
and not just an individual star. Therefore there is no concern about a circular
dependency in correcting these stellar spectra using the degradation correction
derived from the stellar measurements themselves. Furthermore, the amount of
degradation over the SORCE mission is fairly small for most wavelengths. In the
FUV channel, the total degradation after 8 years is about 10 %. The MUV channel
on SORCE SOLSTICE has degraded even less, suffering only about 2 % change in
responsivity over the mission. Such low degradation rates are a tribute to the care
taken during preflight handling of the instrument to prevent contamination.
SOLSTICE Preflight Calibration

The fundamental measurement equation for SOLSTICE is given by a modified
version of Eq. 1 of McClintock et al. (2005b):
C(λ, τ, Dc , Sl , St )
E(λ) = (7.2)
RC (λ, T )F OV (λ, Ω, θ, φ)AEntrance ΔλBP DEG(t, λ, Ω, θ, φ)
S(λ)N (τ ) − Dc − Sl (λ) − St
C(λ, τ, Dc , Sl , St ) = (7.3)
Δt
where C is the instrument count rate computed by correcting the observed signal,
S(λ), for nonlinearity, N (τ ), where τ is the electronics dead time. Dc is the dark
current, Sl is the contribution from scattered light and St is the amount of stray
light. Δt is simply the integration time of the observation, and R is the instru-
ment responsivity at the beginning of life. The responsivity of the instrument was
measured before flight at the Synchrotron Ultraviolet Radiation Facility (SURF;
Arp et al. 2000) and is shown in Fig. 7.2. R is a function of wavelength and de-
tector temperature. In the case of stellar observations, the field-of-view is fully
SOLSTICE 197
Figure 7.2: Responsivities of SOLSTICE A and B in both the far ultraviolet (FUV)
and middle ultraviolet (MUV) channels. These responsivities were determined
based on calibration at the SURF-III synchrotron source before the spacecraft was
launched (From McClintock et al. 2005b)
filled and the integration over the angular size of the target described in McClin-
tock et al. (2005b). Equation 7.3 can be simplified because the contribution from
scattered light, stray light, and nonlinearity are all negligible in the SOLSTICE
stellar observations. It then becomes:
S(λ) − Dc (t)
C(λ) = . (7.4)
Δt
Dc is measured by pointing the instrument at dark space on a daily basis.
Figure 7.3 shows the contributions to the uncertainty in the SURF calibration.
The SURF beam has a well-known absolute irradiance (0.75 %), but since it is
a featureless continuum, it cannot be used to confirm the instrument wavelength
scale the way a line source would. The uncertainty in the wavelength scale and the
counting statistics of the observation represent about equal contributions to the
final uncertainty. The largest source of uncertainty in the preflight calibration is
the geometrical correction shown as “Gamma” in Fig. 7.3. This correction factor
enters into the calculation of the uncertainty in the responsivity because the SURF
beam does not fill the field-of-view of the instrument. Instead, it is more like a point
source. The calibration at SURF included a map of the instrument response to
the synchrotron point source over the field-of-view. The uncertainty in converting
this grid of map positions to the smooth field-of-view, Γ, is shown in the plot.
Figure 7.3: (top) Major sources of uncertainty in the SOLSTICE absolute cal-
ibration. The largest component, Gamma, is the uncertainty in the geometric
correction between the point source SURF beam and the full instrument field-of-
view. (bottom) The combined standard uncertainty of the SOLSTICE absolute
calibration from the components is shown in the top panel. For most wavelengths,
the uncertainty in responsivity is around 2 % (From McClintock et al. 2005b)
The lower panel of this figure shows the rms combination of these effects. For most
wavelengths, the combined uncertainty in the responsivity is around 2 %.
SOLSTICE Instrument Degradation
Unlike most instruments which measure stellar fluxes, SOLSTICE spends the
vast majority of its life staring directly at the Sun. This exposure to the solar
flux causes the instrument to degrade over time, so correction for changes in the
instrument responsivity are a necessary and important step in the data reduction.
The full description of the SOLSTICE technique for tracking degradation is avail-
able in Snow et al. (2005), but the following summary should provide the reader
with a sufficient understanding of the process to give confidence in the uncertainty
estimates that follow.
During the eclipse portion of each orbit, SOLSTICE observes stars that are
available based on celestial geometry. Rather than measuring the entire spectrum,
the typical observation is at a single wavelength in order to increase the signal-
to-noise ratio. Observations are planned to yield at least 104 counts in order to
SOLSTICE 199
Figure 7.4: Individual fixed-wavelength stellar observation used to determine long-

term degradation of the SOLSTICE instruments. (top) Raw count rate observed
during 200 integrations of 1 s each. (bottom) Histogram of integrations used to
determine mean count rate. A Gaussian curve is fit to the histogram for each
observation (From Snow et al. 2005)
reduce the fractional uncertainty due to counting statistics to less than 1 %. This
may take anywhere from 200 to 500 integrations of 1 s each, depending on the
star’s brightness. The wavelengths used to track changes in the responsivity curve
are evenly spaced throughout the wavelength range of each channel. There are 18
wavelengths in the FUV and 22 in the MUV. Five of the FUV wavelengths are
contaminated by geocoronal Lyman-α emission and a special technique is required
to isolate the stellar signal in this range. As described in Snow et al. (2005), the
Lyman-α airglow is removed by observing a dark region of the sky adjacent to the
target star before and after the stellar observation. The dark region measurement
includes the geocoronal emission, so the pair of observations can be used to estimate
the contribution of the airglow to the stellar signal taken between the two dark
region measurements. The analysis in this paper will be limited to the range 130–
300 nm to avoid the contribution of the airglow.
Figure 7.4 shows a typical fixed-wavelength observation of α Cru. A series
of 1 s integrations are analyzed statistically to determine the mean and standard
deviation as well as the likelihood that the data represent a normal distribution.
Observations passing these statistical tests are then used to determine the long-
term degradation of the instrument.
Figure 7.5: Measurement of long-term degradation in SOLSTICE. (upper left)

The raw irradiances for all stars used in the fit. Each symbol is the result of
a measurement such as the one shown in Fig. 7.4. (lower left) Multi-parameter
least-squares fit using all the stars. Each star is normalized based on the observed
brightness and an exponential function is fit in time to the ensemble. (upper right)
Histogram of stellar irradiances after correction by the fit determined in the lower
left. The residuals after the fit show a constant mean value with a scatter that is
entirely consistent with Gaussian noise. (lower right) Legend showing the name of
each star used in the fit as well as the mean observed count rate. Stars with lower
count rates are observed for longer times to give them equal statistical weighting
The top left panel of Fig. 7.5 shows the individual stellar irradiances from the
fixed-wavelength observations at 167 nm. The gradual decrease of irradiance with
time is at the same rate in all the stars, and therefore must be due to the degra-
dation of the instrument’s responsivity. The bottom left panel shows the same
irradiances normalized and fit with an exponential decay model. The relative
brightness of each star (i.e. the normalization factor) and the decay model are
determined simultaneously through a least-squares fit. Statistical outliers are re-
moved and the fit is repeated to produce the final degradation function for a given
wavelength. The whole procedure is repeated for each of the 40 wavelengths that
span the SOLSTICE sensitivity range.
The panel on the upper right shows the histogram of normalized stellar observa-
tions after correction for degradation. It is well fit by a Gaussian and is statistically
consistent with a single mean and an uncertainty due to random noise (χ2ν ∼ 1).
SOLSTICE 201
The bottom right panel identifies the stars used in the other plots and also shows
the average count rate. Since the detector dark rate is only a few counts per second
for the FUV channel, all the stellar irradiances used in the degradation analysis
are well above the background. The dark rate in the MUV channel is significantly
higher, and the stellar count rates at the long wavelength end of the spectrum
(λ > 270 nm) are only a few times larger than the dark rate. Stellar observations
above 270 nm are therefore less certain than those at shorter wavelength due to
statistical uncertainty in the dark rate correction.
SOLSTICE Uncertainty
The combined standard uncertainty in stellar irradiance for a SOLSTICE stellar
measurement is the combination of a variety of sources of error, some random
and some systematic. The main sources of systematic error are uncertainty in
the preflight calibration, uncertainty in the dark current subtraction, and inflight
correction for instrument degradation. These three sources of uncertainty are much
larger than the random uncertainty due to statistical noise in each observation,
although an error in the calculated degradation correction will appear to be a
larger spread in observed count rates from one observation to another.
Comparison of SOLSTICE to IUE Standard Stars

IUE has been the most widely used source of ultraviolet stellar spectra. It was
launched in 1978 and operated throughout the 1980s and 1990s observing a very
large number of targets. Originally, IUE’s calibration was based primarily on the
spectrum of a single star, η UMa (Bohlin et al. 1980), but over the course of the
mission, improvements in our understanding of stellar atmospheres led to changes
in the calibration reference. The ultimate calibration of IUE is based on models
and observations of white dwarfs (Bohlin et al. 1990; Bohlin 1996). Instruments
on the Hubble Space Telescope (HST) also use the white dwarf scale for their
calibration (Bohlin et al. 1990, 2001, 2011). Throughout this paper, we will refer
to “IUE Standard Stars” as those whose fluxes have been converted to the white
dwarf scale.
Reference spectra based on this white dwarf scale are available from the HST
calibrated observations (CALOBS) archive http://www.stsci.edu/hst/observatory/
cdbs/calobs.html. Of the stars in this archive, only three are in the SOLSTICE cat-
alog (Table 7.1) and also have IUE flux measurements that have been corrected to
the white dwarf scale: η UMa, α Lyr, and α Leo. Figure 7.6 shows the SOLSTICE
spectra for those three stars in the top panel and their ratio to IUE observations in
the bottom panel. Over the full wavelength range, the ratio is essentially unity, in-
dicating that the irradiances are in agreement to within their uncertainties. There
is no systematic difference in calibration between the SOLSTICE instrument which
is tied to the NIST SURF standard and the IUE measurements tied to the white
dwarf scale.
In the CALOBS archive, there is also a spectrum for β Cen, but it has not been
corrected to the white dwarf flux scale according to the header of the FITS datafile.
It differs from SOLSTICE by about 20 %. A user of the CALOBS archive must
Figure 7.6: Comparison of SORCE SOLSTICE and IUE spectra for three stars
calibrated to the white dwarf scale. The top panel shows the SOLSTICE spectra
for the three stars, while the lower panel shows the ratio of each observation to
the corresponding IUE spectrum from the CALOBS archive. The SOLSTICE
calibration based on a NIST standard is in good agreement with the IUE calibration
based on the white dwarf scale. The IUE spectra themselves are shown in Figs. 7.17,
7.18, and 7.27 in the appendix section
SPICAM-UV 203
take care to use only the spectra which have been corrected to the white dwarf
scale. The CALSPEC archive on the STScI web page also contains well-calibrated
spectra, but the only star which overlaps the SOLSTICE catalog is α Lyr. The
CALSPEC and CALOBS data files for α Lyr are identical in the 130–300 nm range.
SPICAM-UV
SPICAM-UV (referred to as simply SPICAM hereafter) is an instrument onboard
the Mars Express mission which was launched in 2003 (Bertaux et al. 2006) and
is currently in orbit around Mars. The primary science objective of the SPICAM
instrument is to make measurements of stars occulted by the Martian atmosphere.
These data are used to retrieve the density profiles of various components of the
Martian atmosphere. Because this only requires a relative measurement of changes
in the star’s spectrum as the line-of-sight passes through the atmosphere, it was not
necessary to have an accurate absolute calibration of the instrument before flight.
Linearity is important for an occultation measurement, but that can be tested
on the ground by varying the integration time while observing a stable source.
Absolute calibration can be a challenging task that may require taking the entire
instrument to a specialized facility. However, instead of using a ground-based
source tied to international standards, one can use a calibrated stellar reference
spectrum and transfer the calibration of that instrument to SPICAM. This section
will describe the process of using stellar reference spectra to determine SPICAM’s
calibration.
Stellar spectral measurements are performed by SPICAM during occultations of
stars by the atmosphere of Mars. These observations are fully described in Bertaux
et al. (2006). During the observation, the instrument boresight is pointed towards
a chosen star. The stellar spectrum is measured for about 10 min with the line-
of-sight well above the atmosphere of Mars. These data are used to compute the
reference spectrum. As the spacecraft travels along its orbit around Mars, the star
moves behind the planet and the line of sight approaches the surface of Mars. The
part of the observation where the star is occulted by the Martian atmosphere lasts
only a few minutes. The variation of the transmission is used to characterize the
composition and density profile of the atmosphere (Bertaux et al. 2006).
Initially, the SPICAM team had selected 39 stars to provide sufficient longi-
tudinal and latitudinal coverage for the study of the atmosphere of Mars. More
recently, the catalog has been extended by more than 200 stars and dedicated ob-
servations have been performed outside of the atmosphere. The stars observed by
SPICAM are listed in Table 7.4 in the Appendix section. At the time of publication
of this paper, SPICAM has observed 147 individual stars. It should be noted that
similar observations have been performed by the SPICAV-UV instrument of the
Venus-Express mission (Bertaux et al. 2007).
SPICAM Instrument Description

Figure 7.7 shows the basic optical layout of the SPICAM instrument (Fig. 1
of Bertaux et al. 2006). The SPICAM detector is an intensified charge-coupled
Figure 7.7: Optical layout of the UV channel of the SPICAM instrument on Mars
express. (1) aperture blend of the UV channel, (2) off-axis parabolic mirror, (3)
slit, (4) concave UV grating, (5) intensifier, (6) CCD, (7) flat mirror, (8) solar
opening (closed by a shutter when not viewing the Sun) (Adapted from Fig. 1 of
Bertaux et al. 2006)
device (CCD), i.e. an image intensifier coupled to a CCD through stacks of fiber
optics. The Hamamatsu image intensifier has a CsTe photocathode deposited on
a MgF2 window. Photoelectrons created at the photocathode by UV photons are
accelerated by the high voltage applied to a microchannel plate (MCP), creating
a cloud of electrons that is again accelerated to a phosphor screen. The resulting
pulse of green light is transmitted through the fiber optics to the CCD pixels. The
charge on each pixel is read and digitized to 12 bits, giving a maximum of 4,095
ADU (Analog to Digital units). Therefore the detector is used in a pseudo-counting
mode (Sandel and Broadfoot 1986). The gain of the MCP may be adjusted so
that a single photoelectron created at the cathode will produce 1–40 ADUs in the
CCD, spread over ≈ 6 pixels on the CCD matrix.
The CCD is an array of 408 by 288 pixels. The 288 lines are perpendicular to
the slit axis and parallel to the dispersion plane of the grating. The spectra are
spread along the lines of the CCD. Of the 408 pixels in each line, 384 are used to
measure the spectra. The remaining pixels, some of which are masked, are used to
determine the dark current of the CCD.
Each measurement has an integration time of a few hundred milliseconds. A
typical value is 640 min. The gain of the MCP is adjusted according to the bright-
ness of the source. Due to limited telemetry bandwidth, only five bands of 408
pixels are transmitted to Earth. The five bands are formed by summing individual
lines by 1, 2, 4, 8, 16 or 32 pixels. Typically, for stellar occultations, each band
is the sum of 16 physical lines of the CCD. Most of the stellar signal is included
in the central band. There has been very little degradation of the MCP over the
course of the Mars Express mission, likely because the voltage applied to it is fairly
low (Bertaux et al. 2006).
SPICAM-UV 205
SPICAM Inflight Calibration

The absolute photometric calibration of the SPICAM UV spectrometer (110–
320 nm) onboard Mars Express was first done on the ground. The reflectivities of
all optical elements were measured separately, and the whole instrument was cali-
brated, but with a limited accuracy because of inherent difficulties in the vacuum
UV such as cleanliness, lack of portable standard sources, etc.
The purpose of the inflight calibration procedure is to obtain a more accurate
determination of the instrument responsivity and to follow its performance or sta-
bility. Stellar observations have been used to achieve this goal. First, they were
used to check the wavelength assignment and the dependence of the gain of the
SPICAM image intensifiers as a function of high voltage level. They have also
been used to derive the Point Spread Function (in the spatial axis) or Instrumental
Spectral Response Function (in the spectral axis) of the SPICAM instrument. Fi-
nally, we have performed comparisons between stellar measurements collected by
the SPICAM UV channel and stellar spectra obtained by SOLSTICE and IUE. The
comparison of SPICAM spectra with these reference spectra was done to obtain
a more accurate determination of the SPICAM sensitivity as a function of wave-
length. It also provided a check for possible changes in the instrument following the
integration on the Mars Express spacecraft and the launch due to contamination,
misalignment, shifts of optical elements inside the instrument, and other sources
of change.
SPICAM Wavelength Assignment
The wavelength assignment is the relation between wavelength and pixel num-
ber on the spectral axis. Because the instrument behaves as an objective grating
spectrograph, the position of the source in the field-of-view influences the posi-
tion of the spectrum on the detector. Therefore the wavelength-pixel relation is
determined for a reference position, i.e. the center of the slit, but must be cor-
rected for the position of the source. This relation was initially measured on the
ground. However, this assignment had to be verified in space due to the possibility
of a change in alignment during or after launch. The wavelength assignment after
launch was measured by comparing SPICAM observations with an IUE spectrum
of γ Vel (HR 3207, see Fig. 7.8). The spectrum of this star shows various narrow
emission or absorption lines that allow for a good calibration of the spectrum posi-
tion on the CCD. The star spectrum can also be used to check whether the relation
between the photon wavelength and the pixel number is linear. For SPICAM-UV,
the wavelength assignment after launch is given by
λ = 321.935 − 0.549216 ∗ pn (7.5)
where λ is the wavelength in nm and pn is the pixel number. The change between
ground and inflight calibration was very small. Figure 7.8 shows the spectrum
of γ Vel as determined by IUE. The spectrum obtained by SPICAM with the
correct wavelength assignment is shown by the red curve. The blue curve shows
the spectrum derived with the wavelength assignment measured on the ground.
There is a small shift between the two curves. This reference star can be used on a
Figure 7.8: The wavelength assignment was performed using the spectrum of γ
Vel (HR 3207) that shows narrow emission and absorption lines between 100 nm
and 300 nm. This allows us to check and correct the ground measurements of line
positions. The black curve shows the IUE spectrum. The spectrum obtained by
SPICAM with the correct wavelength assignment is shown by the red curve. The
blue curve shows the spectrum derived with the wavelength assignment measured
on the ground. The red and blue curves diverge at shorter wavelengths, but are
in close agreement at longer wavelengths. These spectra have been normalized to
arbitrary units for ease of comparison
regular basis to check for variations in wavelength assignment. So far, no long-term

changes have been detected within the SPICAM wavelength resolution element.
SPICAM Detector Gain Curve
SPICAM data are expressed in units of ADU (Analog to Digital Units), the
output of the CCD readout. When a photoelectron is created in the photocathode
(a photoevent), it eventually results in a pulse of light on the phosphor screen at
the back of the MCP. This pulse of light is distributed over a few pixels of the CCD.
It is detected by the CCD readout electronics as a total charge which is converted to
ADUs (NADU ). The high voltage (noted as HT hereafter) of the image intensifier
may be adjusted to give more or fewer electrons (hence ADUs) per photoevent, a
parameter that we note G, the gain of the UV detector. Therefore, the number of
counts, or NADU , depends on the gain of the detector, on the spectral resolution
Δλ and on the integration time Ti . We have the relation:
NADU = G × Np × Δλ × Ti (7.6)
SPICAM-UV 207
Figure 7.9: Linear relation between the signal SADU and its variance when observ-
ing a star at a given high voltage setting. The absolute gain is obtained from the
slope of the linear relation between the signal SADU and its variance
where Np is the number of photoevents detected by SPICAM per second and per
nanometer.
In other words, the gain G of a CCD camera is the conversion between the
number of photoevents (Np × Δλ × Ti ) collected at the photocathode during an
integration time Ti , and the number of digital units (NADU ) contained in the CCD
image.
There are two steps used to determine the absolute gain for all high voltage
settings. The first step is the determination of the gain Gk for a given high volt-
age. The reference gain Gk can be obtained inflight by an independent statistical
method. This is done by determining the mean number of photoevents contained
in the whole band devoted to the star signal outside of the atmosphere. The total
ADU signal fluctuations are analyzed, and the gain Gk is determined as:

a.b V ar(SADU ) V ar(SADU )
Gk = = 0.65 , (7.7)
c SADU SADU
where SADU is the mean value of a series of measurements and V ar(SADU ) is the
variance of the series. The scaling factor depends on different effects that modify
the statistical distribution of measurements. A detailed analysis by Dimarellis
(private communication, 1998) has shown that the scaling factor is affected by
• Factor a : The non-Gaussian distribution of the number of electrons coming

from a single photoevent (Sandel and Broadfoot 1986).
• Factor b : The spatial distribution of the cloud of electrons reaching the

phosphor screen (the spot spreads over more than one pixel).
• Factor c : The number of pixels used to determine the signal (binning of lines
of the CCD).
From numerical simulations, it was shown that the ratio a.b

c
is equal to 0.65
in the case of the SPICAM stellar observations used in Fig. 7.9. This ratio is
determined by the geometry of the instrument, and does not change with the value
of the high voltage. The determination of Gk is shown in Fig. 7.9 for a series
Figure 7.10: Variation of the number of counts (NADU ), for a given source, as a
function of the digital level of the high voltage setting. This curve (from Eq. 7.9)
shows the relative gain of the intensifier. This curve was obtained by comparing
stellar counts measured with different high voltage settings. This curve is very
stable over the whole dataset
of measurements collected at HT = 20. A reference gain G20 = 1.54 has been

obtained.
Taking the gain Gk as a reference gain, a relative gain Gi /Gk as a function
of HTi can be derived by computing the ratio NADU i /NADU k . The inflight de-
termination of the relative gain curve can be obtained by comparing the number
of ADUs recorded by the instrument at different high voltage level, when looking
at the same source, a star for instance (Fig. 7.10). The amplifier gain may be ad-
justed by telecommand with a high voltage level from 500 to 900 V, commanded
by a digital level HT from 0 to 255. With a gain Gk corresponding to a digital
level HTk , a number of counts NADU k is recorded. With another gain Gi , corre-
sponding to a digital level HTi , a number of counts NADU i is collected. Assuming
that the source observed is identical (same number of photoevent Np detected by
SPICAM per second and per nanometer, no time variation), we have the relation
(Eberhardt 1979):
NADU i Gi
= (7.8)
NADU k Gk
An absolute gain curve Gi = f (HTi ) can therefore be obtained by comparing
the NADU i number of counts collected at different digital levels HTi with the
reference NADU k number of counts collected at the reference digital level HTk .
Equation 7.9 results from assuming a linear relationship between the output voltage
and the level of the high voltage.
Gi (HTi ) = exp{7.46113 ln(500 + 1.57HTi ) − 46.3864} (7.9)

SPICAM-UV 209

G V
ln = ln
Gk Vk
and
Vi = a + b × HTi
SPICAM Instrumental Spectral Response Function

The SPICAM-UV Point Spread Function (PSF, in the spatial axis) and Instru-
mental Spectral Response Function (ISRF, in the spectral axis) were determined
inflight using stellar observations. Initially, it was assumed that these two functions
are very similar when expressed in pixel units. Therefore the PSF (or ISRF) was
modeled by a two-dimensional Voigt function. The fraction of a signal centered at
point (xi , yi ) that is seen at point (x, y) of the CCD is given by
PSF(x, y, xi , yi ) = Wi × H( ai , ki .r ) (7.10)

r= (x − xi )2 + (y − yi )2 (7.11)
2
a +∞
e−u
H(a, x) = du. (7.12)
π −∞ a + (u − x)2
2
H(a, x) is a Voigt function, i.e. a convolution of a Lorentzian and a Gaussian

function.
The coefficients (ai , ki ) appearing in Eq. 7.10 were fitted sequentially using a
Levenberg–Marquardt algorithm in order to minimize the discrepancy between
the measured PSF along the spatial axis and the computed PSF. Since this bi-
dimensional PSF at any given spectral index i depends not only on (ai , ki ), but also
on (aj , kj ) for j = i, multiple iterations—usually less than 5—of the Levenberg–
Marquardt fit were needed to converge on a stable solution. Stringent constraints
on the smoothness of ai and ki variations with respect to i were also provided for
numerical stability purposes. This PSF determination was designed (although not
detailed) by Marcq et al. (2011).
The coefficients (Wi ) were computed so that the two-dimensional integration
of Eq. 7.10 over (x, y) is equal to 1. The determination of coefficients (ai , ki ) was
made with a series of 90 individual images (384×288) of star ζ Pup. The result
of the fit of the Point Spread Function (PSF) is shown in Fig. 7.11. The Gaussian
part of H(a, x) is defined by parameters ki and characterizes the width of the core
(i.e. within 5 pixels from center position and then the focusing of the beam at a
given wavelength). The Lorentzian part, with coefficients (ai ), defines the level of
the wings and applies outside of the core but within about 20 pixels from center
position.
Following this determination of the PSF, we checked that the assumption that
the ISRF was very similar to the PSF by convolving a high resolution spectrum
of a star (α Vir, SPICA, O55) observed by UVVS/MASCS (McClintock, private
communication) with the SPICAM PSF. The resulting modeled spectrum and the
Figure 7.11: Point spread function measured in the spatial axis for star ζ Pup
centered on line 144 of the CCD (black line). This figure was obtained by binning
all columns between pixels 100 and 200 on the CCD image. The blue curve shows
the result of the fit of the function given in Eq. 7.10. The same binning between
pixels 100 and 200 has been applied. The fit is the convolution of a Gaussian (core)
and a Lorentzian (wings). Outside 20 pixels from the star position the signal is
dominated by the background noise
spectrum observed by SPICAM were in a good agreement thus validating the as-
sumption.
Calibration of SPICAM’s Effective Area Using Stellar

Reference Spectra
Selection of SPICAM Calibration Stars
Calibration of the SPICAM effective area using stars observed by SPICAM
requires calibrated reference spectra of the same stars from past or current UV
instruments. These stellar fluxes can be used to convert the SPICAM count rate
spectrum into physical units. The stars observed by SPICAM are listed in Table 7.4
in the Appendix. Of the large number of stars observed by SPICAM, 17 have been
observed by both SPICAM and SOLSTICE. SOLSTICE observations can provide
reference spectra in the range of 130–300 nm. For shorter wavelengths, we can use
the IUE spectra in the CALOBS archive discussed in the section “Comparison of
SOLSTICE to IUE Standard Stars,” http://www.stsci.edu/hst/observatory/cdbs/
calobs.html. There are eight stars listed in CALOBS which have also been observed
by SPICAM but have not all been observed by SOLSTICE. All the stars used in
the calibration of SPICAM are listed in Table 7.3.
SPICAM-UV 211
Table 7.3: SPICAM-UV calibration stars. The columns show the SPICAM catalog
number, the bright star catalog number, the star’s Bayer designation, spectral type,
the level of the high voltage (HT) used for the observation, the number of repeat
observations for that star, and the source for the reference spectrum. The column
listing figure numbers refers to the spectra shown in the appendix
SPICAM HR Star Spectral HT Nobs Source Figure
number number name type
9 1220 Per B0.5V 90 3 SOLSTICE 7.25

25 2294 β CMa B1II-III 30 2 SOLSTICE 7.22
28 2491 α CMa A1Vm 20 46 SOLSTICE 7.14
36 3165 ζ Pup O5f 20 132 CALOBS 7.34
41 3734 κ Vel B2IV-V 60 3 SOLSTICE 7.28
43 3982 α Leo B7V 140 3 CALOBS 7.17
46 4621 δ Cen B2IVn 80 3 SOLSTICE 7.23
48 4730 α Cru B0.5I 1 16 SOLSTICE 7.15
55 5056 α Vir B1III 1 3 SOLSTICE 7.20
57 5191 η Uma B3V 60 29 CALOBS 7.27
59 5231 ζ Cen B2.5I 40 4 SOLSTICE 7.32
60 5267 β Cen B1III 1 30 CALOBS 7.21
71 5953 δ Sco B0.3I 80 2 SOLSTICE 7.24
76 6165 τ Sco B0V 60 2 SOLSTICE 7.30
77 6175 ζ Oph O9.5V 160 4 CALOBS 7.33
88 7001 α Lyr A0V 100 3 CALOBS 7.18
89 7121 σ Sgr B2.5V 80 3 SOLSTICE 7.29
91 7790 α Pav B2IV 60 5 SOLSTICE 7.19
94 8425 α Gru B7IV 130 2 SOLSTICE 7.16
99 153 ζ Cas B2IV 80 48 CALOBS 7.31
122 1641 η Aur B3V 140 1 CALOBS 7.26
The main characteristics of the SPICAM stellar observations are also given in
Table 7.3. The High Voltage setting (HT) is low when the star is bright in the UV.
The number of observations shown in the table lists the number of separate obser-
vations with unique dates. For each distinct observation, the number of individual
spectra can go from a few hundred to a few thousand. Each spectrum has an inte-
gration time of 450 min or 640 min and the sampling rate during one observation
is 1 s. The star ζ Pup has been observed 132 times. This corresponds to more than
35,000 individual spectra. This star has been used to show the excellent stability
of the SPICAM-UV channel between October 2004 and April 2010. During that
period of time all spectra of that star show a variability lower then 5 % with no
systematic trend.
A visual inspection of the ratio between the SPICAM count rate and the SOL-
STICE irradiance shows that three stars, α Vir (O55), β Cen (O60) and α Cen
(O48), are outliers and will therefore not be used in the analysis. The other 14 stars
observed by both SOLSTICE and SPICAM will be used to derive the SPICAM
effective area. Six of the stars in the CALOBS archive—η UMa, ζ Oph, α Lyr,
α Leo, ζ Cas, and η Aur—have IUE measurements converted to the white dwarf
scale using the FLXCOR procedure from Bohlin (1996). Two other CALOBS
stars, ζ Pup and β Cen, are based on OAO-2 measurements that have been cor-
rected to the IUE scale but not the white dwarf scale (Bohlin and Holm 1981).
Since these two stellar spectra have not been converted to the white dwarf scale,
we have decided not to use them in our analysis.
The spectra in the CALOBS archive combine measurements from both the short
and long wavelength IUE cameras. The short wavelength channel covers the range
115–197.5 nm, while the long wavelength channel includes 191–330 nm. We note
that the long wavelength section of the spectrum for ζ Oph does not appear to be
consistent with the short wavelength portion.
As shown in Fig. 7.6, the calibration between SOLSTICE and IUE is consis-
tent to within their combined uncertainties. For wavelengths above 150 nm, the
uncertainty quoted for SOLSTICE is less than the nominal 3 % uncertainty quoted
for IUE spectra (Bohlin 1996). There are also 14 well-calibrated stars observed
by both SOLSTICE and SPICAM, while there are only six well-calibrated IUE
stars observed by SPICAM. Therefore we have chosen to use SOLSTICE irradi-
ance spectra to determine the SPICAM effective area for wavelengths greater than
150 nm and IUE irradiance spectra for wavelengths shorter than 150 nm.
Method of Determining SPICAM Effective Area

We can define the effective area of the detector, Aeff , with the following equation:
Np = Aeff Φ. (7.13)
In this relation, Np is the number of photoevents per second per nm reported by
the instrument electronics and Φ is the incident stellar flux in photons/s/cm2 /nm
through the aperture.
If we assume that an input spectrum for a given star measured by another
instrument is well calibrated, then we can use it to derive the effective area. Aeff is
therefore only a function of the instrument and the ratio Np /Φ should be the same
for all stars. Figure 7.12 shows the derived Aeff using IUE spectra below 150 nm
and SOLSTICE spectra above 150 nm. There is excellent agreement between the
average of the six IUE spectra and the average of the 14 SOLSTICE spectra at
150 nm where the crossover occurs. This implies that there is internal consistency
in all three instruments for all the measured stars. The red and blue curves in
Fig. 7.12 represent the mean effective area computed for each wavelength range
respectively.
Comparison of IUE Standard Stars to SOLSTICE

and SPICAM
In this section we will compare the measurements for all stars measured by the
three instruments (IUE, SOLSTICE, and SPICAM) to try to understand differences
caused by calibration and data processing. As described above, the reduction
Comparison of IUE Standard Stars to SOLSTICE and SPICAM 213
Figure 7.12: Effective area of the SPICAM-UV channel as a function of wavelength.

Above 150 nm, this result is based on the ratio of 14 SOLSTICE spectra (photons
cm−2 s−1 nm−1 ) with the count rate (in photoevents per nm per second) of SPI-
CAM. For wavelengths below 150 nm, the effective area is derived from the ratio
with six well-calibrated IUE spectra. The largest uncertainties are found close
to the Lyman-α line where slight error on the saturated line width causes large
variations in the ratio
of SPICAM spectra includes a normalization to match the IUE calibration for

wavelengths shorter than 150 nm and to match SOLSTICE above 150 nm. If the
reduction of all stars for SPICAM is done consistently, the fluxes would be in
good agreement for all stars, assuming that the same consistency is also true for
SOLSTICE and IUE spectra.
As shown in Fig. 7.6, the agreement between SOLSTICE and IUE is quite
good when the appropriate spectra are used. Only those IUE spectra which have
been corrected to the white dwarf scale match the SOLSTICE calibration, namely
η UMa, α Lyr, and α Leo. Other spectra in the IUE archives can differ from this
calibration by up to 20 % (e.g. the MAST archive http://archive.stsci.edu/iue).
On IUE, there are two apertures (large and small), plus options for moving the
star along the slit during the observation to increase the signal-to-noise ratio. For
bright stars such as those observed by SOLSTICE, details of the observing strat-
egy can result in deviations from a well-calibrated irradiance spectrum during data
reduction. Users of the IUE archive should be aware of these issues and use spectra
calibrated to the white dwarf scale if absolute calibration is important to the user’s
science goals.
Figure 7.13 shows the ratio of SOLSTICE and SPICAM fluxes to the IUE
spectra. In general, the ratios are very close to unity with only occasional outliers.
There are no systematic trends with either wavelength or stellar brightness. Stars
with relatively low ultraviolet irradiances agree just as well as the brightest stars.
The ratio of SOLSTICE to SPICAM (top panel) is quite consistent as expected
since the calibration of these SPICAM measurements is based on the SOLSTICE
observations.
The curves in the lower panel of Fig. 7.13 are not quite as flat as those in the
top panel. At the short end of the wavelength scale, the ratio appears to have a
Figure 7.13: Irradiance ratio of SOLSTICE to SPICAM (top panel) and SPICAM
to IUE (lower panel) for stars observed by both instruments. The SPICAM to IUE
ratios are only for stars in the CALOBS archive that have been converted to the
white dwarf scale. The large ratio near Lyman-α is likely due to a mismatch in
spectral resolution rather than a true divergence in calibration
large variation with wavelength. This is due to the sharp features near Lyman-α
in all the spectra. Slight errors in either the wavelength scale or in matching the
spectral resolution before taking the ratio will lead to the large variance in the
ratio. There is no reason to believe that the true calibration changes so rapidly
with wavelength.
The other interesting feature in the SPICAM/IUE ratio is the broad feature
near 190 nm. In the 185–205 nm range, the SPICAM to IUE ratio, with a value of
1.06, is significantly larger than the mean ratio of SOLSTICE to SPICAM, which
is equal to 1.00 in the same range. This is the wavelength range where both IUE
channels have greater uncertainty in their responsivity and correspondingly lower
signal-to-noise ratio in all their spectra. Note that the transition between the two
SOLSTICE channels occurs at 180 nm. The two SOLSTICE channels appear to
be consistent since there is no discontinuity when switching from one to the other
at that wavelength. So we are confident that the region around 190 nm is less well
calibrated, and that explains the “bump” in the SPICAM/IUE ratio.
Discussion 215
Discussion
It has been more than 30 years since IUE was launched, yet it still forms the basis
of absolute calibration for ultraviolet astronomy. IUE observed so many targets
during its mission that nearly all current missions can find an IUE measurement
of any target that it observes. SPICAM and SOLSTICE observe many of the
same stars, and the IUE observations of those stellar spectra allow them to be
intercompared.
SPICAM uses SOLSTICE and IUE reference spectra from a limited number of
stars to determine its inflight calibration. The effective area calculated from the
average of those few spectra is then used to calibrate the observations of the rest of
its stellar catalog. This system produces spectra that are consistent to about 5 %.
That is presumably the uncertainty of the reduction method.
SOLSTICE, on the other hand, uses a preflight calibration at the SURF-III syn-
chrotron source (McClintock et al. 2005b) and then tracks changes to the instru-
ment responsivity by making regular observations of the ensemble of catalog stars
(Snow et al. 2005). The absolute uncertainty in the preflight calibration shown in
Fig. 7.3 is estimated to be about 3 % or better. The measurement uncertainty and
degradation correction are no larger than 2 %. These two sources of uncertainty
are independent and can therefore be added in quadrature.
The uncertainty of the IUE absolute calibration was initially estimated to be
10 % when based on only η UMa (Bohlin et al. 1990). Subsequent work using
model atmospheres of white dwarfs has reduced the estimated uncertainty of the
IUE absolute fluxes to just a few percent (Bohlin 1996). The repeatability of an
IUE observation is similarly at the few percent level (Bohlin et al. 1990). The
two panels of Fig. 7.13 confirm that the consistency from star to star for the three
instruments is on the order of a few percent. The average ratio between IUE and
SOLSTICE is close to 1 with a 1-σ spread of less than 5 %. The IUE spectra in
the CALOBS archive which have been converted to the white dwarf scale have
irradiances that match SOLSTICE’s, but the user should be alert to the fact that
some spectra in that archive have not been converted. Careful inspection of the
headers of the files in the CALOBS archive is strongly advised.
SORCE SOLSTICE’s absolute stellar irradiance is derived from a measured
calibration (McClintock et al. 2005b), while IUE’s ultimate calibration is from
a theoretical understanding of white dwarf atmospheres (Bohlin 1996). These
two methods are in good agreement over the full wavelength range considered here
(150–300 nm), which provides independent validation of the white dwarf model.
The ground calibration of SPICAM-UV was not very accurate. This study has
provided a new absolute calibration based on SOLSTICE’s absolute calibration
and which is therefore in agreement with the white dwarf absolute scale. This
calibration can then be used to produce more spectra in absolute units based on
SPICAM’s stellar observations. The SOLSTICE and SPICAM spectra available
on the FONDUE web service, http://bdap.ipsl.fr/fondue/, are consistent with the
best calibrated IUE spectra.
Figure 7.14: α CMa
Figure 7.15: α Cru
Figure 7.16: α Gru

Discussion 217
Figure 7.17: α Leo
Figure 7.18: α Lyr
Figure 7.19: α Pav

Figure 7.20: α Vir
Figure 7.21: β Cen
Figure 7.22: β CMa

Discussion 219
Figure 7.23: δ Cen
Figure 7.24: δ Sco
Figure 7.25: Per

Figure 7.26: η Aur
Figure 7.27: η Uma
Figure 7.28: κ Vel

Discussion 221
Figure 7.29: σ Sgr
Figure 7.30: τ Sco
Figure 7.31: ζ Cas

Figure 7.32: ζ Cen
Figure 7.33: ζ Oph
Figure 7.34: ζ Pup

Acknowledgements 223
Acknowledgements
The work was supported by NASA contract NAS5-97045 (SORCE) at the Uni-
versity of Colorado. We would like to thank the referees for their assistance in
understanding the IUE white dwarf scale and their helpful comments on the initial
draft of the paper. The authors would also like to thank the International Space
Science Institute, Bern, Switzerland for their support of the working group.
Appendix
For each of the 19 stars listed in Table 7.3, we show spectra from each of
the three instruments—IUE, SOLSTICE, and SPICAM—if good spectra exist for
that star. The ratio between the spectra is also shown in the lower panel of each
figure. The corresponding data are available on-line from the FONDUE web service,
http://bdap.ipsl.fr/fondue/. The spectra are in FITS files including statistical
uncertainties.
Table 7.4: List of all stars observed by SPICAM. Columns give the SPICAM ob-
serving number, the bright star catalog number, star’s Bayer designation, spectral
type, celestial coordinates, and visual magnitude
SPI HR Name and constellation Spectral type RA2000 DEC2000 Mag
1 39 γ Peg B2IV 3.31 15.18 2.83

2 264 γ Cas B0IVe 14.18 60.72 2.47
5 472 α Eri B3Vpe 24.43 −57.24 0.46
8 1203 ζ Per B1Ib 58.53 31.88 2.85
9 1220 Per B0.5V 59.46 40.01 2.89
12 1713 β Ori B8Ia: 78.63 −8.2 0.12
13 1788 η Ori B1V+B 81.12 −2.4 3.36
14 1790 γ Ori B2III 81.28 6.35 1.64
15 1791 β Tau B7III 81.57 28.61 1.65
16 1852 δ Ori O9.5I 83 −0.3 2.23
17 1879 λ Ori O8III 83.78 9.93 3.54
18 1899 ι Ori O9III 83.86 −5.91 2.77
19 1903 Ori B0Ia 84.05 −1.2 1.7
20 1948 ζ Ori O9.7I 85.19 −1.94 2.05
21 2004 κ Ori B0.5I 86.94 −9.67 2.06
23 2088 β Aur A2IV 89.88 44.95 1.9
24 2282 ζ CMa B2.5V 95.08 −30.06 3.02
25 2294 β CMa B1II- 95.68 −17.96 1.98
26 2326 α Car F0II 95.99 −52.7 −0.72
27 2421 γ Gem A0IV 99.43 16.4 1.93
28 2491 α CMa A1Vm 101.29 −16.72 −1.46
29 2618 CMa B2II 104.66 −28.97 1.5
30 2653 ω2 CMa B3Iab 105.76 −23.83 3.02
32 2827 η CMa B5Ia 111.02 −29.3 2.45
33 2891 α Gem A1V 113.65 31.89 1.98
34 2943 α CMi F5IV- 114.83 5.22 0.38
36 3165 ζ Pup O5f 120.9 −40 2.25
37 3207 γ2 Vel WC8+O 122.38 −47.34 1.78
38 3307 Car K3III 125.63 −59.51 1.86
39 3485 δ Vel A1V 131.18 −54.71 1.96
40 3685 β Car A2IV 138.3 −69.72 1.68
41 3734 κ Vel B2IV- 140.53 −55.01 2.5
43 3982 α Leo B7V 152.09 11.97 1.35
44 4199 θ Car B0Vp 160.74 −64.39 2.76
46 4621 δ Cen B2IVn 182.09 −50.72 2.6
47 4656 δ Cru B2IV 183.79 −58.75 2.8
48 4730 α1 Cru B0.5I 186.65 −63.1 1.33
49 4731 α2 Cru B1V 186.65 −63.1 1.73
51 4798 α Mus B2IV- 189.3 −69.14 2.69
52 4844 β Mus B2.5V 191.57 −68.11 3.05
53 4853 β Cru B0.5I 191.93 −59.69 1.25
54 4905 UMa A0pCr 193.51 55.96 1.77
55 5056 α Vir B1III 201.3 −11.16 0.98
56 5132 Cen B1III 204.97 −53.47 2.3
Appendix 225
SPI HR Name and constellation Spectral type RA2000 DEC2000 Mag

57 5191 η UMa B3V 206.88 49.31 1.86
58 5193 μ Cen B2IV- 207.4 −42.47 3.04
59 5231 ζ Cen B2.5I 208.88 −47.29 2.55
60 5267 β Cen B1III 210.96 −60.37 0.61
62 5440 η Cen B1.5V 218.88 −42.16 2.31
66 5571 β Lup B2III 224.63 −43.13 2.68
67 5576 κ Cen B2IV 224.79 −42.1 3.13
68 5695 δ Lup B1.5I 230.34 −40.65 3.22
69 5776 γ Lup B2IV 233.79 −41.17 2.78
70 5944 π Sco B1V+B 239.71 −26.11 2.89
71 5953 δ Sco B0.3I 240.08 −22.62 2.32
73 5984 β1 Sco B1V 241.36 −19.81 2.62
74 6084 σ Sco B1III 245.3 −25.59 2.89
76 6165 τ Sco B0V 248.97 −28.22 2.82
77 6175 ζ Oph O9.5V 249.29 −10.57 2.56
79 6247 μ1 Sco B1.5V 252.97 −38.05 3.08
80 6453 θ Oph B2IV 260.5 −25 3.27
81 6462 γ Ara B1Ib 261.35 −56.38 3.34
82 6508 Sco B2IV 262.69 −37.3 2.69
83 6510 α Ara B2Vne 262.96 −49.88 2.95
84 6527 λ Sco B2IV+ 263.4 −37.1 1.63
86 6580 κ Sco B1.5I 265.62 −39.03 2.41
87 6879 Sgr B9.5I 276.04 −34.38 1.85
88 7001 α Lyr A0Va 279.23 38.78 0.03
89 7121 σ Sgr B2.5V 283.82 −26.3 2.02
90 7557 α Aql A7V 297.7 8.87 0.77
91 7790 α Pav B2IV 306.41 −56.74 1.94
92 7924 α Cyg A2Ia 310.36 45.28 1.25
93 8238 β Cep B1IV 322.17 70.56 3.23
94 8425 α Gru B7IV 332.06 −46.96 1.74
96 8728 α PsA A3V 344.41 −29.62 1.16
98 15 α And B9II 2.1 29.09 2.06
99 153 ζ Cas B2IV 9.24 53.9 3.66
107 1122 δ Per B5III 55.73 47.79 3.01
122 1641 η Aur B3V 76.63 41.23 3.17
125 1756 λ Lep B0.5IV 79.89 −13.18 4.29
132 1855 Ori B0V 82.98 −7.3 4.62
138 1910 ζ Tau B2IV 84.41 21.14 3
139 1931 48σ Ori O9.5V 84.69 −2.6 3.81
152 2538 κ CMa B1.5IVe 102.46 −32.51 3.96
186 4133 ρ Leo B1Iab 158.2 9.31 3.85
191 4590 TY Crv B2IV 180.21 −19.66 5.26
202 5190 ν Cen B2IV 207.38 −41.69 3.41
203 5248 φ Cen B2IV 209.57 −42.1 3.83
215 5708 Lup B2IV-V 230.67 −44.69 3.37
221 5948 η Lup B2.5IV 240.03 −38.4 3.41
247 8353 γ Gru B8III 328.48 −37.36 3.01
Bibliography
U. Arp, R. Friedman, M.L. Furst, S. Makar, P.S. Shaw, SURF III—an improved storage
ring for radiometry. Metrologia 37, 357–360 (2000). doi:10.1088/0026-1394/37/5/2
J-L. Bertaux et al., SPICAM on Mars express: observing modes and overview of
uv spectrometer data and scientific results. J. Geophys. Res. 111, E10S90 (2006).
doi:10.1029/2006JE002690
J-L. Bertaux et al., SPICAV on Venus express: three spectrometers to study the global
structure and composition of the Venus atmosphere. Planet. Space Sci. 55, 1673–1700
(2007)
R.C. Bohlin, Spectrophotometric standards from the far-uv to the near-ir on the white dwarf
flux scale. Astrophys. J. 111, 1743–1747 (1996). doi:10.1086/117914
R.C. Bohlin, A.V. Holm, New software — absolute calibration — iue. IUE ESA Newsletter
11, 18 (1981)
R.C. Bohlin, W.M. Sparks, A.V. Holm, B.D. Savage, M.A.J. Snijders, Photometric calibra-
tion of the international ultraviolet explorer (IUE): low dispersion. Astron. Astrophys.
85, 1–13 (1980)
R.C. Bohlin, A.W. Harris, A.V. Holm, C. Gry, The ultraviolet calibration of the hubble space
telescope iv. absolute IUE fluxes of hubble space telescope standard stars. Astrophys. J.
Suppl. 73, 413–439 (1990). doi:10.1086/191474
R.C. Bohlin, M.E. Dickinson, D. Calzetti, Spectrophotometric standards from the far-
ultraviolet to the near-infrared STIS and NICMOS fluxes. Astrophys. J. 122, 2118–2128
(2001). doi:10.1086/323137
R.C. Bohlin et al., Absolute flux calibration of the IRAC instrument on the Spitzer space
telescope using hubble space telescope flux standards. Astrophys. J. 141, 173–185 (2011).
doi:10.1088/004-6256/141/5/173
E.H. Eberhardt, Gain model for microchannel plates. Appl. Opt. 18, 1418–1423 (1979).
doi:10.1364/AO.18.001418
E. Marcq, D. Belyaev, F. Montmessin, A. Fedorova, J-L. Bertaux, A.C. Vandaele, E. Neefs,
An investigation of the SO2 content of the venusian mesosphere using SPICAV-UV in
nadir mode. Icarus 211, 58–69 (2011). doi:10.1016/j.icarus.2010.08.021
W.E. McClintock, G. Rottman, T.N. Woods, Solar stellar irradiance comparison ex-
periement II (SOLSTICE II): instrument concept and design. Sol. Phys. 230, 225–258
(2005a). doi:10.1007/s11207-005-7432-x
W.E. McClintock, M. Snow, T.N. Woods, Solar stellar irradiance comparison experiment II
(SOLSTICE II): pre-launch and on-orbit calibrations. Sol. Phys. 230, 259–294 (2005b).
doi:10.1007/s11207-005-1585-5
D. Mihalas, J. Binney, Galactic astronomy structure and kinematics (Freeman, New York,
1981), p. 135
G. Rottman, The SORCE mission. Sol. Phys. 230, 7–25 (2005). doi:10.1007/s11207-005-
8112-6
G. Rottman, T.N. Woods, T. Sparn, Solar stellar irradiance comparison experiment
1. I — instrument design and operations. J. Geophys. Res. 98, 10667 (1993).
doi:10.1029/93JD00462
B.R. Sandel, A.L. Broadfoot, Statistical performance of the intensified ccd. Appl. Opt. 25,
4135–4140 (1986). doi:10.1364/AO.25.004135
M. Snow, W.E. McClintock, G. Rottman, T.N. Woods, Solar stellar irradiance comparison
experiment II (SOLSTICE II): examination of the solar-stellar comparison technique. Sol.
Phys. 230, 295–324 (2005). doi:10.1007/s11207-005-8763-3
—8—
Absolute Ultraviolet Irradiance of the

Moon from the LASP Lunar Albedo
Measurement and Analysis from
SOLSTICE (LLAMAS) Project
Martin Snow∗ , Gregory M. Holsclaw

William E. McClintock, and Tom Woods
Boulder, CO, USA
Abstract
The Moon has been shown to be an extremely stable radiometric reference for
calibration and long-term stability measurements of on-orbit sensors. The majority
of the previous work on characterizing the lunar reflectance has been in the visible
part of the spectrum using ground-based lunar images. The SOLar-STellar Irradi-
ance Comparison Experiment (SOLSTICE) on the SOlar Radiation and Climate
Experiment (SORCE) can be used to extend the lunar spectral irradiance dataset
to include the 115–300 nm range. SOLSTICE can directly measure both the solar
and lunar spectra from orbit, using the same optics and detectors. An observing
campaign to map out the reflectance as a function of phase angle began in mid 2006
and continued through 2010. The geometry of SORCE’s orbit is very favorable for
lunar observations, and we have measurements spanning a range 0–170◦ in phase
angle. In addition to Earth Observing Systems using the Moon for calibration, re-
cent planetary missions have also made ultraviolet observations of the Moon during
Earth flyby, and these SOLSTICE measurements can be useful in calibrating the
absolute responsivity of those instruments as well.
Introduction
One of the many challenges to quantitative, space-based, absolute remote-
sensing measurements is tracking long-term instrumental changes. Comparison
to a standard source is one way to monitor such changes. Some missions may carry

228 8. Ultraviolet Irradiance of the Moon
a comparison source onboard, while others may rely on cross-calibration with other
observations such as a rocket underflight. But if a suitable celestial target whose
radiometric properties are well known is available, then it can be a valuable weapon
in the calibration arsenal. The Moon is such a radiometric standard.
Over 25 years ago, it was recognized that the Moon could be used for on-orbit
calibration (Kieffer and Wildey 1985, 1996; Pugacheva et al. 1993), and a ground-
based observatory program to determine the radiometric properties of the Moon
in the visible and near-infrared part of the spectrum began in 1996 (Kieffer and
Anderson 1998; Stone and Kieffer 2002). The goal of the LASP Lunar Albedo
Measurement and Analysis from SOLSTICE (LLAMAS) project is to extend the
measurements of the lunar irradiance to the ultraviolet using data from the SOLar-
STellar Irradiance Comparison Experiment (SOLSTICE) (McClintock et al. 2005a)
on the SOlar Radiation and Climate Experiment (SORCE) spacecraft (Rottman
2005).
The first comprehensive campaign to characterize the disk-integrated
wavelength-dependent phase curve of the Moon was conducted by Lane and Irvine
(1973). The geometric albedo was determined for each bandpass filter across the
visible region and an empirical phase curve was derived over the phase range of
6–120◦ . The results from that study effectively demonstrated the phase-reddening
effect. This is the phenomenon where the color of the Moon tends to become
increasingly red at larger phase angles. In this context, color is the ratio of the
brightness of the Moon at a longer wavelength to a shorter wavelength.
The RObotic Lunar Observatory (ROLO) was a program within the United
States Geological Survey (USGS) to accurately determine the exoatmospheric spec-
tral irradiance and phase curve of the Moon from 350 to 2450 nm (Stone and Kieffer
2002). The Moon was observed every clear night throughout a six-year period cov-
ering phase angles of −90 to +90◦ (negative phase angles refer to lunar phases prior
to full Moon) through multiple bandpass filters in the visible and near-infrared. An
empirical model with 18 fitting coefficients in each of 32 bands for the lunar disk-
integrated reflectance was developed. Eight of these coefficients are common to all
bands for a total of 328 coefficients in the ROLO photometric model. The residual
from this fit was about 1% (Kieffer and Stone 2005).
The Moon has been observed at ultraviolet wavelengths by many other space-
based missions, and a careful comparison of their results with the SOLSTICE
measurements is planned for future publication. Figure 8.1 shows the primary
datasets for the ultraviolet wavelength range as a function of wavelength and phase
angle coverage.
Starting at the shortest wavelengths, the Hopkins Ultraviolet Telescope (HUT)
(Henry et al. 1995) and Apollo 17 (Lucke et al. 1976) determined the albedo from
radiance measurements of relatively small regions of the lunar disk. The Nozomi
ultraviolet scanner took an image of the Moon at a phase angle of about 130◦ ,
but the only wavelength that had good signal was Lyman-α (Taguchi et al. 2000).
Mariner 10 (Wu and Broadfoot 1977), Galileo (Hendrix 1996), Cassini (Hendrix
et al. 2009), and the MErcury Surface, Space ENvironment, GEochemistry, and
Ranging (MESSENGER) spacecraft (Holsclaw et al. 2010) are all planetary mis-
sions that took data during lunar flybys. The footprint of their fields-of-view (FOV)
on the Moon depend on their distance from the Moon during the observation.
Instrument Description 229
Signal levels also vary widely among datasets, and we have just begun to work
on intercomparing them. The International Ultraviolet Explorer (IUE) took sev-
eral spectra of the Moon, but those results are still unpublished (Michael Combi,
private communication).
In the middle ultraviolet (MUV, 180–300 nm), Galileo and the Student Nitric
Oxide Explorer (SNOE) made partially overlapping measurements at a phase angle
of about 20◦ (Hendrix 2002). These were also radiance observations over varying
FOVs. The Global Ozone Monitoring Experiment (GOME; Dobber 1996, 1997)
and the Shuttle Solar Backscatter Ultraviolet (SSBUV; Janz et al. 1996) made spec-
tral irradiance measurements of the Moon, but since these were specially planned
events, they span a very small range of phase angle. The Advanced Camera for
Surveys (ACS) on the Hubble Space Telescope (HST) obtained measurements of
the lunar radiance at 250 nm and a few visible wavelengths (Robinson et al. 2007).
Finally, in the near-UV, dedicated observational campaigns by Lane and Irvine
(1973) using the Harvard College Observatory and Kieffer and Stone (2005) using
ROLO have filled in large ranges of phase angle over these wavelengths and into
the visible. The ROLO dataset is by far the most complete.
The Lunar Reconnaissance Orbiter (LRO) has been in orbit around the Moon
since June 2009 and includes the Lyman Alpha Mapping Project (LAMP) instru-
ment (Gladstone et al. 2010). In addition to its primary mission of searching for
ice in the permanently shadowed polar regions, it is also taking measurements of
the lunar albedo in the ultraviolet and will ultimately map out a large fraction of
the Moon on the dayside. The Wide Angle Camera on the Lunar Reconnaissance
Orbiter Camera (LROC) has a number of filter bandpasses in the near ultravi-
olet, but the shortest wavelength they observe is 320 nm (Robinson et al. 2010).
Although there is no overlap between LROC/WAC and SOLSTICE, trends in spec-
tral and photometric properties vary continuously with wavelength, so trends seen
in the two instruments will provide useful insight into the surface properties of
the Moon.
As Fig. 8.1 clearly shows, there have been many ultraviolet lunar observational
datasets, but none have the wavelength and phase angle coverage that the SOL-
STICE measurements provide. Prior to SOLSTICE, there was very little overlap in
phase and wavelength between the various sets of observations, so intercomparisons
were generally quite difficult and had fairly large uncertainties. Our new dataset
can be a bridge between them all and will greatly improve our knowledge of the
lunar albedo throughout the UV for all observing geometries.
Instrument Description
SOLSTICE is a grating spectrometer that can measure ultraviolet irradiance
over an extremely large dynamic range (McClintock et al. 2005a). Its primary
mission is to measure the absolute solar irradiance on a daily basis. It is on board
the SOlar Radiation and Climate Experiment (SORCE) (Rottman 2005) which is
a small satellite in low earth orbit. During orbit day, SOLSTICE measures the
solar spectral irradiance, but during the eclipse portion of the orbit, the spacecraft
turns and SOLSTICE measures the irradiance from a variety of celestial targets
Figure 8.1: Catalog of lunar ultraviolet albedo measurements. The shaded areas
represent the phase and wavelength coverage for each dataset. An underlined
dataset name indicates that the measurement comes from images. Dataset names
in parentheses are not full-disk measurements. (From Snow et al. 2007)
including the Moon. The altitude of the orbit is 580 km with an inclination of 40◦ .
This 96-minute orbit allows lunar viewing on a routine basis, including nearly the
full range of phase angles (0–170◦ ).
Before launch, SOLSTICE was calibrated using Beam Line 2 at the Synchrotron
Ultraviolet Radiation Facility III (SURF III) at the National Institute of Standards
and Technology (NIST) in Gaithersburg, MD (Arp et al. 2000). This synchrotron
beam is a standard source in the ultraviolet and the uncertainty in its irradiance
is only 0.75% (Arp et al. 2000). The final preflight accuracy of SOLSTICE’s re-
sponsivity is on the order of 5% (McClintock et al. 2005b). To monitor long-term
degradation of the instrument response, SOLSTICE uses an ensemble of early-type
stars as its calibration reference source, thus the need for a dynamic range of 108 .
The wavelength range measured by SOLSTICE is from 115 to 300 nm using two
channels. In solar mode, the entrance aperture is a 0.1 mm square and the exit slit
for each channel is 0.375 mm wide. The spectral resolution in this mode is 0.1 nm.
The instrument can exchange these small entrance and exit slits with much larger
apertures to convert to stellar mode (i.e. lunar mode).
Replacing the small solar entrance aperture with a large 16 mm circular aper-
ture and increasing the exit slit width to 0.75 mm decreases the resolution dur-
ing a stellar measurement to 1.1 nm in the far ultraviolet (FUV) channel (115–
180 nm). The exit slit in the middle ultraviolet (MUV) channel (180–300 nm) is
1.5 mm for a spectral resolution of approximately 2.2 nm in the 180–300 nm range.
Determination of the spectral resolution of a lunar measurement is somewhat
Instrument Description 231
Grating Drive
Entrance Slit Assembly Sphere-Sphere
Assembly Kinematic Mount
Purge/
Fill Assy
Sunshade Fold Mirror
Assembly Assembly
Door Mech &

EMI Filter Assy Sphere-Rigid
Kinematic Mount
Sphere-Translate
Kinematic Mount
Camera Mirror Exit Slit Detector Head

Assembly Assembly Assembly
Figure 8.2: Layout of SORCE SOLSTICE. (From McClintock et al. 2005b)
more complicated and will be explained below. The rest of the optical system is
unchanged in the switch from solar to stellar mode. For part of the MUV solar
measurement, a neutral density filter is placed into the optical path just before
the exit slit to reduce the detector count rate. In-flight tracking of this filter has
shown no measureable changes to its transmission since launch. Figure 8.2 shows
the opto-mechanical layout of the SORCE SOLSTICE instrument.
In stellar mode, the instrument operates as an objective grating spectrometer.
While this mode is well-suited to lunar observations, it does introduce a few com-
plications to the data processing. The two major issues are wavelength shifts as
the target moves within the FOV and vignetting of the incoming rays. Both of
these effects will be discussed in detail below.
SOLSTICE has the unique ability to measure both the solar and lunar irradi-
ances with the same optics and detectors. The entrance and exit slits are different
between the two modes, but the elements that disperse, reflect, or detect light, i.e.
the optical elements that might change properties over time, are the same for both.
The ratio of the aperture areas for the two modes was measured by NIST before
flight and has an uncertainty of 0.5% (McClintock et al. 2005b). A detailed un-
certainty analysis will be included in the section “Uncertainty” below. Most other
albedo measurements rely on either model solar spectra or must include the un-
certainties from relative calibration of two separate instruments. The ratio of solar
and lunar irradiances can be calculated directly from the SOLSTICE observations
and a large number of calibration factors in this ratio cancel out.
In the discussions that follow, we will need to make reference to the x, y, and
z coordinate frame of the instrument. The z-axis is the direction of the optic axis,
often referred to as the boresight. The x-axis is in the dispersion direction for
SOLSTICE B, and the y-axis is the cross-dispersion direction. The plane defined
by the instrument, the Moon, and the Sun is known as the scattering plane. The
roll angle of the instrument will be defined as the angle between the x-axis and the
surface normal to the scattering plane.
Lunar Observing Campaign

As described in Snow et al. (2007), SOLSTICE began a campaign of repeated
observations of the Moon in June 2006. The schedule of eclipse calibration exper-
iments is determined by an expert system (Pankratz et al. 2005) that fills in the
observing time with targets from our catalog. We simply added the Moon as if
it were one of our calibration stars. The planning system adds activities to the
observing schedule based on how many times each target has been observed in the
past, as well as how long since the most recent observation. This strategy gives
the system the freedom to schedule observations as efficiently as possible. With
a large enough number of observations, all phase and libration angles will be ob-
served eventually. From 1 July 2006 through 1 July 2010, we acquired 1733 FUV
and 1783 MUV spectra. That is an average of more than one spectrum in each
channel every day.
The lunar observing campaign on SOLSTICE stopped near the end of 2010. The
batteries on the SORCE spacecraft have been slowly degrading over the past few
years, and the operations team has been pro-active in preserving their capacity for
as long as possible. One action has been to power off both SOLSTICE instruments
during the majority of eclipses. There are two identical SOLSTICE spectrometers
on the SORCE satellite (SOLSTICE A and SOLSTICE B). SOLSTICE B remains
on during the shortest eclipses, but then it must devote all of its observing time
to stellar calibration experiments. These measures will help to extend the overall
SORCE mission for another few years, but the lunar reflectance portion of the
mission has ended. The entrance aperture mechanism on SOLSTICE A suffered
an anomaly in January 2006, and it has been used only in solar mode since then.
All lunar observations discussed in this paper have been taken with the SOLSTICE
B instrument.
Phase Angle Coverage

The standard observation mode for SOLSTICE is that during the daylight
portion of the orbit, the instrument measures the solar irradiance. During eclipse,
the spacecraft rotates to observe calibration stars, dark regions on the sky, and the
Moon (McClintock et al. 2005a; Snow et al. 2005, 2007). Figure 8.3 shows the phase
angle coverage of the SOLSTICE lunar dataset. Note that these observations span
nearly the entire range of possible phase angles. The limitation at large phase angle
is that the Moon must be visible for at least a few minutes before it is occulted
by the Earth in order to scan the spectrum. There is no intrinsic limitation at low
phase angle, although in practice it is rare for the conditions of near-zero phase to
occur.
Data Reduction 233
Figure 8.3: Phase angles of SOLSTICE lunar observations
Libration Coverage
Although only one side of the Moon generally faces the Earth, there is a small
oscillation of the selenographic longitude and latitude of the sub-observer point
due to the eccentricity of the lunar orbit and the inclination of the Moon’s axis
of rotation relative to the plane of its orbit around the Earth. Similarly, the
sub-spacecraft point on the Moon varies as a function of time as the celestial
geometry evolves. With regular observations from mid-2006 until the end of 2010,
SOLSTICE has achieved fairly good coverage of the full range of possible libration
angles. Unlike a ground-based observatory, the selenographic coordinates of the
observation can sometimes change by a significant amount on short timescales
due to the orbital motion of the spacecraft. Figure 8.4 shows the position of the
sub-spacecraft point on the Moon at the beginning and end of each of the FUV
observations (diamonds). The start and stop points for each scan are connected
by a line. Occassionally, the sub-spacecraft point will have moved by as much as
a degree during the observation, but in general the change is much smaller. The
MUV observations have a very similar distribution of selenographic positions, so
we only show the FUV observations in Fig. 8.4.
Data Reduction
In order to create a photometrically accurate model of the lunar reflectance
over the entire wavelength and phase angle region of interest, the raw data must
be carefully corrected and reduced. Some of the reduction steps are the same
as for a standard SOLSTICE stellar observation, but many of them are specific
Figure 8.4: Selenographic coordinates (in degrees) of the sub-spacecraft point. Due
to the orbital motion of the spacecraft, the sub-spacecraft point changes throughout
the observation. Pairs of diamonds (connected by lines) show the start and end
selenographic position for each scan
to the lunar observing conditions. We will briefly describe the additional correc-
tions needed to fully reduce the lunar observations in this section. The discussion
in the following sections extends the basic data reduction process described in
Snow et al. (2007).
Pointing
The SORCE spacecraft has only two tracking modes: inertial targets and the
Sun. Unfortunately, the Moon is neither of those. In order to observe the Moon,
we schedule and execute the observation as if it were an inertial target. Parallax
due to spacecraft orbital motion causes the Moon to drift away from the center
of the FOV during an observation. To minimize this problem, we have developed
an observing strategy where we re-point at the current position of the Moon every
2 min, so it never gets more than 10 arc minutes from the center. Figure 8.5 shows
the position of the Moon in the FOV for typical observations. The two axes are the
pitch and yaw directions in the instrument frame of reference. When the spacecraft
re-points at the current position of the Moon, it also makes adjustments to the roll
angle. The roll angle refers to rotations about the spacecraft z-axis. Therefore the
drift direction of the Moon can be slightly different after each pointing correction.
Because SOLSTICE operates as an objective grating spectrometer, offsets in the
position of the target from the instrument boresight have the effect of displacing
Data Reduction 235
Figure 8.5: Position of the Moon in the field-of-view of the instrument for a few
typical observations. The axes are the pitch and yaw directions for the instrument.
The crosses show the position of the center of the lunar disk as it drifts away
from the center of the FOV during the observation. The spacecraft is periodically
re-pointed to the new position of the Moon to re-center it, but during each such
maneuver, the roll angle of the spacecraft is allowed to change, so the new direction
of the Moon’s drift can be different after each re-pointing
the effective wavelength sampled by the detector. This shift can be positive or
negative and changes continuously throughout an observation depending on the
attitude of the spacecraft. Uncorrected wavelength shifts would lead to errors in
the reflectance that would be particularly large when there are sharp features in the
spectrum. This effect can be characterized analytically from the grating equation:
mλsolar = 2d sin(θS ) cos(φG ), (8.1)
where m is an integer representing the diffraction order, and θS is the grating angle,
given by
β+α
θS = , (8.2)
2
and the deviation angle is defined as
β−α
φG = . (8.3)
2
α and β in these relationships are the incidence and diffraction angles, respectively.
The stellar-mode wavelength equation is given by:
mλstellar = 2d sin(θS + /2) cos(φG − /2) (8.4)
where is the offset angle of the star (or in our case, the Moon) from the boresight
in the grating dispersion plane. The nominal value (i.e. when the target is located
on the optic axis of the instrument) of φG is fixed by the instrument geometry
and has a value of 2.32◦ (McClintock et al. 2005a). For all wavelengths such that
|β| > |α|, both incident and diffracted beams lie on the same side of the grating
normal, and the adopted sign convention is such that m = +1 with both α and β
positive. The angle is effectively a change in the grating incidence angle, α, but
does not change the diffraction angle, β.
McClintock et al. (2005a,b) provide the stellar-mode wavelength equation as a
function of pointing offset. It should be noted that Eqs. 2 and 3 of McClintock
et al. (2005a) are meant to be equivalent to Eqs. 27 and 28 of McClintock et al.
(2005b); however, the definition of the constant deviation angle (φG ) is different in
these two papers which leads the stellar-mode wavelength equation to be incorrect
in both. The wavelegth relationships presented here use a consistent definition and
should be used instead of the equations in the two McClintock papers.
Using the reconstructed spacecraft attitude, the angular offset of the Moon
center from the instrument boresight is known. However, the attitude derived from
the spacecraft star trackers is sampled at a different cadence and less frequently
than the SOLSTICE detector readout. The quaternions representing the spacecraft
attitude need to be interpolated to allow a determination of the effective wavelength
of each detector sample.
There is a secondary wavelength shift that must also be established. The grating
position reported by the instrument may be in error by an amount that is fixed
for each observation but will vary from one observation to the next. This grating
offset is described in Snow et al. (2005) and can produce a shift in the wavelength
scale of up to a few nanometers, although it is typically less than 1 nm. In data
processing, the magnitude of this offset is determined by an iterative least-squares
procedure to align the spectrum to the solar spectrum.
Vignetting
SOLSTICE was designed to accommodate either the diverging beam from the
solar entrance aperture (a pinhole) or the collimated beam from the stellar aperture
(16 mm diameter) and all of the optical surfaces were sized appropriately. Table 8.1
lists the sizes of each optic in the SOLSTICE optical path along with the size of the
lunar beam as determined from raytracing. For lunar observations, we must use
the large stellar entrance aperture, and the diverging beam from the Moon nearly
fills the first mirror. If the Moon is not properly centered in the FOV, some of the
light will be vignetted. As is shown in Table 8.1, the lunar image very nearly fills
the first optic. Therefore, the pointing errors during lunar observations caused by
parallax discussed in the section “Pointing” become important for even relatively
small angles.
Data Reduction 237
Figure 8.6 shows the loss of signal as a function of pointing offset angle as
determined by a raytrace model developed with ZEMAX optical design software.
The upper panel shows the loss of signal in the cross-dispersion direction and the
lower panel shows the loss in the dispersion direction. The loss of signal due to
vignetting in the dispersion direction is a function of wavelength. The vignetting
correction becomes more significant at longer wavelengths because the projected
area of the grating becomes smaller as the grating is rotated to a larger angle
relative to the incident beam. If this effect were not taken into account during
data processing, it would introduce a wavelength dependent artifact into the lunar
reflectance.
Since the vast majority of the measurements are made within 5 arc minutes of
the spacecraft boresight, this is a small correction for the MUV channel (less than
5 %). But in the case of the FUV channel, there is a 20 arc minute offset between
the spacecraft and instrument boresights. Therefore, we must make a significant
correction for vignetting to the FUV data.
Spectral Resolution
The spectral resolutions for the various observing modes of SOLSTICE men-
tioned in the section “Instrument Description” and discussed in McClintock et al.
(2005a) are applicable for the Sun or for a stellar point source. The situation for
lunar observations is somewhat more complicated. Given the optical layout of SOL-
STICE, the image of the full Moon at the exit slit has a diameter of approximately
1.57 mm (wavelength dependent). The width of the exit slit in MUV stellar mode
is 1.5 mm, and 0.75 mm in the FUV channel (McClintock et al. 2005a). Convolving
this circular image with the rectangular exit slits for each channel produces the
two instrument profiles shown in Fig. 8.7. These profiles assume ideal imaging con-
ditions and will be assumed to be lower limits to the true spectral resolution while
observing the Moon. We discuss an estimate of the uncertainty of these profiles in
the section “Uncertainty”.
We have calculated the spectral resolution for lunar observations as a func-
tion of phase angle and the angle between the projection of the Moon’s rotational
axis on the sky and the spacecraft y-axis based upon geometrical analysis of the
instrument. The lunar axis angle relative to the instrument dispersion direction
becomes particularly important at large phase angles because a crescent Moon ori-
ented parallel to the exit slit will yield a very different spectral resolution than
if the lunar axis is perpendicular to the slit. A non-zero angle will also lead to
an asymmetric profile. The roll angle of the spacecraft is part of the telemetered
Table 8.1: SOLSTICE optic sizes

Optic Size of optic Distance from aperture Size of lunar beam
Mirror 1 20.5 475 20.3
Grating 31.5 × 26 920 24.3
Mirror 2 39.75 1,368 28.4
Camera 50.4 1,775 32.1
Mirror
The lunar beam dimensions are for full Moon conditions. All sizes are given in mm
cross-dispersion plane
1.00
0.95
0.90
fraction
100nm
0.85 150nm
200nm
250nm
0.80 300nm
350nm
0.75
0.70
-30 -20 -10 0 10 20 30
angle (arcmin)
dispersion plane
1.00
0.95
0.90
fraction
100nm
0.85 150nm
200nm
250nm
0.80 300nm
350nm
0.75
0.70
-30 -20 -10 0 10 20 30
angle (arcmin)
Figure 8.6: Lunar vignetting correction. As the Moon moves away from the center
of the field-of-view, the image on the optics is vignetted. This curve shows the
fraction of rays which reach the exit slit as a function of angle based upon a
raytrace model of the instrument. The left panel shows the vignetting for offsets in
the cross-dispersion direction, the right panel is for the dispersion direction. The
wavelength dependence is only significant in the dispersion direction
spacecraft quaternion. Information about the lunar axis can be determined from
the ephemeris, and the mean geometry is used for each observation. Although
it is possible to convolve the lunar image with the exit slit for each sample, it is
much more computationally intensive. The celestial geometry generally changes
very slowly over the orbit, so using the mean geometry does not introduce a large
systematic uncertainty. However, we recognize that this effect should be quantified
to determine if it is worth the extra computation time to implement in the next
Data Reduction 239
Figure 8.7: Instrumental spectral profiles for FUV and MUV channels. These
curves come from the convolution of a circular lunar image (full Moon) with the
rectangular exit slit for each instrument configuration. The exit slit in FUV mode
is about half the diameter of the full Moon image, while the MUV exit slit is about
equal to the size of the image. (From Snow et al. 2007)
version of the data processing system. These spectral profiles are used as the convo-
lution kernel for the high-resolution solar spectrum in order to match the resolution
of the lunar observation.
Data Reduction Summary

For each lunar observation, the current data processing system takes the
spacecraft telemetry and corrects the lunar count rates for detector dead time
and dark rate (McClintock et al. 2005b; Snow et al. 2005). It then finds the solar
spectrum taken closest in time and convolves it with the appropriate kernel. This
imaging kernel is then used to degrade the resolution of the solar spectrum to match
the lunar spectrum. The wavelength scale of the solar spectrum is determined from
comparison to a reference spectrum. Since the solar spectrum is at relatively high
resolution (0.1 nm), the uncertainty in the solar wavelength scale is not significant
for the lunar processing.
The next step in the data reduction is to calculate an initial wavelength for each
sample of the lunar spectrum based on the pointing and grating angles described in
the sections “Pointing” and “Spectral Resolution”. The lunar count rates are then
corrected for vignetting. The final step is to form the ratio of the lunar to solar
spectra and empirically derive the grating offset, , by minimizing the residuals in
the solar/lunar ratio.
The improvements in the convolution kernel described in this document have
so far only been implemented for the MUV channel data. The FUV channel data
has a 20 arc minute offset between the spacecraft boresight and the instrument
boresight. We need to do some further raytrace analysis to ensure that no addi-
tional corrections are needed in the vignetting calculation due to this misalignment
of boresights. Another planned update to the processing system will be to itera-
tively solve for both the vignetting and grating offset together since the vignetting
correction will be slightly adjusted by the final wavelength scale.
Lunar Irradiance and Albedo

Irradiance
Figure 8.8 shows an absolutely calibrated lunar irradiance spectrum measured
by SORCE SOLSTICE for each channel. The FUV observation was at a phase
angle of 2.0◦ , while the MUV observation was from 3.4◦ . The observed irradiance
will not only be a strong function of phase angle, but it will also depend on solar
variability and the distances between the Sun, Moon, and spacecraft. The solar
irradiance changes by a few percent over the solar cycle in the MUV wavelength
range, but it changes by about 20% from solar maximum to solar minimum in the
FUV (Rottman 1999). This is a much larger variation than seen in the visible and
cannot be neglected in the UV.
Thermospheric and geocoronal hydrogen near the line of sight to the Moon will
scatter solar Lyman-α (121.6 nm) photons. This emission is known as airglow, and
it complicates measurement of the lunar irradiance within about 10 nm of Lyman-
α. Although the scattered emission is in a very narrow line, the emission is coming
from everywhere in the FOV. In stellar mode, SOLSTICE is essentially an objective
grating spectrometer, therefore rays entering the entrance aperture from off-axis
are shifted in wavelength relative to rays entering along the instrument boresight.
The result is that the diffuse geocoronal Lyman-α emission appears to be a broad
Lunar Irradiance and Albedo 241
Figure 8.8: (Top) FUV lunar spectral scan at full Moon, acquired on 7 October
2006 at a phase angle of 2.0◦ . Data points marked with X contain contribution
from airglow. Error bars shown are due to counting statistics and do not include
systematic uncertainties. (Bottom) MUV lunar irradiance spectrum at full Moon,
acquired on 2 April 2007 at a phase angle of 3.4◦ . Uncertainty due to counting
statistics is smaller than plot symbols. (From Snow et al. 2007)
feature in the spectral scan data. Correction for airglow contamination is fairly
straightforward, but the quantitative analysis has not been completed in time for
this publication.
Albedo
Since the spectrum of the Sun changes with solar activity, the irradiance
spectrum of the Moon will also change. For purposes of using the Moon as a
calibration target, a more useful quantity than the irradiance may be the disk-
equivalent albedo (Kieffer and Stone 2005), which is derived from the ratio of lunar
to solar irradiance. This quantity is a property of the Moon alone, and it is there-
fore stable over long time periods (Kieffer 1997). The relationship between albedo
and irradiance is given by:
EM /ΩM
AM = . (8.5)
E /π
In this equation, AM is the disk-equivalent albedo, ΩM is the solid angle subtended
by the Moon, EM is the irradiance of the Moon, and E is the irradiance of the
Sun. In our notation, ΩM varies with the Moon-spacecraft distance. Since the
solar and lunar observations are taken close in time, there is no need to correct the
irradiances to 1 AU; such a correction would cancel out in the ratio. The ratio of
lunar to solar irradiance for the SOLSTICE instrument is simply:
EM CM a Δλ
= Γ. (8.6)
E C aM ΔλM
In this expression, C and CM are the solar and lunar count rates, a and Δλ are
the entrance aperture area and radiometric bandpass of the two instrument modes.
The function Γ is a geometrical correction factor which accounts for differences
in responsivity for targets of different angular size. The correction factor is a few
percent at most wavelengths [Figure 4 of McClintock et al. (2005b)]. Both C
and CM include correction factors for dark current, scattered light, temperature,
filter transmission. We believe that these instrumental corrections have all been
accounted for, but it is possible that there could still be subtle systematic differences
between the solar and lunar observing conditions that have not yet been identified.
Figure 8.9 shows the measured disk-equivalent albedo for the FUV and MUV
channels on SOLSTICE. For simplicity, we show only the albedo for observations
in the range of 9–11◦ phase in the figure. The feature at about 215 nm in the lower
panel of Fig. 8.9 is likely an artifact. There is a neutral density filter that is inserted
at that wavelength during the solar scan, so it is suspicious that there is a feature
in the reflectance at that same wavelength. Further investigation of the properties
of this filter will resolve this uncertainty.
Since SOLSTICE directly measures both solar and lunar irradiance with the
same optics and detectors, this ratio is independent of solar variability and long-
term instrument degradation. Random and systematic errors in this ratio will
be discussed in the section “Uncertainty”. Most importantly, this ratio is also
independent of the absolute calibration of SOLSTICE since it is a direct ratio.
The ratios shown in Fig. 8.9 can be measured at all phase angles. Figure 8.10
shows the typical behavior of the albedo as a function of phase angle for one
wavelength (282 nm). Near full Moon an effect known as the opposition surge causes
a rapid increase in brightness as the phase angle decreases. The behavior of the
lunar integral photometric function at large phase angles is less well studied. The
only other dataset of a similar surface with comparable phase coverage is the visible
Lunar Irradiance and Albedo 243
Far Ultraviolet
0.10
0.08
Disk-Equivalent Albedo
0.06
0.04
0.02
0.00
130 140 150 160 170 180 190
Wavelength (nm)
Middle Ultraviolet
0.10
0.08
Disk-Equivalent Albedo
0.06
0.04
0.02
0.00
180 200 220 240 260 280 300
Wavelength (nm)
Figure 8.9: (Top) FUV disk-equivalent albedo for phase angles between 9 and 11◦ .
Measurements have been averaged over 2 nm (approximately two resolution ele-
ments). Data from eleven SOLSTICE scans were used in this plot. (Bottom) MUV
disk-equivalent albedo for phase angles between 9 and 11◦ . Measurements from
sixteen SOLSTICE scans were averaged over 2 nm intervals (about one resolution
element)
Figure 8.10: Disk equivalent albedo as a function of phase angle for 282 nm.
Negative phase angles indicate a waxing Moon
phase curve of the planet Mercury as measured by the Large Angle Spectroscopic
COronograph (LASCO) onboard the Solar and Heliophysical Observatory (SOHO).
The phase curve of Mercury exhibits a similar sharp decline in albedo at large phase
angles (Mallama et al. 2002). This SOLSTICE dataset will further understanding
of the lunar surface in this regime.
Uncertainty
In order for the SOLSTICE lunar observations to be a useful calibration dataset,
a thorough understanding of the uncertainties is critical. We present here a pre-
liminary analysis of the major contributors to uncertainty. From McClintock et al.
(2005b), the irradiance measurement equation for SOLSTICE is given by:
C
E= , (8.7)
R · FOV · AEntrance · ΔλBP · DEG
where C is the count rate (corrected for dead time, dark current, and scattered
light), R is the responsivity, FOV is the variation in responsivity over the field-of-
view of the instrument, and AEntrance · ΔλBP is the aperture area and radiometric
bandpass. DEG is a function of time and wavelength that accounts for changes
in the instrument responsivity. From Eq. (23) of McClintock et al. (2005b), the
uncertainty of the ratio between a solar and stellar irradiance measurement is:
2 2 2 2 2
σE/E = σC
+ σC
+ σG/G + σR A−BP
+ σf2AU + σΓ2 . (8.8)
In this equation, σC and σC are the uncertainties in solar and stellar (or lunar)
count rates. These are sources of random error. σG/G is the uncertainty in ratio
Uncertainty 245
of the gain factors for the two modes. Γ is a correction factor that accounts for
the slightly different illumination between solar and stellar configurations. When
comparing solar to stellar irradiances, a correction to 1 AU is necessary to account
for the eccentricity of the Earth’s orbit, and σfAU is the uncertainty of that correc-
tion. However, there is no 1 AU correction needed in the lunar case so σfAU can be
neglected. Finally, σRA−BP is the uncertainty in the responsivity due to errors in
the aperture and radiometric bandpass. These four terms are systematic sources
of error and will be discussed first.
Systematic Error
The gain and FOV (i.e. Γ) terms contribute 0.1% and 1.0% respectively
[Table III of McClintock et al. (2005b)]. These two terms are wavelength-
independent sources of systematic error. The illumination of the optics during
lunar observations is slightly different than for stellar mode. The stellar illumina-
tion is basically a collimated beam. The lunar illumination is a diverging beam
that illuminates a somewhat larger portion of each optic than the stellar beam,
but the difference in responsivity for the two cases is very small. Therefore we are
using the stellar value for Γ in the lunar processing. We do not expect that the
contribution to systematic error from an error in Γ will ever be larger than the 1%
that we have already added to the error budget. The correction to 1 AU can be
neglected in the albedo measurement since both solar and lunar count rates are
affected by the same factor if the observations are made close in time. For the
case of lunar observations, we also must account for systematic uncertainties in our
vignetting correction and convolution kernel.
The vignetting correction is derived from a raytrace model of the instrument,
using the measured sizes of the optical elements. We can test the model for the
stellar case with observations. Once a week, the SORCE spacecraft performs a
cruciform alignment maneuver while SOLSTICE observes a star. The maneuver
covers the range ±5.5◦ in both dispersion and non-dispersion planes. Our raytrace
model agrees quite well with the stellar alignment data, which gives us confidence
that it is also accurate in the lunar case. We will assume for now that we know this
correction factor to the 5% level. A quantitative estimate of the accuracy in the
stellar case is planned for the next version of the analysis. The vignetting correction
for the MUV is typically less than 0.5%; therefore uncertainty in this correction to
the MUV data makes a negligible contribution to the overall error budget. In the
case of the FUV, where the magnitude of the correction is aproximately 7%, we
estimate a 0.35% contribution to the systematic uncertainty that is independent of
wavelength.
For the uncertainty in the convolution kernel, we can empirically estimate the
effect of an error in its assumed value. The lunar albedo is presumed to be slowly
varying with wavelength at this spectral resolution, so an error in the convolution
kernel will cause the features in the solar spectrum to show up as features in the
albedo. The uncertainty in the ratio due to counting statistics then sets an upper
limit on our knowledge of the convolution kernel. The size of an artifact due to
error in the kernel must be larger than the random errors in the measurement. We
estimate that the lunar convolution kernel is accurate to 5%. This estimate will
likely decrease as we analyze the data further. The profile that we are currently
using is from the simple convolution which assumes ideal imaging conditions. A
more detailed model may result in changes to our assumed instrument profile.
Since these sources of systematic error are independent, we can add them in
quadrature. By far, the uncertainty in the convolution kernel is the dominant
source of systematic error. The total contribution from systematic errors in the
lunar albedo measurement is likely no larger than 5.1%.
Instrumental artifacts that could influence the value of EM /E include sys-
tematic differences in temperature, libration angle, etc. since the solar and lunar
observations are not done simultaneously. However, we think that we have made
conservative estimates for the contribution to the total combined uncertainty for
these corrections and included them. We will continue to make improvements to
our data processing system and update the results.
Random Error
Random error sources also contribute to the uncertainty of the lunar albedo
measurement. These errors can be reduced in magnitude by repeated observations
at each phase and libration angle. This category of error includes counting statistics
and uncertainties in the lunar wavelength scale. The former is a straightforward
calculation of the square root of the number of counts. The latter is somewhat
more complex. In the current version of data processing, we are neglecting the
variation due to libration angle and averaging all observations from a given phase
angle together. The random variation due to statistics is likely much larger than
the systematic variation due to libration angle. Based upon the ROLO photometric
model, variations due to libration are at most a few percent for phase angles smaller
than 90. We have not yet quantified the magnitude of the change in reflectance
with libration angle in the ultraviolet. A future version of the data processing will
take libration angle into account.
As discussed in the section “Pointing”, the SOLSTICE wavelength scale is very
repeatable, although there can be an offset introduced when the instrument is re-
set, i.e. in Eq. 8.4 (Snow et al. 2005). But more importantly, since SOLSTICE
acts as an objective grating spectrometer in stellar mode, the wavelength scale will
be continually changing during an observation due to the parallax motion of the
Moon in the FOV. A change of 5 arc minutes in position is about 16 of the lunar
diameter. This translates to about a 0.4 nm shift in the wavelength scale. The cur-
rent processing system calculates the appropriate wavelength shift for each sample
based on spacecraft telemetry and applies it to the lunar spectrum. Our estimate
of the residual error in this method is about 0.1 nm. The uncertainty shown in
Fig. 8.11 is the change in irradiance for a wavelength scale offset of that size.
Polarization
Understanding the polarization of the lunar irradiance is important for both
calibration and scientific investigation of the lunar surface. In this section we will
briefly discuss both aspects. The SURF beam used during SOLSTICE’s preflight
Polarization 247
Figure 8.11: Sources of random error in a single SOLSTICE lunar albedo measure-
ment. The top panel shows the FUV channel while the bottom panel shows the
MUV. The contribution from these errors to the final albedo measurement can be
reduced by averaging multiple observations. A systematic error of about 5% needs
to be added in quadrature to the random errors to get the final uncertainty in each
measurement
0.025
sensitivity 0.020 H
0.015
0.010 V
0.005
0.000
180 200 220 240 260 280
wavelength (nm)
0.8
0.6 (H-V)/(H+V)
polarization
0.4
0.2
0.0
180 200 220 240 260 280
wavelength (nm)
Figure 8.12: Response of SOLSTICE MUV channel to polarized light. H and V are
the signal levels detected by the instrument when the dispersion direction is either
Horizontal or Vertical relative to the SURF polarization plane. The difference of
these two orientations divided by the sum gives the fractional polarization detected
by the instrument
calibration is highly polarized and responsivity measurements were taken in two

orthogonal orientations (McClintock et al. 2005b). Figure 8.12 shows the differ-
ence between the signal levels in horizontal and vertical orientations divided by
the sum of the two orientations during preflight calibration at SURF. If the in-
strument were not sensitive to polarization at all, this ratio would be zero. If it
were completely polarizing, the ratio would be 1 (or −1). As can be seen in the
figure, the MUV channel of SOLSTICE shows little sensitivity to polarization for
wavelengths shorter than 200 nm, but between 200 and 250 nm the sensitivity rises
to about 60% and remains near 60% for wavelengths longer than 250 nm. The
FUV channel responsivity showed less than 10% change for the two orientations
at all wavelengths. Since SOLSTICE is partially sensitive to polarization, we can
use the measurements of lunar irradiance at different roll angles to determine the
polarization of the disk-integrated Moon.
Lyot (1929) showed that the lunar polarization exhibits a strong phase de-
pendence, with a positive maximum for all wavelengths near 100◦ and a negative
minimum near 10◦ . Observations of the disk-integrated Moon indicate that the po-
larization fraction increases toward shorter visible wavelengths (Coyne and Pellicori
1970). This trend has been found to also continue through the middle ultraviolet,
based on observations of small, resolved regions on the Moon (Fox et al. 1998).
For all SORCE SOLSTICE lunar observations, we can calculate the orientation
of the scattering plane (the plane defined by the Sun, the Moon, and SORCE) as
a function of time based on the ephemeris and spacecraft telemetry. Therefore, it
Polarization 249
0.25
WUPPE (not full disk) [8] waning
waxing
0.20
polarization SOLSTICE
0.15
0.10 Coyne & Pellicori [7]
0.05
0.00
200 300 400 500
wavelength (nm)
Figure 8.13: Percent polarization of the lunar irradiance 55◦ phase angle. Box and
X symbols show the results from Coyne and Pellicori (1970). Diamonds mark the
SOLSTICE observations. The SOLSTICE measurement is derived from two spec-
tra taken at orthogonal spacecraft roll angles. The middle ultraviolet wavelength
band appears to be very consistent with measurements from the visible band
is straightforward to determine the orientation of the polarization and measure its

effect on the albedo.
We have analyzed a pair of reflectance spectra of the Moon, both with a phase
angle of 55◦ but at perpendicular spacecraft roll angles. The product of the in-
strument polarization and the lunar polarization should be equal to the normalized
difference between the two spectra. Therefore we can solve for the lunar polariza-
tion. Figure 8.13 shows the disk-integrated polarization of the Moon, interpolated
from Coyne and Pellicori (1970), at the spectral position of their photometric filters.
The two curves represent the polarization of the Moon before and after full Moon
at a phase angle of 55◦ . The polarization of the Moon has been found to depend
on phase angle (Dollfus 1998) and to be inversely proportional to albedo (Dollfus
and Bowell 1971). The western nearside region of the Moon contains a relatively
greater abundance of dark mare material, resulting in larger polarization in the
waning phase. Figure 8.13 also shows the binned polarization measurement from
SOLSTICE which continues the trend of increasing polarization towards shorter
wavelengths. The increase at shorter wavelengths is again due to the inverse depen-
dence on albedo. The SOLSTICE measurements appear to be consistent in mag-
nitude with the longer wavelength observations from Coyne and Pellicori (1970).
Because the dominant scattering mechanism changes from volume scattering at
visible wavelengths to surface scattering at ultraviolet wavelengths (Wagner et al.
1987) these measurements may provide unique insight into the physical properties
of the lunar regolith. The SOLSTICE polarization measurements are the first disk-
integrated observations in this wavelength range and only the second measurement
of lunar polarization in the ultraviolet (Fox et al. 1998).
Summary
We have presented preliminary results of measurements of the lunar irradiance,
and more importantly, the lunar albedo at a large range of phase angles throughout
the UV (115–300 nm). The current processing system accounts for the dynamic
observing geometry from a satellite in low earth orbit and accurately produces lunar
reflectances. Improvements to this system are continuing as part of the LASP Lunar
Albedo Measurement and Analysis from SOLSTICE (LLAMAS) project funded by
NASA and final data products are scheduled for delivery in 2013.
This dataset will be used to create a photometric model of the lunar irradiance
that can then be used for both calibration of any Moon-viewing instrument and
for scientific investigations of the surface properties of the Moon. A photometric
model of the lunar observations for UV wavelengths over the full range of phase and
libration angles will allow Hapke photometric parameters (Hapke 1981, 1984, 1986,
1993) to be derived and compared with previously published results in the visible
(e.g. Helfenstein and Veverka 1987). This comparison will help us to understand
the physical mechanisms for scattering from the lunar surface at ultraviolet wave-
lengths. Discriminating mineralogic features are known to exist in the ultraviolet
(Wagner et al. 1987) and their identification in this data would demonstrate an
additional method of detection with remote sensing instruments.
Initial analysis of the polarization properties of the lunar irradiance indicates
that trends observed in the visible continue through at least the middle ultraviolet.
Measuring the polarization of the sunlight reflected off the Moon will help us to
understand the physical state of the lunar surface.
The current status of the data processing system is that the MUV channel data
is nearly ready for photometric modeling, but the FUV dataset processing is not
yet ready for routine production. We have summarized the major sources of both
systematic and random error in the measurements. Further analysis and additional
measurements will likely further reduce these uncertainties in the future.
The data from SOLSTICE overlap the observations from a wide variety of
previous experiments, none of which had the full range of wavelengths and phase
angles contained in the new dataset. This overlap will allow the cross-calibration
of a large number of planetary missions once the SOLSTICE data has been fully
reduced.
The surface reflectance properties of the Moon are known to be stable on very
long time scales (Kieffer 1997). The solar irradiance in the ultraviolet varies much
more than it does in the visible, so accurate solar measurements (even if based on
a proxy) will be necessary to make the Moon a useful UV calibration target in the
future. Beyond the SORCE mission, there will likely be MUV of the Sun observa-
tions from an instrument such as the Spectral Irradiance Monitor (SIM) which is
currently flying on SORCE and a similar instrument is presently manifested as the
Total and Spectral Irradiance Sensor (TSIS) on the Joint Polar Satellite System
(JPSS) scheduled for launch later this decade (http://www.nesdis.noaa.gov/jpss).
Alternatively the MUV solar spectrum can also be fairly well-modeled through
correlations with the Mg II index (Chandra et al. 1995). This index is currently
measured on a daily basis by the National Oceanic and Atmospheric Adminis-
tration (NOAA) and will certainly be continued by future missions. Uncertainties
Bibliography 251
inherent in solar models based on proxies are larger than those from measurements,
but since the amount of solar variability in the MUV is only a few percent, the
model irradiance may be adequate. The FUV solar irradiance measurement after
SORCE will also have to come from models and proxy measurements. The current
NASA plan does not include a FUV solar measurement. The SORCE primary
mission ended in January, 2008, but it is continuing to operate on an extended
mission. Spacecraft operations on SORCE have curtailed the lunar observations
during eclipse, but during the lunar observing campaign, we acquired a very exten-
sive dataset of lunar irradiance that will produce valuable scientific and calibration
results.
Acknowledgements
This research is supported by NASA grant NNX09AQ60G (LLAMAS) and contract
NAS5-97045 (SORCE) at the University of Colorado. The authors would like to thank
the International Space Science Institute, Bern, Switzerland, for their generous support
of the working group.
Bibliography
U. Arp, R. Friedman, M.L. Furst, S. Mkar, P.-S. Shaw, SURF III—an improved storage
ring for radiometry. Metrologia 37, 357–360 (2000). doi:10.1088/0026-1394/37/5/2
S. Chandra, J.L. Lean, O.R. White, D.K. Prinz, G.J. Rottman, G.E. Brueckner, Solar uv
irradiance variability during the declining phase of the solar cycle 22, Geophys. Res. Lett.
22, 2481–2484 (1995). doi:10.1029/95GL02476
G.V. Coyne, S.F. Pellicori, Wavelength dependence of polarization. xx. the integrated disk
of the moon. Astron. J. 75, 54–60 (1970). doi:10.1086/110940
M.R. Dobber, Moon observations with gome. Proc. SPIE 2831 154–168 (1996)
M.R. Dobber, GOME moon measurements, including instrument characterization and moon
albedo. In: Guyenne T-D, Danesy D (eds) Third ERS Symposium on Space at the service
of our Environment, ESA SP-414, p. 743, Florence, Italy, March 1997
A. Dollfus, E. Bowell, Polarimetric properties of the lunar surface and its interpretation.
part 1. telescopic observations. Astron. Astrophys. 10, 29–53 (1971)
A. Dollfus, Lunar surface imaging polarimetry: i. roughness and grain size. Icarus 136,
69–103 (1998)
G.K. Fox, A.D. Code, C.M. Anderson, B.L. Babler, K.S. Bjorkman, J.J. Johnson, M.R.
Meade, K.H. Nordsieck, A.J. Weitenbeck, N.E.B. Zellner, Solar system observations
by the wisconsin utraviolet photopolarimeter experiment—iii. the first ultraviolet lin-
ear spectropolarimetry of the moon. Mon. Not. R. Astron. Soc. 298, 303–309 (1998).
doi:10.1046/j.1365–8711.1998.01633.x
G.R. Gladstone et al., LAMP: the lyman alpha mapping project on nasa’s lunar reconnais-
sance orbiter mission. Space Sci. Rev. 150, 161–181 (2010). doi:10.1007/s11214-009-9578-
6
B. Hapke, Bidirectional reflectance spectroscopy. 1. theory. J. Geophys. Rev. 86, 3039–3054
(1981)
B. Hapke, Bidirectional reflectance spectroscopy. 3. correction for macroscopic roughness.
Icarus 59, 41–59 (1984)
B. Hapke, Bidrectional reflectance spectroscopy. 4. the extinction coefficient and the
opposition effect. Icarus 67, 264–280 (1986)
B. Hapke, Theory of reflectance and emitance spectroscopy. (Cambridge University Press,
Cambridge, 1993)
P. Helfenstein, J. Veverka, Photometric properties of lunar terrains derived from Hapke’s

equation. Icarus 72, 342–357 (1987). doi:10.1016/0019-1035(87)90179-5
A.R. Hendrix, The galileo ultraviolet spectrometer: in-flight calibration and ultraviolet albe-
dos of the moon, gaspra, ida, and europa. (Dissertation, University of Colorado, 1996)
A.R. Hendrix, in The lunar phase curve in the near ultraviolet, ed. by R. Strom, P. Bo,
M. Walker, N. Rendong. Solar System Remote Sensing, Astrophysics and Space Science
Library, vol 278 (2002), p. 29
A.R. Hendrix, G. Holsclaw, L. Esposito, W.E. McClintock, The ultraviolet reflectance
of the moon as measured by Cassini UVIS. Lunar Science Forum (2009). http://
lunarscience2009.arc.nasa.gov/node/103
R.C. Henry, et al. Ultraviolet albedo of the moon with the hopkins ultraviolet telescope.
Astrophys. J. Lett. 454, L69–L72 (1995). doi:10.1086/309771
G.M. Holsclaw, W.E. McClintock, D.L. Domingue, N.R. Izenberg, D.T. Blewett, A.L.
Sprague, A comparison of the ultraviolet to near-infrared spectral properties of
mercury and the moon as observed by messenger. Icarus 209, 179–194 (2010).
doi:10.1016/j.icarus.2010.05.001
S.J. Janz, E. Hilsenrath, R.P. Cebula, T.J. Kelly, Observations of the lunar geomet-
ric albedo during the atlas-3 mission. Geophys. Res. Lett. 23, 2297–2300 (1996).
doi:10.1029/96GL01122
H.H. Kieffer, Photometric stability of the lunar surface. Icarus 130, 323–327 (1997).
doi:10.1006/icar.1997.5822
H.H. Kieffer, J.A. Anderson, Use of the moon for spacecraft calibration over 350 to 2500
nm. Proc. SPIE 3498, 325–336 (1998)
H.H. Kieffer, T.C. Stone, The spectral irradiance of the moon. Astrophys. J. 129, 2287–2901
(2005). doi:10.1086/430185
H.H. Kieffer, R.L. Wildey, Absolute calibration of landsat instruments using the moon.
Photogramm. Eng. Remote Sens. 51, 1391–1393 (1985)
H.H. Kieffer, R.L. Wildey, Establishing the moon as a spectral radiance
standard. J. Atmos. Ocean Tech. 13, 360–375 (1996). doi:10.1175/1520-
0426(1996)013¡0360:ETMAAS¿2.0.CO:2
A.P. Lane, W.M. Irvine, Monochromatic phase curves and albedos for the lunar disk. Astron.
J. 78, 267–277 (1973). doi:10.1086/111414
R.L. Lucke, R.C. Henry, W.G. Fastie, Far-ultraviolet albedo of the moon. Astron. J. 81,
1162–1169 (1976). doi:10.1086/112000
B. Lyot, Research on the polarization of light from planets and from some terrestrial sub-
stances, vol. 8 (Annales de l’Observatoire de Paris, Paris, 1929)
A. Mallama, D. Wang, R.A. Howard, Photometry of mercury from soho/lasco and earth. The
phase function from 2 to 170 deg. Icarus 155, 253–264 (2002). doi:10.1006/icar.2001.6723
W.E. McClintock, G. Rottman, T.N. Woods, Solar stellar irradiance comparison ex-
periement ii (solstice ii): instrument concept and design. Sol. Phys. 230, 225–258 (2005a).
doi:10.1007/s11207-005-7432-x
W.E. McClintock, M. Snow, T.N. Woods, Solar stellar irradiance comparison experiment
ii (solstice ii): pre-launch and on-orbit calibrations. Sol. Phys. 230, 259–294 (2005b).
doi:10.1007/s11207-005-1585-5
C.K. Pankratz, B.G. Knapp, R.A. Reukauf, J. Fontenla, M.A. Dorey, L.M. Connelly,
A.K. Windnagel, The sorce science data system. Solar Phys. 230, 389–413 (2005).
doi:10.1007/s11207-005-5008-4
S.G. Pugacheva, V.V. Novikov, V.V. Shevchenko, The moon as a natural standard for
calibration of spectrofotometric under-sputnik observations. Astron. Vestn. 27, 47–64
(1993)
M.S. Robinson, B.W. Hapke, J.B. Garvin, D. Skillman, J.F. Bell, M.P. Ulmer, C.M. Pieters,
High resolution mapping of TiO2 abundances on the moon using the hubble space tele-
scope. Geophys. Res. Lett. 34, L13203 (2007). doi:10.1029/2007GL029754
M. Robinson, S.M. Brylow, M. Tschimmel, D. Humm, S.J. Lawrence, P.C. Thomas, B.W.
Denevi, E. Bowman-Cisneros, J. Zerr, M.A. Ravine, M.A. Caplinger, F.T. Ghaemi, J.A.
Schaffner, M.C. Malin, P. Mahanti, A. Bartels, J. Anderson, T.N. Tran, E.M. Eliason,
Bibliography 253
A.S. McEwen, E. Turtle, B.L. Jolliff, H. Heisinger, Lunar reconnaissance orbiter camera
(LROC) instrument overview, Space Sci. Rev. 150, 81–124 (2010). doi:10.1007/s11214-
010-9634-2
G. Rottman, Solar ultraviolet irradiance and its temporal variation. J. Atmos. Sol. Terr.
Phys. 61, 37–44 (1999). doi:10.1016/S1364-6826(98)00114-X
G. Rottman, The sorce mission. Sol. Phys. 230, 7–25 (2005). doi:10.1007/s11207-005-8112-6
experiment ii (solstice ii): examination of the solar-stellar comparison technique. Sol.
Phys. 230, 295–324 (2005). doi:10.1007/s11207-005-8763-3
M. Snow, G. Holsclaw, W.E. McClintock, T.N. Woods, Absolute ultraviolet irradiance of
the moon from sorce solstice. Proc. SPIE 6677, 66770D (2007). doi:10.1117/12.732498
T.C. Stone, H.H. Kieffer, Absolute irradiance of the moon for on-orbit calibration. Proc.
SPIE 4814, 211–221 (2002)
M. Taguchi, G. Funabashi, S. Watanabe, Y. Takahashi, H. Fukunishi, Lunar albedo at
hydrogen lyman α by the nozomi/uvs. Earth Planets & Space 52, 645–647 (2000)
J.K. Wagner, B.W. Hapke, E.N. Wells, Atlas of reflectance spectra of terrestrial, lu-
nar, and meteoric powders and frosts from 92 to 1800 nm. Icarus 69, 14–28 (1987).
doi:10.1016/0019-1035(87)90003-0
H.H. Wu, A.L. Broadfoot, The extreme ultraviolet albedos of the planet mercury and of the
moon. Geophys. Res. Lett. 82, 759–761 (1977). doi:10.1029/JB082i005p00759
—9—
Lyman-α Observations of Comet Holmes

from SORCE SOLSTICE and SOHO
SWAN
Wayne R. Pryor∗
Central Arizona College,
Coolidge, AZ, USA
(also at LASP, University of Colorado and
Space Environment Technologies, Palisades, CA, USA)
Martin Snow
Boulder, CO, USA
Eric Quémerais
LATMOS-IPSL
Université Versailles-SaintQuentin, Guyancourt, France
Stéphane Ferron
ACRI-ST,
Guyancourt, France
Abstract
In November 2007, the Solar Wind Anisotropy (SWAN) instrument on the Solar
and Heliospheric Observatory (SOHO) and the SOLar-STellar Irradiance Compar-
ison Experiment (SOLSTICE) on the SOlar Radiation and Climate Experiment
(SORCE) both made observations of comet 17P/Holmes at Lyman-α wavelengths.
The intent of these observations was to compare the calibration of SWAN and
SOLSTICE, rather than to study the properties of comets. The two datasets re-
quire significant processing to remove interplanetary and geocoronal backgrounds
which increased the uncertainty of the result. However, both instruments measured
similar Lyman-α irradiances, confirming their calibration agreement.

256 9. Comet Holmes
Introduction
Comet 17P/Holmes brightened in a spectacular outburst on 23–24 October 2007
and remained unusually bright for several months. Comets can provide a localized
source of Lyman-α photons which are potentially easy to distinguish from geo-
coronal and interplanetary hydrogen sources. This paper uses nearly simultaneous
Lyman-α observations of this bright comet from two independent satellite instru-
ments to compare their calibrations. The Solar Wind Anisotropy (SWAN; Bertaux
et al. 1995) instrument on the Solar and Heliospheric Observatory (SOHO) maps
the ultraviolet (UV) emission from the entire sky on a daily basis. It has observed
Lyman-α emission from over 20 comets since launch on 2 December 1995 from
its location near the L1 Sun-Earth Lagrange point (Combi et al. 2011, 2012). The
SOLar-STellar Irradiance Comparison Experiment (SOLSTICE) instrument on the
Upper Atmosphere Research Satellite (UARS; Rottman et al. 1993) observed the
spectrum of comet Hale-Bopp (Woods et al. 2000), but did not measure its Lyman-
α irradiance due to the complications of the airglow. In this paper, we will ana-
lyze observations of comet Holmes from the SOLSTICE (McClintock et al. 2005a)
instrument on the SOlar Radiation and Climate Experiment (SORCE; Rottman
2005). The absolute calibration of SORCE SOLSTICE is also discussed in Snow
et al. (2012, this volume). The SORCE observations of comet Holmes used a differ-
ential technique to remove the geocoronal contribution to the comet signal. This
technique will be described in the section “SOLSTICE Comet Observations”.
In mid November 2007, there was a region about 0.6◦ in diameter around the
comet (larger than the Moon) which was bright in visible light. However, the
Lyman-α emitting region was likely much smaller than 0.6◦ since it had a fairly
modest gas production rate in November 2007 (Schleicher 2009).
The brightening of comet Holmes provided an opportunity to compare the cur-
rent SWAN Lyman-α calibration to the Lyman-α calibration of the SOLSTICE
on SORCE. SORCE was launched on 25 January 2003 into low Earth orbit. In
addition to observing the Sun, SOLSTICE is designed to use bright stars to track
changes in the instrument over time (Snow et al. 2005, 2012). Comet Holmes was
reported to be as bright as some of SOLSTICE’s calibration stars in November
2007, so observations were specially planned and executed.
Both SOLSTICE and SWAN must make significant corrections to the data
to account for the contribution from local or interplanetary hydrogen in order
to determine the emission from the comet alone. These corrections are highly
variable in time and depend on the detailed observation geometry. The uncertainty
introduced by these corrections is significant for the case of 17P/Holmes, but both
instruments did get a reliable measurement of the flux of Lyman-α photons coming
from the comet.
SWAN
SWAN provides regular Lyman-α maps of the sky on a 1◦ × 1◦ grid (Bertaux
et al. 1997). It uses twin 2-mirror periscopes mounted on opposite sides of the
spacecraft (+Z and −Z) to construct these maps. The detectors measure 5◦ by 5◦
SWAN 257
on the sky at a time, using a Hamamatsu 5 × 5 multi-anode micro channel plate

array detector tube with a MgF2 window. The CsI photocathode of the detec-
tor renders the instrument solar-blind above 200 nm, but leaves the instrument
with sensitivity to starlight in the FUV between 120 and 200 nm (see Fig. 13.4
of Quémerais and Bertaux 2002). Each unit has four unilluminated side pixels
suitable for instrumental dark current subtractions.
The initial published sensitivity of SWAN to Lyman-α radiation in photometer
mode was 0.75 counts per sec per Rayleigh per 1◦ × 1◦ pixel (Bertaux et al. 1995).
SWAN’s ground calibration was based on comparison with an aluminum standard
photodiode from the National Institute of Standards and Technology (NIST) at
a synchrotron light source in Orsay. Inflight cross-calibration with Hubble Space
Telescope (HST) Goddard High Resolution Spectrometer (GHRS) data from 9
March 1996 led to slightly revised values that depended on which sensor was used
and its high voltage setting (Bertaux et al. 1997).
SWAN Comet Observations

Figure 9.1 shows a sample SWAN image with a band of stars visible along
the galactic plane. The blacked-out regions in the figure are the region inacces-
sible to the periscope because of spacecraft obscurations and the solar avoidance
zone. SWAN’s L1 orbital location places it in an ideal setting to study heliospheric
Lyman-α without geocoronal Lyman-α contamination due to Earth’s hydrogen ex-
osphere. SORCE SOLSTICE, in contrast, from its low Earth orbit vantage point
always looks through the geocorona at more distant sources.
Our analysis used SWAN all-sky maps from 18 October 2007 to 31 December
2007. JPL’s online HORIZONS program provided the comet locations used. SWAN
observed an initial Lyman-α outburst from the comet of about 700 R in late October
that later dimmed to less than 100 R by December. On 22 November 2007, the
comet center was at ecliptic latitude 30.36 and ecliptic longitude 61.65. The closest
SWAN pixel center is nominally at ecliptic latitude 30.5 and ecliptic longitude 61.5.
By 24 November, the comet center had shifted slightly to ecliptic latitude 30.29
and ecliptic longitude 61.18. Based on these small offsets, we can assume that
the comet’s Lyman-α output was captured by this single pixel. We will refer to
this pixel as (62,120), corresponding to its x and y coordinates in the data array.
One issue that complicates the comparison between SWAN and SOLSTICE is that
the comet drifted through a region containing UV-bright stars during the cross-
calibration period. Figure 9.2 shows the time series of the irradiance measured by
pixel (62,120). SWAN’s 1◦ by 1◦ pixels can in principle image a monochromatic
point source in a single pixel (Bertaux et al. 1997). The crosses in Fig. 9.2 show
the irradiance for this pixel after subtraction of the time-varying interplanetary
background (shown in the lower curve). The interplanetary Lyman-α background
was obtained from a 16 pixel patch of sky adjacent to the comet but away from
stars, with derived values of 285 R and 292 R for November 22 and November 24.
The signal that remains is a combination of the comet plus background stars.
The average irradiance on the left and right sides of Fig. 9.2 are quite different.
The irradiance in this pixel before the comet entered its field of view and afterwards
should be the same. However, the two levels are different due to a spacecraft roll
258 9. Comet Holmes
Figure 9.1: SWAN Lyman-α all-sky map from 1 November 2007. Comet Holmes
is visible as a bright patch near ecliptic latitude 30◦ and ecliptic longitude 60◦
maneuver on November 20 (marked by the vertical line). After the roll manuever, a
different detector was used for the (62,120) pixel measurement. The two detectors
have somewhat different sensitivity to starlight. On the left side of the vertical line,
the irradiance initially appears almost level until the comet begins drifting into the
1◦ × 1◦ field-of-view (FOV). On the right side of the vertical line, the irradiance
initially declines as the comet outburst fades away, then levels off.
One should note that after the roll maneuver, the irradiance measured by this
pixel remained very stable throughout November and December. The signal during
October had a larger variance, but was also relatively constant before the comet
entered the field of view. Therefore, the SWAN detectors show a stable value for
the irradiance of the background stars, although the level is different at each end
because of the two different detectors used. To estimate the comet brightness on
the cross-calibration days 22 and 24 November, we subtracted a straight-line fit to
the last 19 days of the observation period to obtain estimated comet brightnesses
for 22 November and 24 November of 146.2 R and 99.7 R, or irradiance values of
3,545 and 2,417 photons/s/cm2 /nm.
Uncertainty in SWAN Irradiance Measurement

We also examined the behavior of neighboring pixels. Our analysis found that
scattered light from the spacecraft was not important in the comet pixels. However,
a pixel registration offset was identified by observing the diagonal region of stars in
SOLSTICE 259
Figure 9.2: The signal from the SWAN pixel containing the comet as a function
of time (+ symbols) and after subtraction of the estimated Lyman-α background
(lower solid curve). The two diamonds after the spacecraft roll maneuver (vertical
line) are the two points used in the cross-calibration
which the comet was embedded. After the spacecraft roll maneuver on 20 Novem-
ber, over the course of several days, a diagonal group of pixels at one edge of the
band of stars went from low brightness levels to a higher steady level. A diagonal
group of pixels at the other edge of the band of stars went from a high initial level
to a lower level over the same period of time. This behavior can be understood as
a slight temporary shift in pixel positions with respect to the star field. The size
of the brightness shift is of the same order of magnitude as the comet brightness
found for the cross-calibration period.
SOLSTICE
SOLSTICE is a grating spectrometer whose primary mission is to measure
the ultraviolet solar spectral irradiance. It also observes the irradiance from an
ensemble of stable, bright, early-type stars in order to monitor changes in the
instrument response. As detailed in Snow et al. (2005, 2012), SOLSTICE can
make measurements over such a large dynamic range simply by changing apertures
and integration times. All other optical elements are used for both solar and stellar
observations.
Unlike SWAN, SOLSTICE is on a spacecraft in low Earth orbit. SORCE orbits
at an altitude of about 580 km with an orbital period of 97 min. The inclina-
tion of the orbit is 40◦ . This orbital geometry means that comet observations
at wavelengths near Lyman-α are contaminated by geocoronal airglow emissions
(Snow et al. 2005). In order to correct for contributions to the measured irradiance
from the airglow, comet observations are combined with observations of the nearby
260 9. Comet Holmes
sky background before and after each comet radiance measurement. This technique
is more fully described in Snow et al. (2005) and is the same observing strategy
used by SOLSTICE to measure stellar irradiances near Lyman-α on a routine ba-
sis. All the observations described here were conducted while the spacecraft was
in the eclipse portion of its orbit, so only the nightglow need be considered.
SOLSTICE Comet Observations

Figure 9.3 shows a typical observing sequence. The observation begins with
a measurement of the airglow plus interplanetary background by pointing the in-
strument at a region of the sky which contains no bright stars (i.e. a dark region).
The spacecraft then slews to point at the comet and SOLSTICE measures the
comet plus background. Finally, the spacecraft slews back to the dark region.
The observed airglow background changes quickly as a function of time due to the
spacecraft orbital motion. Over the course of an observing sequence (15 min), the
solar zenith angle typically changes from 125 to 155◦ . The orbital geometry for
all of these observations was very similar: in each case, the observation began at
the start of the eclipse period. This background is modeled with a simple polyno-
mial function plus an offset during portion of the observation including the comet.
This background thus accounts for the geocoronal airglow as well as detector dark
current and contributions from interplanetary hydrogen. The dark region used in
these observations is at Right Ascension 3h 44m , declination +28◦ .
Before and after the observation of the comet, the instrument also rotates the
grating to measure the position the zeroth spectral order. The purpose of the
zero order scan is to determine the current offset of the grating drive. As shown
in Snow et al. 2005, the offset varies by a few Angstroms after each hardware
reset. Figure 9.4 shows that the offset determined by zero order scans for stellar
observations was very stable during the time period where comet observations took
place. The large change in the offset during the comet observations is simply due
to peak brightness of the background not being at the center of the FOV of the
instrument. The precise position on the sky of the spacecraft boresight during
these observations is given in Table 9.1.
For these comet observations, the final count rate, C, is the total count rate
(Craw ) minus the modeled geocoronal emission (Cairglow ), which includes the dark
rate. Additionally, the observed count rate must be corrected for long-term degra-
dation (DEG). At the time of these observations, the degradation factor at Lyman-
α was 0.89, meaning that the responsivity has decreased by 11 % through the first
four and a half years of flight (see Snow et al. 2005 for a description of how this
correction factor is determined).
A typical airglow spectrum from SOLSTICE is shown in Fig. 9.5. In this ex-
periment, the grating is rotated through an angle θ to direct Lyman-α wavelengths
to the detector from throughout the FOV. SOLSTICE is much like an objective
grating spectrometer in stellar mode, so the emission from different parts of the
FOV appear shifted relative to the center (Snow et al. 2005). This shape arises
from the convolution of the circular entrance aperture with the primary mirror
(M1) as a function of grating angle. The grating angle used in these observations
was equivalent to a shift of 1.7 nm away from the center of the distribution, based
SOLSTICE 261
Figure 9.3: Typical SOLSTICE comet observation. In order to correct for geo-
coronal airglow emission, a region of the sky with no background stars is observed
before and after the comet observation. The top panel shows both the raw counts
of the background airglow and the comet and the polynomial model fit to the
background alone during the comet observation. The diamonds are observed back-
ground counts and the plus symbols indicate signal from the comet plus background.
The small circles in the top panel are the counts from the comet after subtracting
the background estimated from the model. The middle panel shows the observation
after the polynomial background has been removed. The bottom panel shows the
histogram of the final count rate measurement for the observation
on the zero order position derived from the stellar measurements before and after
the cometary observations to determine the absolute offset. Given the shape of
the airglow spectrum in Fig. 9.5, a 1.7 nm shift in grating angle produces a signal
that is reduced from the peak by a factor of 0.96. Therefore the observed count
rate must be divided by an additional factor of fλ = 0.96. The final count rate is
therefore
Craw − Cairglow
C= . (9.1)
DEG · fλ
The final count rate was then converted to irradiance using the SOLSTICE
measurement equation in McClintock et al. (2005b). Lyman-α irradiances for all
observations of comet Holmes are shown in Table 9.1.
262 9. Comet Holmes
Figure 9.4: Wavelength stability during the time period of the comet observations.
The position of the zero order offset for stellar observations throughout the day
changed very little in November 2007. The large offset between the stellar and
cometary datasets is due to the offset between the instrument boresight and the
peak of the background Lyman-α radiance
Uncertainty in SOLSTICE Irradiance Measurement

The uncertainty in the irradiance measurements given in Table 9.1 are a com-
bination of random and systematic errors. The random errors are due to simple
photon counting statistics. The width of the histogram in the bottom panel of
Fig. 9.3 is a measure of this variation. Since there are 200 samples taken during the
on-comet portion of the observation, the uncertainty of the mean due to random
error is actually fairly small.
The more complex and significant source of uncertainty is from the systematic
errors of the observation and data reduction. There is about a 1 % uncertainty in
the degradation correction. The uncertainty in the polynomial fit to the background
is probably no more than a few percent. One test of the accuracy of the polynomial
model is to calculate the residual slope after the correction for each of the three sub-
observations. As shown in the middle panel of Fig. 9.3, the slope of the corrected
comet counts with time is zero to within the uncertainty of the linear fit. A
conservative estimate would be that the polynomial model accurately removes the
background to within five percent.
The greatest source of uncertainty is in the fλ correction used in Eq. 9.1. It
not only represents the commanded grating angle, but also represents pointing
errors between the instrument boresight and the true position of the comet in
the FOV. During the campaign of observations for comet Holmes, the ephemeris
position of the comet was well known. However, due to instrument misalignment,
SOLSTICE 263
Figure 9.5: Spectral scan of airglow background including model fit
Table 9.1: SOLSTICE Lyman-α irradiance measurements of comet Holmes

Time RA Dec Net Count Rate Irradiance
YEAR/DOY-HH:MM:SS (deg) (deg) (counts/sec)
2007/325−18:58:56 49.7648 49.8046 78.0 3,690
2007/325−20:36:04 49.7653 49.8037 72.1 3,420
2007/325−22:13:12 49.7665 49.8019 63.7 3,020
2007/325−23:50:20 49.7671 49.8012 75.6 3,590
2007/327−08:13:06 49.7828 49.7806 68.4 3,240
2007/327−09:50:14 49.7839 49.7792 67.1 3,180
2007/327−11:27:23 49.7846 49.7785 71.7 3,400
2007/327−13:04:31 49.7856 49.7773 58.7 2,780
2007/327−14:41:40 49.7862 49.7765 70.0 3,320
2007/327−16:18:49 49.7873 49.7754 60.0 2,850
2007/327−17:55:57 49.7884 49.7740 68.8 3,260
2007/327−19:33:06 49.7889 49.7733 62.8 2,980
2007/327−21:10:15 49.7900 49.7723 60.5 2,870
Irradiances are given in photons/s/cm2 /nm
the SOLSTICE FUV boresight is offset by 20 arc minutes from the spacecraft z-
axis. It is this z-axis which is commanded to point at the ephemeris position of the
comet nucleus. The position angle of the comet in the instrument FOV will depend
on the roll angle of the spacecraft. For each observation, the spacecraft planning
system chooses a roll angle which minimizes obscuration of the star tracker’s FOV.
Therefore the roll angle may vary from one observation to the next. In the worst
case scenario, i.e. two roll angles separated by 180◦ , the comet’s wavelength scale
264 9. Comet Holmes
Figure 9.6: Time series of SOLSTICE Lyman-α irradiances from comet Holmes.
The error bars shown on the plot only include the statistical error for each obser-
vation. Systematic errors due to pointing are discussed in the text
could shift by up to 0.7 nm. This would change fλ by about 10 %. Based on the
preflight calibration, the absolute uncertainty of SOLSTICE at 121.6 nm is about
four percent (McClintock et al. 2005b), so the uncertainty in the fλ correction
factor is the dominant source of uncertainty in the SOLSTICE measurement.
Since these three sources of systematic error are all independent, they can be
added in quadrature to give a combined standard uncertainty of about 12 %, which
is consistent with the observed spread in the irradiance measurements shown in
Fig. 9.6.
Comparisons and Conclusions

Observations of comet Holmes were conducted in the hopes of cross-calibrating
SWAN and SOLSTICE. As detailed in Table 9.1, SOLSTICE obtained accurate
background-subtracted Lyman-α measurements of Holmes on 21 November 2007
(day 325) with an average value of 3,400 photons/s/cm2 /nm and on 23 November
2007 (day 327) with an average value of 3,000 photons/s/cm2 /nm. The SWAN
measurement for the intervening day is approximately 3,300 photons/s/cm2 /nm.
However, the comparison between the two sets of measurements needs to take
account of the difference in spectral bandpass between the two instruments. The
SWAN measurement includes contribution from wavelengths other than Lyman-α.
It has some sensitivity to the entire FUV range. We can use SOLSTICE spectra
of comet Holmes taken earlier in November 2007 to estimate the contribution from
the rest of the FUV. Fig. 9.7 shows the average of 30 spectra taken on 13 November
2007. The airglow dominates the spectrum below 130 nm. Longwards of 130 nm,
Comparisons and Conclusions 265
Figure 9.7: SOLSTICE spectra of comet Holmes from 13 November 2007. The
top panel shows the average count rate from 30 co-added spectra after background
subtraction. The lower panel shows those counts converted to irradiance units
(photons/s/cm2 /nm). The signal-to-noise ratio was very low during these obser-
vations, so the uncertainty in the derived irradiance is significant
there is a weak signal from the comet. The top panel shows the residual count rate
after background subtraction. The signal-to-noise ratio is only about 1 for each
spectrum, so even co-adding 30 observations yields an irradiance with about a 30 %
uncertainty. Integrating over the entire FUV band gives a contribution of about
1,300 photons/s/cm2 /nm to the signal measured by SWAN.
If the comet did not change UV brightness over the 9 days from 13 November
to 22 November, then the rest of the FUV band would contribute about a third of
the spectral irradiance measured by SWAN. However, based on Fig. 9.2, the total
irradiance in the SWAN band decreased by about a factor of two over those 9 days.
Therefore the non-Lyman-α contribution to the comet flux on 22 November is
probably also half as much as it was on 13 November when the SOLSTICE spectra
were acquired. Therefore the FUV contribution to the SWAN measurement on 22
November was probably no more than 15 % of the total signal. If so, the corrected
SWAN value on 22 November would be about 2,800 photons/s/cm2 /nm (Table 9.2).
These observations of Lyman-α emission from a comet by both SWAN and
SOLSTICE confirm their similar calibration. As discussed in Snow et al. (2012,
this volume), the SOLSTICE calibration is in good agreement with observations
from the International Ultraviolet Explorer based on the white dwarf flux scale
(Bohlin et al. 1990; Bohlin 1996). This white dwarf scale is also used to calibrate
266 9. Comet Holmes
Table 9.2: Summary of SWAN and SOLSTICE irradiance measurements of comet

Holmes
Year/DOY Month Day SOLSTICE SWAN
2007/324 Nov 20 – –
2007/325 Nov 21 3,400 –
2007/326 Nov 22 – 3,500
2007/327 Nov 23 3,000 –
2007/328 Nov 24 – 2,400
Irradiances are given in photons/s/cm2 /nm
the instruments on the HST, and the inflight calibration of SWAN used the God-
dard Hight Resolution Spectrograph on HST (Quémerais and Bertaux 2002). So
agreement between SOLSTICE and SWAN is to be expected.
In summary, the comparison technique of using a target of opportunity such as
a comet was moderately successful. However, due to the relatively low signal levels
in these observations, the uncertainty in the Lyman-α measurement is significant.
The agreement between SWAN and SOLSTICE is not tightly constrained by these
observations due to the large uncertainty in the SWAN measurement caused by
the contribution of the background stars. Comet Holmes, a rare naked-eye comet,
apparently was not a particularly strong Lyman-α source. Two other comets were
easier to see in the SWAN data from this period. This may indicate a larger role for
dust than water in the 2007 comet Holmes outburst. Other comets, not necessarily
naked-eye comets, might actually be better targets for cross-calibration.
Unfortunately, no follow-up observations will be possible with SORCE SOL-
STICE. Spacecraft systems on SORCE, particularly the batteries, have forced op-
erations during the eclipse portion of the orbit to be curtailed. Starting in late
2010, only a few of the standard stellar calibration observations can be scheduled,
and only during orbits with eclipse durations less than 30 min.
Bibliography
J-L. Bertaux et al., SWAN: a study of solar wind anisotropies on SOHO with Lyman-alpha
sky mapping. Solar Phys. 162, 403 (1995) doi:10.1007/BF00733435
J-L. Bertaux et al., First results from SWAN Lyman alpha solar wind mapper on SOHO.
Solar Phys. 175, 737–770 (1997) doi:10.1023/A:1004979605559
R.C. Bohlin, Spectrophotometric standards from the far-uv to the near-ir on the white dwarf
flux scale. Astrophys. J. 111, 1743–1747 (1996) doi:10.1086/117914
R.C. Bohlin, A.W. Harris, A.V. Holm, C. Gry, The ultraviolet calibration of the hubble space
telescope iv. absolute IUE fluxes of Hubble Space Telescope standard stars. Astrophys.
J. Suppl. 73, 413–439 (1990) doi:10.1086/191474
M.R. Combi, Y. Lee, T.S. Patel, J.T.T. Mäkinen, J-L. Bertaux, E. Quémerais,
SOHO/SWAN observations of short-period spacecraft target comets. Astrophys. J. 141,
128 (2011) doi:10.1088/0004-6256/141/4/128
M.R. Combi, Z. Boyd, Y. Lee, T.S. Patel, J-L. Bertaux, E. Quémerais, J.T.T. Mäkinen,
SOHO/SWAN observations of comets with small perihelia: C/2002 V1 (NEAT), C/2002
X5 (Kudo-Fujikawa), 2006 P1 (McNaught) and 96P/Macholtz 1. Icarus 216, 449–461
(2012) doi:10.1016/j.icarus.2011.09.019
Bibliography 267
W.E. McClintock, G. Rottman, T.N. Woods, Solar stellar irradiance comparison experiment
ii (SOLSTICE II): instrument concept and design, Solar Phys. 230, 225–258 (2005a)
doi:10.1007/s11207-005-7432-x
W.E. McClintock, M. Snow, T.N Woods, Solar stellar irradiance comparison
experiment ii (SOLSTICE II): pre-launch and on-orbit calibrations. Solar Phys. 230,
259–294 (2005b) doi:10.1007/s11207-005-1585-5
E. Quémerais, J-L. Bertaux, Radiometric calibration of the SWAN instrument. In The
Radiometric Calibration of SOHO, ed. by A. Pauluhn, M.C.E. Huber, R. von Steiger.
(ISSI Scientific Series, SR-002, 2002)
G. Rottman, The SORCE mission. Sol. Phys. 230, 7–25 (2005) doi:10.1007/s11207-005-
8112-6
G.Rottman, T.N. Woods, T. Sparn, Solar stellar irradiance comparison experi-
ment 1. i—instrument design and operations. J. Geophys. Res. 98, 10667 (1993)
doi:10.1029/93JD00462
D.G. Schleicher, The long-term decay in production rates following the extreme out-
burst of comet 17P/Holmes. Astrophys. J. 138, 1062–1071 (2009) doi:10.1088/0004-
6256/138/4/1062
experiment ii (SOLSTICE II): examination of the solar-stellar comparison technique. Sol.
Phys. 230, 295–324 (2005) doi:10.1007/s11207-005-8763-3
M. Snow, A. Reberac, E. Quémerais, J. Clarke, W.E. McClintock, T.N. Woods, A new
catalog of ultraviolet spectra. In Cross-Calibration of Past and Present Far UV Spectra of
Solar System Objects and the Heliosphere, ed. by E. Quémerais, M. Snow, R.M. Bonnet.
(ISSI Scientific Series, SR-013, 2012) (this volume)
T.N. Woods, P.D. Feldman, G.J. Rottman, Ultraviolet observations of comet Hale-
Bopp (C/1995 O1) by the UARS SOLSTICE. Icarus 144, 182–186 (2000)
doi:10.1066/icar.1999.6262
Index
ACE. See Advanced Composition 42, 57, 68–71, 79–86, 97,

Explorer (ACE) 100, 101, 103, 104, 106,
ACO. See Annual check out (ACO) 121–125, 143, 165, 168, 170,
ACR. See Anomalous cosmic ray 179
(ACR) Chromosphere, 73, 192
Advanced Composition Explorer Circumheliospheric, 16
(ACE) Circumheliospheric interstellar
Solar Wind Electron Proton medium (CHISM), 16
Alpha Monitor Cloudlets, 13
(SWEPAM), 98 Clouds, 10, 11, 13, 70, 71, 142, 145,
Solar Wind Ion Composition 204, 207
Spectrometer (SWICS), CME. See Coronal mass ejection
108, 172 (CME)
AE-E. See Atmospheric Explorer-E Collisionless, 25, 69, 70
(AE-E) Comets
Albedo, 164, 227–251 Hale-Bopp, 256
Anisotropic, 71, 99, 104 17P/Holmes, 256
Anisotropy(ies), 12, 24, 31, 57, 74, Composite, 72–74, 78, 79, 107, 127,
92, 97, 104, 105, 156 165
Annual check out (ACO), 179, 180, Cone, 48, 113
183, 184 Convolution, 209, 210, 239, 240, 245,
Anomalous cosmic ray (ACR), 14, 15 246, 260
Atmospheric Explorer-E (AE-E), 73 Core-to-wing, 72
Axicylindrical, 155 Corona, 73, 74, 86
Axisymmetric, 15, 21, 29, 30, 52, 155 Coronal, 92, 98, 100, 106, 114
Coronal mass ejection (CME), 98
Co-rotating, 9
bi-Maxwellian, 99
Crosswind, 30, 35, 36, 52, 53, 55
Blobs, 9
Bubble, 14 Cruise, 1, 107, 108, 144, 151, 164,
166, 168, 170, 173, 177–186
Calibration observation (CALOBS), Damping, 171

201–203, 210–212, 214, 215, Database, 3, 36, 37, 87, 89, 98, 100,
266 101, 117, 144, 165, 167, 194
Cassini, 3, 26 Degradation, 171, 193, 194, 196,
Ultraviolet Imaging 198–201, 204, 215, 230, 242,
Spectrograph (UVIS), 164 260, 262
Center-to-limb, 72 Diffraction, 102, 235, 236
Charge-exchange, 2, 9–13, 15, 16, Diffusion, 25
18–21, 23–26, 28–30, 35–38, Disk-averaged, 73

DOI 10.1007/978-1-4614-6384-9, © Springer Science+Business Media New York 2013
270 Index
Disk-equivalent, 242–244 233, 237, 239–243, 245, 247,

Disk-integrated, 73–77, 228, 248, 249 248, 250, 251, 257, 263–265
Dispersion, 204, 231, 236–238, 245, Fast-latitude scan, 100, 108, 110,
248 115, 117, 118, 120
Distribution, 1–3, 7–59, 68–71, 73, Fluorescence, 104
74, 80, 81, 85, 86, 93, 94, Flyby, 3, 166, 186, 228
97, 99, 109, 113, 126, FONDUE. See Fully ONline
142–146, 148, 152, 155, 156, Database of Ultraviolet
158, 164, 165, 167, 168, Emissions (FONDUE)
180, 185, 199, 207, 233, 260 Force-free, 179
Downwind, 12, 13, 20, 28, 44, 48, Full-disk, 230
50–56, 78, 84, 97, 145, 152, Full-sky, 26, 107, 155, 167, 173
153, 155, 156, 164, 165, Fully ONline Database of Ultraviolet
172, 179 Emissions (FONDUE), 195,
Dust, 266 215
FUV. See Far ultraviolet (FUV)
Early-type, 192, 230, 259
Ecliptic, 11, 36, 41, 71–73, 83, 84, Galactic cosmic ray (GCR), 14, 15,
89–92, 97–103, 106–108, 57
112–115, 117, 151, 153, 154, Gasdynamics, 18, 23
156, 157, 164, 165, 167, GCR. See Galactic cosmic ray
179, 180, 184, 185, 257, 258 (GCR)
Ejections, 98 Geocentric, 36
Electron-impact, 2, 15, 18, 25, 68, Geocorona, 8, 11, 145, 257
70, 83–85, 93–97, 104, 125 g-factor, 165
Emissivity, 11, 46, 48, 97, 113, 125, Global Ozone Monitoring
143, 146, 156 Experiment (GOME), 229
Energetic neutral atom (ENA), 9, Gravity-assist, 108
10, 18, 20, 24–27, 68, 79, Great-circle, 145, 166, 180–183, 186
81, 86 Groove, 104, 105, 107, 167, 168, 170,
EUV. See Extreme ultraviolet 195, 196
(EUV) Ground-based, 3, 192, 203, 233
EUV81, 89 g-value, 165, 166
EUV flux model for aeronomic
calculations (EUVAC), 89 Halo, 94–96
Exoatmospheric, 228 Halo-to-core, 95
Exosphere, 1, 8, 76, 142, 146, 166, H alpha, 8
169 Heliocentric, 3, 18–20, 24–26, 35, 38,
Extraheliospheric, 106, 107 40–43, 86, 95, 109, 167
Extraterrestrial, 8, 10–12 Helioglow, 69–71, 74, 76, 86, 102,
Extreme ultraviolet (EUV), 68, 69, 104, 106, 108, 109, 112,
72, 73, 79, 88, 89, 92, 97, 113, 118, 125–126
100, 107, 126, 165 Heliographic, 42, 43, 106, 109
Heliolatitude, 2, 29, 36, 39, 43, 68,
F10.7 , 72, 73, 89–92, 108 71, 73–75, 92, 102, 104–110,
Far ultraviolet (FUV), 1, 166, 112–116, 118–126
193–197, 199, 201, 230, 232, Heliolongitude, 42, 43
Index 271
Heliopause, 14, 16, 18, 19, 21, 22, 25, 164, 165, 167, 172, 177,
26, 28, 30, 57, 70, 80, 144 179, 180, 182–184
Heliosheath, 9, 10, 14–21, 25–28, 70, Interstellar magnetic field (IsMF), 9,
71, 82, 143 14, 21–24, 57, 58, 142
Heliotail, 58 Interstellar medium (ISM), 15, 16,
High-energy, 9 21, 24, 26, 148, 177
Hopkins Ultraviolet Telescope IP. See Interplanetary (IP)
(HUT), 228 IPH. See Interplanetary hydrogen
Hot-type, 28, 58, 59 (IPH)
Hubble Space Telescope (HST) IPM. See Interplanetary medium
Goddard High Resolution (IPM)
Spectrograph (GHRS), 145, IPS. See Interplanetary scintillation
153–155, 169, 257 (IPS)
Space Telescope Imaging ISM. See Interstellar medium (ISM)
Spectrograph (STIS), 2, IsMF. See Interstellar magnetic field
145, 153, 155 (IsMF)
HUT. See Hopkins Ultraviolet Isothermal, 146
Telescope (HUT) Isotropic, 9, 12, 24, 25, 96, 155–158,
167
IMP-8. See Interplanetary IUE. See International Ultraviolet
Monitoring Platform Explorer (IUE)
(IMP-8)
Inflow, 68, 85, 103, 106, 108 Joint Polar Satellite System (JPSS),
Intercalibration, 71, 98, 109, 250
126 JPSS. See Joint Polar Satellite
International Ultraviolet Explorer System (JPSS)
(IUE), 2, 3, 179, 192, Jupiter, 107, 108
201–203, 205, 206, 210,
212–215, 229, 265 Kinetic, 1, 8, 18, 19, 24, 25, 27,
Interplanetary (IP), 1, 2, 8–10, 12, 29–36, 38, 40–44, 46, 51,
14, 24, 58, 59, 67–127, 52, 55–57, 59, 80, 93, 94
141–159, 167, 168, 170, 178, Kinetic-continuum, 57, 58
257, 260 Kinetic-gasdynamic, 23
Interplanetary hydrogen (IPH), 8, Kinetic-MHD, 23, 27, 28
20, 149, 151, 152, 154–158,
164, 169, 170, 256, 260 LASP Interactive Solar IRradiance
Interplanetary medium (IPM), 1, 75, Datacenter (LISIRD), 127
142, 148, 177–186 LASP Lunar Albedo Measurement
Interplanetary Monitoring Platform and Analysis from
(IMP-8), 98 SOLSTICE (LLAMAS),
Interplanetary scintillation (IPS), 227–251
102, 103, 105, 107, 111–117, LIC. See Local interstellar cloud
123, 125–126 (LIC)
Interstellar, 1, 7–59, 68–73, 77–83, Lightcurves, 106
85, 86, 93, 96, 97, 100, 101, Limb, 116
103, 104, 108, 109, 126, Line-center, 8, 45, 76–78, 148, 152,
142, 143, 145, 148, 156, 156, 159, 165, 178, 179, 185
272 Index
Line-integrated, 76, 77 Visible-Infrared Spectrograph

LISIRD. See LASP Interactive Solar (VIRS), 166
IRradiance Datacenter Mesosphere, 165
(LISIRD) Mikulski Archive for Space
LISM. See Local insterstellar Telescopes (MAST), 213
medium (LISM) Moon, 1, 3, 164, 227–251, 256
LLAMAS. See LASP Lunar Albedo Multi-component, 15, 27, 57, 58
Measurement and Analysis Multi-fluid, 18, 23, 86
from SOLSTICE Multi-population, 146
(LLAMAS)
Local interstellar cloud (LIC), Near-Earth, 10, 99, 165, 179
13–15, 19, 28, 30, 57–59, Near-infrared, 228
71, 142, 145 Near-UV, 229
Local interstellar medium (LISM), Neptune, 178, 183
7–9, 27, 31, 33, 57, 167 New Horizons (NH)
Loss-function, 11, 50 Alice, 2, 141, 142, 151–153, 157,
Lunar Reconnaissance Orbiter 179–186
(LRO) Nightglow, 260
Lunar Reconnaissance Orbiter Non-equilibrium, 25–27, 57
Camera (LROC), 229 Non-Gaussian, 71, 207
Lyman Alpha Mapping Project Non-isotropic, 12
(LAMP), 163–174, 229 Non-Maxwellian, 24, 30, 31, 35, 36,
Wide Angle Camera (WAC), 46, 51, 53–55, 57
229 Non-radial, 85
Non-stationary, 13, 38–44, 56
Non-thermal, 26, 59
Magnetosonic, 23 Nowcast, 36
Magnetosphere, 97 Nozomi, ultraviolet scanner (UVS),
Mariner 2, 97 228
Mariner 10, 12, 228
Mars, 2, 3, 159, 191, 192, 203–205 OAO-2. See Orbiting Astronomical
Mars-7, 12 Observatory (OAO-2)
Mars-Express, 2, 3, 192, 203–205 Occultation, 203, 204
MAST. See Mikulski Archive for OGO. See Orbiting Geophysical
Space Telescopes (MAST) Observatory (OGO)
Mercury, 159, 164, 166, 173, 228, 244 OMNI, 36, 98–100, 109, 112, 115,
MErcury Surface, Space 117, 120, 121, 165, 173
ENvironment, OMNI-2, 83, 87, 98, 101, 117
GEochemistry, and OMNIWeb, 36, 98, 165
Ranging (MESSENGER) One-component, 32, 33, 47, 52–54
Mercury Atmospheric and One-fluid, 16
Surface Composition Orbiting Astronomical Observatory
Spectrometer (MASCS), 2, (OAO-2), 212
159, 163–174 Orbiting Geophysical Observatory
Ultraviolet-Visible (OGO), 11, 142
Spectrometer (UVVS), 166, Out-of-ecliptic, 103, 112
209 Ozone, 229
Index 273
Particle-background, 90 Scintillation, 102–103, 112–118, 126

Phase-reddening, 228 Secular, 84, 97, 100, 103, 119
Photochemistry, 178, 183 Selenographic, 233, 234
Photoelectrons, 204, 206 Shock, 9, 10, 14–23, 25, 28, 29, 51,
Photoionization, 2, 10–12, 15, 25, 29, 57, 58, 70, 85, 99, 164, 172,
36, 38, 57, 68, 75, 79, 186
84–93, 96, 97, 100, 104, Shuttle Solar Backscatter Ultraviolet
112, 125, 165 (SSBUV), 229
Photometry, 2, 107, 125, 126, 205, Sidewind, 13
228, 229, 242, 246, 249, 250 Sky-maps, 11, 26, 42, 107, 171, 257,
Photosphere, 192 258
Pickup ion (PUI), 14–16, 20, 57, Slot, 107, 179–181
85, 86 SME. See Solar Mesosphere
Pile-up, 21, 144 Explorer (SME)
Pioneer-Venus, 12 SNOE. See Student Nitric Oxide
Pitch-angle, 24 Explorer (SNOE)
Plage, 166 SOLAR2000, 36, 89–92
Pluto, 151, 179, 183, 185, 186 Solar-blind, 169, 257
Point spread function (PSF), 205, Solar Mesosphere Explorer (SME),
209, 210 73, 151
Polarization, 246–250 Solar Orbiting Heliospheric
Porosity, 164 Observatory (SOHO)
Post-shocked, 9 Charge, Element, and Isotope
Prognoz, 11, 12, 167 Analysis System
Pseudo-counting, 204 (CELIAS)
Pseudo-temperatures, 143 Solar EUV Monitor (SEM),
PSF. See Point spread function 88–90
(PSF) Large Angle and Spectrometric
PUI. See Pickup ion (PUI) Coronagraph (LASCO),
244
Quasar, 102 Solar Ultraviolet Measurements
Quasi-neutral, 15, 94 of Emitted Radiation
Quasi-neutrality, 94 (SUMER), 143, 159
Quasi-perpendicular, 25 Solar Wind ANisotropies
Quasi-thermal, 94 (SWAN), 2, 3, 8, 12, 23, 24,
Quaternion, 236 37, 53, 58, 59, 142, 153,
156–158, 163–174, 255–266
Redistribution, 143, 144, 146 Solar-origin, 24
Reflectance, 164, 166, 228, 232, 233, SOlar Radiation and Climate
235, 237, 242, 246, 249, 250 Experiment (SORCE)
Regolith, 249 Solar Irradiance Monitor (SIM),
Remote sensing, 68, 109, 113, 186, 250
250 SOLar-STellar Irradiance
Responsivity, 154, 196–200, 205, 214, Comparison Experiment
215, 242, 244, 245, 248, 260 (SOLSTICE), 3, 73, 152,
RObotic Lunar Observatory 156–158, 193–196, 202, 215,
(ROLO), 228, 229, 246 231, 240, 248, 255–266
274 Index
Spectrograph, 145, 164, 166, 179, eta Aur, 211, 212, 220
180, 205, 266 eta UMa, 194, 201, 211–213,
Spectrometer, 1–3, 142, 144, 145, 215, 220
151, 152, 159, 164, 166, kappa Vel, 194, 211, 220, 224
169, 172, 192, 193, 205, sigma Sgr, 194, 211, 221, 225
229, 231, 232, 234, 240, tau Sco, 194, 211, 221, 225
246, 257, 259, 260 zeta Cas, 211, 212, 221, 225
Spectroscopy for Investigation of zeta Oph, 211, 212, 222, 225
Characteristics of the zeta Pup, 209–212, 222, 224
Atmosphere of Mars Strahl, 94
(SPICAM), 2, 3, 159, 192, Streamlines, 22, 23, 97
203–215, 224 Student Nitric Oxide Explorer
Spectroscopy for Investigation of (SNOE), 229
Characteristics of the Subsonic, 9, 16, 70
Atmosphere of Venus Sunspot, 72, 89
(SPICAV), 159 Suprathermal, 27
SPICAM. See Spectroscopy for Synchrotron Ultraviolet Radiation
Investigation of Facility (SURF), 196–198,
Characteristics of the 201, 246, 248
Atmosphere of Mars
(SPICAM)
SPICAV. See Spectroscopy for Temperature, 11, 12, 16–18, 22,
Investigation of 30–32, 35, 36, 38, 40–46,
Characteristics of the 51–57, 70, 71, 93–96, 98,
Atmosphere of Venus 99, 126, 146, 164, 178, 192,
(SPICAV) 196, 242, 246
SSBUV. See Shuttle Solar Thermodynamics, 24, 25
Backscatter Ultraviolet Thermoelectric, 183
(SSBUV) Thermosphere Ionosphere
Stars Mesosphere Energetics and
alpha CMa, 194, 211, 216, Dynamics (TIMED), 92,
224 165
alpha Cru, 194, 199, 211, 216 Solar EUV Explorer (SEE), 73,
alpha Gru, 194, 211, 216, 225 88–90
alpha Leo, 194, 201, 211–213, Three-component, 73, 94, 143
217, 224 Three-dimensional, 21, 31, 36, 86,
alpha Lyr, 194, 201, 203, 155, 156, 158
211–213, 217, 225 Three-Gaussian, 76
alpha Pav, 194, 211, 217, 225 TIMED. See Thermosphere
alpha Vir, 194, 209, 211, 218, Ionosphere Mesosphere
224 Energetics and Dynamics
beta Cen, 194, 201, 211, 212, (TIMED)
218, 225 Tomography, 103, 112–117
beta Lup, 225 Total and Spectral Irradiance Sensor
delta Cen, 194, 211, 219, 224 (TSIS), 250
delta Sco, 194, 211, 219, 225 Two-component, 15, 33, 35, 47,
epsilon Per, 194, 211, 219, 224 52–57, 73, 143
Index 275
Two-dimensional, 46–52, 209 Upwind, 12, 13, 17–20, 26, 28–30,

Two-parameter, 104, 168 32, 33, 35, 36, 38, 40–44,
Two-population, 30, 53, 55, 57, 71 46–56, 58, 71, 78, 84, 85,
Two-shock, 57 97, 126, 144, 145, 152, 153,
155, 156, 164, 165, 167,
UARS. See Upper Atmosphere 172, 179, 180
Research Satellite UV-background, 142
(UARS)
Ulysses Venus, 159
Interstellar Gas (GAS), 12, Venus-Express (VEX), 3, 26, 27,
165 159, 203
Solar Wind Observations Over Voyager, 1, 18, 23, 24, 58, 85, 142,
the Poles of the Sun 145, 183, 185
(SWOOPS), 110, 112 Ultraviolet Spectrometer
Upper Atmosphere Research (UVS), 2, 146–152, 157, 158
Satellite (UARS), 194, 195,
256 WIND, 95, 100
SOLar-STellar Irradiance Solar Wind Experiment (SWE),
Comparison Experiment I 98, 99
(SOLSTICE), 72, 73, 151 Wind-interstellar, 24
Upstream, 15–18, 20, 21, 85, 99, 180,
182, 183 Zenith, 260

(ISSI Scientific Report Series 13) Eric Quémerais, Martin Snow (Auth.), Eric Quémerais, Martin Snow, Roger-Maurice Bonnet (Eds.) - Cross-Calibration of Far UV Spectra of Solar System Objects and The H

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

(ISSI Scientific Report Series 13) Eric Quémerais, Martin Snow (Auth.), Eric Quémerais, Martin Snow, Roger-Maurice Bonnet (Eds.) - Cross-Calibration of Far UV Spectra of Solar System Objects and The H

Uploaded by

Copyright:

Available Formats

ISSI Scientific Report 13

For further volumes:

ISBN 978-1-4614-6383-2 ISBN 978-1-4614-6384-9 (eBook)

Library of Congress Control Number: 2012956300

© Springer Science+Business Media New York 2013

Printed on acid-free paper

Springer is part of Springer Science+Business Media (www.springer.com)

Bern, Switzerland Roger-Maurice Bonnet

Part I: Interstellar Hydrogen and the Sun 5

3 Solar Parameters for Modeling the Interplanetary

Part II: Interplanetary Hydrogen 139

5 Lyman-α Models for LRO LAMP from MESSENGER

6 New Horizons Cruise Observations of Lyman-α Emissions

Part III: Instrument Cross-Calibration 189

8 Absolute Ultraviolet Irradiance of the Moon

9 Lyman-α Observations of Comet Holmes

E. Quémerais et al. (eds.), Cross-Calibration of Far UV Spectra of Solar System 1

between 10 to 20 AU from the Sun. The hydrogen distribution at that heliocentric

Distribution of Interstellar Hydrogen

E. Quémerais et al. (eds.), Cross-Calibration of Far UV Spectra of Solar System 7

Introduction: A Brief Historical Review

μ = |Frad |/|Fg |. (2.1)

The Solar Wind/Local Interstellar Medium

1.0 102 Vp, km/s

logarithm of plasma density, log10(cm-3)

-200 -200 -200

-400 -400 -400

-600 -600 -600

d Logarithm of (plasma+magnetic) pressure,

-13.3 -13.0 -12.7 -12.4 -12.0 -11.9 -11.8 -10.7

-200 -200 -200

-400 -400 -400

-600 -600 -600

Table 2.1: Positions of the TS in the directions of Voyager 1 and Voyager 2

Cross MT MT MT MT Steb. Steb MT MT MT MT MT

Summary of the Section

Advanced Model of the Interstellar Hydrogen

Description of the Model

df (r, w, t)/dt = −β(r, t, λ) f (r, w, t).

where fb (r0 , w0 , t0 ) is the velocity distribution function of hydrogen atoms at the

Kinetic Properties of the Hydrogen Distribution at 90 AU from the Sun

1. The existence of two populations of interstellar hydrogen atoms (primary and

2. Spatial inhomogeneity of the hydrogen distribution at 90 AU. All moments

3. Anisotropy of kinetic “temperatures”, i.e. Tz = Tρ , where

4. The correlation coeﬃcient is given by:

especially for the secondary interstellar atoms.

nH = 0.54 · nH,LISM , VH,z = −0.79 · VH,LISM, TH = 15, 000 K.

Figure 2.10: Illustration of the asymmetry of f (r, w) of the secondary interstellar

Table 2.2: Summary of the heliospheric models discussed in this manuscript

nH,prim = 0.22 · nH,LISM , Vz,prim = −1.06 · VH,LISM , TH,prim = 6, 840 K.

The values for the secondary interstellar atoms are:

nH,second = 0.32 · nH,LISM , Vz,second = −0.63 · VH,LISM , TH,second = 18, 126 K.

As shown by Katushkina and Izmodenov (2010), model 2 leads to signiﬁcant

Here DR = mkH TR , Dϕ = mkH Tϕ , and Dz = mkH Tz . k is the Boltzmann constant

there are qualitative diﬀerences in the behavior of temperatures Tz in upwind and

Modeling of Time-Dependent and Heliolatitudinal Eﬀects

where σ(wsw,E ) is the charge-exchange cross-section (Lindsay and Stebbing 2005).

βtot,E (t, λ = 0) = βex,E (t, 0) + βph,E (t, 0).

Since the time-resolution of βtot,SW AN (t, λ) is sometime smaller than 27 days to

where FB is a function described in Bzowski (2008).

Results of the Advanced Hot Model in the 3D Non-Stationary Case

Figure 2.13: Total ionization rate as a function on time and heliolatitude

Figure 2.14: (a) Parameter μ as a function of heliolatitude λ for wr = 0; (b)

3. Anisotropy of kinetic “temperatures”, i.e. Tz = Tρ , where

where .Carr marks averaging over the Carrington rotation period.