PROPAGATION OF

REFRACTION ERRORS
IN TRIGONOMETRIC
HEIGHT TRAVERSING AND
GEODETIC LEVELLING

G. A. KHARAGHANI

November 1987

TECHNICAL REPORT
NO. 132

PREFACE
In order to make our extensive series of technical reports more readily available, we have
scanned the old master copies and produced electronic versions in Portable Document
Format. The quality of the images varies depending on the quality of the originals. The
images have not been converted to searchable text.

PROPAGATION OF REFRACTION ERRORS
IN TRIGONOMETRIC HEIGHT TRAVERSING
AND GEODETIC LEVELLING

Gholam A. Kharaghani

Department of Surveying Engineering
University of New Brunswick
P.O. Box 4400
Fredericton, N.B.
Canada
E3B5A3

November 1987

Chrzanowski will be greatly appreciated. . thesis of the same title. Any comments communicated to the author or to Dr.Sc.PREFACE This report is an unaltered printing of the author's M. The thesis supervisor was Professor Adam Chrzanowski. which was submitted to this Department in August 1987.

. This thesis investigates the propagation errors in trigonometric height traversing.ABSTRACT The use of alternative to considerable trigonometric geodetic levelling attention. It is concluded that the accumulation trigonometric of the refraction height traversing effect is within in short-range the limits Canadian specifications for the first order levelling. As in geodetic levelling. levelling is sought to uncertainty height due to traversing has recently replacement A reduce the cost the refraction for as an been given geodetic and to reduce the and other systematic errors. the atmospheric refraction can be the main source of error in the trigonometric method. for the temperature profile up to propos. of accepted Kukkamaki 's have shown of fit that the and are easier new to utilize. A computer simulation of the influence of refraction. in trigonometric height traversing accumulation of the refraction suggests that the effect becomes randomized to a large extent over long traverses.ii - of .ed and models give The results better precision refraction Three new models 4 m above the ground are compared with the widely temperature model.

. . • Angle of refraction error • • • • • • • • • • . . Profile of mean potential temperature gradients The Angus-Leppan equation for refraction correction • . .. Comments on the Discussed Methods • • • • • • . LIST OF FIGURES • . ..iii - 20 22 27 27 30 31 35 36 38 .. 1 2. • . Chapter 1... . . . • • • • • • • • .. refraction correction • • • • • • • Refraction Correction Formulated in Terms of Sensible Heat Flux . . • 3. Kukkamaki's equation for geodetic levelling . . . .. . . .. . . . . .. . Thermal stability parameter • • • • • • • • • .TABLE OF CONTENTS . . . . . . . .. . . . Investigation by Holdahl • • • • • • • • • Comments on the Meteorological Methods • • • • . . . . . . . . Angle of Refraction Derived from the Variance of the Angle-of-Arrival Fluctuations • Determination of the Vertical Refraction Correction Using the Reflection Method • . . • . .. Determination of the Vertical Refraction Angle Using Lasers of Different Wavelengths • . . . . . .. . . INTRODUCTION . . Review of the meteorological parameters • • • . . 7 7 9 13 15 16 17 REFRACTION CORRECTION IN GEODETIC LEVELLING USING THE METEOROLOGICAL METHOD • • • • • • • • • • • 19 Refraction Correction Based on Direct Measurement of Temperature Gradient • • • • • • . . vi LIST OF TABLES .. . . • • • viii ACKNOWLEDGEMENTS . . • ii .. . • . . . • • • • • • • • • • • . .. . . .. • • . . A REVIEW OF METHODS FOR THE DETERMINATION OF THE REFRACTION CORRECTION • • • • • • • • • • • 6 Determination of the Vertical Refraction Angle by the Meteorological Approach • • • • • • Refractive index of air • • • • • . . .. • xi ABSTRACT .. . ... .

• • • • . Description of the Field Equipment • • • • • . . . . • • • . . . . Leap-Frog Trigonometric Height Traversing Formulae • • • • • • • • • • • • . . . . . . . Proposed method for the calculation of refraction correction • • • • • • • Refraction in Leap-Frog Trigonometric Height Traversing • • • • • • • • • • • • • • • • . . . . . • Head-Hall test line • • • • • • • • • • . . 67 Choice of models • 67 Temperature gradient measurement • o • • o • • 69 Determination of the coefficient of temperature o models . . . . 74 Comparison and field verification of the results • Computed Versus Measured Refraction Effect . . Achievable accuracy using reciprocal trigonometric height traversing • • . . 83 93 93 95 100 112 115 116 120 . . . . . . . . o o • • . . • o 6o • • • o o • o o • • • • • 0 • • • • • • • • • • • o • • • SIMULATIONS OF REFRACTION ERROR IN TRIGONOMETRIC HEIGHT TRAVERSING • • • • • • • • o . • • . .4. . . • 5. . . . . Temperature gradient • • • • • • • • • • • • . 41 42 44 47 49 52 53 55 56 . . • 59 Description of the Test Areas and Scope of the Tests . . Precision of refraction corrections in reciprocal method • • • • • • • • • .iv - • • • o 0 o 0 • • . . . 59 TEST SURVEYS AT UNB • • • • • • Background of Trigonometric Height Traversing at UNB • • • • • • • • • • • • • • • • • • . . • • • Tests on 29 July 1985 and estimation of standard deviation of vertical angle measurements • • • • • • • • • • • Comments on South-Gym test surveys • • • Tests on 06 August 1985 • • • • • • • . 61 64 66 66 67 . . . Tests on 20 June 1985 • • Tests on 19 July 1985 Tests on 23 and 24 July 1985 • . . • • Achievable Accuracy Using Leap-Frog Trigonometric Height Traversing ••• Precision of refraction correction in leapfrog method • • • • • • • • • • • • . . . . REFRACTION CORRECTION IN TRIGONOMETRIC HEIGHT TRAVERSING • • • • • • • • • • • • • • • • 39 Reciprocal Trigonometric Height Traversing • • • • Formulae of reciprocal trigonometric height traversing • • • • • • • • • • • • • • . Trigonometric height traversing • • • • • • • • Investigation of Temperature Models as Function of Height . . . 61 South-Gym test lines • • • • • • • • • . . . . .

..... .......... ................... . CONCLUSIONS AND RECOMMENDATIONS • Conclusions Recommendations REFERENCES .v - 145 145 149 151 ..Simulation Along a Geodetic Levelling Line on Vancouver Island • • • • • • • • • • • • • 120 Computation of the refraction error in geodetic levelling • • • • • • • • • • • • • • • 122 Refraction error in trigonometric height traversing • • • • • • • • • • • • • .............31 7. ... . 124 Results of simulations • • • • • • • • • • • 125 Simulation of the Refraction Error Using other Values of Temperature Gradient Measurements 127 Simulation on the Test Lines at UNB • • • • 1...

.. .LIST OF TABLES Table ~ 5 . . 73 Mean standard deviations . . . i1 in Table 5.4. • • • • • • • • • • • • • • • • • • • • 93 Discrepancies between the results obtained using trigonometric height traversing and geodetic levelling for BMl to BM2 • • • • • • • • • • • . 75 5. • . The Time Averaged Temperatures on Grass Line 5. .8.14. . . • . Refraction effect [mm] computed using the seven models versus the measured value (BM1-BM2). . 5. 96 .10. • 87 Refraction effect [mm] computed using the seven models versus the measured value (BM3-BM1). .11. Curve fitting with test of the significance of coefficient (Kukkamaki's model. Preliminary test measurements using UNB trigonometric method at South-Gym area from BMl 5. .6. 71 . • . The Time Averaged Temperatures on Gravel Line 5. 88 5.5. 4) . 80 5... . i1 in Table 5. Curve fitting and coefficient of refraction computations (model. • • 81 5. .... . . 78 5.12..4).vi - . . . .13. 86 Refraction effect [mm] computed using the seven models versus the measured value (BM2-BM3).7.3. 72 The Time Averaged Temperatures on Asphalt Line .2. over asphalt) • • • • • • • • • • • • • • • 77 5. . Correlation Coefficients Matrices • 89 5. 5.4. . Curve fitting with the significance of coefficient test (model i3 in Table 5. ..4.1. .. Curve fitting and coefficient of refraction computations (Kukkamaki's model. over asphalt) • • • • • • • • • • • • • . to BM2 5.9. i4 in Table 5. . .

Computed refraction using measured temperature gradient . .16. 98 5. Discrepancies between the results obtained using trigonometric height traversing and geodetic levelling for BM3 to BMl • • . b and H along levelling routes in United States (after Holdahl [1982]) • • • • 130 .1. t-test on the significance of the correlation coefficients • • • . .2. .17.5. Average 6t. Average 6t.19. • • • • • . • • • • • 111 5. b and H along the levelling routes • 128 6. Average 6t.B •• 129 6.vii - . • • • .15. Computed refraction effect versus value for Head-Hall test line • 119 the measured 6. . b and H in Fredericton. Discrepancies between the results obtained using trigonometric height traversing and geodetic levelling for BM2 to BM3 • • • • • • • • • • • • 97 5.18. . • • • • • • • • • • • • • • • • . 99 5. • .3. N.

. 54 4.1. 79 . 2. . 65 5. Refraction effect computed using the seven models versus the measured value for BM1-BM2 • • • • • 90 5. • • . . . Refraction effect computed using the seven models versus the measured value for BM2-BM3 • • . 63 5. • • . . . • .. .2... • • • • • • • . 92 .. • .. • • • • • • • • • • • • .7..6..5. 49 4. • .4.1.3. Standard deviation of refraction correction in leap-frog height traversing as a function of distance.. • .. Profile 4....2.. Refraction effect computed using the seven models versus the measured value for BM3-BM1 .4. •.. 91 5. . Standard deviation of refraction correction in reciprocal height traversing as a function of distance • • • . . . .1.viii - . Ellipsoidal section for reciprocal trigonometric height traversing • • • • • • • • • • • • • • ..4 . • • .. .. . Vertical Refraction Angle . .. Principle of refraction by reflection Refraction effect in a geodetic levelling setup . • • • • • • • • • • • 58 5 . Plan and profiles of South-Gym test lines .1. • • • • • • • • • • .LIST OF FIGURES Figure 1. Plan and profile of Head-Hall test line . Test of the significance of coefficient for models in Table 5. .3. 2 . Refraction Coefficient Contours 5. .1. 42 4. . 3 . . Ellipsoidal section for leap-Frog trigonometric height traversing • . • • • • • . • • 82 5.2. • Methods of trigonometric height traversing • of mean potential temperature 5 10 17 23 e • • 32 3.2. . .

. . . • • . .9. • • • • • • • • . . 135 Accumulation of refraction error in geodetic levelling and trigonometric height traversing along line i3 • • • .1. 5. • • • • • • . . .12. .15.3.3) . 5. . Fluctuations of observed vertical angles • • 107 5. Accumulation of refraction error in geodetic levelling using equations (3. . . 103 5.10. . • • • • • • • • • • 138 6. Fluctuations of point temperature gradient • 106 5. . . 6. • • • • .5. (6.ix - . . 137 6.and fore-sight magnitude of refraction difference • • • • • • • .16. 133 Accumulation of refraction error in geodetic levelling and trigonometric height traversing along line i1 • • • • . • • • • • • . . . • . . . Back. . . 6. . • • • • • • • 136 Accumulation of refraction error in geodetic levelling and trigonometric height traversing along line i4 • • • . 134 Accumulation of refraction error in geodetic levelling and trigonometric height traversing along line t2 • • • . . • . . • • • • • • • 110 5. . • . • • • • • • • • • . . • • • • .8. 139 6. .1) and (6. . . . .7. .2. . . The measured refraction effect [mm]. The discrepancies of height difference determined by trigonometric height traversing and geodetic levelling. Measured refraction error versus the computed value. . • • • . . • . 117 6. between BM2 and BM4 at Head-Hall teat 1 ine. .5. • • • . . . . • • . • • • • 104 5.14. . • • • • . • • • • • • • • • . • . . Computed refraction effect versus the measured value. • .6. • • • . • • • .13. Accumulation of refraction error in geodetic levelling and trigonometric height traversing (line i2) • .4. .11. .19). 114 5. • • • • • • • 100 Measured refraction effect versus the computed value. . 6. • • • • • 108 Linear correlation between the computed and measured refraction error. . Variations of turbulent heat flux along the levelling line +2 • .

9.6.12. Refraction correction for line BM2-BM3 • • 6. Refraction correction for the Head-Hall test line 144 - X - .11. • • • • • • • • • • 140 6.. Refraction correction for line BM3-BM1 • 6.8. 142 143 . Refraction correction for line BM1-BM2 • 141 6.. Accumulation of refraction error in geodetic levelling and trigonometric height traversing (line i4) • • • • • .10.

xi - long days . Wolfgang Faig for his critical review of this thesis and for his guidance and sound advice. to thank Dr. Dr. and continuous encouragement throughout the course of this work were highly appreciated. in the His topic and constructive suggestions guidance. A. . Adam Chrzanowski. I would like to express my sincere thanks to Mr. Yong-Qi Chen from Wuhan Technical University of Surveying and Mapping for his many hours spent in manuscript and for on leave University at the of New discussion and reading the original his constructive criticism while Department Brunswick. In addition. Secord who spent many long hours to discuss. from Poland for his help measurements possible Jarzymowski a visiting scholar in making the meteorological and his assisstance during of field work. James to read and to render my difficult script into a readable form. of Surveying I also wish he was Engineering. I am indebted to Dr. immeasurable support sincere His interest were invaluable.ACKNOWLEDGEMENTS I wish to express my deepest gratitude and thanks to my supervisor. I would like to thank Dr.

P. Thanks are also due to the personnel at the Geodetic Survey of Canada for providing the data used in this study. Kornacki. Mantha. J. Research Council the National and by the . Shi and M. thesis has been financially of Canada. z. Katekyeza are particularly thanked. Romero.xii My thanks for their are extended also to many of generous assistance with my colleaques the field work. them. Among Donkin. C. K. Messrs J. Faig. The work described in this supported by Science and the Geodetic Survey Engineering University of New Brunswick.

A suggested remedy is to shorten the sight length. 1971]. developement years to increase error effects trigonometric height traversing method . because the influence of refraction is proportional to the square of the sight distance levelling.g. 1985]. an accumulation in the order 20 mm per 100 m of height difference can be expected [Bamford. refraction influences systematically because which Along inclined the measurements the horizontal line of sight passes obliquely through different isothermal layers of air. Angus-Leppan. levelling is confined by its terrain.Chapter 1 INTRODUCTION The refractive properties of the atmosphere have placed a limit on the accuracy Geodetic of conventional geodetic levelling. Because intiated in reduce the of these the last few systematic reasons.1 - has been production and by using the as an alternative to . Bamford [1971] recommends sight under lengths. [e. a slow survey procedure horizontal line of sight. This even though restriction makes For precise keeping the length of the slope may allow longer the survey progress even slower and more expensive. certain extreme conditions such as is the long Under easy gradient along railways. 30 m.

2

geodetic

In

levelling.

differences
vertical

in

the

elevations

are

from

determined

measured

the new

state-of-art

theodolites and compact and

accurate EDM

angles and

modern electronic

the

trigonometric methods,

distances using

(Electromagnetic Distance Measuring) instrumentation.
Two

types of

trigonometric height

traversing can

be

distinguished (see Figure 1.1):
1.

Simultaneous

reciprocal:

the

zenith

angles

are

measured in both directions simultaneously.
2.

Leap-frog:

the

instrument is set up midway

between

two target-reflector stations.
At the

University of New

Brunswick,

the

leap-frog method

with elevated multiple targets was developed and tested from
1981 to 1985.

This variant of the

leap-frog trigonometric

height traversing is called the "UNB-method".
In

the

observations

trigonometric
are

affected

methods,
by

the

vertical

long-term

angle

temperature

gradient variations which cause vertical displacement of the
target

image.

fluctuations

cause

dancing).
refraction

The

As

in

can

be

trigonometric

short-term
the

blurring

geodetic
the

methods,

main

temperature
of

the

levelling,
source

though its

gradient

image
the

of

(image

atmospheric

error

systematic

in

effect

the
is

expected to be much smaller than in geodetic levelling.
Many
theoretical

authors
aspects

have
of

investigated

both

refraction

error

practical
in

and

geodetic

3

levelling [e.g.
1979b,l980].
for the
have

Kukkamaki 1938, Holdahl 1981,

These investigations have arrived at formulae

refraction correction,

shown

Angus-Leppan

that

similar and are

the results

and practical
from

various

close to actual values

experiments
formulae

[Angus-Leppan 1984,

Heer and Niemeier [1985), Banger 1982, Heroux et al.
These formulae

are

are generally based on

1985].

estimated (modelled)

or measured temperature gradients.
In

the trigonometric

correction

can

compatiable

methods,

be derived

with

the

if

a similar

the

length of

lengths of

sight

levelling, i.e. not exceeding 100 m.
are longer,

then the correction

refraction

used

sight
in

are

geodetic

If the lines of sight

for refraction

becomes a

more complicated task.

On the other hand,

as

shown

the

refraction

in

this

thesis,

influence

trigonometric

height

traversing becomes

large extent,

if the lines of

of

it will be

randomized

sight are short,

in

to

i.e.

a

less

than 100 m.
This thesis investigates the
errors
methods

in

the

with

traversing.
1.

To

optical
more

height

emphasis on

difference
the

determination

trigonometric

height

The objectives can be summarized as follows:
determine

temperature profile
basis

propagation of refraction

of several

an

optimal

up to 4 m
long

term

typical ground surfaces (gravel,

model

for

above the ground
test surveys

the
on the

over

grass and asphalt)

three
and

4
profiles;
these

to investigate the

surfaces and

influence of refraction on

to compare

the measured

refraction

effect against the computed refraction correction.
2.

To

the

develop

available

new models
models

and to

such as

compare them against

Kukkamaki's

and

Heer's

temperature functions.
3.

To confirm in practice

the designed precision of the

UNB-method under controlled field
the

understanding

of

the

conditions,

refraction

to add to

effect

and

to

compare the UNB-method against the reciprocal method with
regard to the influence of refraction.
4.

To

simulate

the

refraction

effect

in

the

trigonometric methods along a line of geodetic levelling,
to assess the dependence of
profile

and to

trigonometric

compare

the refraction errors on the

the

methods versus

refraction effect
the

in

the

refraction effect

in

geodetic levelling.
An overview of the solutions
in optical height difference

to the refraction problem

determination methods is given

in Chapters 2, 3 and 4. Chapter 2 reviews the method already
developed for

the refraction correction

computations based

on the evaluation of the temperature gradient, the so called
"meteorological method" and three other approaches namely:
1. the dispersion (the two wavelength system),
2. the variance of the angle-of-arrival, and
3. the refraction by reflection.

of results. of the effect a new in the Chapter 5 deals and discussion simulations is given in Chapter 6. Figure 1. approach in solving Chapter 4. their analysis. in details the meteorological approach in the optical height difference determination methods. Chapters 3 and better performance [Brunner.5 The development of these methods depends on further advances in technology than the Angus-Leppan. 4 review 1979a. and they are promising a meteorological method 1983].1: Methods of trigonometric heioht traversing (a) leap-frog (b) reciprocal . for the refraction reciprocal method proposed by the with the 1985 test surveys. The outcome also summarizes author.

1979a. Determination of derived from the the vertical variance of angle of the refraction angle-of-arrival fluctuations. These methods are [e. as well as in geodetic levelling. This approach to refraction angle computation is referred the to. Determination of the vertical angle of refraction using using the dispersive property of the atmosphere.6 - . that has been applied in temperature gradients the geodetic levelling of air are The most popular method which can be obtained direct measurements is the is based on either through temperature at different heights or by modelling the atmosphere using the theories of atmospheric physics. discussed in the following various three other literature. vertical here. Determination of the vertical refraction correction using the "reflection method" [Angus-Leppan. Several solutions suggested by different researchers. 1983]. as the meteorological method. 2. effect 6f atmospheric refraction.g. Besides the approaches are above method. Angus-Leppan. 1983]: 1.Chapter 2 A REVIEW OF METHODS FOR THE DETERMINATION OF THE REFRACTION CORRECTION The most significant source of error in trigonometric height traversing. 3. . Brunner.

25 (1 + a t) -8 10 (2.03% of 1013.1 Refractive index of air Determination of using the errors due to meteorological method refraction properties of the atmospheric refraction requires knowledge atmosphere.1). index n of a medium is defined The depends on humidity.-------.1.25 mb and and is given by with a [e.7 This chapter summarizes meteorological approach the above The methods. 1971]: 1 p 4. formula was pressure adopted by and the International Association of Geodesy in terms of temperature t [°C].2 e (n. .g. which is [Bamford. 1969] with a pressure content of 0. 2.--------(1 + a t) 1013. will be discussed with more detail in Chapters 3 and 4. refractive as the ratio of the velocity of light in a vacuum. co.00366 is the thermal expansion of air and no is the refractive index of light in standard air at a temperature 0 oc carbon dioxide Hotine. pressure p [mb] and partial water vapour pressure e [mb]. of the of light in the Variation in the refractive index of air the variation of In 1960 a temperature.1) = (no . to the velocity c medium: n = C 0 /C.1) where a = 1/273 = 0.1 Determination of the Vertical Refraction Angle hY the Meteorological Approach 2.---------.

1 e • 10 (2.--M (2.00366 t The vertical gradient of the refractive index expressed by differentiating equation can be (2.4) the absolute temperature [.iq. A = 0.6288 A -4 A is the wavelength [pm] of monochromatic light in in which a vacuum.g.25 -8 (2. of ( dp/dz) The In the second (de/dz) and vertical in normal it can be gradient of pressure is approximated by [e.00366 into equation -6 (n - (2. Bomford. than 2% 1971].604 -2 + 1. T is term.14 e -----------T ) ---dz l 10 -6 (2..56 pm for Substituting an average value of white light and a = 0.14 condition is less neglected [Bomford.3) with respect to z dn dz 78. 1 --------------1 + 0.5) T .3) 1 + 0.0. is negligible and 0.9 = -----T [ p ( where. 1971] dp dz g p = .14 e. 0.8 1) 10 (no - 6 = 287.00366 t 1) = 293 x 10 4.1) yields p 1013.0136 ). dp de .14 ( ) dz - - dz dT 0.2) + 0.

6) shows that under such conditions. density is independent of Equation (2.2 Anqle of refraction error Considering Figure 2.7) .9 p = ------- dT 0. s s = z = s ~ dn/dz (S - X) the chord length AB.6) dz atmosphere. The the lapse rate rate of [Shaw and decrease Smietana. ) 10 -6 (2. the zenith angle.0342 K/m is necessary to compensate for the decrease in atmospheric pressure with height. In this equation..1 the vertical refraction angle is the angle between optical path AB.78. g/M has This value is known as the numerical value of 0.. If the chord dn/dz and the is known at tangent to oo the all points along AB. and dx (2.0342 + ( 2 T a homogeneous height. 1982].1... 2.9 where g is the gravitational acceleration and M is the specific gas constant for dry air. a lapse rate of -0. of temperature simplified formula for vertical with gradient of the refractive index is then given by dn dz In .0342 K/m • the autoconvective Lapse rate is height. the vertical refraction angle can be calculated from the eikonal (optical path length) equation [Brunner and Angus-Leppan 1976] 00 sin z ------= where.

.10 x = the distance along the chord.9 p ------2 { 0. gives -6 s 10 00 ~ = ------ s ( dT -78.1.8) From Figure 2 .7) asaumin<J sin Z = 1.6) into and (2.0342 + ) dz T l (S - X) dx {2._A p \ \ ---~ Figure 2. z A - -. the refraction correction is c = R 00 • s (2.9) .1: Substituting Vertical Refraction Ancle equation {2.

11) and dT 0.10) (A) R .9) c = in R = 6371000 m gives 502.(dn/dz) .--R In the simple (2.11) into (2.11) I p R Substituting equation (2.14) . may be also light the calculated path and in terms the of the coefficient of The curvature is given by 1/p = .6) assuming = -------2 T ( - (2.11 The correction curvature of refraction. the refraction angle is given by s (A) = ----2 p Substituting p from equation (2.0342 + Then equation (2.10) The coefficient of refraction is defined as the ratio of the radius of the earth R to the radius of the curvature of the light path = k (2.s 1 = .13) of refraction is constant along the line of sight AB.14). sin Z (2.12) ) dz can be written as s ~ k .7 p k and (2. gives (2. (S-x) dx case when the coefficient (2.

15) 2 R Then.----.12 s k 00 = ----- (2.16) k 2 R Which means that the refraction error is a function of the square of the sight length. Equation (2.2). shows that. dT/dz. has The temperature gradient can be obtained either by observing the temperature of air at different heights above the ground and then fitting these observed values to a temperature function (see section 3. Please refer to Chapters 3 and 4 for a detailed discussion refraction correction using meteorological method.• R (2. along the line of sight. one needs in to order to know gradient. to be known as a function the Thus.8) refraction angle. the refraction correction for a circular refraction path (constant k along the line of sight) is s 2 c = . compute the temperature dT/dz of height above the ground. of the . or by modelling in terms of sensible heat flux and some other meteorological parameters •.

two time delay photomultiplier signals related to the red and blue arriving refracted beams. 1984] oo = V • Aoo in which (2. 0. on this measurement.9 .18) -1 where w is the wavenumber [pm ] of the For red and blue colours with wavelengths of 633 nm and 442 nm respectively.17) V = N/AN. and N is the refractivity at standard values of pressure and temperature. 1984] as N = [ 24060. air at 288.w 159. by a known oo [Williams and Kahmen. This Aoo has to be multiplied obtain the angle of refraction Aoo.000004 while N for the .+ 2 130 .2 Determination of the Vertical Refraction Anale Usina Lasers of Different Wavelengths The dispersive property to determine the angle of fact that of the atmosphere can refraction. 10 -6 (2.15 K and 1013 mb The value of N for dry is given by [Williams and Kahmen.97 -----------2 38. factor V to A between the arriving beams wavelengths by considering the difference between be used Based the angle between arriving beams.w l .3 83. AN is about light in vacuo. blue light is bent when propagating dispersometer measures the angle of two different or phase In this approach the slightly more than through a dispersive red light medium is used.13 2. can be computed.4213 + ---------.

14 mean of the two wavelengths is around that the value of V is close refraction angle can be found 0.5" be expected in for the according to Brunner vertical refraction the near future under favourable observation conditions using the dispersion method.000279 which means The variance to 70. of Aoo has to be of 0. [1979a].19) Aoo 00 According to equation (2. performance of this method.17) 2 a = v 2 2 a (2. by applying of the the propagation law of variances to equation (2. the precision than the an accuracy angle can required precision of the This requirement puts a limit on the however. a number of tests were Autumn of 1978 and January of The tests were made over a 4 km line using two bench marks with a known height difference. carried out in the Spring and 1980 by Williams [1981]. . along with a dispersometer to and its corresponding refraction On average. Using this dispersion method.19) about 70 times higher angle of refraction oo.9" in 1980. A T3 theodolite was used measure the vertical angle angle.6" in 1978 and by about +0. the observed refraction effect deviated from the estimated value by about -1.

~. Brunner [1980] of the angle-of-arrival fluctuations as the variance of the image fluctuations. the image of a suitable light source can be the telescope using a continuously recorded in photo detector connected to a data logger [Brunner. of ~ Angle of Refraction Derived from Angle-of-Arrival Fluctuations the Variance of the propagation in the was first proposed by Brunner This method gives the angle of refraction in terms variance of angle-of-arrival the fluctuations ~ 2 caused by atmospheric turbulence.3 A method baaed on studies of light atmosphere (turbulent medium) (1979a]. refers to the variance 1970]. of the formula for the angle of standard and some deviation the meteorological parameters. Since this formula given here. ~2 could be dancing. The angle-of-arrival fluctuations correspond to the fluctuations of the normal to the wavefront. inferred estimated by telescope [Brunner. 1980. The major advantage of this method over established methods is that the computed angle of refraction . of 1979b. needs a detailed background.15 2. Brunner [l979a. it is not 1982. spread observations of the image through the For a precise determination of the mean and the variance of the angle-of-arrival. arriving at the telescope [Lawrence and Strohbehn. from the visual 1979a]. 1984] the other provides a complete treatment of the subject. Brunner [1979a] has derived a refraction in terms angle-of-arrival. 1980].

be also refracted to final image will be When the coefficient along the The the same direction seen lower or higher If image of line of than the of refraction sight (circular point A' and the cross hair will be C is the magnitude of refraction affecting the levelling observations. reflected ray will and the cross-hair at happens to point A'. sight is refracted. the incident angle longer a normal but makes an angle. be constant refraction path) separated by 4C. the length of the refraction effect is proportional to the square of the distance for a circular refraction path. same elevation as the cross-hair of the level instrument. in refraction varies along the line of the separation general the coefficient of of sight and the magnitude could be smaller or larger than 4C which . 2. However. The factor of 4 is not unexpected since the the line and as it was ray has traversed twice shown before. where the line of to the mirror is no to the normal.16 is a better representation of the whole optical path. it is derived from measurements along the actual since line of sight [Brunner. When there the reflected is no refraction. Q.4 Determination of the Vertical Refraction Correction Using the Reflection Method Figure 2. then.2 illustrates the principle of the reflection method with an exaggerated scale The can be target a point in the vertical light source with angle the Q. target would be seen on the cross-hair. 1980].

17
makes the method inaccurate.

This is the major drawback of

the method.

'

A

T
c

Level

i
Vertical
Mirror

Target A

Figure 2.2:

2.5

Principle of refraction by reflection (after
Angus-Leppan [1985])

Comments on the Discussed Methods
.Among

the

discussion,
has been

four

approaches considered

the

above

the meteorological method is the only one which

developed and applied

in practice

corrections in geodetic levelling.
variance of

in

for refraction

The dispersion and the

the angle-of-arrival methods are

promising and

18
they may show a better performance

in the near future since

they both rely on further advances in technology.
Because the

angle between the

two receiving

very small in the dispersion method,
a

very high

performance

accuracy.
of

the

This

dispersion angle

dispersometer.

demands on
Although

the

recent

to measure the differential

with a good precision,

have shown that atmospheric
limitations and

it must be measured to

makes severe

technology has made it possible

beams is

test measurements

turbulence imposes considerable

good measurements

are only

possible under

favourable conditions.
The variance of
developement

stage and

precise measurement
still to

be built.

useful approach,
of

the angle-of-arrival method is

refraction

the

of the
But it

instrumentation
fluctuations of
has the

in its

for the

very

the image

has

potential of

being a

since it takes into account the variations
effect

due

to

the

refractive

index

fluctuations along the actual line of sight.
The main disadvantage of the
for

a non-circular

possible to

refracted

estimate the total

reflection method is that

line of

sight,

refraction effect

it is

not

and some

residuals remain in the results of measurements.
Further discussion

in this thesis

is based

the application of the meteorological method.

mainly on

Chapter 3
REFRACTION CORRECTION IN GEODETIC LEVELLING
USING THE METEOROLOGICAL METHOD
Geodetic
principle,

levelling,

is

though

an inherently

which has remained

precise measurement

Over a long distance

great

of

systematic error

refraction.

with

in each set-up that

These

systemtic errors which

both

height

a

very

small

accumulates steadily.

errors are due to
are

turn of

its results depend on a

instrument stations

The most troublesome

in

approach

practically unchanged since the

the century.
number

simple

remarkably

rod calibration and
gradient

may not be detected

correlated

in loop closure

analysis.
Error in

rod calibration can

be controlled

combination of field and laboratory procedures.

through a
Refraction

error is less easily controlled and is more complex because,
in

addition to

height

temperature gradients

difference,
and the

it

square of

is

a function

the sight

of

length

[Vanicek et al., 1980].
In this

chapter,

methods of refraction

geodetic levelling derived

correction in

from meteorological measurements

are discussed.

- 19 -

above the z [m]. and .1 Refraction Correction Based 2n Direct Measurement of Temoerature Gradient The error first important problem was step taken in solving the in 1896 by refraction Lallemand when he suggested a logarithmic function for temperature. since greater errors up to a few some the were never applied in decades ago there involved such as of theoretical Lallemand's investigations in geodetic refraction practice. 2. in terms of height. b and c are constants for any instant. error was 1961] in applied in with research respect horizontal angle to work the by lateral observation on a sideward slope. like errors in were other poorly designed rods and instruments. corresponding the following assumptions: 1. ground [Angus-Leppan. About forty 1939b] formulated refraction years after Lallemand. the refraction coefficient of temperature since the effect of air depends mainly on humidity is negligibly small for optical propagation. Lallemand's Kukkamaki refraction model [1939a.20 3. isothermal surfaces are parallel to the ground. t [°C]. In geodetic levelling. Heer [1983] has shown that Lallemand's recently model proposed works almost models.1) where a. 1984] t = a + b ln(z + c) (3. his temperature correction which Kukkamaki model and was based on [1938.

with t i c c 2 1 c c 3 2 z .21 3. then.z At = ln 3 2 c c z .z =a + b c z i ) =b ( z .z 2 z = ln 3 z - 2 c 1 1 c I z I 1 z 1 c 2 . The constant a does refractive coefficient is a not play any role since the function of the temperature gradient and constants b and c can be easily computed using three different heights temperature sensors at arranged such that z = z 1 z I Z 1 2 At = 2 - t = t 2 3 3 b ( 1 - t = t 2 1 At 2 . b and c are constants for any instant and vary with time. c 2 ln At 1 c c z . The Kukkamaki temperature model is an exponential function of height =a t c (3.2) + b z Where t [°C] is the temperature at height z (m] above the ground and a. the terrain slope is uniform in a single set-up of the instrument.z ) so.

5) 1 Differentiating ex with respect to the refractive index n and results in 1 = (dnln) cot a dcx 1 (3. 1938] = constant. Priestley.g.t 3 c I I t:.6) 1 .22 c z Replacing c = ln ( t:.t 1 I < z 2 ) 2 I ln ( z 1 2 c I z z ) I yields 1 (3.g.z 2 c z by (3. the line of sight starts out horizontally at the instrument while making an angle a 1 with the assumed surfaces of constant refractive index that are parallel to the around. This makes a ground surface. when only two temperature sensors are used. 1959 Holdahl [1981] agrees with the in the lower atmosphere and Webb. Kukkamaki's equation f2£ aeodetic levelling refraction correction 3 .1 In geodetic levelling. Kukkamaki.4) ) 1 In a simple case. 1964].3) 1 c z . the following relation Then.1. an average value can be recommends a value of -113 used for c. for c which theories of turbulent heat transfer [e. 1 also the slope of the holds [e. n cos a (3.t 2 c b= t:.

The angle oo is given by integration along the line of sight (1. oo (see Figure 3. levelling the line of the line of sight inclines at an angle.1: Due to Refraction set-up the change sight at a point. at a geodetic index along distance x.7) 1 0 where n and n are the refractive indices at the instrument 0 . dn (3. effect in a in refractive P.1).) n 1 n n = f cot ex .23 Figure 3.

t is the temperature [°C] and pis the pressure [mb]. If dt = (t - t ) and dn = (n .00366 -6 10 2 • dt (3.[ 0.3) t after neglecting the e term gives dn = p 293 X 0. with sufficient accuracy [Kukkamaki.10) 1013. the 1 angle of the slope of the ground surface. and at the point P.7) we have oo = .20 ) ] --------1013.n )/n 1 0 shows (3. gives .4) into equation (3.25 10 • dt (3.cot a Equation (3. 1938] oo = .cot a (3.24 respectively. From equation (3.8) ln (n/n ) 1 0 or.11) and assuming dt ~At.931 .25 ( 1 + 0.9) differences of 0 that oo the two refractive is a function indices and of of the a .n ) 0 to be infinitely small are considered 0 increments and substituting equation {3.00366 t) or. with sufficient accuracy [Kukkamaki.0. Differentiation of equation with respect to (2.0064 ( t .9) (n .11) where. 1938] -6 p dn .

and z =the instrument height [m]. z = the rod reading [m].16) i i 0 Assuming c 2 cot ex = 1. 1938] .13) --~------. d n (z n a c z 1 ) dz (3.n c c = d • b • { z . the refraction correction in the back- sight is found to be [Kukkamaki.25 where.0.0064 (t -20)] (3.931 . cot ex ( z 1 n c . cot a 1 then C1 = b . d = .10 -6 p [ 0.1 z ) cot ex x = (z - i (3.15) 1 and from equation (3.14) 0 From Figure 3.15) dx = dz .000.z ) dx i (3.Z ) (3. i In Figure 3.12) i 0 in which.1 the vertical refraction effect at a distance x is given by integrating ro along the line of sight c d • b X Cl = ~ ro dx = . 1013.25 n .

models suggested g~ound based is assumed.b.cot ex simila~ exp~ession C2 = b c + 1 1 A C+l z . 1 d. 1939b]. Z and Z b C a~e backward and forward rod readings [m).d. [Kukkamaki.z z + ----b c + 1 i can be obtained for the fore-sight cot 2a 2 • d • b • [ --~-­ C+1 z c + 1 The total refraction correction f C . b • [ R C+l c + 1 ( z b (3.26 1 2 C1 = .5 ern. [ C C . f is the refraction error [m] and =a a R The temperature profile adopted on direct temperature measurement the surface.C1 C R 2 c = cot a . on by researchers di~ect such as There temperature Garfinkel .z z i fo~ C + ----f c + 1 one instrument set-up is given by = C2 .19) where. 2 by Kukkamaki was based at different heights from studies temperatures measured by Best in 1935 are =a 1 utilized the at heights of 2. His empirical 30 em and 120 em above the some other measurements.

Webb [1969] was the first who explained at a conference in 1968 that it could be feasible to evaluate an approximate gradient vertical relationship of with mean other through temperature meteorological its parameters. Potential Temperature a Potential temperature is defined as the temperature that a body of dry air would take if brought adiabatically (with no exchange of heat) to a standard pressure of 1000 mb . was conducted at to Niemeier [1985]. 3.1 Review of the meteorological parameters 1.27 [1979] and [1985] Heer and have given Kukkamaki's model.2.g. There is extensive literature available in this field and for comprehensive treatment one can refer to Webb [1984]. The Angus-Leppan. Subsequently a number of papers were written on this subject [e. a summary of eight In the last few years. the University of New the development of new models which Heer and Niemeier models including a research study Brunswick that lead are discussed in Chapter 5. following section is a review of the 1971]. 3.2 Refraction Correction Formulated in Terms of Sensible Heat Flux The second group of models is based on the laws of atmospheric physics. 1971 and Angus-Leppan and Webb. meteorological parameters.

0098 [Kim] is the adiabatic lapse rate. T [K]. and is given by u* = k U I ln ( Zv I Zr ) (3. U. 1971]. For more details see e.20) [K] shows that for pressure near the standard the difference between the potential temperature and the absolute temperature is very small.28 [Angus-Leppan and Webb. The gradients of absolute and potential temperature are related by d91dz = dTidz + f (3. (3. This roughness length.21) [Kim] Where f = 0. Priestly [1959]. is the height at which the wind velocity is equal to zero.g. at a pressure.g.22} [mls] where k is von Karman's constant with numerical value 0.4. mean height of the this value is trees [Webb. between 6% 1984]. Zr is about 10% of the grass height. and Zr is Zr. For grassland. 1977] e = T ( 10001P ) 0. the roughness length. 2. to 9% and for of the pine forests. U is measured at height Zv.20} (1000mb). Fraser. Potential temperature can be related to the absolute temperature. Friction velocity u* Friction velocity is a reference velocity which is related to the mean wind speed.286 Equation (3. p [mb]. using Poisson's equation [e. .

G.24) + Ld . net radiation. Sensible heat flux H Sensible heat flux forms one element of equation at the surface of the other elements. Sd is not (0.Su (3. Sd = the downward short-wave radiation flux 3 pm) from sun and sky. with A. Su = the short-wave radiation reflected from the surface.29 3. energy balance equation. Munn. being the latent heat of the vapourization of water and E is the rate evaporation.g. the energy balance earth where it combines with O. radiation flux emitted by surface. 1966) H 2 = 0 . and evaporation flux.A. and A. of . Ld = the downward long-wave radiation flux (4 pm) received by Lu = the upward the long-wave to 60 earth from the atmosphere. heat flux into According to the the sensible heat flux is given by [e.E [ W/m (3. AE.E = the latent heat flux of evaporation or condensa2 tion in [W/m ].Lu where.G .3 to present at night. namely: the ground.23) ] in which Q = Sd . 2 G = the heat flux into ground (W/m ].

1946]. which a very governs the regarding the the temperature degree significant element of thermal [Obukhov. where g is the acceleration due to gravity Three . time after sunrise Under this with height occurs when = 0. occurs This condition appears when Ri > 0 when the surface is cooled. Priestley.30 3.25) 2 [m/sec ].2. It a short time Ri and condition the distribution of is adiabatic (no exchange of .g. Neutral stratification appears a short before sunset. stability At each height condition. There exists one governing nondimensional parameter which is height dependent. temperature heat). 2. This it indicates the thermal parameter is the well known Richardson number Ri that has the following appearance [e. Under this condition. 1959] Ri = (g • d9Jdz ) I (9 • ( dU/dz ) 2 ) (3. the thermal buoyancy forces suppress the turbulence and cause the downward transfer of heat.2 Thermal stability oarameter According to meteorological literature distribution of the average gradient parameter stability is wind velocity. regimes of thermal stability be can distinguished: Stable 1. stratification (inversion).

2. 1964] (3.p = 1200 (k = 0.26) Ri = z I L where L is the Obukhov scaling length [m].26). Using the above equations the following expression can be found for 3 c U* L e p p L (3. another form of stability parameter. Ri < 0 Unstable stratification occurs when It appears typically on a clear heated by day when the incoming solar radiation.27) shows fluxes and constants which as constant may be regarded as throughout the can be surface a characteristic height which determines the thermal structure of the surface layer.3 and k is density von of the air Karman's (C p constant . ground is the heat is being of air and the turbulence will tend to be increased by thermal bouyancy forces. carried upwards by the current (lapse). Profile of mean ootential temperature oradients Accprding to equation (3.27) = g k where C and p p H are respectively the specific heat at cons- tant pressure and the 3 [J/K m ]). .31 3. 3.4). ~or conditions near to neutral when the Ri value is small [Webb. that L is a function of momentarily considered layer. then L z/L can be regarded as Equation (3.

g. In this Figure. (b) STABLE (d) UNSTABLE upper reg1on intermittently ~:: 0 1 dZ middle region ~ <L z-t.2: Profile of mean potential temperature e (a) unstable and (b) stable conditions. e* = H 1 ( C p p u* ). intervals of several minutes in (a) or between 30-min runs in (b). Broken lines indicate variablity over time.Je r --e . 1964]./3 az 0·03 0·03 lm.32 In other words. reg1on ae C( az z-1 ln z f - 8 Figure 3. and the whole structure of the behavior expands contracts in height according to the magnitude of L [e. Webb.. (after Webb [1984]). . and a = 5.

Figure 3.33 a. buoyancy effects) range of or free (interaction of convection (in caused by density differences within the The potential temperature gradient profile in this region is given by dS dz -l l H = -------c p p 2!3 f/3 [ a g -4/3 z (3.03ILI < z region < ILl. temperature is found to be inversely proportional to height d6 dz =- The l H ------------c p k U* p middle 0.03 IL I.28) • z extends over a height Heat transfer in this region is mostly kind of composite convection wind and thermal calm conditions) moving air. according to behavior and These three height L rather than absolute terms. Unstable regime. governed by a l -1 (3.2a shows the three regions of different physical the profile of e in the ranges are defined unstable case. the gradient to of potential z = 0. region extends in height The lower In this region.29) . conditions the mean Within potential the unstable temperature turbulent profile takes a different form in three different height ranges.

34 This equation is uniquely dependent independent of friction velocity u* on H and z and which means the middle region is independent of wind speed. Within of the the following profile forms are found [Webb.81 [m s -2 ]. by substituting p and C p p = 1200 -1 [j K -3 m ] then. e and C p g = 9.29) can be simplified approximate values for g. the middle region starts at a height 0.0. .75 m to 1. b. e = 290 [K].35 m above the ground. Stable potential conditions Figure temperature stable regime. 1969] 3.2b in stable shows the profile conditions. On a typical clear day when the between 25 m and 45 m. 2/3 d9 dz = .0274 H The upper region -4/3 (3. Equation (3. of geodetic Obukhov length varies optical measurement This means that in most the critical part of the sight-line lies within this region.30) z begins at where the gradient of e is a height approaching ILl often averaging near zero over a period of several minutes.

e. s B [m] (3. and the with increasing height. the be simply carried out the Kukkamaki formula. The Anqus-Leppan equation 12£ refraction correction 3.33) .32) In the lower region ( z < L ) the gradient changes rapidly with height for very small z. The refraction effect on a back-sight in the unstable case is given by C1 = 10 -6 2 p I T 2 .31) and d9 dz 6 H [ c----------- =- u* k p p l . close to the surface. equation (3.4 Once the temperature gradient refraction correction computation can by using some equation similar to is determined.2. -1 for z > L z (3.19). gradient dependence on height becomes weaker.35 [ d9 dz = - [ H ----------p c U* k p 1 5 z 1 + ----L l . z-1 <L for z (3. i.

precipitation.z ) b i and s is the length of line of sight [m] and the rest of the variables have back rod already been reading by forward defined above. first presented by Angus-Leppan This equation was [1979a]. However.5 Holdahl historical [1979] levelling measurements.36 in which. B = z 2/3 1 . 1981] differences the ground by . H.3 H -1/3 2/3 z + 2 • i i 2/3 z z . cloud cover and ground reflectivity from many locations across the United heat flux between can be used to two heights say States.3 b b 2 (z . he suggested [Angus-Leppan. He developed a method observations was meteorological parameters able for without obtained to model for estimation correcting the Llt required of sensible heat flux. because data for estimating H for stable and neutral conditions is not yet adequate. 1980].2. obtain Z and The estimated sensible temperature z above integrating equation (3. similar A expression was given by him for the stable and neutral cases [Angus-Leppan. Replacing the reading in this equation will give the refraction effect on fore-sight.3. Investigation gy Holdahl 3.29) [Holdahl. 1984] that further investigation is needed. by using the historical records of solar radiation.

by letting is at [Holdahl.5 m. be neglected because AZ Then.0098 f z ).34). 1981] 2 b H = 3 T 1 c = .34) p where = 0. suggested by equation (3.34) is compatible in equation applied for (3.19) Kukkamaki can be correction computation with At In addition.f AZ 1 p ) g (3. refraction obtained by equation (3. is the adiabatic lapse rate. by the National During when the condition is near neutral. cover (3. of course of the The sun to the correction have traditionally been levelling United factor States. (3. AZ=(Z 2 and T is the absolute temperature 1 In equation [K] of air.37 2 H At = t .t 2 1 T (z = 3 -1/3 z 2 2 (C -1/3 ) .--3 and 2 p ) g (C p it can be with seen that At from equation the form Hence. the At can also be affected by wind and the influence is taken into . factor is based sun correction on sun codes which recorded during the Geodetic Survey transitional stage.4).34) the adiabatic lapse rate has very little influence on the estimated At and it can the most 2. Holdahl takes into account the effect of cloud by multiplying a predicted temperature differences.

3. Whalen [1981] approach against Holdahl's net reduction an in refraction method and error of least 70% is achieved using either of the methods. Due to the direct large fluctuations measurement of over of a few minutes a route of precise these temperature gradients levelling done the time the to be in every carried out meteorological levelling.3 Comments on the Meteorological Methods The two meteorological methods for determination of the refraction correction based on either measured or modeled temperature gradient were reviewed in this chapter. at .38 account by considering another code for wind similar to the one for sun. correcting For example. the results of either may be geodetic levelling measurements. The correcting the in the corresponding period of temperature The second approach tends to smooth fluctuations which can be considered as advantage of this method over the direct approach. compared Kukkamaki's reported that a mean of can be used for gradients measurements. temperature gradient has a period station along in satisfactory for However. out temperature.

X ) 2 1 2 3 l + •••• 6 n •••• + 2 [ k (S .Chapter 4 REFRACTION CORRECTION IN TRIGONOMETRIC HEIGHT TRAVERSING A numerical integration of equation (2.X ) + k (S .39 - (S n+1 -xnll (4.X n ) + k n-1 . . .13) is I = k (S - X) dx S 1 = 2 [ k 1 s k (S 2 + - X ) 1 l 6 2 + ---2 [ k (S . s S 2 3 n with corresponding refraction coefficients k 1 k 1 1 •••••• k k 2 3 n then the integration part of equation (2.13) the magnitude of refraction correction. .1) . out using the will give This can be carried trapezoidal rule by dividing the line into n sections of length S 1 S 1 1 ••.

trigonometric height traversing is The line longer than in geodetic levelling and one cannot assume that the terrain slope is uniform.1). X n = refraction be calculated + S S 1 n = S coefficients along the by using terms of the vertical + ••• + S 2 equation (2.40 where = X 1 S X I 1 2 = S 1 + S 2 In equation (4. which are equation (3. line of sight can which is in . in terms of sensible b) from equations heat flux and for the case of geodetic levelling of sight in or are dicussed in Chapter 3.. terrain. The temperature gradient at a height z above the ground can be either determined: a) from a temperature function such as Kukkamaki1 s model.. choose practice. On the other hand. model and the refraction for . meteorological information. correction is needed. a trigonometric meteorological possible during profile of the when the refraction the period location height measurements when the To as vertical According to this gathered the temperature the coefficient of the points along the line can be computed. be carried overcome characteristic traversing frequent as the meteorological measurement cannot out in more than this and one location in problem. for make a one may up of set the angle observations take place.2).12) gradient of temperature. . .

41 The following sections discuss the required temperature gradient accuracy correcting and refraction the difficulties in reciprocal involved and for leap-frog trigonometric height traversing. as or impossible. very accurate temperature gradient is a needed to compute the refraction error. For the sake of height difference reviewed below. solution to the refraction problem. and in optimal conditions of overcast and mild wind speed. error analysis. Usually a combination of asphalt at the centre of the road. computation in the formulation the reciprocal of method is . in reciprocal trigonometric the refraction effect height traversing is more or less symmetrical. A simultaneous reciprocal vertical angle observation is considered by many researchers as the only reliable. gravel at the These make the using equation (2. the refraction effect is minimal on such terrain. only at one point if the of the it will be shown below. But terrain changes in slope and in texture of the surface.9) meteorological data line of sight. side and vegetation come into evaluation of the refraction error very difficult is gathered In addition.1 Reciprocal Trigonometric Height Traversing On a moderately uniform terrain. 4. yet only partial. effect.

1 Assuming a circular refracted path AB. in term of the angle between centre of the earth.1.1 the angle V is . the refraction angle. v.IB I I I ELLIPSOID I I I I I 1P IB ~p..2) 2 A and from Figure 4. I I \ I I \ \ \ \ \ \I Figure 4. A and B subtended at the and the refraction coefficient at point A is given by v I k w A (4.1: Ellipsoidal section for reciprocal trigonometric height traversing Formulae of reciprocal trigonometric heiqht traversing 4 . w.

the ellipsoidal height difference from·A to B is given by Brunner [1975a] as ~h AB = S cos S sin Z A ( e A (4.4) A A which is the same as equation (2. the radius of refraction.oo A A = the deflection of the vertical at point A.1.5) gives 1 ~h = S cos AB z .5) . e where z - A Subsitution of and oo V from equations (4.3) and (4.7) . the earth and In this equation is k the R coefficient is of Considering Figure 4.6) A expression can be written for the height difference from B to A = s cos ~h 1 z BA 2 R B S sin 2 (s sin z ( B -e ) B z ) B (1 .4) A into (4.V/2) .15).k ) B (4.--A 2 R z S sin similar z ) 2 (1 .k ) A A ( e ) A A (S sin (4.3) A then = 00 A s sin k 2 R z (4.43 s v = sin z R (4.

8) the law of variance to equation [Brunner. 4.44 sin z Assuming s Ah = 2 z ( sin D = S equation.k A B is treated as a random error and cos z A is assumed to be equal to -cos z for the B purpose of error . 1975a] 2 a = of propagation (COS 2 z 2 a s ) A 1 + --2 2 D 2 a z 1 + ------2 16 4 D Ak 2 R (4. In this correction due to the effect of the deviation neglected for lengths of sight in moderately accuracy [e.g.k A B is The third term is less than k ( 4 R of the vertical which can be of 2 1 - + E A z sin ) B E A 2 where z .8) term and loss of - ) B ) is the horizontal 500 m .1. D without any gives second (4.9) where Ak = k .9 and B z cos ( and combining 4.cos A s 4. hilly topography Rueger and Brunner.8 sin z '::::= A the distance. the variance of a measured height difference can be found by applying (4.2 Achievable accuracy usino reciprocal trigonometric height traversing Ignoring the effect of the deviations of the vertical. 1982]. the refraction.

0 mm with a proper calibration and use of the EDM. the estimated precision is 5. traverse 2. precision average 3. Using equation (4.3.0" in zenith angle and 5 mm in slope distance and 0.K Brunner [1981] for non-simultaneuos 0.3 mm~K (K in km) in a practical non-simultaneous trigonometric average sighting distance of angle of 88° non-simultaneous coefficient of observations is about By using 310 m and observations refraction test of reciprocal height traversing This 30'.57.9) and assuming uncertainties of 1. 4 mm ~K 300 m and (K in 500 m. Rueger and Brunner [1981] have 4. 4 mm ~K legs of and 4.1 mm~K of slope angle Under the same (K in km) 0 and of 10 is a expected over traverse legs assumption with traverse leo an of 300 m. distance measurement If the uncertainty is reduced to due to the slope 3.5" if performing four sets of measurements 1984]. . lengths of 500 m. recently developed precision electronic theodolites to measure zenith angles.3 in the coefficient of refraction (for simultaneous observation). the standard error can be as small as [Chrzanowski. then the achievable accuracy in the above cases will be km) for the respectively. 0. 0 mm YK (K in km). .O. estimated by Rueger and reported a precision of an average result the is more with shows zenith that in in the uncertainty than an The 0.45 analysis.

standard error of a mean double run of was achieved with lengths of sight of up 1985]. The area was inclination of feasibility study carried accuracy the University less than after rejection of 1984]. deviation of 1. Brunswick moderately flat with a out a year earlier. in km) IGN with claims errors of lengths in the of The United 400 height results are very 1 mm v'x sight to 1985]. was expected to be around better [Chrzanowski. respectively [Kaeser. general 1985]. Paris has trigonometric 1982 to 1985.46 The National carried out Geographic Institute (IGN) traversing tests from impressive. m. (NGS) reciprocal extensive in and 3 mm v'x (K m and 1500 The National Geodetic Survey States has tried reciprocal trigonometric height traversing on a 30 kilometer loop. 8 mm v'x (K in km) as the expected value. 0 5 • According to gave the a the overall 1. 1 mm v'x The (K in km) to 148 m [Whalen.5 mmv'K (K in km) An adjustment of one line. Reciprocal trigonometric height traversing was used to determine heights in a network with a total length of the interconnecting lines of over 70 km by the Department of Surveying Engineering at [Chrzanowski. of New or the network estimated standard which was almost the same .

7 p A horizontal =T distance.2 p = ----- T A p A ---------2 + T where dT A --------2 4 R R 502.47 4.7 p B p dz A 17.1.11) yields 2 D (J c = ( 4 R 502. The difference in the refraction coefficients between direct and reverse measurements is correction.7 p --------) 2 T 1/2 [ 2 (J dA 2 + (J dB 1 (4.12) refraction into a~d (4. Assuming simplifies equation (4.2 p 2 D c = B 2 dz 2 B (4.10) to B A 2 502.11) 4 R 2 T Applying the error propagation law of variance to equation (4.3 Precision of refraction corrections in reciprocal method In equation (4. in the term is the magnitude of reciprocal method.8) the refraction the second effect.12) .7 p D R R (4.9) considering only the refraction correction term. needed to Substituting compute the equation (2.10) B z is D = S sin = p B T T the T = T and B A dT 502. yields 17.

Figure 4.48 dT dT B A where and dA respectivily and stand for dB dz dz and is U c the deviation U = (]' = CJ .13) (]' R where is the standard deviation of the measured or modeled temperature gradient. it is . 100 m.2 shows the standard deviation correction of the refraction versus sighting distance D. when the lines of sight are becomes randomized to correction limiting shorter than 100 m. Although a more reliable refraction correction. for is sighting distances temperature gradient required Measuring to compute the the temperature gradient along the line of sight is neither practical nor economical. as it will be shown later. equation (4.12). p dA dB Assuming refraction correction. is not the line a large required of sight solution than earring out the not economical. According to greater than along the equation (4. line a of very accurate sight refraction correction. standard of a = 1013 mb and T=300 K. the extent (see seems effect of refraction and the Chapter to be refraction 6).12) will be simplified to 2 2 D u = c (4. On the other hand.

5) a similar expression can be written for the ellipsoidal height difference from B to A 6h = BA S cos z B S sin z ( B .05 °C/m.3 oc. assuming P=l013.1. believes methods can that these be used of the The author indirectly compute the refraction effect in the reciprocal method.14) .1 °C/m and 0.oo -E B . Curves show standard deviation of refraction correction for temperature gradient precision of 0. to The following section is a summary of this new proposed method. mb and T=300 K. 0.4 Considering equation (4.m. Proposed method for the calculation of refraction correction 4.V/2) B (4.49 ~ [mm] 6 4 50 100 150 200 250 300 Distance [m] Figure 4. other methods for the determination refraction angle were discussed in Chapter 2.2: Three Standard deviation of refraction correction in reciprocal height traversing as a function of distance.

Then with sufficient the total of the accuracy by refraction angles can be written as = + 00 00 A B s (COB D Z + COB Z ) + A B s R sin z A (4. = S sin z = D S sin A The effect of the distances (less z B deviation than 500 m) of the vertical can be neglected for short without any loss of accuracy. = + 00 00 B A s + COS Z (COB Z D A ) + V - ( B e e ) B A (4.V ] = A B A B A B 0 then. because the total of two refraction angles are affected by . 1975a] 0 or S ( COB Z + COB Z ) .16) .D [ e .15) where.e + 00 + 00 .3).6h BA AB = two direct and reverse height [Brunner.6& = & - A e B The angle V can using equation be estimated (4.6h .50 Theoretically the sum of differences must be zero + .

the dispersion method divide the total of different weights for each A weight can be estimated or the discussed reflection or in one has to Chapter the 2. of Brunner a detailed discription of the angle-of-arrival application for meteorological for mentioned in 1980 and 1982] gives variance which can The angle-of-arrival [1979a] only measurements refraction angle computation by observations.11).51 According to this equation the total of refraction angles can be computed in terms of either measured or approximately known quantities. compute individual refraction angles or refraction angles by considering refraction angle. As it was of observation. [1979b. In a and its using some reciprocal mode of the amplitude of image dancing can be measured sides and these two measured amplitudes can be . (4. in order to make corrections to reciprocal observations. the main concern as it can be seen from equation is the difference of the two refraction angles 6oo = w A oo B which affect the result of measurements. using either angle-of-arrival Here angle-of-arrival approach is considered appropriate for conditions and does not unstable special instrumentation. In reciprocal levelling. proposed by Brunner refraction. Therefore. from both the be more need any method was estimation of angle Chapter 2.

2 Refraction in Leap-Frog Trigonometric Heiqht Traversing An alternative approach to be leap-frog over a the reciprocal method could trigonometric height terrain that is uniform material with which the surface by refraction fore-sight. is described by Brunner further has fluctuations (spread of the second of a [1950] using a However.52 used as the weights to split the total refraction angle into two separate refraction measured the amplitude of image image dancing) through the from accurate method of photo detector for 10 telescope Kukkamaki angles. both in slope is covered for the This method and the can be affected and back- the when levelling along a highway the slope of the route is not uniform and the back-sight may pass over a ground covered by material different from the fore-sight (e. information because sensitivity of the human the frequency is not necessary beyond eye (15 Hz) [Brunner. 4. grass) These are the differential refraction . This method is recommended for observations during clear days under unstable conditions. situations that can magnify the (e. only the visual measurement may be sufficient. long visual level observations measuring the image fluctuation investigation on the more A instrument. 1979b]. total proposed of refraction angles method.g. asphalt). symmetrically But in practice traversing. [1980].g.

Leap-Frog Trigonometric Height Traversing Formulae 4.6h MB D A distances.17) A V /2 ) B (4. and the second term is the effect of .19) I = S sin z are horizontal B The contents within the first bracket represents the height difference.and the fore-sight in leap-frog arrangement: = 6h S MA COB Z A = S 6h MB - COB Z B sin S A A (-E - B A E - 00 - B /2 ) V 00 - A sin Z ( .S B Z (4.18) B A combination of these two equation gives 6h = 6h (D MA D oo B B where = (S . In a careful reconnaissance mountainous terrain to find the mid-point is a difficult and time consuming task. leap-frog method is the necessity weakness of of finding the mid-point (theodolite station) before the actual measurement takes place and it has to be located to better than 10 m.53 effect in the the The greatest leap-frog method.S B .1/2 (D V B and COB Z A D B B ) - A . = A S COB B oo ) A sin Z A Z .D A £ (D +D B V ) ) A (4.1 Two expressions similar to equation (4.5) can be written for the back.2.

Since the lenoths of sioht in the leap-frog method are usually short (less than 300m). the effect neglected. of the deflection of the vertical can be The third term is due to refraction and the last term is the effect of earth curvature which can be ignomd as long as the point (leas theodolite station is close enough than 10 m) or it can be computed to the mid without any difficulty. B Figure 4.54 the deflection of the vertical.3: Ellipsoidal section for leap-Frog trigonometric height traversing .

2 By applying the law of (4.21) 2 4 R 6k = k . of refraction added to Brunner [1982] for details).k ) B A Achievable Accuracy Usinq Leap-Froq Trigonometric Heiqht Traversing 4. 5 mm for slope coefficient distances.21).and fore-sights are equal.2. with standard deviations of 1" for zenith angles difference from and with a back- to .s B cos A D z ( k 2R A (4.20) .20) is valid when the lengths of the back.20) the variance the leap-frog propagation of variances to equation of a measured height difference using method can be obtained as 4 2 a = 2 < 2 cos z 2 2 a > + 2 s 2 o D a + ----- z 6k 2 (4.k B A in which Equation (4. terms proportional to If they are not equal then two other their differences should be the equation (see Rueger and According to equation (4. omitting the D= D = D A B third and the last terms and assuming yields 2 6h = s cos B z .4) into (4.19).55 Substituting equation (4.

2.5.21). The network gave the estimated (K in km) [Chrzanowski.20) correction we get By for the refraction .8 mm VK found for average sight 250 m respectively average slope of 0. 3. 4. The with lengths 0. 1985].20) leap-frog is the refraction correction to leap-frog trigonometric height traversing.2 lines.12) into (4. and 3.8 mm~K (K in km) for sight lengths of 200m and 250m were found by using equation (4. The National Geodetic Survey in United States tested leap-frog trigonometric height traversing loop. was achieved 1985] .56 fore-sight 0.66 mmv'K of sight up UNB leap-frog method was used determine heights in the network mentioned in Section with a total of 70 km of interconnecting least-squares adjustement of the standard deviation of 1 mm~K to 85 m to 4.3 Precision of refraction correction in method The last term of equation {4. of the accuracies were 4.9 mmv'K and mentioned conditions.6 mm VK (K in km) of 200 m and 0 properly calibrated angle and over a terrain of EDM with accuracies lengths with theodolites and Using electronic of 10 . on a 30-kilometre A standard error of a mean double run of (K in km) [Whalen.5" for the 3 mm for the distance observation under the above the standard errors of 2. substituting equation (2.1.

23) R cr is the precision of Graphical temperature.57 c = R dT 2 D 502.13) is cr c 2 4 D cr = where (4.7 p --------- 2 R 2 ( dT A ---. the measured or representation of for different temperature precisions error of refraction corrections being lengths in the temperature with the standard doubled (over a twice Equation (4.dz B (4.23) shows that the sight the leap-frog lengths in the equation (4.2 (see Figure 4.23) and sighting distances is similar to Figure 4.4) longer traverse leg). obtain the . modelled method should reciprocal method or a gradient is needed be half the sight higher precision for in order to same refraction correction in both methods.22) ) dz T and the standard deviation of refraction correction with the same assumptions as in equation (4.

.05 °C/m.58 [mm] <{ 6 0"=0. assuming P=l013.3 °C/m.1 4 0"=0. Curves show standard deviation of refraction correction for temperature gradient precision of 0.1 °C/m and 0. 0.05 2 50 100 150 200 250 300 Distance [m] Figure 4.4: Standard deviation of refraction correction in leap-frog height traversing as a function of distance. mb and T=300 K.

multiple targets were used to changing the clearance of in the traversing field by comparing the This method using "UNB-method". initiated a because The testa carried out in the summers of were described as encouraging by Chrzanowski [1983]. method were inside the UNB campus. (UNB) the feasibility of traversing for precise approach with elevated The multiple and to have a results from targets was tests quick check height named the of the levelling. results of preliminary 1981 and 1982 A to reduce the effect of pointing errors. other test lines have been established on the campus. . trigonometric research implementing elevated targets randomize the refraction error by the line of sight. to Head-Hall test line. but is renamed here.Chapter 5 TEST SURVEYS AT UNB 5.1 Background of Trigonometric Heiaht Traversing g1 UNB In 1981. the Department of Surveying Engineering at the University programme of New Brunswick to investigate trigonometric height modified leap-frog was chosen. different targets. carried of leap-frog multiple The out at elevated first practical a site chosen which used to be called the UNB-test line.59 - .

In the summer of 1984. a test network of over 70 kilometres using the of lines was measured leap-frog UNB-method reciprocal height and the traversing. determination of an optimal model of the temperature profile up to 4 m above the ground. Besides the test network. .60 In the over a test summer of 1983 six kilometre area) and Fredericton. The study concentrated mainly on: 1. computerized and traversing. order to have refraction a better effect in The test surveys involved the long term observation of changes of the refraction angle over different types of surface coverage. kilometres of height traversing once second time total A twice: of using over 140 was completed [Chrzanowski. A the test surveys route outside into some loop complete were extended Fredericton (Mactaquac surveys description in of the city the of projects. In the summer of that year. 1985]. attempts were made to develop a motorized system of trigonometric The results and descriptions height of the system are given by Chrzanowski [1984] and Kornacki [1986]. including a statistical analysis and dicussion of results is given by Greening [1985]. The computerized system of levelling was further improved in 1985. surveys were understanding number of individual test carried out of atmospheric the in a trigonometric height traversing.

and 4. legs This sunlight and All three their maximum 2. confirming the UNB-method in practice the precision of designed under controlled field conditions and to add to the knowledge of the refraction effect. distances of station. The results of the 1985 test surveys have been analyzed by the author and are discussed in this chapter. differences measured five between times before bench and marks were after the test . 5. variation of refraction error under unstable. entirely over either combination about 200 m from As can be BM2 and BM3 the central seen from have equal instrument forming three lines each of which lies almost gravel or of different grass or asphalt. of comparison available the such models as Kukkamaki's temperature function against three new models which were suggested by the author. 3. three bench marks BMl.2 Description of the Test Areas and Scope Qf the Tests 5. ground surfaces investigate situations when one sight was selected are open inclination The precisely to is about height direct to passes over a surface ground texture which is different from the other.1 the UNB campus. IS.61 2.1 South-Gym test lines The South-Gym area is located at the top of the hill at the south side of Figure 5. neutral and stable conditions.2.

temperature gradients refraction error. at marks were stable to within the surveys were carried out The UNB-method test was conducted four times in the South-Gym area. 4 hours between 10:20 and 14:30 19/20 July.62 surveys.2 mm during the period that [Chrzanowski. tigonometric the height be considered as true errors. gradient of temperature. The stability of the bench marks was also verified by several reference points established near BM2 and BM3. 38 hours between 09:30 and 23:40 29 July. The surveys showed that. These geodetically differences are assumed to be assessment of results. only of air were speed and utilized to direction useful in interpreting some of the results. compute the of wind were . 1985]. 6 hours between 11:40 and 17:30. outcomes of the geodetic levelling the one sigma level. all bench ±0. on: 20 June. and the of the ground Out of the above measurements. barometric temperature pressure and surface were measured. discrepancies between determined errorless for According to this results of traversing and geodetic levelling can mean height a quantitative assumption. 13 hours between 11:10 and 00:30 23/24 July. speed and direction of wind. In most of the tests surveys performed in the summer of 1985.

63 0 ' E=- IS BM1 8M2 m l_----~----~----~----~----~-----.----~----~BM1 0 Figure 5.1: 200m Plan and profiles of South-Gym test lines .

were especially selected of sights open to direct buildings which Out of the six in the would experience lengths It is mostly bounded by speed. These two so that the high refraction while can bench marks only summer of 1985.2). for this test survey as . The same meteorological observations that were mentioned for the South-Gym area were carried out well (see section 5.64 Head-Hall 5. line passes over asphalt or 0 of about 5 is partialy diminish the wind BM2 and clearance along one remain lower than sight is 0.5. The the minimum about 35m distance the other sight is well above 3 m for the effective part of the line i. Thomas University. the half closer to the instrument.e.2.7 m at from the instrument station. sun-light and The lines of effects while 230 m. The UNB trigonometric method was hours of tests between 12:15 and 20:15.6). BM4 were used bench marks sight concrete with the average slope (see Figure 5.2 ~ The Head-Hall line of a test line total length of 600 m consists of 3 bench marks located near Head Hall and 3 bench marks near the main building of St. used to conduct 8 on 06 August 1985.

2: Plan and profile of Head-Hall test line .65 Contour Interval Sm 6 5 0 1 Figure 5.

according to the the resolution of a temperature indicator is 0. thermilinear resistor stainless 2 m.1 Temperature aradient A 4-metre wooden rod was temperature sensors could 0.6 m. Inc. The sensors Spring Instrument. Instrument Co. The differences and determination indicator's resolution of 0. constructed Yellow and type in Spring This arrangement of sensors was chosen temperature profiling up to 4 m height with denser temperature points in the first two metres above the ground. The sensors were shaded height from direct and indirect (reflected from the ground) radiation by the sun.2 m.3 DescriPtion of the Field Equipment 5. The sensors were calibrated by the manufacturer and guaranteed with linearity of for a much longer 2 parts period than the test surveys' duration.01 oc sensor (see also Chrzanowski the absolute temperature resolution) quartz temperature [1985]).13 oc oc. be mounted at the 3 m and 4 m.1 oc. were within the . The accuracy manufacturer is of the sensors better than 0.1 per thousand or better.66 5. housing. (combination set [Yellow steel to facilitate of so that thermistor produced were of composite 1985]) by six heights 0. 1. At UNB the sensors were compared against a precision (0.3.3 m.

4).2) and 2 t = a + b z + c z (5.3.1 m and A Wild Kern prisms placed at 2 m height T2000 electronic theodolite with the Kern E2 in the test on 29 were located within two Distances were measured with aluminum rods. the vertical angle observations were made using a Kern E2 electronic theodolite to two targets long located at heights 2. DK502 EDM to 3. 1984] t = a + b • log(z + c) Kukkamaki [1979] reports that found other (5. 5.5. on the rods.4 Investigation of Temperature Models Height 5. models proposed by different researchers.1 ~ Function of Choice of models It was mentioned earlier (see Chapter 3) that the first temperature function in terms of height was adopted by Lallemand in 1896 [Angus-Leppan.2 In all refraction teat surveys.1) during his investigation.4.5 m on 3.5 m systems could was also July. used together Both theodolites a few metres of each other be compared conditions (see section 5. two of such models are t =a 2 + b z (5.3) . a Kern so that the under strongly correlated For more details about the instrumentation see Chrzanowski [1985].67 Trigonometric height traversing 5. he For example.

1.7) is Niemeier. z 2 3 a + b z + c z + d z t = t =a 2 (5. added besides testing the following three models -1/3 t =a + b z t =a + b z -4/3 + c z 2 -1/3 t = a z + b z (5. two polynomials of order 3 and 4.9) for z > 0 (5.5) 4 3 + b z + c z + d z + e z (5.68 In 1937 Kukkamaki suggested his temperature model.10) . (5.7) (5. some of the above equations.6) t = a + b • exp(c z) Equation (5. Heer's temperature According to the model [Heer and investigations of Heer and polynomials of greater than second degree failed to work properly. [1983] and which was reviewed in Heer and Niemeier [1985]. Niemeier [1985].8) + c z c .1. as =a + b z t c (5. 1985]. In the UNB investigations. and one exponential function of height.4) and his refraction correction formula Three other models have been tested by Heer section 3 . the author.

4 and 5. of observations the sensor at 4 m measured at heights After about 22 hours malfunctioned and the remaining 16 hours of observations were conducted using only five sensors. (5.4) when c = -1/3 t = a + b z -1/3 (5.8) (5. The absolute temperature was 0. the sensor at 0.2 m. During this period.3 m elevation.10) and (5.6 m was .11) From all the above models seven were investigated by fitting them to observed significance data of their and statistically coefficients. 0. 3 m and 4 m. It should be mentioned in here that extrapolation using the above equations part closer is not recommended.11). when subjected to the least squares estimation of the exponent c. models are given by equations (5.2 Temperature gradient measurement The most complete field 23/24 July 1985.6 m. 1. were suggested. (5.8). Equations (5. 2 m. to the ground which especialy in this case is for the for under 0.7.10) That is why equations (5. testing The seven (5. sometimes fail to converge. the selected (5.7). (5.9) and are expansions of Kukkamaki's model (5.3).4).9) and (5.3 m. (5.9).4.8). The main observations were conducted on concentration in this section will be on the collected information on these two days.69 The author has found that models 5. 5.

and so on.2 the South-Gym test lines were described. above all grass and three asphalt.3 show the time averaged measured temperatures. the new sets of readings were started. Gradients types of temperature of ground were determined surfaces: gravel. . coming back grass after about 60 to 70 minutes.0 m for about 5 hours and then it was returned to its original height. 5. completed immediately the system up the sensors was sensing the temperature of ten sets within 10 of readings to 12 was transferred to the of all minutes and next spot on asphalt and after ten minutes.2 and 5. sensors were Then. In section 5. all three measurements took 63. angle of refraction for every every hour.3 minutes which means that one complete round of observation was made almost every hour. to the intial spot on the In average. Thus. Tables 5.1. the measurements of temperature commenced at least 10 minutes after setting to make sure that the system the new environment. the line could be determined once For every one-hour interval.70 transferred to 4. Starting in the grass field. one set of average temperature values is computed out of ten or more individual observations.

66 25.51 16.64 19.79 19.14 20.36 18.32 <1 22.65 23.66 25.0 20.55 27.01 19.46 18.46 15.71 18.82 12.23 22.78 <1 24.04 20. 10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00 18:00 19:00 20:00 21:00 22:00 23:00 24:00 01:00 02:00 03:00 04:00 05:00 06:00 07:00 08:00 09:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00 18:00 19:00 20:00 21:00 22:00 23:00 REMARKS wind c.42 16.24 25.25 13.53 22.95 24.26 13.3 0.53 <1 22.63 19.98 1 18.66 15.58 20.26 13.23 16.23 13.15 22.63 <1 0 69 13.92 18.63 21.40 20.99 23.27 20.96 21.73 20.94 22.04 26.61 12.27 13.71 20.49 20.53 18.02 21.91 20.75 26.33 15.52 22.54 22.97 20.15 22.77 <1 23.26 21.82 21.01 18.79 23.42 1 20.25 20. m/s % % -.48 16.12 24.47 <1 20.60 12.71 TABLE 5.38 20.85 20.45 25.74 25.76 19.71 24.90 14.73 19.64 21.17 21.19 24.58 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 - c.06 20.6 1.69 20.0 4.83 16.89 17.58 16.87 19.36 16.63 16.42 <1 25.89 20.71 18.95 16.1 The Time Averaged Temperatures on Gravel Line ZONE TIME NO (ADT) TEMPERATURES [•C] AT HEIGHT [m] 0.11 18. hu.39 1 24.19 22.49 21.41 16.54 20.66 15.18 23.15 22.07 20.66 <1 18.93 14.63 12.82 21.53 26.86 14.36 15.34 20.10 23.94 24.44 21.65 20.58 16.85 <1 15.67 15.08 14.66 24.44 20.59 15.33 20.73 15. hu.67 12.63 20.94 21.27 21.09 2 22.85 25.47 16.88 18.07 21.2 2.32 24.92 22.40 21.84 <1 0 0 0 0 0 0 0 68 58 48 26 18 10 8 25.09 19.54 19.50 16.74 20.80 18.63 24.92 <1 24.0 3.62 18.60 21.92 14.98 20.61 23.33 12.83 20.05 22.99 15.95 24.19 20.38 15.77 18.22 12.18 15.46 24.57 24.11 21.39 19.09 20.06 20.4 0 41 4 0 28 4 50 32 6 50 24 6 50 20 1 50 12 6 50 22 6 100 21 4 75 28 4 75 30 <1 0 49 <1 0 57 <1 0 62 <1 0 62 <1 0 64 <1 0 66 <1 0 66 <1 0 68 12.68 19.95 20.22 20.-.73 12.17 20.88 15.39 1 0 30 90 85 80 80 8 8 22 34 32 46 53 56 58 relative cloud cover relative humidity ·so 80 80 ----- .49 15.29 20.35 18.58 19.c.50 17.63 15.26 13.20 19.65 21.69 18.37 15.28 13.58 20.81 18.43 18.48 <1 17.33 24.70 24.04 24.99 14.60 17.c.99 20.10 19.51 21.

30 17.64 15.6 1. 10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00 18:00 19:00' 20:00 21:00 22:00 23:00 24:00 01:00 02:00 03:00 04:00 05:00 06:00 07:00 08:00 09:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00 18:00 19:00 20:00 21:00 22:00 23:00 25.45 20.10 19.01 14.74 13.87 22.66 12.2 2.40 17.22 20.36 21.75 21.43 12.32 12.58 11.44 20.93 20.78 24.39 20.40 17.46 16.09 17.28 11.80 20.45 REMARKS wind c.39 24.47 25.27 20.27 12.98 20.17 17.41 25.48 21.59 21.16 18.40 14.02 13.20 15.66 21.84 24.05 25.53 23.03 20.72 TABLE 5.78 25.16 18.40 15.11 19.66 13.73 20.94 25.48 21.82 25.83 14.24 22.60 25.48 12.32 19.58 relative cloud cover relative humidity 22. hu.75 20.04 15.84 25.63 27.50 20.59 14.16 21.65 18.37 25.08 20.06 13.45 21.62 24.11 19.c.34 21.85 21.30 17.20 12.96 12.80 20.86 21.21 14.84 19.63 22.39 23.66 19.62 20.10 25.14 19.65 15.40 24.07 12.69 18.87 18.46 12.41 21.78 23.47 21.84 26.19 25.93 23.95 11.38 21.15 20.25 20.54 21.77 15.80 16.69 12.14 19.16 14.45 21.90 21.39 19.20 24.17 19.36 21.14 21.91 15.0 19.72 20.91 14.73 14.74 20.52 26.28 21.98 20.13 20.65 15.22 20.97 19.58 20.40 14.92 17.88 23.80 16.53 14.97 20.87 25.95 18.97 23.75 21.00 15.66 18.83 23.27 21.07 19.70 19.64 21.93 15.05 23. hu.45 26.97 13.35 19.07 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 - c.21 19.76 20. m/s % % -.49 14.48 12.07 21.26 14.09 21.42 17.c.3 0.45 15.71 23.31 22.41 12.06 19.89 19.16 14.53 18.88 11.24 20.2 0 43 4 0 36 4 50 20 6 50 20 6 50 24 4 50 20 4 50 14 6 50 26 6 0 20 4 75 41 <1 0 38 <1 0 51 <1 0 60 <1 0 62 <1 0 62 <1 0 62 <1 0 66 <1 0 66 <1 0 69 <1 0 69 <1 0 70 <1 0 70 <1 0 68 <1 0 54 <1 0 38 <1 0 22 <1 0 12 <1 0 9 <1 7 0 2 20 10 1 40 12 1 70 22 1 20 30 2 0 35 <1 20 48 <1 20 54 1 20 58 2 20 59 ---.67 21.99 23.08 22.21 15.55 25.90 20.24 17.04 20.38 24.68 15.0 3.34 12.89 20.17 22.- .66 20.63 14.44 19. 89' 21.-.31 18.2 The Time Averaged Temperatures on Grass Line ZONE TIME NO (ADT) TEMPERATURES [°C] AT HEIGHT [m] 0.05 18.75 16.37 11.0 4.88 22.08 13.83 17.76 12.67 16.54 18.22 26.02 20.44 20.02 14.04 22.12 17.28 19.99 17.20 27.38 15.65 12.85 16.15 19.44 16.16 24.

29 <1 <1 <l <l 0 0 0 0 65 68 70 69 14.51 23.30 24.19.79 28 13:00 24.34 15.12 20. relative REMARKS wind cld.21 20.65 35 20:00 23.3 0.77 22 07:00 23 08:00 14.28 16.48 20.88 17.30 20.17 20. relative hu.0 - 1 10:00 19.76 16.31 19. hum m.34 5 14:00 21.34 20.52 15.78 24.89 12.93 20.25 23.32 22.39 18.48 19.39 15.11 20.56 18.17 13.58 20.c.82 14 23:00 16.20 ---- .77 15.02 20.55 21.89 25.56 20.86 26 11:00 20.16 19 04:00 13.12 21.71 19.68 18.45 26.24 11 20:00 20.37 16 01:00 15.15 24.35 18.48 18.77 21 06:00 12.11 26.51 21.82 20.55 16.0 3.80 34 19:00 24.56 22.10 25.92 7 16:00 23.92 25.49 33 18:00 25.04 21.62 18.92 17.44 25.31 14.27 15.30 10 19:00 21.69 20.22 13.02 24.33 20.36 17.82 20.35 20 05:00 11.42 22.19 13.01 17.42 20.53 <1 <1 <1 1 <1 <1 1 <1 l 2 2 <l <l 1 0 0 0 0 0 0 0 80 70 20 0 30 30 30 52 30 16 11 7 7 8 16 22 32 41 51 55 58 cloud cover humidity 22.04 17 02:00 18 03:00 15.41 25.79 25.08 22.44 11.41 18.81 25.06 20.19 29 14:00 26.14 26.31 19.84 22.3 The Time Averaged Temperatures on Asphalt Line ZONE TIME NO (ADT) TEMPERATURES [°C] AT HEIGHT [m] 0.75 9 18:00 21.66 15.26 4 13:00 20.01 37 22:00 19.20 19.46 19.23 21.35 19.33 15.69 15.70 21.89 25.37 20.25 20.35 24.16 38 23:00 18.73 TABLE 5.51 12.04 12.s % % -.96 12.20 22.2 2.74 15 24:00 15.27 15.08 16.27 19.58 17.11 30 15:00 26.20 25.93 19.53 22.24 15.73 16.19 21.93 17.27 19.03 12.41 25.19 20.77 14.87 20.08 12.47 25.44 12.36 c.70 19.53 19.21 23.77 20.71 20.28 21.25 21.37 19.69 2 4 4 6 6 5 6 6 4 4 <1 <1 <1 <1 <1 <1 0 0 50 50 50 50 50 50 35 35 0 0 0 0 0 0 42 30 26 27 22 20 14 20 32 41 45 57 60 60 62 65 15.35 23.14 23.31 19.54 27 12:00 22.67 24.22 15.46 20.28 20.95 19.0 4.21 <1 0 69 17.93 17.25 15.6 1.08 25.40 24.72 20.67 15.82 21.23 13.35 20.03 6 15:00 20.80 15.92 20.63 19.31 19.36 11.15 13.19 18.59 19.34 3 12:00 20.41 36 21:00 21.85 '·· 25.98 12 21:00 19.03 32 17:00 26.66 16.27 18.93 25.09 18.21 20.-.74 2 11:00 21.47 14.92 20.20 23.50 18.38 20.92 20.19 8 17:00 21.69 31 16:00 27.20 13 22:00 16.97 20.52 12.86 24 09:00 25 10:00 17.65 20.

failed to to table i7 in because converge. . Figure 5. 5. the weight matrix as factor for every Table 5. According c = -l/3 makes Kukkamaki model the asphalt. height observations.5 and 5.74 Determination of models 5. Tables 5. il and for The coefficients different height above the measured of refraction ground show very little change from one model to another. standard all one-hour except in the number 35 averaging the The number of curve fittings of intervals.6 show a detailed computation of these curve fittings temperature over computed for of models asphalt.4. temperature were +1 and reduced. and a unit value models were sometimes the mentioned earlier.4 leas a as constant flexible with larger standard deviation in comparison to other models with the exception of the second order polynomial (model i6). which average the 1 °C) of 38 a posteriori identity (a vector for the a intervals in all is equal to the number case of intervals solution and deviations. During difference determination intervals were coefficient Qf temoerature These values and the gravel and priori standard error of priori variance cases.3 The observed one-hour ~ temperatures interval.4 lists the tested are calculated by variance factors assuming for of continuous and temperature completed for models averaged 38 hours intervals for the grass field.3 shows the contours of coefficients of refraction with respect to time and height above the ground.

076 0.047 0.047 0.088 0.063 0. The solid line contours show the zero refraction coefficient which is associated with the .061 0.060 0.066 0.058 0.4 Mean standard deviations Mean Standard Error [oC] South-Gym Area MODEL NO Gravel Grass Asphalt 0.046 0.055 0.115 0.048 0. neutral and unstable conditions can be detected from these figures.053 c 1 t =a + b z 2 t =a + b z 3 t = a + b z - - -1/3 -1/3 + c z -3/4 -1/3 2 +C Z 4 t = a + b z 5 t = a z + b z 6 t = a + b z + c z 7 t = a + b exp c for z >0 2 (C Z) Model t4 is refraction.066 0.090 0.068 0. used for computation of the coefficients of The stable.068 0.057 0.054 0.75 TABLE 5.043 0.

to model a solution. main problem with Heer'e model. the case over gravel. same. The time of the seems respect to change with particularly true for There is to neutral condition height.7 and 5.4 and Figure choose either of the modele 13 or and precision of fit. The c ~ 0 where c t statistic is given by e. Tables 5.8 show the test of the significance of the coefficients at the 95% confidence level (1-a level) as well as the probability of being insignificant (a level) the coefficients.4 shows all models presented because in this model in Table only two coefficients are involved). fit to shows the that observed 45 and 46 are almost the Table 5. Considering both best 14.76 neutral condition time. Figure 44 or 15. a slight change from This is Figure 5.3 a. 5. it sometimes does not 5. hypothesis that c of These tables show the testing of the null =0 versus the alternative is the estimated coefficient.4 converge Kukkamaki's temperatures. Draper and Smith [1981] as t = c I Sc where Sc is the standard deviation the results 5. model 45 gives the closet results to these figures. of c.g. but. as it was mentioned earlier. Figure 5. using one model to another. is not the and models 13. The best fit according to this figure is The to Heer's model.4 of the tests on (except model 42. is that in least squares estimation of the exponent coefficient c.4. one can for their ease .

08 -0.11 0..06 0.00 -0.5 -1. 6 .73 -1.0 0.09 -0.00 17.00 -0.00 -0.11 0.22 0.2 -3. over asphalt).4 -4.59 ••••• -0..04 -0.4 0. 1 .19 20.00 o.02 -0.00 -0.6.1 0.1 0..0 METRES .90 22..2 0.03 -0.0 METRES.5 K2.00 0.02 -0.1 2.3.5 -2.00 0. 3 .oo -0.02 -0.02 0.00 16. 7 .3.03 -0.00 -0..05 0.54 -2.04 -0.8 -0...1 -2.11 4.03 -0.TABLE 5.03 0.3 0.14 -0.52 -1.0011 0.00 5..16 .00 13.aa -----------RESIDUALS---------- VARNCE A B c POINT COEFFICIENT OF REFR.05 0.01 0.1 1.00 0.6 -1.03 -0.0026 0.52 -0.1 -0.02 0.2 0. FIVE TEMPERATURE SENSORS.3 0.00 23.07 -0.to .02 -0.01 0.02 -0.00 0.5 -1.05 -0.0009 0.03 0.1 0.1 -1.4 0.0.26 -0.05 -0.1 -0.02 -0.01 -0.74 -9.01 -0.oo o.1 0.00 -0.08 39.11 0.02 0.9 -0.26 -1.2 o.60 19.0034 11.0002 0.1 -4.0006 0.0.03 -0.21 0.03 0.5 1.1 -4..09 0..99 0.24 15. 2 . 1. 8 .01 0.03 -0.9 -0.04 -0.2 -1.06 -0..04 16.3.0089 0.01 0.0 K1.6 -5.05 0.0004 0.04 0.0002 0.46 23.29 0.0 ANO 4.05 0.4 -4.01 0.07 0.03 -0.00 12.2 -0. SENSORS ARE AT HEIGHTS 0.79 -0.00 0.01 -0.01 0.8 -1.01 -0.74 19.9 4.05 0.00 "'18 "19 •20 •21 •22 t23 924 825 f26 927 928 t29 •30 •31 •32 .04 -0.85 -0.0 -3.01 0.04 -0.04 3.02 0.9 -0.2 0.1 -0..J .4 -0.4 -0.49 -4.84 -2.0 AND 3.1 0.01 -0.J .6 -0.0003 0. 24 ••••• 1. 10 0..10 20.02 0. 13 6 5 13 3 5 8 8 3 4 4 13 9 18 8 17 16 21 11 6 7 10 7 23 35 8 9 35 9 11 12 9 29 .40 18.03 0.90 17.10 -2.92 25. -1.6 -0. 3..00 21.SO 0.00 0.3 -5.15 -1.01 -0.10 0..04 24.0020 0.1 o.3 0.1 0.00 -0.3 -0.3 -0.0058 0.3 -6.1 -3. 14 "'15 .55 33.08 -0.01 0. KO.6 -0..31 0. .10 -0.01 -0.4 -0.4 1.6 -0.. f FIVE TEMPERATURE SENSORS.01 -0.8 -4.0034 0.7 -0.0 ANO 4. 58 0.6 -3.21 18.36 -0.04 -0.3 -0.6 -0.02 -0.77 23.01 -0.4. 5 ITEt -0.00 0.03 18.00 13.94 19.04 0.2 1..1 -1.2 -0.5 -6.3 1.046 ----------------------------------------------------------------------------------------------.00 14.00 19.2 -1.03 -0.0007 0.17 12.3 0.2 -0.8 -1.64 -0.01 0.4 0..27 16.00 22.4 0.5 -1.00 -0.0011 0.6 0. • TIME .00 -0. 2.00 o.25 19.50 -0. 1.03 -0.01 0.2 0.00 4.00 11.02 0.31 -0.oo 20.02 -0.73 -2.a 0.4 -1.0042 0.0006 0.0 -0 .. 3.3 -0.38 0.95 23.13 -0.00 0.2 0.13 0.01 -0.s -0.00 8.1 -0.00 -0.02 0.71 -1.6 0.03 -0.15 0.03 -0.2 0.01 -0.37 -0.3 -0.4 -o.05 -0.03 -0..0006 0.0007 0.0 -1.6 -0.7 -0. SENSORS ARE AT HEIGHTS 0.oo -0.20 ••••• AVERAGE STANDARD DEVIATION : -0.02 0.7 -6.08 0.oo -0.0 0.8 -0.02 o.35 -0.05 0..02 0.9 -1.2 -7.02 -0.3 0.1.2 -0.5 -0.2 -0.2.10 32.03 0.00 6.00 15.00 12.01 -0.0082 0.6 -1.0 -2.04 0.09 1.00 21.88 15.5 -1.02 -0.3 -1.03 0..3 -2.00 0. 0.2 0.0007 0.00 10.5 -0. 9 -.00 22.10 0.0041 0.00 16. 11 ->12 "'13 .2 0..4 0.8 -2.07 0.05 -0.5 -2.05 -0.2 0.00 24.2 -0.4 0.00 19.02 -0. 4 ..1 -0. 00 3.9 1.01 -0.01 -0.8 -0.0068 0.0028 0.41 0.0064 0.0034 0.3 7.00 11.2 0.01 -0.82 19.00 18.01 -0.75 1.5 1.00 17.1 0.2 -2.01 0.oo 0.5 K4.8 1.6 1.03 0.02 -0.01 0.3 -0.3 0. 2.0 0.34 -0.03 -0.78 12.1 -0.07 -0.26 0.01 -0.4 -1.7 -2.03 -0. #l in Table 5.3 -1.0089 0.00 15.55 19.01 o.0023 0.01 -0.00 -5.02 -0.37 2.01 0.03 0.01 -0.2 -2..0 -0.02 -0.7 -4.00 0.0 -0.5 -1.10 0.5 -0.05 -0.11 0.0013 0.6 -5.01 0.06 0.56 31.48 36.1 -0.00 14.7 -0.2 -0.1 -1.6.02 -0..2.03 0.o 0.00 1.0 METRES.2 0.17 10. 1.1 0.1 -0.01 -0.05 0.5 0.01 0. 5 . 0017 0.0 -1.0005 0.00 0.8 -5.03 0.02 -0.98 7.9 1.5 -1..4 0.35 -1.01 0. SENSORS ARE AT HEIGHTS 0.0012 0...0012 0.30 19.01 -0.04 -0.4 -0. .56 -5.5 -0.49 15.7 -1.6 -0. 0.9 -6.3 0.3 -0.0 -1.20 0.01 -0.4 -1.2. 2.02 0.5 -0.02 0.4 -0.5 Curve fitting and coefficient of refraction computations (Kukkamaki's model.73 8..00 18. SIX TEMPERATURE SENSORS.00 0.2 0.39 ••••• 28.02 -0.5 K3.

4 -0.]8 .69 0.04 -0.79 0.00 18.03 -0.01 0.00 13. SENSORS ARE AT HEIGHTS 0.04 0.00 -0.5 -0.04 0.00 14.03 -0.03 -0. c K0.0007 0.07 -0.3 0. 1 0.0003 18.9 -1.6 -0. 19 .9 0.02 0.5 -2.4 -0.3 0.03 -0.4 0.01 -0.0001 0.3 0.00 16.15 .01 0.00 21.5 1.00 14.0025 0.04 -0. 2 .0058 0.09 24.03 0.87 16.8 -6.50 -1.5 -4.0.3 0.06 0. SE~SORS ARE AT HEIGHTS 0.oo 0.69 21.03 -0.0102 0.0 0.03 -0.9 -1.o8 0..02 -0.5 K4.4 0.2 -1.00 -0.03 0.0 -1.2 -4.01 -0. 3.5 0.7 -1.0 -1.9 -0.1 -0.03 0.00 24.94 24.5 -0.04 0.00 22.04 -0.00 -0.7 -6.4. 6 -0.59 25.28 19.02 -5.6 1.7 -1. 1 -0.57 12. 00 3.8 -1.20 18. 7 -1.os -0.oo -0.4 0.00 16.92 0.01 0.43 24.1 o. 10 ~11 ~12 ~13 ~14 -..00 0. 15 0.5 0.2 -7.00 -0.1 -3.2 6.1 -3. 2.02 0.2 1. 1 0.02 -0.36 ].07 -0. 5 ~ 6 ~ 7 .02 0.03 0.00 -0.0012 0.1 0.7 I .9 -1.51 15.79 13.09 0.0011 0.03 -0.00 -0.55 0.15 -0.64 0.04 0.6 -0.03 0.3 1.2 -6.0 0.7 0.00 0.38 19.01 -0.00 4.2 0.01 0.01 -0.25 0.8 -1..09 16.00 12.0 0.7 -0.0050 0.00 6.56 18.08 20.02 0.01 0.79 16.28 0.o3 0.01 0.01 -0.6 -0.3 -7.08 0. 1 0.0 AND 4.01 0.00 0.00 0.0 AND 3.1 -0.oo 0.02 -0.0062 0.01 -0.00 11.01 -0.00 0. 16 .2.94 0.07 -0.00 1. 10 -0.0012 0.59 0.03 0.3.14 0.3.02 -0.01 -0.. 0.9 -0.51 1.1 -2.2 0.01 -0.0 METRES.2 0.87 19.00 12.01 -0. 5 -2.7 -0.0024 0.3 -0.0011 0.01 0.09 -0.01 0.02 0.02 0.9 -1.0006 0.7 -0.2.01 o.6 -5.0 METRES. 1.3 -3.1 1.28 1.01 -0.0034 0.50 ~ 1 .14 -0.8 1.0081 0..5 0.00 -0.76 1. 3.02 0.11 -0.92 0. 5 K2.8 1.5 0.02 AVERAGE STANDARD DEVIATION : -0.3 0.1 -1.4 -0.01 -0.03 0.0 AND 4.01 -0.6 0.00 10.02 -0.2 0.5 -0.03 -0.0021 0.19 1.02 0.04 0. 1 -0.00 -0.0 1.9 -0.5 1.5 -0.03 -0.00 -0.17 1.6 1.2 -1.5 -3.1 -0.00 -0.03 0.5 0.05 -0.03 0.44 18.02 o.81 22.00 11.0005 0.00 -0.01 0.2 Kl.04 -0.8 -5. 3 ~ 4 .03 -0.02 -0.2 -0.02 -0.05 -0.08 0.2 -1. 8 ..00 2o.6 0.03 -0.02 -0.04 0..3 -1.09 0.03 -0.0047 0. -1 .0 2.2 0.03 -o.76 21.00 0.00 23.7 -0.0008 0.01 0.06 -0.01 -0.2 -0.05 0.89 19.o1 -0.7 -0..6.05 0.0024 0.3 -0.04 1.3 -5.0005 0.01 0.00 -0.9 -0. 9 -1.39 -0.31 0.7 -0.7 -0. 1 -1.09 -0.4 -4.6 -0. 3 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 -.6 -2.0008 o.02 -0.02 -0.04 -0.4 -2.4 -0.0123 0.5 -0.TABLE 5.54 18.0 -1..3 0.02 -0.5 . 1.5 -0.7 -1.02 0.12 19.07 -0.00 23.6 -6. (model.8 -1.o2 0.05 0..02 0.0031 0.04 0.4 0.01 -0.00 15.00 22.00 5.6 -0.2.5 0. 4 -1 .4 -0.02 0.6 0.00 -0.03 -0~00 o. G FIVE TEMPERATURE SENSORS.01 0. 1.53 1..11 0.30 0.09 0. 1 0.02 -0.02 -0.67 21.56 19.8 0.00 o.98 1.5 ITEf -0.6 -2.048 ----------------------------------------------------------------------------------------------t FIVE TEMPERATURE SENSORS.04 -0..50 19.4 -0.03 -0. 20 •21 •22 •23 •24 i25 i26 927 928 829 t30 •31 •32 t33 •34 •35 10.4 -0.01 -0. SENSORS ARE AT HEIGHTS 0.00 17.00 11.6 refraction computations Curve fitting and coefficient of 5.0.J co .01 -0.0004 0.03 0.39 0.4 0.3 0.8 -1.24 21. 9 .00 0.00 0.3 0.88 18.00 19.3 0. over asphalt).5 -1. 2.2 0.02 -0.5 -2.4 0.36 11.2 -0.01 o.03 o.04 0.00 19.06 -0.3 0.01 0.48 -0.05 -0.3 -0.0006 0.1 1.3.00 -0.03 -0.00 18.8 -2.56 19.02 0.01 -0. 2.oo 21.01 o. 0002 0.01 -o.03 0.6. ~ SIX TEMPERATURE SENSORS.00 20.04 -0.01 0.3 -0.01 -0. 5 K3.8 -1.2 -2.4 0.01 -0.0 -0.01 -0.79 0.00 -0.4 -7.0019 0.02 0.79 0.04 -0.04 -0.0022 0.0 2.5 -0.05 -0.01 0.08 -0.14 13.02 -0.1 0.5 0..4 -0.60 24.6 0.55 1.1.01 -0.01 0.00 15.01 0.o1 o.00 0.0 METRES.oo 0.6 -1.1 -1.0 -1.58 23. 0.00 0.24 15.7 -1.03 0.00 8. 17 -.5 -1.00 13.00 0.8 -4.3 0.0015 0.0005 0.05 0.02 0.0121 0. #4 in Table t TIME -----------RESIDUALS---------- VARNCE A B POINT COEFFICIENT OF REFR.11 -0.04 -0.03 0.65 0.2 -0.02 0.01 -0.01 -0.2 -3.0 -0.00 0.0017 0.32 1. 17 -0.8 -1.0093 0.01 -0.5 0.00 D.4 -1.04 -0.7 4.05 -0.

. . .' . gravel. • 0. b..._..•• '.'> i . . ._ TI!~l t!. 7 5 ~.. __ .u.I\ Ir

-·.\.I .3: Refraction Coefficient Contours a.I'-_I_..::.----~~-~~ l ~'\ \ 1---. -o c ~ ~ 2..\·-.ceJnl Figure 5....' I 2 .. l I . ! _ _ _____:___: II L.... ~ l. \ il:-1:.. \/ .' ../ .. .. _ _ _ I · I I ' :. i ....\ ~~::! i~:v ! \: ' I .c Q) <1> :r: 0 • 75 "· i \ i I .\ c·. ' -. -' ..-\~-..___"'-L...J._ __ ._ \ '. i \ . • I •I• 1~ iii i !j I: .' ... : :._ . I'' ... 'lJ -..D _.. _ ·.---(a-)-----"T---.... \ ~' :. /j. 3..2 5 <. \ l I ./ .... grass and c.. '· . \!1 1.oo.•·11 J •\ . ..-q------\----. ' "'t. ~"~J. • .I.J''-''-'-'''-'--.:'\·.._· l/ I' I I ':I: ''· 1: !· I : I i: f' _ ~ ~.l ·.. ... / I' :I '\:.::> <1.'/ __ '.'::1~ I I'\ f· I:' • I ..•\ ·~-· ..: \ If\ ·.~-.'-.' I \I \ I '.. i4 in Table 5. \ ... / \ 3. 1\ \.1::. . _ __J__.50 : ". \ I~ \ I\ I r: \ " ' I: I... \ ·.' 1'f i. \/\ I..·~-~~-------..-l-----~-'-'I: 20:00 12:00 Zone Time -10 -2 12:00 04:00 -a I . 20:00 [h] -6 0 2 -4 4 Rerra<:llon Coerr. !('/ .J-.. \ t\ ~ \ \~ /: · ~ ...25 ~ <.79 ...4 is used}. .'----------.. \! \ · ~~ ~ .::> \~ <1.I:.../(·.1 i.. .. ....1 (L> 1..._ \ . <! . I~ ..1\ 0 .. asphalt (model.... \ ~ ~ ! .. =·:~.c..'• •\ /' . .. ~ \ ' I•~!\ i! j I.~\ .I ~ : ./ > '1 !l i \l ) .' 1 i..U.\

Related Interests

v\ I~'\ • ..l_ _ __ ...-\·-·. ~..J...i/ v \ y ·· '\.' : ~ \ I 4' I.__.oo-..u. "I' 'l· ~~ '"': . . \ ~If: I ./ ..!'! jl -·:~: :1 1!:1: !·_I(· ./_:'.._..\.: ' : _:r• ·..:..50 <! --+- . ··.

0043 15.7 1.0768 0.00 0.00 0.3.0093 13.2 1.93 -7.13 .35 13.2.00 0..85 T-VALUE FOR A 5. 4 .0000 0.2262 0.0007 0.33 -0..7 0.82 16.8 61.3 3.05 22.00 0.5313 0.0 0.16 0.2215 0.32 24.2 3.2545 0.0000 0. 8 FIVE TEMPERATURE SENSORS.47 -1.2590 0.8 -0.2 3.03 -0.5 -1.83 -1.00 .19 -0. 1. 3..84 24.0005 24.0820 0.20 -0.0010 3..79 0.0709 0.73 -0.7 -3.5846 0.0043 14.2 3.03 0.49 1. 0.0000 0.7 129.4 386.44 58.5 2.00 0. OF BEING INSIG.6 0.0000 0.68 -0.2 3.3 12.00 0.29 0.35 -0.2 3.8 3.0 METRES.0012 1.2 3.0000 0.3 53.0008 17.7 3.6 2.0012 20.4 -4.0 -1.14 -0.2 3.6889 0. 10 ..0352 0.2 3.74 20.09 17.0022 20.2 9.9 182.1 524.0004 23..2 3. 1.2 3.2 -5.1 787..3846 0. 12 .88 14.2057 0.0004 0.00 0.2 3.0 AND 3.92 25.08 18.0004 12.1 37.2 3.2 3.1 2.0967 0. 2.0021 0.02 c -0.57 -0.3.70 3.09 0.oooo 0.00 0.0183 11 .2 3.17 0.G 1.90 16..7 -1.060 ----------------------------------------------------------------------------------------------• FIVE TEMPERATURE SENSORS.49 -0. #1 in Table 5.69 -0.00 0.69 19.00 0..2.6 -5.2 3.0795 0. 5 .4103 0. 2 .6 2.9 42.8 1.13 -0.8 -0.2 3.2720 0.2 3.9 532.0000 0.9 711..9 215.9 39. 1 95% TEST TABLE VALUE 3.23 0.0000 0.2 3.0 METRES.0069 0.2716 0.0006 0.68 ••••• 32. SENSORS ARE AT HEIGHTS 0.0010 16.0433 0.01 0.2257 0.2 3.6.5 136.00 0.9 -1.52 -0.2 THE MEAN STANDARD DEVIATION IS : F F F F F F F p p F F p F F F F F F F F F F F F F F F F p PRO.2 3.4940 0.2424 0..0283 0.0982 0.7 22.4903 0.18 18.1435 0.0000 0.6793 0.5800 0.86 31.35 -0.00 0.0023 19.37 B 2.0009 0.7 2.0 METRES • .4 0.0048 16.0225 0.0 AND 4.16 24.1922 0.94 0.0967 0.15 2.6.0.9540 0. SIX II 3 4 10 3 8 7 7 11 5 5 16 10 20 13 7 5 2 3 4 11 25 30 13 10 4 5 12 9 21 (X) 0 .4 16. 3.3734 0.1 1.0000 0.0000 0. 1. SENSORS ARE AT HEIGHTS 0.00 0.0002 0.0196 0.8 1.1 334.38 -1.0124 10.3369 0.38 15.4 -2. 8 .8 B c 0. 2.0 -10.40 3.60 0.0 -2.0 -3.6 -0.00 0.1931 0.00 0.5 -2.96 -5. 7 .TABLE 5.2 3.3 96.0857 0.54 16.6979 0.4249 0.66 0.0076 15.3 1. 11 .2296 0.0137 18.4 22.00 0.55 -2.91 0. TEMPERATURE SENSORS.1761 0.5827 0.0024 9. c B A 0.12 0.51 1.2.05 0. 1 .16 •17 •18 •19 t20 @21 622 823 ~24 825 •26 •27 t28 t29 TIME VARNC FACTR 10.2 0.50 12.0302 12.7 0.5 -2..2 3.3 1.1 1.3198 0. 0.0006 8.0 AND 4.5923 0.2 -0.00 0.0000 0.3 1.09 -0.5 0.7 -1.0945 0.2 3. 0.2313 0.00 0.0.00 0.1027 0.0907 0.00 o..0 3.3 -1..2784 o..3 -1.07 19.0878 0.3182 0.0876 0.4 -1. 6 .2 3.14 .0044 13.2 3.5 4.3 50.1104 0.0 -3.01 0.46 -0..0119 0.00 0.2 -2.0028 22.2 3.33 -1.76 18.5054 0.0023 0.28 -0. 1 -1.0 -4. 9 .54 0.59 6.00 0.0023 A 17.60 24.6 -0.42 0..2 1.00 0.84 16.43 -1.00 0.0002 0.0086 11. 00 0.09 0.7 Curve fitting with test of the significance of coefficient (Kukkamaki's model.0018 0.62 0.2 3.3.0108 0.1 3.00 0.0020 18.2994 0.4 -l.14 19. • .0031 17.00 0.4945 0.00 0.oooa 7.22 -4.9 33.1 0.2 3. 3 . SENSORS ARE AT HEIGHTS 0.8878 0.8975 0.03 -0.33 21.0010 19.7 -1. 2.8 0.15 -0.65 15.0035 0.54 25.3444 0.4).49 3.5 -0.7 10.0000 0.38 1.15 .0000 0.0753 0.

14 4..88 20.2 3.3761 0.3 126.5 3.0028 0.2 3.0000 0.59 22.3.22 2.04 -0. .2 3.3.0147 0.0 -1. 3.75 -0.0003 0.4 67. TEMPERATURE SENSORS.5 147 . 2.5 -5. SENSORS ARE AT HEIGHTS 0.65 14.6.0114 0.73 -1.0015 0.2 521. 14 .7 65.0002 0.0216 0.2 3.0002 0. .18 -0.02 1.58 24.3 82.0000 0.9 10.0043 0.0079 0.49 15.40 20.8 -31.91 1...49 1.0 AND 4.2 THE MEAN STANDARD DEVIATION IS : F F F F F F F F F p p p p p F F F F F F F F F F F F F F p p p p p p p p p p PRO.01 0.51 12.0647 0.2 3.a -0. ----------------------------------------------------------------------------------------------• FIVE TEMPERATURE SENSORS.0451 0.0018 0. SENSORS ARE AT HEIGHTS 0. .0016 0.0019 0.4 279.00 11.03 24.30 -0.0000 o.09 -0.0002 0.8 -1.6 161.054 1.0163 0.3 48.08 2.17 -0.20 -1.9 -1.0007 0.08 2.0917 0.86 2.2 -9.2 3.04 0.1 1.0000 0.00 11.10 1.2 3.0023 0.0679 0.3704 0.00 3.7 -3. c B A o.oooo o.0010 0.0004 0.19 1.0054 0.6 -6.00 5.2.81 24.0012 0.00 6.0000 0.0000 0.00 4.0049 0.0000 0.46 19.00 19.63 19. TIME t .4 -9.0006 0.2 3.7368 0. 2.4 4.37 3.2 3.37 3.4 -0.2 3.0253 0.0001 0.00 23.0024 0.00 16. 12 .2671 0.1074 0.2 61.0010 0.OOOO 0.00 21.0003 0.00 12.00 24.0000 0.0057 0.01 0.00 22.0014 0.1 9..87 -0.4 -4..2 3.1 90.00 21..1 -1. .2 3.2 9.0016 0.3 4.0 94.2 3.9 7.9 -4.18 18.00 17.5 48.0000 0.0230 0.3 -0.3 2.0000 0.15 -0.0456 0.0.00 15.4 -4.2319 0.0000 0.1 1.02 -0..2 -3.0000 0.00 1.0046 0.06 -0.10 0.0714 0..5 -4.0 -10.29 20.3 -0.0143 0.00 23.7 2.24 21.2 3.0000 0.04 -0.33 -0.4 30.8 -17..25 -0.7 -1.3 266.0000 0. 0.16 25.0556 0.0 3.TABLE 5.7 -4.9 146.6 235.2 3.2 3.1 0.16 "'17 "'18 .0122 0.35 c 0.4 -2.9 -1.0126 0.0031 0.2 3.oooo 0.86 14.21 13.07 0.50 -0.6 -5.7 5.2 1.4 190.8 of coefficient test Curve fitting with the significance (model #3 in Table 5.0061 0.50 -1.0422 0.1047 0.84 ]3.0024 0...2.5 2.00 16.8 191.6 -2.2 3.09 0.9 0.8 7.1 128.0000 0.00 10.0 AND 3.0383 0.1 95!1.2 3.0019 0.4).23 -0..0000 0. 1.08 -0.as •36 •31 •38 10.05 0.16 -0.00 18.0001 0.2 3.0128 0.40 -0.9815 0.5514 0.4 432.44 -0.0208 0.1 53.0005 0.26 -0.1 5.oooo o.97 16.1479 0.2 3.0436 0.8 -6. 19 .6 154. SIX It 3 3 3 3 3 3 3 3 3 3 3 3 3 2 3 3 3 2 3 3 3 3 3 2 3 3 3 3 3 2 3 3 2 3 2 3 3 3 00 .50 -1..00 19.00 15.36 12.2 3.2 3.4298 0.0102 0.5 -2.2281 0.0014 0.0107 0.2 3.0434 0.05 -0.7 -2.4 155.0480 0.1 3.3 -9.07 -0.6.0000 0.0156 0.0000 0.4 65.0011 o..09 0..00 22.53 -0.2528 0.03 -0.0226 0.20 -0.2 3.0000 0. 20 "'21 .3 0.05 T-VALUE FOR c B A - 2.00 12.2 3.89 0.8 ••••• 150.4 350.1028 0.9 -20.2 3.0.0000 0.16 -0.35 -0.45 20.71 5.0000 o.85 20.00 13.88 21.7 31.0544 0.07 -0.. .64 -0.0007 0..00 20.1596 0.7 -18.0165 0.2 3. 19 0.48 18.0091 0.07 -0.3 45. 0.8 -10. 1.2 3.09 -0.oooo 0.10 -0. e FIVE TEMPERATURE SENSORS.2 3. 1 -3..0439 0. 37 1.11 -0.to "'11 .0110 0.3783 0.00 VARNC FACTR 0.2 3.76 20.97 3..6943 0. 2. 1 -4.9 190.0002 A 18.5040 0.00 2.30 18.00 9. OF BEING INSIO.0257 0.92 -1.0003 0.2 -10.5 -3.07 15. TEST TABLE VALUE 3.3. SENSORS ARE AT HEIGHTS 0.0023 0.0009 0.2 3.1 -0.2 3.0 143.12 19.00 14.00 14.oooo 0.2 3.17 -0...oooo 0..0 56.. . 22 •23 •24 •25 •26 827 828 629 830 831 •32 •33 •34 .0010 0.0075 0.21 -0.22 19.0263 0.3189 0.2 3.0935 0.49 19.89 20. 1 2 3 4 5 6 7 8 9 -. 15 -.75 -0.6 44.0008 0.09 19.0 METRES.9288 0.9 -6.4 4..0007 0.49 18.7 -5.0006 0.8 26.0 METRES.5 -3.5 -3. .2 77.00 17. 3.00 13.08 0.02 0.7 0.09 B 0.0002 0.0020 0.4236 0.00 20.28 -0.00 18.2.0001 0. .90 2.0179 0.0 METRES • .0084 0.82 -0.02 -0.2 3.0096 0.0 AND 4.4407 0..43 17.21 -0.0467 0.20 16.7 -1.9 80.0186 0.0510 0. 13 .0000 G.7 95.07 -0. 0071 0.2 3.0147 0.4 163.9 2.0642 0.8 222.5 1.9 -2.91 15.0230 0.5 -4.0007 0.0707 0.0000 0.0012 0. .07 23.00 8.0000 0.2 219.6537 0.0 39.0000 0.0000 0.00 7.

.JP c c r!i' c: ~c c e 18:00 G:e 18:00 a . c c c e c: 8 c:.4: c . c== e 7 18:00 c c 08:00 e c " ac <. ..........:.82 ..... c: c: c: e c: CB . c "'a "'9% c "' a c: e c:: c a 0.20 0. c c 0.10 0.. c 08:00 i c CA c 18:00 .... . . a c: E c: c: c 8 ... c "' " c:: ea Ce . • c: c: • e"' c: 18:00 c: - c a c: "' c: a '!!I . Test of the significance of coefficient for models in Table 5.. 4 c: .. . a:: c .. S' cC ·:: c..c:: Cl) I!! e c c 08:00 c c: a"' ....30 .. a c: c: F" 1!!1 c e 08:00 a II a c c:EI . c c B c "' c e E . . ... 6 .e . a B 18:00 c:: . •c c .e c c c !:! c c c:B c c c c c c: c: c c 0 5 N A 08:00 c .. c: c:: ..25 Probability of Being Insignificant Figure 5. c .4 0.. model ca C:& a 08:00 eeoc 1 c: ce e . ....... c c:: c ....15 0.. a a a <c .00 c c c E c:: .. c j=_cc .. c: c c a c: Cl) ~ fill a a 3 .05 0. ad!' c e c:ac c.

5. and iS produce not it very close can be results in spite of their different appearance. For results (all 35 developed least in order intervals for to compute squares estimation c of For model converge.3) to UNB trigonometric measured method on between each pair of the three bench marks at the South-Gym area. then c software was the based on either the coefficients 0.7 show the spline fitted plots and the models i4 the a and b were estimated by least i7.3 m.6 m and 1.10 and 5.6 and From these seen that 5. = -1. corresponding tables.discrepancies between effect versus the geodetic levelling and the UNB-method).9.2 m). the results results were used to compute the section using the ~ Tables 5. full range using equation (3. whenever the was assummed solution did and the other coefficients were estimated through least squares.5.4. squares. model i1. Models i1 and i3 also give i6 gives very similar results. Model some discrepancies with the measured refraction effect. or (using the 0. a fixed value The other two coefficients.11 show computed refraction <.5 of comparison).4 field verification 2f The seven models in Table 5. 1985.83 ~ Comparison 5. large The same . -1/3 for exponent c.3) observed temperatures at heights assumption of have the the correction computation of exponent or the to two Figures of these results.2. 5.4 refraction correction (see height differences July 23 and 24. of the measured values 4. 5.

12 c) seen from the for that correlation does not exist corresponding Figure will be discussed the correlation coefticient factor) the alternative hypothesis later. The t statistic with In (n-2) correlation The value 0.12 shows the The three cases. coefficient among correlation coefficient matrices high values of the results of expected (except for model 17.44 that can degrees of pass the freedom is given by e.g. the correlation are as the seven models asphalt-gravel line).44. the correlation coefficients between the measured refraction effect and the individual computed values using the seven models are generally significant. the third case (Table as can also be The reason 5. first two matrices.82 which is greater than t value from corresponding table of percentage points: . The In the highest correlation exist between models 14 and 15.44. Hamilton [1964] as t = I 2 (n - 2) r 2 1 . using = 2. is the smallest correlation coefficient test. the above equation: t For r = 0. for refraction effect similar to other models.84 Model i2 also results in values can be seen for model i7.7 • At the 1% hypothesis that r = 0 level of significance the null that r (r is can be rejected against ~ 0 if the coefficient is equal to or larger than 0. Table for the 5. 5.r where n is the number of observations.

85
t = 2.82

For

r = 0.43,

value.

t

>

t

33,0.005

= 2.75

= 2.74 which is smaller

than the

table

In total, the r values for models fl, 43, f4 and f5

are larger than

0.44 in the two first cases,

Table 5.12 a

and b, and for models f2, f6 and f7 are smaller than 0.44 in
the first case, Table 5.12 a.
The above discussion shows that the new proposed models
can safely replace

the Kukkamaki model when

temperature sensors is equal to or
number
choice.

of sensors

is 2,

then the

the number of

greater than 3.
model f2

is the

If the
only

To draw a firm conclusion about which model is the

best reperesentative of the temperature profile in the lower
atmosphere up to

4 m height and to confirm

the validity of

the new models, more investigations are needed.

86
TABLE 5.9
Refraction effect [mm] computed using the seven models
versus the measured value (BM1-BM2).

__ ZONE
TIME
NO (ADT)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
- 29
30
31
32
33
34
35
36
37
38

M0 D E L
i1

i2

i3

i4

i5

i6

i7

meas

----------------10:00 -2.0 -1.0 -1.7 -1.7 -1.8 -1.9 -2.6 o.o

11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
24:00
01:00
02:00
03:00
04:00
05:00
06:00
07:00
08:00
09:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00

5.3
1.9
-1.8
0.3
0.1
-0.6
-2.2
1.3
-2.3
-2.7
0.5
-1.7
-3.7
-2.4
-2.3
-2.3
-3.9

3.6
1.7
-2.0
-0.5
0.6
-0.9
-2.0
1.4
-2.1
-3.3
0.1
-4.0
-5.8
-3.9
-4.3
-2.4
-6.0

5.6
2.5
-1.1
1.4
0.7
-0.8
-2.2
1.2
-2.1
-2.8
-1.3
-1.8
-3.7
-2.4
-2.1
-2.2
-4.0

4.3
1.5
-1.2
-0.2
-0.2
-0.7
-2.1
1.5
-2.4
-2.9
-0.3
-3.1
-4.9
-3.1
-2.9
-2.3
-5.2

4.9
1.8
-1.0
0.3
0.0
-0.6
-2.2
1.7
-2.4
-2.9
-0.5
-3.1
-4.9
-3.0
-2.8
-2.4
-5.2

5.9
2.2
-2.9
-1.1
0.2
-1.6
-3.6
2.2
-3.5
-5.1
0.0
-6.0
-8.8
-5.9
-6.2
-3.7
-9.2

5.9
-0.6
-0.5
-3.0
5.2
-0.4
-2.5
2.2
-3.0
-2.7
0.1
-0.8
-2.8
-1.8
-2.7
-2.5
-3.0

-0.7
-1.7
-1.4
-1.5
-1.9
-2.1
-2.2
-2.4
-1.8
-3.4
-2.6
-2.0
-1.8
-1.9
-2.0
-1.7
-1.8

-2.6
0.8
2.6
1.5
1.1
1.4
2.3

-3.4 -2.8 -3.0 -2.9 -4.1 -2.0 -1.2
-0.3 0.5 1.1 1.2 1.5 1.4 -0.5
-0.4 2.2 1.6 2.1 1.6 1.4 -0.3
1.5 2.2 1.0 1.3 1.4 1.2 0.1
-0.1 0.1 1.4 1.2 -0.3 0.3 -0.5
0.0 2.3 0.6 1.1 -0.4 0.7 -0.5
1.1 3.4 1.5 2.0 1.5 6.4 -0.7

-0.5
0.1
-0.2
-0.7
-2.7
-0.5
-2.9
-3.2
-2.1

1.3
-0.7
-0.7
-1.2
-3.1
-3.4
-4.0
-2.5
-2.6

Mean

-0.8 -1.5 -0.7 -0.9 -1.0 -1.9 -0.6 -1.7

-3.4 -3.9 -3.1 -3.5 -3.6 -5.9 -4.0 -2.4

-0.7
0.2
-0.2
-0.8
-2.6
-1.9
-3.0
-1.8
-2.3

0.1
-0.2
0.6
0.0
-2.4
-2.1
-2.6
-0.9
-1.8

0.0
0.0
0.7
0.0
-2.3
-1.9
-2.6
-1.0
-1.8

1.4
-1.1
1.1
-0.1
-3.0
-3.1
-3.9
-1.6
-2.6

4.7
0.5
-0.1
-0.2
-2.4
-1.1
-4.8
-3.1
-3.1

-2.1
-1.2
-1.8
-2.5
-2.5
-2.5
-2.5
-1.7
-1.7

------------

87
TABLE 5.10

Refraction effect [rnm] computed usino the seven models
versus the measured value (BM2-BM3).

ZONE
TIME
NO (ADT)

M0 D E L
#1

#2

#3

#4

#5

#6

#7

meas

---------------------------1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38

9:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
24:00
01:00
02:00
03:00
04:00
05:00
06:00
07:00
08:00
09:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
Mean

1.6
-3.2
-3.1
1.5
-3.0
-6.1
0.3
2.7
-1.1
1.2
3.4
0.1
2.1
3.5
2.2
2.1

0.3
-2.9
-1.9
0.5
-0.8
-4.3
0.5
1.4
-1.7
1.9
3.4
0.7
3.7
5.3
3.7
3.3

1.5
-3.9
-3.1
0.4
-3.5
-6.2
0.2
2.3
-1.0
1.2
3.4
2.0
2.1
2.9
2.1
2.1

1.0
-2.4
-2.2
0.9
-2.2
-4.8
0.8
2.6
-1.8
1.3
3.5
0.6
3.7
3.9
3.0
2.9

1.1
-3.0
-2.7
0.7
-3.0
-5.6
0.5
2.6
-1.9
1.2
3.6
0.8
3.8
3.8
2.9
3.0

0.6
-4.2
-2.8
1.1
-1.8
-6.8
1.1
3.1
-3.0
2.4
5.4
0.9
6.3
7.5
5.6
5.2

1.5
-6.4
-1.0
0.4
-0.7
-6.8
0.0
-1.1
-1.8
1.1
3.4
0.1
1.7
2.7
1.8
2.0

-1.4
-1.4
0.2
-0.5
-0.6
0.6
-0.3
-1.8
-1.2
0.0
1.1
1.7
1.7
1.9
1.9
1.1

3.2
2.6
2.8

5.3
4.4
3.0

3.2
2.7
2.8

4.2
3.6
3.1

4.3
3.6
3.3

7.9
6.7
4.9

2.5
2.6
3.4

2.0
1.6
1.0

1.3

2.7

1.4

2.0

1.8

3.8

0.5

0.4

0.2
-1.3
-2.1
-2.3
-3.2
-5.2
-0.9
0.7
0.8
0.7
2.6
0.4
3.8
0.4
3.0

0.0
-1.3
-1.6
-2.1
-2.2
-3.1
-2.9
-0.5
-0.8
1.1
2.8
3.4
3.7
1.6
2.7

0.2
-1.6
-2.5
-4.9
-4.0
-6.5
-1.1
0.4
0.8
0.8
2.5
1.7
3.7
1.3
2.6

0.5
-0.7
-1.4
-0.1
-2.2
-3.8
-1.2
0.8
0.4
0.0
2.6
1.5
3.8
0.1
1.5

0.2
-1.2
-1.9
-1.8
-2.8
-4.7
-1.3
0.6
0.6
-0.1
2.4
1.4
3.9
0.5
1.8

-0.9
-2.4
-2.5
-3.1
-3.5
-5.3
-3.9
-0.1
0.2
-0.5
2.9
2.2
5.5
1.0
2.2

-0.4
-1.9
0.4
0.2
-4.0
-5.9
-6.5
-3.8
1.9
0.3
2.3
1.3
4.2
0.4
3.2

-1.0
-2.0
-1.6
-1.5
-0.5
-2.0
-0.5
-1.7
0.2
0.4
-0.3
1.5
2.3
2.6
1.7

0.3

0.8

0.2

0.7

0.5

1.0 -0.1

---- --- --- --- --- --- ---

0.3
--- - - -

8 1.0 0.9 -1.7 -0. 1 -1.1 -0.9 0. 9 0.5 0.0 -0.2 .2 -0.6 0.9 0.8 -1.5 -1.8 -1.1 0.-- -2.2 -1.2 0.9 2.8 0.2 0.5 0.3 -.2 1.2 0.3 -0.4 0.2 -0.6 0.5 2.5 1.0 0.0 0.1 0.9 0.9 -1.3 1.5 -1.0 0.5 1.8 1.6 -0.8 0.1 0.2 1.7 -1.1 -0.2 0.-1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 9:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00 18:00 19:00 20:00 21:00 22:00 23:00 24:00 01:00 02:00 03:00 04:00 05:00 06:00 07:00 08:00 09:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00 18:00 19:00 20:00 21:00 22:00 Mean 0.0 0.1 4.0 1.6 0.9 -0.0 -0.1 2. 2 1.11 Refraction effect [mm] computed using the seven models versus the measured value (BM3-BM1).9 0.0 0. 7 0.1 6.3 2.2 -2.0 -1.0 -1.5 2.8 o.5 0.3 1.1 0.0 0.2 1.4 0.2 0.9 0.2 -0.9 0.4 0.3 -1.2 -1. 0 1. 5 1.2 me as 0.2 0.2 0.-.6 -0.9 -0.6 1.0 -0.1 -0.9 -0.1 0. 3 0.1 0.1 -0. -0.0 0.6 -2.1 0.9 0.4 -0.4 0.2 0.1 -1.6 -0.9 -0.4 -0.1 0.-.1 0.7 0.2 -1. 1 -0.0 -1. 6 -0.9 -0.2 1. 9 0.0 0.3 0.2 1.1 -0.5 -2.5 -0.2 -0.8 -1.2 -0.3 1. 3 2.2 -0.0 -0.1 0.3 0.6 1.0 -2.3 -0.1 0. ZONE TIME NO (ADT) M 0 DEL i1 i2 i3 +4 45 +6 i7 -.-.8 1. 0 0. 7 2.8 0.4 2.8 2.3 -0. 2 1.7 0.0 -0.5 0.2 0.3 -1. 6 0.5 0.9 -1.9 0.6 0.3 0.2 -1.6 0.1 0.2 -1.2 0.-.4 1.5 2.7 -0. 9 2.7 0. 3 2.6 -0. 8 2. 1 0.0 -0.2 -1.3 -0.4 -0.4 2.-- -.9 2.6 0.0 2.1 1.2 0.0 -2.0 2.0 -0.0 -0.1 1.1 -1.4 0.1 0.1 2.0 3.2 0.0 0.7 -0.2 0. 1 1.7 0.1 0.6 1.3 -0.9 1.8 -0.0 0.1 1.1 0.4 0.1 0.9 2.0 0.1 0.2 1.88 TABLE 5. 3 2. 3 0.3 -0.9 3.3 -0.1 -0.5 0.3 0.2 2.2 -2.6 -1.5 0.2 0.1 0.3 0.9 2.3 0.8 1.5 -0. 1 1.3 5.5 0.3 0.4 1.1 -2.3 -2.0 -1.9 0.9 -0.3 -1.2 1.7 -0.9 -0. 8 -1.2 0.4 -0.3 5.8 -2.4 -1.1 -1.2 0.0 -0.3 2.3 0.8 1.3 0.-.4 -0.8 2. 8 -0.0 -2.0 0.9 0.5 0.2 0.5 -0.6 -0.1 0.7 -0.9 -0.5.2 1.-.1 0.5 0.7 1.3 -0.1 0.4 0.2 -0.0 0.0 -0.5 0.8 2.5 0. 4 1.4 0.-.8 -1.7 5.0 0.2 -0.

86 0.59 1 0.97 0.48 1 0.58 1 1 -.97 0.94 0.34 -.56 1 0. 1 1 a: gravel-grass (BM1-BM2) line model #1 model #2 model #3 model #4 model #5 model #6 model #7 measured 1 0.94 0.82 0.94 0.25 .99 0.95 0.75 0.12 Correlation Coefficients Matrices model #1 model #2 model #3 model #4 model #5 model #6 ·model #7 measured #1 #2 1 0.92 0.79 0.10 1 0.95 0.49 #6 #7 0.92 0.82 0.96 0.92 0.94 0.98 0.92 0.93 0.97 0.36 1 1 0.74 0.81 0.77 0.03 1 measur.50 1 0.94 0.91 0.02 c: asphalt-gravel (BM3-BM1) line .68 0.89 0·.89 0.90 0.93 0.42 #3 MODEL #4 #5 0.84 0.78 0.67 b: grass-asphalt (BM2-BM3) line model #1 model #2 model #3 model #4 model #5 model #6 model #7 measured 1 0.96 0.97 0. 79 0.10 1 0.97 0.18 -.56 1 1.38 -.97 0.87 0.95 0.89 TABLE 5.68 0.52 1 0.91 0.89 0.05 1 0.75 0.79 0.99 0.89 0.94 0.00 0.96 0.82 0.89 0.77 0.10 -.74 0.95 0.02 1 0.88 0.41 1 0.60 1 0.04 1 0.71 1 0.38 -.

~ \: -2 0 ·n .' ..5: t-1odel #4 ···---· Model · . #3 #7 (h] ...._I /' I ../ 1-1 -4 ~ 4-l Q) 0:: . I .. .. /..-. .. ... I ' 1 2 ~ \ I ./ I '..0 0 \ '.. ' ~ -o... \.Model #I #5 12:00 ----.Model ----Model Refraction effect computed using the seven models versus the measured value for BM1-BM2. \ \ t!l 4-1 4-1 I J 1... ~ ~ ...'.~..jJ 0 () Q) ·.. -.... \_/ I' -6 I I I .jJ C) ctl I ~~-~\~ \ i ..6 I I' I 1 I I 1 I I I 4 . ..I I I \ I ·.. .1 II .·. \ ' ·./' -o 0 12:00 LEGEND• 04:00 20:00 o-El--a Me115ured Figure 5.. .Model -----· Model #2 #o 2o:oo Zone Time ..

. ..1-l 1.. I fr... \ ' ' \ 'l!l. cO ~ ~ (!) ~ -4 I I ·6-:J I \/ -6~------~------~-----~~----~--------~-------~------~------~------~ 0 20:00 12:00 04:00 12:00 20:00 Zone Time [h) LEGEND• o-EJ-<3 Me~5ured Model Figure 5.... f.. '" 1 0 ' 1. __ .. \ \ "' h .6 ----..1-l J.. ' ('.: ''.6: #4 Model Model #l #5 Model 11odel #2 #6 Model Model #3 #7 Refraction effect computed using the seven models versus the measured value for BM2-BM3. \• • 0 ..... \ ~' .."' ·. \_../ I• \ I/.'" //. A.. -'' I I~J ·"·' I: ~.. a.\ I i X ~' ). / ~- ...-. ......

Related Interests

\ ' . ..t!l·-\c~ •h ' 6 6 \ ..?" '~ I ' "\h----.. ..'·-:.o-.~ ' _. \ ~-. 1 ' \ \.. \ ' ~' .

Related Interests

' . I ... .:\ I I 4 (!) \ I.' I ~ 0 -2 ·~ . . \. .J I I . I • "' I .: (! \ ~- . . -...:..: --. -p~-... /' .. I .. \

Related Interests

\.. 0 ###BOT_TEXT### . ~ • '19..

...-i 10 •\ '-\..' ~ / ''T:J .l ~ \ li'r:l ###BOT_TEXT### N \~~... +" I . .· .1 I ~ ril P I -.4 ''\ G:l\~ /j':~' : I '.I -2 " t ~ Q) ex. ...-. #3 #7 ..Model ··-··· Model #2 #6 20:00 Zone Time [h) ---Hodel ----Model Refraction effect computed using the seven models versus the measured value for BM3-BM1.Model #l #5 ----... >-! ~.Jill v. • •..:.\ 'I 'f/1•I ''.. •• m I 1 ' 1 1 I I \ I I

Related Interests

\ : :f.. I 1 I 2 I <3·-0·. . 'o •..a. 'q .·•r~·~\: . .7: 12:00 04:00 • • · • · •· Model · . : ~ '\ ~ .· '{ . I I ..._.. '• '0 .... \ 0 a 0 . I \ ' I () ... ' I' () Q) ' I ~. 19.. • lll..I ~ ~ ... \ I I I \ I -4 -E> 20:00 12:00 LEGEND! D·EJ·-!3 Measured Model #4 Figure 5..' \ . .

8 -2.6 -2. [mm] - NO mean of 2 tar LOCAL TIME 2.lm tar 3.6 -1. the between Table 5. windy Sunny periods.5m tar 10:25 10:45 11:02 11:16 11:32 11:54 12:12 12:30 13:01 13:15 13:47 14:00 14:13 0. Only the height difference between repeatedly for four shows the BMl and BM2 was measured hours from 10:20 to discrepancies 14:30. .2 -3. windy Sunny periods.5. 72 -1.9 -1.8 -0.6 -1.2 -0.01 1. TABLE 5.57 -1.93 5.9 -0.9 -2.5 -4.6 0.2 -1.25 1.2 -0.9 -0.4 -1.31 REMARKS cloud cover \ condition 1 2 3 4 5 6 7 8 9 10 11 12 13 Mean s.5 -0.9 50 50 50 70 90 100 100 75 50 50 75 75 75 Sunny Sunny Sunny.5 -1.4 -1.5 1.1 Tests on 20 June The first test ~ was performed on 20th of June as a preliminary observation preparation for the long term tests.0 0.13 Preliminary test measurements using UNB trigonometric method at South-Gym area from BMl to BM2 Geod.1 -2. windy Windy Windy Windy .0 -3.1 -1.8 0.4 -2.1 0.9 -1.3 -0.1 -2.95 1.2 -0.9 -1.Trig. Windy Windy Windy Breeze Windy Sunny periods.4 0. D.0 -2. -0.0 -1.7 -4.3 -1.13 geodetic height differences and UNB method trigonometric height differences.1 -0.5 Computed Versus Measured Refraction Effect 5.

The meteorological were: air temperature measured distances. (the middle region 3. inversion with light below this range.94 The day on which the reported as windy.g. and observations done and pressure on 20 June 1985 for correcting a record of cloud the cover and wind.s. Webb. (see one windy day [1984]). cover.g. e.3 for of of thermal more detail) extends usually to more than 30 m (one Obukhov length) and according to theories temperature. first three rows show positive usually longer The results of the values for the lower targets and negative values for the heigher targets. be due to inversion of but later it was found refraction effect can change sight regardless of On a the windy day.6).1. stability. on a cannot appear days. Obukhov length observations were carried out was In a moderatly windy day of summer.1.2 is Greening. However momentarily inversion first three metres of cloud on air is less of calm than 2 of temperature the atmospheric this does not necessarily mean that the lower targets will result in a sign of refraction effect different from the results of the higher targets (see also section 5. when m. . the defined in section than 30m (see e. of summer may expect But. the horizontal movement of gradients within the layer. thought to gradient. 3.3. see section the temperature that the magnitude of the sign with the sign of the region II This was first elevating line temperature gradients. [1985]).

It was supposed to continue for 24 weather conditions it hours. BM2 and BM3 were measured repeatedly over the whole 13-hour period. For all other tests the wind velocity and direction.15 and using show the obtained the can be interpreted as effect by assuming levelling as being errorless.95 No temperature gradient was measured. and temperature of measured and cloud cover this information was useful the changes of gradient was used to compute the was recorded. the possible their only the temperature magnitude of refraction effect. for understanding atmospheric condition and if correlation with the refraction error.16 differences levelling discrepancies refraction 5. temperature gradient.14.5. Tables between 5. atmospheric pressure. but due to unfavourable was interrupted after 13 hours of continuous observations. 5. The height differences between BM1. the discrepancies using precise UNB-method. ground surface were Although all humidity. These mostly due results of to the geodetic . therefore the magnitude of refraction effect using the meteorological data could not be obtained for the first test.2 Tests Qn l i July ~ The second test was carried out on 19 July 1985. the geodetic height and 5.

96
TABLE 5.14
Discrepancies between the results obtained using
trigonometric height traversing and geodetic levelling for
BMl to BM2.

Geod.

N0
1

2
3
4
5
6
7
8
9
10
11
12

-

Trig. [mm]
mean
of 2
tar

LOCAL
TIME

2.lm
tar

3.5m
tar

11:13
12:09
13:24
14:54
15:49
18:05
18:56
19:31
20:09
21:56
22:47
23:25

-0.1
-2.0
-1.4
-2.7
-3.6
-1.7
-2.4
-0.8
-2.6
-1.7
-1.9
-0.7

-0.9
0.5
-2.7
-2.0
-1.2
-1.5
-0.8
-1.0
-0.3
-2.2
-0.7
-0.7

-1.80
±0.97

-1.13 -1.45
±0.87 ±0.64

Mean
s. D.

Since the profile
line is

the same as the

expect that over

-0.5
-0.7
-2.0
-2.3
-2.4
-1.6
-1.6

-0.9
-1.5
-1.9
-1.3
-0.7

REMARKS
cloud
cover
condition

'
75
100
50
75
100
100
100
100
100

Fair
Windy
Windy, little rain
Little wind

After a period of rain
Cloudy
Cloudy

of surface on the

fore-sight of one

back-sight of the other,

a long period the misclosure

height differences would be near
for the lower targets is

zero.

-0.03 mm

i.e.

one can

of the three

The mean misclosure
as expected.

But

the mean misclosure for the higher targets is -0.85 mm which
is considered as too large.
Column 2 of Table 5.15 shows that the mean discrepancies for
target 3.5 m is a negative

number.

This value was expected

97

TABLE 5.15
Discrepancies between the results obtained using
trigonometric height traversing and geodetic levelling for
BM2 to BM3.

REMARKS

Geod. - Trig. [mm]

N0

1
2
3

4
5
6
7
8
9
10
11
12

mean
of 2
tar

LOCAL
TIME

2.1m
tar

3.5m
tar

11:31
12:28
14:00
15:16
16:04
18:29
19:07
19:41
20:21
22:18
22:59
23:35

-1.5
-1.6
2.2
0.3
0.0
0.3
0.8
2.0
2.3
1.1
0.7
0.6

-1.8
-2.7
-0.8
-0.5
-0.1
-0.1
0.4
0.7
1.7
-0.5
0.6
0.5

0.60
±1.26

-0.22 0.19
±1.18 ±1.13

.Mean
S. D.

cloud
cover

condition

%

shimmers on BM3
75
Shimmers on BM2.
25
Windy
50
Windy, little rain
100
Little windy
100
100
100
100
100
100 After a period of rain
100
Cool
Little windy
100

-1.6
-2.1
0.7
-0.1
-0.1
0.1
0.6
1.3
2.0
0.3
0.7
0.5

to be positive. The computed refraction using the profile of
the lines and

measured temperature gradient in

Table 5.17,

indicates positive mean values for both targets.
This
RB-RF > 0

can

be

AHG-AHT < 0

then

AHG-AHT > 0.

The

lower targets
The temperature

explained

first case

and the second
gradient on

ground (0.3 m to 1.2 m)

by

using

and

Figure

RB-RF < 0

when

had mostly

is larger

then

occurred for

case for the
the grass

When

5.8

the

higher targets.

field close

in magnitude

to the
than the

98
TABLE 5.16
Discrepancies between the results obtained using
trigonometric height traversing and geodetic levelling for
BM3 to BM1.

REMARKS

Geod. - Trig. [mm]
NO

LOCAL
TIME

2.1m
tar

3.5m
tar

mean
of 2
tar

cloud
cover

11:49
12:46
14:17
15:36
16:32
18:43
19:19
19:59
22:35
23:12
00:05

1.5
2.7
3.1
1.2
1.7
-0.2
0.2
0.4
1.1
0.8
0.4

1.9
-0.3
1.8
0.4
0.9
0.6
-0.1
-0.4
0.0
0.0
0.7

1.7
1.2
2.4
0.8
1.3
0.2
0.1
0.0
0.5
0.4
0.5

75
Breeze
Sunny
0
95
100
Little rain & wind
Little windy
100
100
100
100
100 After a period of rain
100
100

1.17
±1.03

0.5
0.83
±0.79 ±0.75

\

condition

-

1
2
3
4
5
6
7
8
9
10
11

Mean
S. D.

temperature gradient over asphalt close
it is smaller when going higher
4 rn).
cannot

This may explain the
explain the

higher targets.
was not

repeated in

nearly zero.

and

above the surface (1.2 rn to

different signa of values,

large miaclosure

However,

to the ground,

of

-0.85 mm for

but
the

this rather irregular misclosure

the three other

cases and

was always

1 1.4 -0.6 0.5 0.5m Mean 2 .5 -0.6 0.83 +2.5m Mean 2.2 0.65 +1.3 0.-.2 -0.8 -1.7 -0.6 2.0 -2.7 -0.2 -0.2 -0.99 TABLE 5.7 4.0 1.5m Mean -.-.5 -1.-.0 0.3 -3.4 -0.2 -0.2 0.4 1.3 0.6 0.1 0.-.11 -0.17 Computed refraction using measured temperature gradient Computed Refraction [mm) LOCAL BMl to BM2 BM2 to BM3 BM3 to BMl NO TIME 2.1m 3. D.7 1.2 0.33 :t0.85 .9 -3.6 0.0 -0.-.0 -0.3 -1.7 0.7 0.7 1.6 -1.1 -1.8 -1.2 -1. 2.7 -0.1m 3.9 4.6 -2.-.-- 1 2 3 4 5 6 7 11:00 12:00 13:00 14:00 15:00 16:00 17:00 8 18:00 9 19:00 10 20:00 Mean s.2 0.2 -0.1 -0.4 -0.4 1.9 -1.6 -3.4 -0.84 4.-.7 -1.9 0.1m 3.2 1.8 0.

RF) < 0 THEN (6HG .6HT) < 0 (RB .100 6HF 6HG 6HT 6H 6HT I) If II) If height difference using geodetic levelling height difference using the leap-frog trigonometric method 6HG = 6HB + 6HF = 6HB + RB + 6HF .5. was measured during profile of the the lines and .8: Back.RF) Figure 5.6HT) > 0 (RB .and fore-sight magnitude of refraction difference 5.RF = 6HG + (RB . the South-Gym area started at of air Knowing the to the next day until continuous observations.RF) ~ >0 THEN (6HG .3 The third test survey in 9:43 on 23 July 1985 23:50 for The a total of and continued 38 hours of temperature gradient entire test period.

is defined by the regression equation (the cr. the refraction effect for all three lines was computed for one hour intervals. The relationship between the refraction.6HT).3 + 1.9 shows: 1.12) .101 using equations (3. and the correlation is more pronounced in the latter for BM2 to BM3.).10 has the expected hypothetical regression mr = cr).10): mr = 0.3.67). (the solid line) A fitted linear regression line in Figure 5.9 parts a and b.2 cr (5. The correlation between the two can be seen in Figure 5. The computed refraction at one hour intervals using the observed temperature gradients.8. The measured discrepancy between using geodetic the refraction the height levelling and UNB-method of effect which difference the results trigonometric is the determined obtained with height traversing (6HG. Figure 5. and computed solid line in Figure 5. small deviations from line (the dashed line: This figure shows the measured refraction effect against the computed refraction effect for the line BM2-BM3. measured. The error bars in this figure the observations and account other errors are based on the precision of they have been increased to take into (this will be briefly discussed in section 5. mr.66) and (3. 2.

Using these graphs and considering the corresponding time of neutral condition when the refraction fluctuation around this refraction and angle effect is expected can be observed at corresponding difference is zero point is computed simply neutral condition measured the to be almost the mostly because by subtracting time vertical angles angle of zero. 5.102 = 0. and from all the other assuming The of angle the of . r =0 of significance Ho : r At the • 1% level r is rejected and thus Ha ~ 0.11 found by detecting 3. Figure temperature 5. i. In order to investigate the computed refraction we instrument station assessment of may difference of measured and consider to the bench the refraction the lines from marks separately.2) from the temperature depicts the approximate time of neutral condition for different sufaces and elevations.4. computation over the A good each ground surface is also possible by examining them individualy. This is almost the The lowest measured and same for the correlation exists the computed refraction between the effects for the line BM3-BM1 (see Table 5.60 with correlation coefficient. can be (see section Figure 5. i.1. refraction.12 gradient shows the crosses the fluctuations of zero observed vertical angles to the lower targets.e. The neutral refraction angle condition measurements.12). correlation exists.e. can be accepted (for details of the test see section line BM1-BM2.4). when the periodic line.

12:00 20:00 04:00 12:00 20~00 Zone Time [h) Figure 5. . ! I ~ -· II ~ I b I ~ 0 ~ "-' 0 m ~ 4-l <l) o:. b. BM2-BM3 and c. BM1-BM2.9: Measured refraction effect versus the computed value. BM3-BM1. a.It m: j 103 computed measured 2 4 e ~ 1 -2 1 I.

10: The measured refraction effect [mm].3. its accuracy due to the about 0. introduced this way is as precise i. This angle of refraction is converted to linear refraction and compared with the computed refraction for all three lines in Figure 5.104 4 :&• .e. refraction found in vertical angle.13 • In all three plots the strong . of the bias detecting the neutral conditon or the zero refraction angle time. The discrepancies of height difference determined by trigonometric height traversing and geodetic levelling for BMl-BM2 line versus the computed refraction error for the same line. The dashed line is the hypothetical regression line. and the solid line represents the actual linear regression.6" can be (see as the measured section lower because uncertaintity for 5. however.8). ~ e e 2 • ~ • •• • 0 ~ ~ ~ ~ ro m ~ • • -2 • • ~ • • ·~ • • • • e -4 -2 -4 4 2 0 computed [mm] Figure 5.

refraction effects. the correlation coefficient. Figure 5. These plots are prepared for the mean corresponding values of the two targets. the regression equations show that the computed refraction effect. cr. According direction of the to Webb [1968] this can . is too large compared to the measured refraction effect.105 correlation between the two computed and measured refraction effects can be seen. 2.18. Specifically over asphalt. The wind was blowing almost in the line of sight to BM3. The site measurements chosen to was too make the temperature close to the gradient ground covered by gravel (about 2 m) and was not characteristic of the road on which reason for the line of sight to choosing that spot transportation to move BM3 was extended. In Table 5.18 gives and measured effect. mr. was the lack of the equipment from one The proper site to another. equations relating and at hypotheses the computed to the the that 1% there level is which exist of zero Table linear regression measured refraction significance the correlation null between the measured and the computed refraction effect (Ho : r = 0). the is measured refraction smaller than the effect computed one.14 gives the linear correlations between computed 5. explanations of why the results more than There are four two times possible of asphalt are so different from the other two surfaces: 1.

11: Fluctuations of point temperature oradient a. gravel....1 .5 2 2.8 E ' 0.106 a -0. grass and c. 0 . b.4 u 0.. · -0.5 3 Height Above the Ground (m] Figure 5. -· c -0..---~--~---r12:00 20:00 04:00 20:00 12:00 Zone Time [h] ----.l.- ._: ·.6~--~--~---r--~--~--~~--...· -.4 ·. ... asphalt...

and c.. .. \ 0 ~ 0 '• 36 b I 12:00 20:00 04:00 • 12:00 20:00 Zone Time [h] Figure 5. BM3 (asphalt). BM2 (grass).. ~ a c:: «3 ... ~ \ \) ~~\ . ... ! \f!\A . I~A~~l 4i' ~ 39'J ~~~ \.107 Ul G) 17 .....12: Fluctuations of observed vertical angles a. b. BMl (gravel).

... '. IS-BM2 (grass) and c.: . a. ~ _.. ' ... ·' : . . -....... .. . ... {a) Gravel . " ' ' ' .: .. . ...13: Computed refraction effect versus the measured value. . (c) Asphalt I ' r I '' 12:00 20:00 04:00 12:00 I ' 20:00 Zone Time [h] Figure 5.• ~~ ...... ' ... . ... ..'. ..' ~ ·. IS-BM3 (asphalt}.. b.108 computed measured mm -4 ... ·... I\ '. .' .'· I .. e e .. -4 ......• f··'\ 4 .. IS-BMl (gravel}.. .·.' 2 (b) Grass " ( .'I I .

for the as can be BM3-BM1 weak correlation equations in Table 5.12 It was Figure previously mentioned that. The correlation coefficient improved to 0.109 result in lower (computed) value. the traffic on the during road could also result in a mixing of the atmosphere and a reducing of the effect of The lower refraction effect over asphalt can refraction. the day. 5.18 computed refraction two equations. The significant refraction effects refraction or trigonometric against the correlations (either derived from discrepancies methods of height computed detailed knowledge that of refraction the from the between differnce is measured angle geodetic of and determination) mostly we have about the two due to the surface profiles and to the long term temperature gradient measurements. refraction effect than the expected Other than the wind direction.9 levelling c. from the small be noticed fluctuations of the vertical angle in comparison to the other two in Figure 5. .12).68 . given cr computed then mr will be the corrected refraction effect. significant improvement in the correlation found between the two after was as A coefficient was the corrections were made. discrepancies and trigonometric versus the asphalt gravel. corrected according to these the seen in the geodetic traversing of have very Using the regression and between height computed refraction (Table 5.

IS-BM3 (asphalt).::s 0. -2 '0 (I) . b.--.. -2 ::l • p. e e 0 0 •• • '0 • Q) ..14: -2 0 2 measured [mm] Linear correlation between the computed and measured refraction error.. -2 e 0 e 0 0 • • •• -4 0 -4 -6 -4 -2 2 0 measured 4 2 measured [mm) -2 (c) Asphalt •• 2 e e 0 -4 [mmJ •• •• 0 '0 Q) . IS-BMl (gravel). IS-BM2 (grass) and c. e 0 0 • -4 • -4 Figure 5.. ==' 0. ...110 (a) Gravel (b) Grass 4 • • • 4 • 2 • 2 g .... a.

31+0.005 -- -- ** gravel BMl 0.75 -* t -- and the table values for t at 0. of the height traversing over a temperature can be very realistic.2.13+0. results 33.3 > 2.7 R asphalt BM3 0.111 TABLE 5. 0 2.0 significantly high.18 t-test on the significance of the correlation coefficients Surface from Corr.05 level are: t = 2. IS Coef.005 work. to * Regression Equat.0. known is accepted. H >t t 33.75 R grass BM2 0.23 cr > 2.0.7 > 2.75 5.36+0. trusted in (see that than the practical and may longer cause too 100 m section simulations in 4. cannot be refraction effect in trigonometric known ~ not accurate enough trigonometric the r = t 33.87 mr= -0.51 cr 10.025 lines traversing proved profile with profile.01 = and 0.3). R gradient along the and useful information may .40 cr 10.87 mr= -0.66 mr= -0.0.75 and ** Ho: r =0 is rejected and Ha: Although the correlations are computed refraction error since large it is corrections for height However.

AH n is the simultaneously taken) height (5. details see Chapter 6. Estimation of the standard deviation of one ~H is given by Chrzanowski [1985] a= [ where ~i ] -1/2 2 I 2n d. synchronized in taken time.4 The last test at the South-Gym area was using two independent theodolites separated 29 July. one may expect higher precision than what has resulted from this experiment.5 Since the their not but hours starting measurements were differences were at not affected by short-term fluctuation of temperature gradient. 11:45 ending were independently a duration for at 17:07. Thus.112 be extracted from For more such simulations. number and An d~H of is average observations (usually the difference of the two standard deviation of . the Two electronic theodolites were used.13) i traversings.5. The main purpose to get sufficient carried out by about 2 m on in making these measurements was data to estimate the actual precision of trigonometric height traversing without the influence of the refraction error. of 5. and observations simultaneously. Tests on 12 ~ ~ and estimation Qf standard deviatipn of vertical angle measurements 5.

18 • Considering the errors involved in both computing the refraction effect. measurements carried out in and 5. 3. five groups indication {only of values that the for these agree one can see that the above closely. equations in particular measuring and lines) Table This can be 5. computed presented in measured on 4.58" for a that corresponds to a zenith angle sight lengths of 200 m.79 mm was found of 0. and corrected Table 5. standard deviation measured in four sets with The bars plotted in Figure 5. extracted from measurements carried out in 23 July 1985 and corrected using equations given in Table 5. measured using electronic theodolite #1.18.15 shows the following refraction errors: 1.9 were computed considering the above estimated standard deviation as well as the contribution of other sources of errors such as a change of the height of targets due to the expansion or contraction of the rods. Figure 5.18 are useful and can improve an the . 2. are in the discussed in distance detail by Chrzanowski These errors [1984] and Greening [1985]. extracted from 24 July 1985 and corrected using equations given in Table 5.18 according and to equations temperature gradient 29 July 1985. measured using electronic theodolite #2. non-verticality of the rods. and errors measurements.113 0.

4). 0 0 0 ~ 08 8.J 0 ~ 0+-----~~=----------=------------~--~~~r-------~0 ~ G) ~ 0 6. f::::P 6 8 "8 Q Q !:. b.. [:] 6 •6 0 S 6 0 0 0 0 08'0 0 0 8. 0 0 8. -2 0 0G 0 ~ 0 G) ~ ~ tll s::: 2 0 0 0 "M +. a. • 0 0 • ~ 8. 1::... BM3-BM1 (a key to this figure is given in section 5. BM2-BM3 and c. . 6 0 0 0 t::.114 A: 1 0:2 0:4 o:s • 2 0 1::. 0 -2 <b 0 • 0 0 El • 2 0 . BM1-BM2...p 8.:. 0 0 08 0 0 -2 12:00 14:00 16:00 Zone Time (h) Figure 5. 0 0 r-o e e .... 6 ' 0 06 68 .5. 08..15: Measured refraction error versus the computed value.

12)) .93 and 0.115 computed values. 5. computed refraction from mid effect for individual point to the three bench than these estimated (previously the knowledge along refraction effect of the variablity the line account. a short time after sunrise. effect from one day to another as long as the measurements are carried out within the same portion of and day over the same profile under similar weather conditions. from one bench mark to another) the over estimations mostly cancelled out (see equation (5. 0.5 Comments on South-Gym The refraction treating the the neutral The effect is surveys successfully measurements carried out condition time as computed correlated ~ values with the using measured correlation coefficients of durino (or free from this by near to) refraction error. BM2-BM3 and BM3-BM1 respectively. a short time before sunset. estimated are strongly effect. Also. can of sight which is larger called "measured" based on detection of neutral condition) values.83 with for lines BM1-BM2. as expected. of However.e. realized that the refraction it is is almost the same. . of marks) lines be The over estimation due to insufficient the temperature could not gradient be taken into in the computation of a full line (i. and in the afternoon. The preferable times for observations durino clear days are early in the morning.e.78. procedure refraction 0. The (i.5.

2.5 m height) are less affected by refraction.5. which.15 and 5.6 ~ The final test survey test line. between zenith angle- started at spot out in each setup and distance 12:15 and ended at Figure 5. effect height levelling method at a about 0. 5. but this does not necessarily always cause randomization of refraction error. Tests on QQ August 5. temperature measurements to BM4.1 m height) different from the higher targets.2. some of the cases the discrepancies listed 5. sian for lower targets (at 2. 1. on asphalt 39 as consisted of the two South-Gym sets of Observations measured discrepancy using geodetic trigonometric 2. .14. given in section and which difference results is the determined of UNB (~HG.~HT).3 m from the about 20:00 on 06 August 1985.16 show a although in in Tables 5. 5. side-walk.13.16 shows: refraction the The concrete were carried observations. The computed refraction at 15 minute intervals using the observed temperature gradients. was carried out on the Head-Hall A description of the line is The change in temperature at different heights from the ground was measured using site for close six temperature sensors.116 Measurements to the higher target (at 3. The was selected setups of measurements in measurements.

profile of the baseline was well known and the temperature gradient was determined every five minutes. between BM2 and BM4 at Head-Hall test line. The average of the measured refraction effect (for the mean of the two targets) came out to be 2.117 o measured • computed .....7 mm. Assuming that the last observation refraction error (since it is the at 20:10 is free of closest possible . 0 • • 0 (. the computed refraction effect observed height differences generally It should be mentioned that the to correct the improves the results..2 mm and the corresponding computed value is 2.j 0 • 0 0 0 Q) ~ ~ • -2 Figure 5. ~ 6 4 0 t:z1 • 2 a 0 • 0 0 • •• ~G e 0 • • ••• 0 • 0 0 0 0 • • •• o c 0 • • 0 0 0 0 0 0 • • • • • • (:) 0 0 0 0 0 0 0 +---------------~~----------------------------~r.-~---­ v • 1-4 ~ Q) 0:: 0 c -~ 6m G{:l.16: Applying 20:00 18:00 16:00 Zone Time [h) 14:00 The discrepancies of height difference determined by trigonometric height traversing and geodetic levelling...

12). The refraction effect on the or less back-sight is oscillating more around the zero value and on fore-sight is always less than zero as expected. fore-sight.93 measurements carried can be assumed and then ground. total of the above two (measured.19 the refraction the was obtained following values of refraction effect: 1. This method of neutral condition detection can be utilized in precise measurements of monitoring the vertical movements of large structures. Table fore-sights shows 5. Excellent values agreement exists (11 and coefficient of 12 0.19) which the two with as a reference (free effect can The neutral temperatures at a condition different measured correlation substantiates out during the neutral the refraction measurements. back-sight. AHG .AHT (measured. 2. 3.118 condition time). tl). computed. observation to the neutral effect on and back-sights separatly. 5. measuring in Table between that the condition time from refraction error) be computed can for other be detected elevations above by the In this case the profile of the line is not needed. 4. . and 5. such as dams.

0 2.84 -2.25 2.5.9 31.65 3.5 -2.60 -2.85 2.70 A key to this table is given in section 5.55 1.20 2.24 -0.0 1.2 2.2 0.1 28.6.2 3.55 2.45 3.3 31.42 2.20 2.96 3.70 5.3 27.3 25.58 -2.4 1.4 31.03 -0.2 -2.13 -0.21 -2.0 3. [mm] [mm] (mm] Temp.29 2.83 -0.20 -2.3 1. -2.06 1.5 27.0 26.85 4.09 0.4 1.4 32.17 -0.66 -0.0 -2.55 2.89 -2.40 -0. Comp.1 30.1 30.6 30.7 2.21 2.71 4.3 1.5 2.0 0.35 3.09 1.0 -2.75 -2. Remarks calm and clear all day .0 31.8 -1. i1 42 Refr.05 2.5 -0.23 -0.91 3.0 -2. Refr.75 2.15 5.42 0.8 31.60 0. F.5 -1.65 3.38 2.15 2.97 1.28 -0.53 2.119 TABLE 5.74 0.5 31.19 Computed refraction effect versus the Head-Hall test line measured value for 2 3 4 5 I I Meas.41 -0.28 -0.22 0. Refr.59 0.4 3.1 1.4 32.92 1.85 2.30 -1.4 29.06 2.10 4.9 28.25 -0.49 -3.2 1.2 1.3 2.5 -2.0 31.2 29.3 29.25 2.09 0.40 -0.35 2.2 4.68 -0.27 0.05 -0.25 3.7 2. [mm] [mm] 0.0 3.24 0.08 0.45 -2.85 -0.23 0.06 -0.45 -0.7 2.03 3.30 2.3 30.16 2.75 2.3 0.25 2.50 3.6 2.74 0.51 0.51 3.25 0.55 -3.15 1.01 2.1 2.35 -0.61 0. 79 1.84 -0.8 32.65 -4.37 1. 1 NO LOCAL TIME 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 12:00 :15 :30 :45 13:00 :15 :30 :45 14:00 :15 :30 :45 15:00 :15 :30 :45 16:00 :15 :30 :45 17:00 :15 :30 :45 18:00 :15 :30 :45 19:00 :15 :30 :45 20:00 :10 Mean B.40 1.20 0. [oC] 29. Meas.6 3.-s.09 0.20 2.43 3.16 1.13 1.69 -0.4 2.85 2.-s.06 -3.53 2.7 29.9 3.8 29.3 31.50 -0.20 -2.3 3.0 -3.5 -1.12 -0.22 0.26 1.7 1.60 2.41 -0.8 31.20 2.9 3.2 -2.2 -1.87 -0.7 2.7 2.0 0.3 31.75 0.1 2.6 3.6 31.7 30.0 31.

.

.Chapter 6 SIMULATIONS Qf REFRACTION ERROR IN TRIGONOMETRIC HEIGHT TRAVERSING 6.120 - is less where K During . levelling and an terrain in 82 km long error in with temperature gradient. A number of new simulations has been done by the author with the same actual purpose as in previous levelling conducted data. 1978]. of the The simulations were cumulative influence height traversing. The results reported by Chrzanowski in details trigonometric of refraction in geodetic trigonometric heighting were simulated profile environmental helpful to understand the of refraction The influence extreme simulated along an assumed average of these simulations are [1984] and Greening [1985]. along a line The new of simulations but using simulations actual geodetic have been levelling of special order (the allowable discrepancy between independent forward and backward levelling between bench marks than +3 mm~K [Surveys and Mapping Branch.1 Simulation Alonq g Geodetic Levelling Line on Vancouver Island In 1984 a number of levelling simulations were carried out at UNB to assess the dependence of the refraction errors on the profile conditions. is the distance between bench marks in kilometres).

2 km km km km Mervi1le Kelsey Bay Merville Nanoose Bay Nanoose Bay Merville Kelsey Bay Merville Lines #1 and #4 are extended. parts for forward and The backward levelling traverses: line #1 #2 #3 #4 to from Forward Forward Backward Backward 108. levelling instrument to back- and fore-sight levelling rods were measured using stadia cross hairs. project on Vancouver The from the above data . The operations.5 115. geodetic levelling was carried out the Geodetic Survey of Canada.5 109.5 m and distances from the temperatures were 2.totally over a parallel to Georgia. Surveys and Resources Canada.7 115. out over a line ~224 into two Mines The Mapping Branch. of the line was generated for 10 m (or less) The profile intervals along the line based on measured stadia distances and height . and started in late May 1984 The middle of October 1984.121 the geodetic observed at heights 0. and was finished in the simulations have been carried km line chosen is divided Island by Department of Energy. Kitchener and Hkusam mountains with maximum height differences of about 300 m. the southern Lines #2 and shore-line #3 pass of flat terrain. the partially Strait over a of hilly terrain south of the Menzies.5 m in every set-ups.

The actual accumulation geodetic levelling has refraction effect been computed and the refraction accumulation was done of the using the same in trigonometric in simulation of height traversing temperature gradient and the same profile.1. the slope height of points was interpolated 6.5 for the profile with along of the lines. 5. Computation 21 the refraction error in geodetic levellina 6. adjacent every two adjacent a straight Assuming line turning points of the as being turning points were connected and then the maximum 10 m horizontal separation See Figures 6.2) . by turning points.122 differences of ground two between constant.50 [ 1 ----- c + 1 C+l ( 50 C+1 250 ) + 150 c . 2 (S/50) . Kukkamaki [1939a] to a form equation was simpilified by more convenient for computation using a hand calculator Cr = 10 -5 G .1) in which.2 to the line. At (6.1 The temperature differences measured during levelling were used in Kukkamaki's formulae for refraction correction using equation This (3.94 G = --------c c 250 . 200 ] (6.19). Ah .

19) given by Kukkamaki [Remmer.1 in all circumstances However. 1980] 2 S Cr = 6 where: the dn d4 (----)( d2 ~h dt dn/dt second and is + 3 ~h given by equation (3.1). c is the refraction difference Ah.5 m above the ground surface.3) ) 80 c-2 d2 and d4 are of Kukkamaki's .1. correction varies Kukkamaki assumes that linearly with the height and the measured temperature difference In equation (6.08 mm can = 50 m and formula which is an adaptation of -1/3 G = 80.5 m and 0. Ah = the levelled height difference in scale division 0.5 mm.25 oc. ~h computed using Remmer's of equation (3. ~t.4 corresponding average value can be used of c = -0. for c = -0. suggests that an G = 69. the author equation assumed a (5.07 mm and assuming ~t constant value obtaining = -0. and ~t = the measured temperature difference between 2. Kukkamaki [1979].11). For Cr = 0. with be found = 2 m. without causing any significant loss of accuracy. Based on Hytonen's [1967] investigations. Cr = 0. S Geodetic refraction for c was also as in c = -1/3. the fourth derivatives temperature function respectively which are d2 = {C-1) C b z (6.2).11).123 where. In equation (6. the exponent in Kukkamaki's model.5.

3) single set-up.4) R in which 6 = = S 2 = = this can be written as 8 I n For a special case. I when = 6 1 (6.5). Figura 6.1).2 to 6.19) and (6. The refraction in geodetic levelling The mean the close agreement. (3.1. 2 The refraction effect in trigonometric height traversing was computed using equation (2.1. Refraction error in trigonometric height traversing 6 .1).13) Cr = 1 I k (S - dx X) R is given by equation (4. were c-4 compute the refraction effect in geodetic levelling along the above lines and as can be seen in Figure the answers 6.g. came out to be in figure shows the effect of along the line comparison with 43. . of the three simulated refraction is used effect in for the trigonometric methods (see e.124 and d4 = (c-3) (c-2) (c-1) c b z S is the sight length and used to ~h is the height difference for a All three equations (6.

(S . Results of simulations Figures simulations.5) 8 n where 502.7) z and S is the sight length of s (S 6 . 3 = divided into n equal subsections n s).1.. special order other hand..125 8 =k I 1 s 2 + k 8 .[ 0. but.0342 + (dt/dz) ] . one can would be expected. to 6. keep the maximum To length show this of sight . the refraction errors in trigonometric height traversing are not correlated with the profile of the route. the geodetic levelling refraction effect is highly correlated with the the levelling route. As can be seen.7 p k = ------.S)+k 8 2 k 2 (6. They are highly dependent on the clearance and the length of the line of sight as it dependency. fluctuates profile of within the limits of On the of Canadian accuracy specifications..2 6.6) i 2 i T = (-1/3) b dt/dz -4/3 (6..2 B)+ 3 2 1 ••••• + (S . (6.5 The starting show the results heights of of the profiles four were chosen arbitrarily.

the author's in motorized A considerable improvement the results of simulations whenever a height of larger than 2. the of these it is cases and substantiated that if the lines leap-frog method are less clearance greater than 1 m. keeping unchanged while increasing the sight length. will result in a higher refraction error. usually more trigonometric height traversing. those presented case of reciprocal method. the limits of specifications ray for the first U ~ 2. is noticed in height of results shown according than 2 m for to 2. . that the assumed for of is all the instrument. where K is in The line of sight should be less than 250 m for the clearance of 1 m in the should also be mentioned instrument is experience with the given by Chrzanowski [1984] in terms of standard deviation as height of sight in The Canadian specification for first order and for one-way levelling is The than 150 m in this the refraction effect is within the Canadian order levelling.126 unchanged while changing the error the clearance.0 mm~K. of refraction will where clearance. Many more simulation were results obtained for different cases of the sight lengths and the ray clearances. kilometres. and decreasing the sight length will help to decrease the refraction effect.2 m for the instrument is assumed. By inspection thesis. increased as well. the clearance the In all be reduced by number of four lines increasing the observations will be Conversely in most of the cases.2 m It the in here.

8) Equating the right hand sides of equations (6.127 6.5 rn above the the ground).5 m height) and higher (at 2.7) and (6. mild. The H using To can look at the the route of the can be estimated gradient profile in the conditions.5 c = -1/3 c . . b is given in terms of ~t according to equation (3.5) b H =[ (6.2 Simulation of the Refraction Error Using other Values of Temperature Gradient Measurements According to the measured temperatures (at height 0.4) b = ~t where c I ( 2. dt 2/3 = -0.10) ~t is the temperature difference between lower (at 0.9) 0. middle region equation (3.5 m height) sensors.32) and neglecting the adiabatic lapse rate r = 0.8).1.23)). H condition during one along (#1 to #4). flux.0274 dz H -4/3 z (6.0. H can be written in terms of b (see section 3.5 m and 2. levelling in investigate the changes of Vancouver changes of sensible heat above levelling lines from the temperature under unstable the weather Island was quite weather.0822 in which.5 ) and (6.0098 (see equation (3.

67 23.5 .75 0.6 along the The line i2 . For other examples of (Fredericton.2 and 6.33) or be slightly figure.1 shows the averaged and its corresponding H values along the four levelling lines.69 28 17 25 24 MEAN -0.B. TABLE Average ~t.3 show other parts of Canada United States. these values in in the in this with the help of coefficient b and negative values may Table 6.128 shows the variations of sensible heat flux Figure 6.36 0.71 0. different from those depicted negative values reflect to be computed (3.53 0. Line 6. one can expect higher H and b values during the summer than those from Vancouver Island. example in N.) Fredericton. according to meteorological obsevations carried out.1 b and H along the levelling routes ~t b [oC] H [W -2 m ] il i2 i3 i4 -0.35 0.39 -0. the they have the corresponding equation 6. since the stable condition and Table (3.37 -0.34).28 -0.

geodetic levelling needs to In be corrected .7 -0.7 -0.57 1.--0.8 -0.7 and 6.9 -0.7 -0. of refraction Figures 6. 53 1. 34 l.7 -0. 34 1. 72 1. b and H in Fredericton.7 l.53 1.129 TABLE 6.0 -0.3 -0. 91 1.2) in the one obtains a set of results which can represent the refraction error in the summer conditions of the Fredericton Figures 6. 72 0.6 -0.5 .1 -0.2 Average Date 1985 Ground Cover Jul-19 Jul-19 Jul-19 Jul-23 Jul-23 Jul-23 Jul-24 Jul-24 Jul-24 Jul-29 Jul-29 Jul-29 Aug-06 Aug-10 AugAug-14 Aug-15 Aug-15 gravel grass asphalt gravel grass asphalt gravel grass asphalt gravel grass asphalt asphalt highway highway highway highway highway MEAN ~t.3 and 6..15 1. the area.7 -1. 34 1.6 -0.15 0.25 (see Table simulations along the Vancouver Island profile. N..7 -0.19 1.8 show the simulation error on the same profiles as in these cases.B.7 -0.. 53 52 18 66 66 52 66 80 66 66 66 95 80 66 112 66 95 4 80 10:25-19:59 10:45-19:35 10:10-20:32 10:10-18:20 09:30-18:40 09:50-19:00 07:00-17:20 07:20-18:48 07:45-19:05 12:10-17:10 11:30-16:30 11:50-16:50 11:00-20:15 07:28-15:23 13:32-15:04 11:13-16:~5 07:28-12:34 13:06-15:36 partially cloudy partially cloudy partially cloudy mostly sunny mostly sunny mostly sunny mostly sunny mostly sunny mostly sunny partially cloudy partially cloudy partially cloudy mostly sunny mostly sunny mostly sunny mostly sunny mostly cloudy partial