anpudwj pur qssnu UOqJBthS woJJ puep uioqd Iw.IIa uo!ssiuI3pu uo!

pIp wopj (gy) sa2mu

UEIV !d5I1!O

OUflSI q'UflS2}1

'P!"1S M8m(OJ

PU UOUJ!S p)100H

iCmiufi jo'uoI2uIqs 'riws UO12URISM 86 1-c61 '01 VSI1 i1doj jojr2ojocj q I2OIOO ''11S JO'Udf 11 '!i1SE!H 'qfl31flL P11J 'SO NYdVI MVdVI AUSI3ARIfl 'OIc)jOjJO Ijfl3j 'UTJOU!2U3JO 1--L 'OUOH 'rn1-OJUfloiCjoj UUOJIjU)T6 '601 YSII uoisjndoij1oTLuOqw '10S-E81

'suotz

, ta

'UPSd

JDVU1SIV

jo U. JsV 1UU3S UO SVSVN 1NV-SO NPIS :qDUnj) (8661 II!M io1Jo AJ SOUUEO JO uu. '-r"' NOEO5 01 W!IS jins upi s31ni1dw pu IC2IAISSIUI '1IdS icjiods 1AO 'ptrej 1LjA% SU!AISST1U3 JU IOU AIJ siuuiinsw nq xis uOLq u UAp 11iAiSsLu1/aImE1dLuj uoiids (sai) s UJJ!P snq oioip Jo wopij SUMOU3JUfl SflOT.mA stpoidd 3A uq psn o uiarisuo tmpuo2szip.iqiq iqi pqsqs 'suiqpo2 si'j uiuiis qi ainioduzi Aq ai2p pzwiou i(WISSIuI 'poqw UE U2 U!Sfl U 01 inzj i(ssiui puq soi uy ouduidsuopi sipid tp wnmiuituiiussux moij S3Sfl AL111!qDO1dd iods suoo -mm) xw :ouaijjtp jo pou.i 'SrqA 2wutuud IAO1 JO Oq IIA!SSRU3 UIfl.fl3dS iAOZO.I smiodui uqu& noq j:j: '0 pu O AOUII pjjoi ijs oupE1J pos uo jioiiuinu 'uonnuis sai. smsstw uqM noq jy suoiiui si.i moij qi oiiidu.i dqsuo!ai IC2!AtSSTUJ SflEA UE j.fl33ds 'isnuoo uosudmoD IOJ iCs 'oupni pu 'uoisioid 'uou1qrJD pu ou3qdsouu U0fl031J03 1u3uIAoJduqJO sai .opq tP1 ii' snooj uo uiuijaiqi awjis diqsuopi

sjr

'i jv

(xj

pjji

sjjy

utq

:qi
4DflW JO

'aznwiadwaj 'JltlSSlU1

lvlw?ZLL

1-''4"I 'JLSV ?WW9I

2usu?g

1 NOUD1XIOI1M

4I jOUEAV uioqEdS LwI3La uO!ss!uIpu UOfl3jJ 1mO!Pa (iaLsv) spn13U! 1Uq0Ag1t113ds!Mnm pau-uuq (JJJ) JUUS SOJ1flS UitOU JO 33JiflS-pUJ DflOUDj1 jins pU 'SOWA!SSIUJ sni S31flE1dUI1 pU2S I3AO1 A!A!SSiUI 3flS 1At p1w-I sini.idmi (j) am 2UUO! U! sJmiadwo snououioq
1OJ IOU

2uEq3-jqoj 'spniS UI 2u1EuI1s uopi sipnq pu Dujq-1q S!A!SSTUIH (3) OlE iJUO11S 'AOIpU UA 'osou2p jo 'uosoduio
pU1 OJ1flS
S1!A!SSIW

IOAO

JO

SC

ipdslbs uwdopAp pu pU iisqns 1q!s sin.idmi m unuzs sunow pu s2uqo ut s smd AW13A 1AO ioj qq' qispuq SflE39fl U3 q p1AOD1 uOJ Xuussiw ids 3!O1O siw qI1U31ARA puissrui spms iinbai s iutu jids spuq m qi uiiijjj iopu s qissod o!nb 'xidwoo 'aaLsv jim q uo uuojejd n qu uiusqcj msA '(ii'sw-soa)pjflpt3S'J qm1 Uflf -aaLsv OS 1AOOi uqo SIfl21OdlU1 UB
Joj oqi oqiS s1EJuiU1 3q:i
St

'S1MtJS

pu S JO1UOO 1OJ 32UIllO

JO

smn 2UUOdW! IOJ

ow dn
JOJ

.s1pOw

S1MUS JO

uO!sO1

1UPU34bPU! JO

J3qWflU JO

'p1-8

p!.Lre3

ISIIJ

JO S,VSVN

UI

'8661 jUE II!M S!A!SSTUI JOJ

fflJJ UOt1flIOS1 Jo ci 'i" LSy s wooz suj .ioj iqo soaui2uii S2UUITJdX UOTflOS31-42!}j-aaLsv I ws 'tui q psn o pusipun qi p21A ssuodsi Jo pu 3 UED q 3LIA q pi siuuq.idx 'pus
UOnnIOs1-iMOI

p
j

P2'['' so2qE3O1 ioj .siuus

flATSSUU3

aiu i-xts WUJq J0 LNV-S03 4l!i E Jfl tIdS UO!flLOS1 J06 UI pim g

dEw Jo

pImI 3EJIflS

1PM

3EJJflS

pu 3 SflA IIP' 1AO31 2UISfl !A!SS!3PUE1WL uo.nd3S 'uupuo sai. ou iC1 SO JO IOAO 'UO!12oA 1M pu 'MOUS Pu (z) 01 sai. ai (j) O uJrIs 31E.m33 pu sizaid 3J.InS S1TflCiOdUI1 sai. in' 3npO1d SSJUIS SUIt -- U! 1q1O 'SpiOM Iq pim siid siussrui .IOJ iuuu sijsqns 1AO31 spoqui sjduis qooidd uouog!ssp pjnoqs q ou 'S2!flU1UO3S!p qons uo q ponpoiui
13LSV

Iq!J

in

jqdEp sqj idd siusid P SaL UIIpIIOIE pu osj sai. m.idw pu I41LA!SSIUI sw! p1EInz[o uIO1J poiBInuI!siaisv 13UU3S (svu) 1s2w! UO!ssnDs!cl IuL1qL parEijul 'SUIT SAjSWfl pA!.Ip UIOIJ jUUiO-XiS sai. ouiuiojid pilE uo JO 042 wq1uoIE 11 1PM AOJ JLjUflJ SOSfl3OJ UO 'T' SJOPJ SPdS UOUU1 iojq TNVSOaqounj

2U1SSUO3 TI!M oi S3.IflOS

(fl Aoq seo i. 0 (puq

JO U42 .IaLsv

c) (spuq II!M

q jqjAs pzpus SOflpOid

WOIJ

SO3 S3L s!

I"P'J

Z9 RIdS JOn LL9Z

Downloaded from SPIE Digital Library on 25 Jul 2011 to 129.137.163.64. Terms of Use: http://spiedl.org/terms

OO9/96/-cOZZi6L9-O

2. THE ASTER IMAGING SYSTEM

ASTER has three bands in the visible and near-infrared (VNIR) spectral range (0.5-0.9 pm) with 15-rn spatial resolution, six in the shortwave-infrared (SWIR: 1.6-2.4 jtm) with 30-rn resolution, and five in the thermal-infrared (TIR: 8-12 jim), with 90-rn resolution.2'3 These 14 bands are collected in three down-looking telescopes that may be slewed (SWIR, TIR) or (VNIR) in the cross-track direction. Combined with the FOV of the maximum 'FIR view angle is thus 1 1. An additional backward-viewing telescope with a single band duplicating VNIR band 3 will provide the capability for same-orbit

stereogrammetric data. The five TW channels (ASTER bands 10-14) have spectral ranges of 8.125-8.475, 8.475-8.825, 8.925-9.275, 10.25-10.95, and 10.95-1 1.65 .tm, respectively. ASTER's estimated TIR radiometric accuracy at 300K is 1K; at 240K it is 3K. Radiometric precision (NEiT) at 300K is 0.3K.4

The ASTER instrument is being provided by the Japanese Government under the Ministry of International Trade and Industry (Mm). The ASTER project is implemented through the Earth Remote Sensing Data Analysis Center (ERSDAC) and the Japan Resources Observation System Organization (JAROS), nonprofit organizations under MITI. JAROS is
responsible for the design and development of the ASTER instrument, which will be built by the Nippon Electric Company (NEC), the Mitsubishi Electric Corporation (MELCO), Fujitsu, and Hitachi.
.

3. BACKGROUND

A surface radiates energy in proportion to its temperature (T) and emissivity (. On the earth, atmospheric opacity restricts radiance measured by spacecraft to spectral windows at wavelengths of 3-5 and 8-14 .tm. ASTER bands 10-14 lie within the TIR window of 8-14 p.m. The basic problem in estimating T and a is that the data are non-deterministic: there are more unknowns than measurements because there is an a for each image band, plus T and atmospheric parameters. This is the case

even if the scene is isothermal and consists of a single material of uniform texture and topographic slope and aspect.
Consequently, even if atmospheric parameters are measured independently, at least one additional degree of freedom must be constrained independent of ASTER. There is a degree of arbitrariness and craft in the way this has been done, resulting in a

variety of approaches and algorithms. Below, the important equations governing hR remote sensing (3. 1) and previous solutions (3.2) are reviewed. The TES algorithm is introduced in its performance is evaluated in

3.1 Conceptual framework for TIR remote sensing

Temperature is not an intrinsic property of the surface; it varies with the irradiance history and meteorological conditions. Emissivity is an intrinsic property of the surface and is independent of irradiance. The radiance from a perfect emitter (i.e., a blackbody for which = 1.00) is exponentially related to temperature, as described by Planck's Law. The radiance R from a

real surface, however, is less by the factor a: R = B, where B is the blackbody radiance and ? is wavelength (jim). ASTER integrates radiance emitted from a number of surface elements. It is attenuated during passage through the
atmosphere, which also emits TIR radiation. Some of this radiance is emitted directly into the scanner ("path radiance"); some strikes the ground and is then reflected into the scanner. For most terrestrial surfaces the reflectivity p and are complements (Kirchhoffs Law): p = 1 - c. A simplified expression for the measured radiance L is:

tX,y,[CX,Y,XBX(TXY)+PX,Y,4S + mLa
x, y Sl, R*

(1)

= position in scene = downwelling atmospheric irradiance = radiance emitted from adjacent scene elements

S'

= atmospheric transmissivity = upwelling atmospheric path radiance

For most terrestrial surfaces

.0, although surfaces with <0.85 are probably restricted to deserts.5 Radiance emitted

at 10 jim from a surface at 300 K is on the order of 10 Wm2sr1 jim1. For a sea-level summer scene, typical order-ofmagnitude values of the atmospheric variables estimated by the LOWTRAN7 atmospheric model are 'r60%, S'=l.5 and S =1.8 Wm2srj.tm 6 One effect of S. is to reduce the spectral contrast of the ground-emitted radiance, because of Kirchhoffs Law. It is necessary to compensate for atmospheric effects, including Sj,, if T and c are to be recovered
accurately. Incident radiance from adjacent scene elements @ixels) varies with terrain roughness but is typically less than S and is usually ignored.7 Therefore, the remote-sensing problem reduces to L='rB(T)+tp Sj, + S'. Equation 1 ignores the effects of the atmospheric point-spread function, as does TES.

SPIE Vol. 2817/83

Downloaded from SPIE Digital Library on 25 Jul 2011 to 129.137.163.64. Terms of Use: http://spiedl.org/terms

Equation 1 describes only the radiance at a single wavelength, and only radiance from homogeneous isothermal surfaces. In practice, the radiance is measured over a band of wavelengths; however, errors due to this integration are small . At the 90m scale of ASTER TIR pixels, many terrestrial surfaces consist of multiple components having different emissivity spectra and temperatures. Strictly speaking, each component adds to the number of unknowns, while the number of measurements is unchanged. ASTER TIR measurements for such complex surfaces are not sufficient to estimate all the unknowns; instead, it is necessary to determine only an effective T and spectrum for each pixel.

3.2 Previous approaches

Inversion of the Tffi equations for T and e have been attempted using deterministic and non-deterministic approaches. The

former are restricted to areas for which one or more of the unknowns is known. Historically the chief reason for 'HR measurements has been to estimate temperatures. This task is deterministic for important scenes for which c is not in
question: the ocean,8 snowfields and glaciers, and closed-canopy forests. However, deterministic solutions require that the atmospheric parameters in equation 1 be measured directly and the measured radiance corrected for them, and this is not always feasible. Most ocean-temperature studies have utilized data from the Advanced Very High Resolution Radiometer (AVHRR), which has two channels, at 10.3-1 1.3 p.m and 1 1.5-12.5 rim, thereby 'splitting" the 'FIR spectral window. Joint analysis of the two "split-window" channels can compensate for atmospheric effects while solving for Split-window algorithms rely on empirical regression relating surface radiance measurements to water temperatures. A version of the splitwindow algorithm has been developed for EOSIMODIS images.'2 Several authors have examined extending the "split-window" technique to land surfaces.1315 They all conclude, however, that large errors arise there due to unknown emissivity differences. Over land, the unknown emissivities are a greater source of inaccuracy than atmospheric effects. Inaccuracy of only 0.01 in c causes errors in T exceeding those due to atmospheric

rr16 In general, land emissivities can not be estimated this closely, and must be measured if accurate kinetic
temperatures are to be recovered. As a result, the usefulness of split-window methods for land is limited and the nondeterministic nature of 'FiR remote sensing must be addressed head-on. Many geologic studies, however, have utilized enhancements such as decorrelation stretching that do not recover T and 17, 18 A spectral-unmixing approach has been used to separate a non-linear measure of T from ., but the separation is imperfect.19

In all, we examined 14 inversion methods for the general land-surface problem in creating TES. These algorithms:
determine spectral shape but not T; require multiple observations under different conditions; assume a value for one of the unknowns; assume a spectral shape; or assume a relationship between spectral contrast and . All require independent atmospheric correction. The temperature-independent spectral indices (TISI) of Becker and Li;2 thennal log residuals and alpha residuals;21 and spectral emissivity ratios recover spectral pe23 The day-night method24 increases the number of unknowns by 1 but doubles the number of measurements; the problem is over-determined. In practice, however, this approach magnifies measurement "noise" greatly and requires "pixel-perfect" registration between the two images. Other techniques have been based on an assumed value for a "model" emissivity at one wavelength,25 or an assumed maximum 27 approaches are emissivity (max) value at an unspecified wavelength (normalized emissivity method, or NEM).26' unsatisfactory for ASTER because inaccuracies tend to be high (3K) and because tilts are introduced into the spectra. One method required only that the emissivity be the same at two wavelengths.28 However, this assumption is commonly violated for ASTER, with only five channels. Finally, the "alpha-derived emissivity" (ADE) method utilized an empirical relationship between the standard deviation and mean emissivity to restore amplitude to the alpha-residual spectrum, thereby recovering T also.21' 29, 30 The ADE method, however, relies on Wien's approximation to invert equation I, thereby introducing slope errors into the spectrum. The Mean-MMD method avoids Wien's approximation and uses a modified ADE empirical relationship based on the minimum-maximum emissivity difference (MMD).31

The MODIS team is considering an approach in which emissivities are specified by classifying VNIR/SWIR data.32 Although important scene types such as vegetation are readily identified in the VNTR and have well known spectra,
classification is ineffective for many geological materials. It also creates sharp boundaries in images of gradual transitions.

4. TES ALGORITHM
The TemperaturelEmissivity Separation (TES) algorithm combines attractive features of three precursors and some new features (Fig. 1). It is most closely related to the Mean-MMD method. Essentially, the TES algorithm uses the NEM to estimate T, from which emissivity ratios () are calculated using the Ratio (RAT) algorithm?-3 The spectrum preserves the shape, but not the amplitude, of the actual emissivities. To recover the amplitude, and hence a refined estimate of the

84 / SPIE Vol. 2817

Downloaded from SPIE Digital Library on 25 Jul 2011 to 129.137.163.64. Terms of Use: http://spiedl.org/terms

temperature, the MMD is calculated and used to predict the minimum emissivity (Emin). TES operates on ASTER "landleaving TIR radiance" data which have already been corrected for atmospheric 'r and Sj .33 The same ASTER standard product reports Sj, , which cannot be removed without knowledge of e. TES removes reflected S.j. iteratively, before estimating the NEM T.34 TES also differs from precursors in: (1) refining the value ofm USd in NEM, pixel by pixel; (2) correcting inaccuracies in Emin for graybodies (e.g., vegetation) caused by errors in MMD due to NEi.VI'; and (3) using first-generation TES T and to refine the correction for S, and the NEM T, leading to a more accurate second-generation TES T and c. Finally, TES estimates and reports pixel-by-pixel accuracies and precisions for T and e, in a QA data plane that is part of the ASTER standard product. Presentation of TES herein is necessarily brief; more complete documentation is available from NASA.35 In Figure 1 and subsequent discussion the TES code is subdivided into modules named for the algorithms they derive from. The significant advance of the 115 algorithm is to produce, for the first time, unbiased and precise estimates of emissivities and, therefore, improved estimates of surface temperatures for the land surface.
For most scenes the TES algorithm can recover temperatures with an accuracy and precision of 1.0-1 .5K, assuming accurate radiometric measurements. Emissivities can be recovered with an accuracy and precision of 0.010-0.0 15. TES's performance over land and sea are comparable. ASTER TES temperature recovery is not as accurate as that of the MODIS split-window

algorithm for sea surfaces because: 1) ASTER resolution is better by an order of magnitude and its SNR is accordingly lower; and (2) none of the ASTER TIR channels is at wavelength at which the atmosphere is absorptive. In any case, ASTER's data acquisition plan is focused on the land surface. Major limitations on algorithm performance arise from two main sources: (1) the reliability of the empirical relationship between emissivity values and spectral contrast; and (2) compensation for atmospheric factors. Measurement accuracy and precision contribute to lBS errors, but to a lesser degree.

INPUT IMAGES:' Land-Leaving 'n'

Radiance (L')
o

NEM RAT

Downwelling Sky

Normalized Emissivity Method - Ratio MMD - Mm-Max Difference QA - Quality Assurance

L. . .Iia.c,L:..:::
NEM Module
O

RAT Module
o

MMD Module
o Calculate MMD o Determine o Calculate Tand o

QA Module
o Estimate accuracies

Subtract

Calculate }3

Flag lBS failures

reflected
O

Estimate

spectrum

Emin

and precisions for

Tand

OUTPUT IMAGES:
Figure 1. Basic design of the TES algorithm.

Tand* .:.:.j

4.1 Theoretical basis for the TES algorithm

Figure 2 is a flow diagram of the lBS algorithm. Below, the steps of the lBS algorithm are presented in sufficient detail to permit regeneration of the processing code. The input image data sets consist of "Land-Leaving TIR Radiance," L', and sky irradiance, S.j,, comprising ASTER standard product ASTO9.33 These calibrated data have been compensated for ASTER filter transmissivities and other instrument effects. Several lBS parameters may be adjusted from their default values as the need arises. These parameters are identified in Figure 1 and below. The output data sets consist of the five emissivity images (ASTO5), corresponding to ASTER channels 10-14, and a single image of surface temperature, I (ASTOS).

SPIE Vol. 2817185

Downloaded from SPIE Digital Library on 25 Jul 2011 to 129.137.163.64. Terms of Use: http://spiedl.org/terms

Figure 2. Flow diagram

of the TES algorithm.

R=L-(1-E) S,

Caption on next page

[NEAT]
I

* Schematic only

86/SPJEVo!. 2817
Downloaded from SPIE Digital Library on 25 Jul 2011 to 129.137.163.64. Terms of Use: http://spiedl.org/terms

Figure 2. Flow diagram of the TES algorithm, schematic in places. L'=cB(T)+p Sj, (i.e., L'=R+p Si,). R' here is the estimated surface-emitted radiance, nominally corrected for Sj, . Tests for divergence (Div) and convergence (Con) determine if the correction for reflected S, will work. Div is lL2R/iM2I<t1 Con is IiR/iI<t2 (iR/i=R'iR'1 The nominal value for em in the NEM module is refined if the spectral variance is small. The value for j found in the MMD module is compensated for measurement error. Finally, the TES c data are used to refine R and the NEM T. Quality Assurance (QA) consists of calculating TES performance and product characteristics and encoding them in a reserved data plane.

4.1.1 Estimating the surface temperature and subtracting reflected sky irradiance (NEM module)
The kinetic temperature is estimated using the normalized emissivity approach. Essentially, Emax is assumed in order to calculate a temperature and the other emissivities. For all materials in the ASTER spectral library, Cmax iS 0.944 .00 and for the NTEM the midrange value of 0.97 is adopted. Therefore, the estimated kinetic temperatures should be within at 340 K, and within at 273 K, provided atmospheric correction is successful. An empirically based process, described below, is used to refine max for nearly flat emissivity spectra, as determined by a low variance for the NEM spectra. For rock and soil spectra, the default value cannot be refined.
Upon entry to the NEM module, radiance R is estimated by R'=L'-(l- Cmax)S . The NEM temperature is taken to be the maximum temperature Tb estimated from the radiances Rb for the different image channels b=10-14:

T = max(Tb);

Tb

]]

Eb

Bb(Tb)

(2)

where ci and c are the constants from Planck's Law. Once Tb is known, NEM emissivities are calculated and used
iteratively to re-estimate R. This process is repeated until the change in R between steps is less than threshold value t2, or until the number of iterations exceeds N. The current default value for t2 is the radiance equivalent to the NEIT, and N=8. If the slope of R vs. iteration increases (exceeds t1 ) correction for S is not possible. Execution of TES is aborted, and the NEM T and are reported along with a warning flag in the QA plane. Correction for S.j. is typically <1K unless the sky is warmer than the ground or humidity is high. Because error introduced during correction depends on , it is possible to recover accurate temperatures from the high-c image channel even if the rest of the spectrum is inaccurate.
For samples with low contrast, the actual Em is typically -0.985, so that the NEM T will be systematically overestimated by l-2K for Emax 0.97. We adjust Emax reflexively to improve accuracy. For near-graybodies, measurement error is the dominant source of spectral variance. The best value for Cm minimizes the variance v of the NEM emissivities. Plotting v

against max yields an upwards-opening parabola. A new value of niax may be found by calculating v (4Eb) for CmaxO.90, 0.94, and 1 .00 in addition to 0.97, and fitting a parabola to the data. If this curve has a minimum for
O.9maxl .00, it indicates the best estimate of m. In this case, the NEM module is executed again to find a new NEM T.

4.1.2 Ratio algorithm (RAT module)

The relative emissivities, 1b are found by ratioing the NEM emissivities, calculated from the NEM T and the
atmospherically corrected radiances, to the average emissivity:
13b =
E

" /=

e(T)

R. BT /;
R Bb(T)

i=

14

14

b=lO

(T)= XBb(T) b=lO

(3)

Because emissivities themselves are generally restricted to O.7<Eb<l.O, O.75<3b<zl.32. The errors in l due to inaccuracy in the NEM T are systematic but less than the random errors due to NEST, for 240<T<340 K. Warping of the f3spectrum is below the threshold of detectability for ASTER data.

4.13 Estimating TES emissivities and temperature (MMI) module)

The 3 spectrum must next be scaled to actual emissivity values, and the surface temperature must be recalculated from these

new emissivities and from the atmospherically corrected radiances. These TES T and values are the reported ASTER
SPIEVo!. 2817/87
Downloaded from SPIE Digital Library on 25 Jul 2011 to 129.137.163.64. Terms of Use: http://spiedl.org/terms

Standard Products. An empirical relationship predicting Emin from MMD is used to convert 3b to Eb. We established this regression using laboratory reflectance and field emissivity spectra,36 documented in
The first step in the TES algorithm is to find the spectral contrast:

MMD=max(f3b)-min(b); b=1O

14

(4)

from which the minimum emissivity is predicted and used to calculate the TES emissivities:

min 0.994 0. 687*

O.737

Cb

bI min 1;
' min(b))

b = 10

14;

(5)

Provided the actual emissivity contrast in a scene element is much greater than the apparent contrast due only to measurement error, MMD is an unbiased estimate. For graybodies, however, MMD is dominated by measurement error and is no longer unbiased. That is, as the true spectral contrast is reduced to zero, MMD is also reduced, but to a positive limit whose value depends on the NELT. It is possible to correct the apparent MMD proforma, as specified by Monte Carlo simulations.

A=O.006-O.98 1 *J4T5558 1

153*MMD3 ;

i=max(, 0); MMD'=rnax(MMD-, 0)

(6)

where MMD' is the corrected contrast and the coefficients were determined for NEiT=O.3K. The nonlinear correction of equation 6 improves the accuracy of TES for graybodies, but at the expense of precision. However, precision can be improved by invoking the physical constraint that EjEx . The optimum pro forina correction has not yet been
determined, and deserves further study before launch.

The NEM T is likely to be in error by up to 3K because the assumed value of Emax may be inaccurate. This error canbe reduced by recalculating T from the measured, atmospherically corrected radiances R and the TES emissivity spectrum:
7tRb*?b*R )) where b* is the ASTER band for which emissivity Eb is maximum (and correction for Sj. is minimum).

T = ELIln1

clcb*

b*

4.1.4 Final correction for sky irradiance and bias in

The TES E and T values are more accurate than the NEM values. Recalculation of the TES and T values improves their accuracy further. To do this, first the YES c values are used instead of the NEM values to make a final single (non-iterative)

correction to L' for reflected S , and then the new estimates of R are used with the TES T instead of the NEM T to
recalculate the f3 spectrum (Eq. 3). Then improved TES and T are calculated as before. Experience shows that there is little gain if this process is repeated more than once (M=1, Fig. 2). For a variety of simulated and real radiance measurements the "refined" 'FES emissivities changed by as much as 0.01; therefore, this final correction is worth doing.

4.1.5 Quality assurance

TES will report on its performance and on the accuracy and precision of its T and products. The report will take the form of a header record and a 3-byte data plane, the formats of which have not yet been finalized. The QA data will include flags to report on the optional paths through TES, described above. Of central interest is the reported reliability of the TES T and data. This is assessed by numerical simulation and by algorithm execution on different images simulated from laboratory and field spectrometer data.35 TES performance depends strongly on the ratio of S.j, to L, and on MMD, indicators that can be assessed pixel-by-pixel during operation. It is anticipated that this information, together with the results of the numerical simulations, will be used to generate the QA data. Additionally, TES performance is affected by the proximity of clouds. This information will be reported in the QA plane and incorporated in the performance assessment.

4.2 Regression of ErninOiltO MMD

The relationship between emissivity and spectral contrast is a key feature of the TES algorithm. It was established by analysis of laboratory reflectance spectra, equivalent to emissivity by Kirchhoffs Law. The data were converted to ASTER was found for each sample. Radiances were estimated, scaling emissivities by blackbody radiances pseudo-spectra and

88/SPIE Vol. 2817

Downloaded from SPIE Digital Library on 25 Jul 2011 to 129.137.163.64. Terms of Use: http://spiedl.org/terms

calculated for T=300 K, and 3 spectra and its MMD were calculated. The cj data were then regressed to the MMD values. They are related by a simple power law (Fig. 3). The regression parameters are insensitive to the temperature assumed in estimating R. Although the regression parameters are defined empirically, the relationship itself is reasonable and physically predictable if deviation from blackbody behavior is due to molecular resonance localized in narrow reststrahlen features.
The critical assumption that this regression applies to the entire gamut of surface materials remains to be proven. We have tested this assumption and, so far, it appears to be valid. A different set of 3 1 of Salisbury's reflectance spectra37'38 yielded nearly identical regression coefficients (H. Tonooka, Ibaraki Univ., pers. comm., 1996), as did field emissivity spectra of

Australian rocks (n=91) collected using the Jet Propulsion Laboratory's iVFIR spectrometer.36 Hundreds of airborne
MIRACO2LAS CO2 laser reflectance spectra, with a narrower window than the five ASTER 'HR bands, yielded a regression having similar overall characteristics (T. Cudahy, CSIRO, pers. comm., 1996).
The regression chosen for the TES algorithm uses Cmin rather than mean emissivity as in the Mean-MMD algorithm, because min was found to improve the correlation. The MMD was used because, for most spectra, it was just as good and faster to calculate than other measures of spectral complexity, such as variance. Use of the variance, however, reduces sensitivity to measurement error for the important class of near-blackbody scene components, and this choice deserves review.

The scatter of the individual samples about the regression line (Fig. 3) results in an irreducible imprecision of -4 .5K in the TES algorithm. Coincidentally, this is about the magnitude of the scatter of data on the Cmjn MMD plane due to ASTER measurement error, evaluated by Monte Carlo techniques. It is also comparable to the predicted inaccuracy of the ASTER TW data of 1K.

1.0

0.9 0.8 07 0.6 0.0

0.2

Figure 3. The empirical relationship between

Emin and MMD, based on 86 laboratory

8min

reflectance spectra of rocks, soils, vegetation, snow and water, provided by J. W. Salisbury in 1995. 95% of the samples fall within emissivity units of the regression line, corresponding to an error in T of about at 300 K. The EminMMD relationship follows
a simple power law:

Emin0.9940.687*MMD737.

0.4
MMD 0.68

0.6

5. PERFORMANCE OF

itS

TES processes an ASTER image ('7OOx700 pixels) in about 5 minutes on a DEC Alpha-3000/900 computer running at 275 Mhz. Half of this time is required to refine the assumed value of em. We have tested the reliability of the TES products by: (1) numerical simulation; arid (2) processing four calibrated, atmospherically corrected TIMS images for which field data were available. In the first approach, radiances estimated from library spectra using Planck's Law are passed through TES. These results give the most insight into the workings of the TES algorithm itself. The TIMS images may provide a more realistic test, but their preparation involves more opportunity for error unrelated to TES.

5.1 Numerical simulation results

Overall, the TES algorithm operating on error-free input radiances can recover temperatures for a wide range of surfaces within 1.5K and eniissivities within 0.015. For numerically simulated radiance emitted from surfaces at 300K, based on the field emissivity spectra in our library, 95% of the recovered temperatures were within 1.5K and 68% (1 std. deviation) were

SPIE Vol. 2817/89

Downloaded from SPIE Digital Library on 25 Jul 2011 to 129.137.163.64. Terms of Use: http://spiedl.org/terms

within 0.3K, for example. TES performance is not related to scene composition in general, but to the scatter about the EminMMD regression line, scatter which is largely independent of MMD (Fig. 3). Monte Carlo simulation shows that the scatter of recovered temperatures due to measurement error is about the same as that due to the inherent scatter about this regression line. Recovered emissivities show little bias, but err systematically by an amount proportional to the error in temperature: if TES overestimates T by 1K at 300 K, it will tend to underestimate c by -0.0l7. Numerical simulation results are discussed at length in the TES Algorithm Theoretical Basis Document, available from NASA.35

Figure 4. TES results from simulated ASTER images acquired by TIMS over Castaic Lake (-900 m amsi; 9

March 1994), California, Lake Tahoe, California (2000 m amsi; 28 May, 1995), and the south coast of Hawai'i (1 October 1988). Left column: cj. The stretched images appear "noisy" because the scene has low MMD. Right column: temperature. Subscene size: 100 x 200 pixels. Resolution is 5.5 - 7.5 rn/pixel;
subscenes are about 1.2 km across.

90/SPIEV0!. 2817
Downloaded from SPIE Digital Library on 25 Jul 2011 to 129.137.163.64. Terms of Use: http://spiedl.org/terms

Figure 5. TES results from simulated ASTER data acquired by TIMS over Railroad Valley playa (-4750 m amsi), 1 June 1996. North is to the left; image is 2O km high.; resolution is .l5 ni/pixel. a. TIMS Radiance data, simulated ASTER band 13 (10.6 pm), showing test areas for which field temperatures and laboratory reflectance spectra were measured: A, playa center; B and C, shore; D, yardangs; E, ponds. b. Temperature image recovered by TES. c. Emissivity image, ASTER band 12, recovered by TES. Due to exposed geologic materials, MMD was as high as 0.15.

5.2 Tests on simulated ASTER images

The airborne TIMS, with NELT O.2K,1 is appropriate for simulating ASTER's TIR images because TIMS bands 1-3, 5 and

6 correspond closely to ASTER bands 10-14. TIMS images are calibrated using internal blackbody measurements, atmospheric data, and surface temperature measurements. Radiosonde atmospheric probes and Reagan Sun Photometer
measurements of total water vapor, together with LOWTRAN 7 and MODTRAN2 atmospheric models,6'39 have been used to estimate t, Sj and S.j. at the time of overflight and as a function of view angle. Surface temperatures are measured with an Everest radiometer. Three simulated ASTER images (Fig. 4) have been prepared from calibrated TIMS overflights of Castaic Lake and Lake Tahoe, both in California, and of the south coast of Hawai'i.41 These images are used to test T and a recovery over water targets (low MMD). An additional image has been prepared over a geologic target, the playa in Railroad Valley, Nevada, and is used to test recovery over land areas having high MMD (Fig. 5). The Castaic Lake image is of a reservoir, the earthen dam that impounds it, and sparsely vegetated hills.40 Two radiosonde K profiles plus sun photometer measurements were made to monitor water-vapor changes. Lake temperatures of 287.9

were measured at 49 locations. Boat trails visible in the image were 2K colder. The TIMS radiances for Castaic Lake indicate a water temperature of 289.01.6 K, about 1.1K too high. TES T was 29O.6 K. The average TES emissivities were correspondingly low (-O.O5). Deviations from laboratory spectra were most pronounced in bands 10 and 14. Emissivities for a stand of trees showed the same pattern, subdued. Correcting for sky irradiance reduced TES T by O.2 K.
The Lake Tahoe image includes the lake and forested, snowy mountains (not shown in Fig. 4). The subscene in Figure 4 is of Dollar Point, a subdivision with roads and houses. Water, air, and melting snow temperatures were 28O.3 K, 283.1 K, and 273.1 K, respectively. Corresponding TES T values were 281.00.4, 283.90.6, and 273.70.4 K, respectively. Assuming

forest and air temperatures were the same, the mean TES T's were systematically 0.4-0.8K too high, but within the
uncertainty predicted from the numerical simulation studies. If the band 10 and 14 radiances were adjusted by empirical gain factors designed to "flatten" the recovered Castaic water spectra, TES a spectra for both Lake Tahoe water and snow were flat. Snow spectra averaged 0.9730.006, 0.011 lower than laboratory values. Recovered water spectra had an rms error of only 0.004 compared to laboratory data. Forest spectra were less well fit, with an mis error of 0.026. The Hawai'i image shows an active lava flow entering the ocean near Kapa'ehu, accounting for the plume-like patterns in the water and for the wide range of temperatures there. A cloud of steam drifts west (left) from the entry point. Concentrations of SO2 are also present nearby. SO2 absorbs strongly in the 8-12 tm region of the spectrum and is not accounted for by the atmospheric corrections. Away from Kapa'ehu, ocean temperatures measured three days before overflight were 296 K

SPIEVo!. 2817191
Downloaded from SPIE Digital Library on 25 Jul 2011 to 129.137.163.64. Terms of Use: http://spiedl.org/terms

(33O K near the entry point). Corresponding TES T = 3050.6 K, 9 K higher than the earlier radiometric temperatures. Recovered emissivities are too low by 0.05 (band 10) to 0.02 (band 12). Uncorrected absorption by SO2 and other gases from the lava may account for the excessive apparent temperatures. Correcting for sky irradiance reduced itS T by 0.5K.

The Railroad Valley image (Fig. 5a) shows the playa, shallow ponds surrounded with reeds, and alluvial fans. Pond
temperatures measured during the T1MS overflight were 295.10.6 K. Emissivity spectra (Fig. 6) were measured for three homogeneous areas on the playa itself, and for a promontory of yardangs nearby and corrected for measured atmospheric

conditions. TES emissivity pseudo-spectra for the ponds were saddle-shaped; atmospheric effects may have been imperfectly removed. With the same empirical corrections applied at Castaic Lake, the TES spectra were flattened and
agreed better with laboratory data (Fig. 6). Thus corrected, TES T for the pond was 297.20.3 K, -2.1K too high. TES c were correspondingly low by as much as 0.022. We observed a general correspondence between averaged TES and field spectra for the playa surfaces; however, the significant discrepancy in ASTER band 1 1 (Fig. 6) resulted in an mis error of 0.034, higher than predicted from numerical simulations. Empirical adjustment as for the shallow ponds did not reduce the error. Precisions for T and , calculated for homogeneous areas on the image, were O.6K and respectively.

1.0 ___________________________
----

=%
__ "I

-##

0.9

c:;

'.

,
10

. TES emissivity pseudo-spectra Figure 6. compared to field emissivity spectra, Railroad Valley playa. Solid line: water spectrum (J. W.
Salisbury, pers. comm., 1995). Dashed line: playa

...

0 .8 .

2
8

spectrum, Site A (Emax0.97). Solid bars: TES values for ponds, Site E. Open bars: TES c values for Site B. ASTER band centers are indicated by
12
nIflh1brd dots on the abscissa.

Wavelength, jim

6. DISCUSSION AND CONCLUSIONS Applying TES to numerically simulated TIR radiance data suggests that TES performs within its design specifications to recover temperatures and emissivities within l.5K and 0.015, with respective precisions of O.4K and 0.003. Analysis of simulated ASTER data acquired by 11MS, however, raises some questions. Although precisions are close to predicted limits, temperatures are less accurate than expected, and there are systematic and significant errors in recovered emissivity spectra. These problems were observed for all four TIMS images for which concurrent atmospheric measurements have been made. As seen for the spectrally flat water targets, the concentration of emissivity inaccuracies in ASTER bands 10 and 14, which are the most subject to atmospheric effects, suggests that the atmosphere correction may be imperfect. There may also be problems with TIMS calibration, field emissivity spectra, or laboratory reflectance measurements. We do not attribute the elTors to 'I'ES itself, because they do not occur in numerical simulations.

its is sensitive to errors in spectral contrast because emissivity amplitude and hence temperature depend on MMD.
Therefore, incomplete removal of atmospheric effects in even a single channel will reduce accuracies of the recovered T and E significantly. This problem may be irreducible, because scatter about the Cj-MMD regression line appears to be a real property of natural surfaces. For high-MMD scenes this scatter is coincidentally equivalent to the ASTER NEAT of 0.3 O}(, and to engineering predictions of 1 K radiometric accuracy; however, measurement noise is random, whereas regression effects are organized spatially. The MMD effects are most serious for graybodies, such as vegetation and water. For these important scene types random measurement errors affecting MMD are amplified in the final T and Eproducts. Although the EminMMD regression now appears to be robust, it does warrant further study before routine use of TES, and improvement of TES before launch of EOS-AM1 will focus on refining the MMD relationship.

Improvements in ASTER sensitivity and noise levels would translate into improved TES precision over natural graybodies. However, even for NEiT=O, TES could not match the accuracy of dedicated sea-surface temperature algorithms, because these presume that E is known and TES does not. For geologic surfaces, the inherent variability on the j1-MMD plane, not the NEST, is the major source of uncertainty. No improvement in this area can be forecast now.
Atmospheric correction may be the effective limit on TES performance with ASTER data during the EOS mission. For most scenes, the critical parameters are atmospheric transmissivity and path radiance. Errors in correction for these parameters

92/SPIE Vol. 2817

Downloaded from SPIE Digital Library on 25 Jul 2011 to 129.137.163.64. Terms of Use: http://spiedl.org/terms

propagate directly into T and . For many scenes, sky irradiance reflected from the scene is of minor significance: for example, at Castaic Lake it accounted for only 0.2 K. For humid atmospheres or warm atmospheres over cold scenes,
however, the correction for reflected sky irradiance may be larger, even the most important factor limiting TES performance.

Mixtures of scene materials fall near the regression line if the endmembers do also. Potential sources of error include admixture of blackbody cavity radiation from rough surfaces, since ideal blackbodies fall above the regression line. Theoretical studies with radiosity models suggest that these effects will be minor because multiple-scattering contributions
rarely exceed 2O% of the total radiance from natural surfaces.

To be effective, TES probably requires at least four or five channels of TIR data. Numerical simulations show that
uncertainties become larger as the number of bands is reduced. It is possible for TES to run, for example, on only four of the five bands acquired by ASTER with little degradation in performance, whereas for two bands the products are only half as precise. Running on four bands is advantageous if, for some reason, one band of data is noisy or is lost during transmission. TES is also readily applied to Tffi data sets acquired by scanners other than ASTER. However, directional effects at high scan angles such as found in MODIS data (45) may create difficulties that have not been considered in writing TES.

TES requires further testing on calibrated, atmospherically corrected TIR data before routine use by the remote-sensing community. However, we conclude that TES is a useful and general algorithm for recovering land surface temperature and emissivities from multispectral TIR imaging systems. The fundamental limitations are the variability in the E,j-MMD relationship for natural surfaces, and atmospheric characterization and correction. Future refinements of TES may be possible, but performance characteristics reported in this discussion are probably close to the ultimate limits achievable with its approach. 7. ACKNOWLEDGMENTS
We thank our colleagues on ASTER's TIE Working Group for helpful advice and criticism, and especially J. W. Salisbury for providing spectra. We also thank Elsa Abbott and Frank Palluconi for their tireless help. This research was a collaborative effort of the U.S. and Japanese EOS/ASTER instrument teams, sponsored by the NASA EOS Project and ERSDAC.

8. REFERENCES
1.

F. D. Palluconi and G. R. Meeks, Thermal Infrared Multispectral Scanner (TIMS): An investigators guide to TIMS

data, JPL Publication 85-32, Pasadena, 1985. 2. A. B. Kahie, F. D. Palluconi, S. J. Hook, V. J. Realmuto and G. Bothwell, The Advanced Spaceborne Thermal Emission and Reflectance Radiometer (ASTER), mt. J. Imaging Systems Tech. 3, 144-156, 1991. 3. Y. Yamaguchi, H. Tsu and H. Fujisada, A scientific basis ofASTER instrument design, SPIE Proc., pp. 150-160, 1993. 4. H. Fujisada and A. Ono, Anticipated performance of ASTER instrument in EM design phase, SPJE Proc., pp. 187-197,
1993.

5. C. Prabhakara and G. Dalu, Remote Sensing of surface emissivity at 9 im over the globe, J. Geophys. Res. 8 1(21),
3719-3724, 1976.

6. F. X. Kniezys, E. P. Shettle, L. W. Abreu, J. H. Chetwynd, G. P. Anderson, W. 0. Gallery, J. E. A. Selby and S.A. Clough, User guide to LOWTRAN 7, Air Force Geophysics Laboratory Report No. AFGL-TR-88-0177, Hanscom AFB,MassOl73l, 1988. 7. W.-H. Li, R. J. Weeks and A. R. Gillespie, Multiple scattering in the remote sensing of natural surfaces, mt. j. Remote
Sens., submitted, 1996. 8. K. Masuda, T. Takashima and Y. Takayama, Emissivity of pure and sea waters for the model sea surface in the infrared window region, Remote Sens. Environ. 24, 3 13-329, 1988. 9. I. J. Barton, Transmission model and ground truth investigation of satellite derived sea surface temperatures, J. Clim. Appi. Meterol. 24, 508-5 16, 1985. 10. L. M. McMillan and D. S. Crosby, Theory and validation of the multiple window sea surface temperature technique, .1. Geophys. Res. 89(C3), 655-3661, 1984. 11. C. Prabhakara, G. Dalu, and V. G. Kunde, Estimation of sea surface temperature from remote sensing in the 11- to 13j.tm window region, J. Geophys. Res. 79(33), 5039-5044, 1974. 12. 0. B. Brown, MODIS Infrared sea surface temperature algorithm, Algorithm Theoretical Basis Document, NASA Contract Number NAS5-3 1361, 10 November 1994. 13. J. C. Price, Land surface temperature measurements from the split window channels of the NOAA 7 Advanced Very High Resolution Radiometer, .1. Geophys. Res. 89, 723 1-7237, 1984.

SP!EVo!. 2817/93
Downloaded from SPIE Digital Library on 25 Jul 2011 to 129.137.163.64. Terms of Use: http://spiedl.org/terms

f7J

S1 V 'WNA 3uqdsowW

9J
L1

61

:z

pu ssuz uouno jo pu sain1E.I3dw pJnsw woij oqs uisn punoi2 uj .f wwaj suag '(zOzi 'O9t'Z-6i1Z 1661 Z UEA pU f 'iizoj jins-pu m.iodtu o iumainsoui wozj :ods jisAq sIdiDuudpu 3S1AUT 'uippoui pS0?9 0W9! I9S '()Lz '2LZ-89Z 6861 V I 'NN TI d 1flp pU 1 'qo jpp pJiJu! iiodsujnui jioii iuuos sisiqu ioj iojoa 'suot13ijddE 7ddy sdO '61 'O6ZZ-6LZZ 0861 puiqwoj JO 'Iq!s! P13LPl pamiju! pilE IuLItp poJLJU! s3Ew! 1oJ 'suiiqv v uoqq ptm y uiddu !!MEH AEI 'sojj j -s(ydoa 3j '(ta)96 'i8P-cLP T661 .irnxiw sisiu piQdsqnmJo 1IUP P.iJU! a1oui! S'U?g .UO.l1U 'Zt 'cN-LET 'oidsojji Z661 j pu z 'q sorput ui wiq piijui 'spuq 3JouII SUag tIOJhtU3 'z 'ELt 0661 s I IOOH V I 'iiO v UOi PE d S 'AJ!N uosudwoJo snbrnqoi 1oJ U1o.Ux I(UA!SSTUIUOflULIOJU! UIOiJ [ULItf1poiiju 1OJ 2OjO 'SijflS ?OW9J SU? UOJMU 'uos j13d OQ1fOIfltU .IOJ2UUflS1U'JUA!SSTUI "1 su3S'Zip 'cEI-zi Z661 '91 Z661 1LOJl4U 'Zip :i suuiainsm io

jo .uds iuissiuzuo sus2:ow! '(01)8 'zzcT-6oc! L861

dw

iuouiinswjo pu ojins idwi

'L

'p

jns

si uloij 'ujps iuJ 1

rj ipo

'jq

'

ids

N 'UOS1M tI

v STUN PU S

'uosi Z 1 Id
9Z V

nsCydoa '(0cc '6L-OL 066!

'uo1c'-1

'11I!N-1Wt1LflH IuLIoLLL p1E1Ju! Uo!1oIdx ui

1 'j'oui'i un.zthS 'z !P' P 1

A f V
SPVLL

pOqIW JOJ UUflSW 'JCTAtSSTUI SlJS U0.ThtU3! 'Zi7 I 'IZT-LI 66T SSAEUy JO s31301 cq E.flOdS poJVJUiUOtSSTUI 8) oi '(suol3tui UOd 7D '09 '9EL-cTL c961 s3I3o1 u!sn 'SWLL lLL SJVLLvwq siasj 'doysjio 81 - '61 'c861 'idsII!9 or2ooqrj U!ddUI JO uonoijqn% .c861 dd 'tt6Z

3ffS

qi U!P3 'pun u13qu0u 'PAN

8Z 6Z

oq jo iru.idmi pu :iUAissnu J(UAISSIUJ3 wru13ds'Uo!z!IEw1oU 30-1c1 pu 'doys,o, 'Idf uonEoqqtt-j 'cc-os 0661 woij 1(pouiai jYSUS sEw! utsn tp '!dd!d 1n1idT.uL pu JCIA!SSTUI iCpo
'pOq2W SUVJj WSOd
\T

sjp

'Idf UOflqM o d 'doys'jiotu s I IOOH uflEidS s pu i1iAiSSflu Ut paii.ju! iudsuinw iuus uonznjduq ioj UUAOO1 pu oj.ins 'sarn.idm szw.ij wsoa aow sues '(9)i i i-cci 'I9I E661 j unsi v uoids potpm usn u jopidw dqsuojo1 uioq qi 'uau' q:t

d s JC)J

IUAISSUUH

ii

)J

'pq9

?JOUI?(

UO11EWflSJo IciAiSstuI 'cc-06 0661

"SU2' '333J Ul uo!s!t19J

pu m11dui22U!Sfl qdJE 'SUTJJOOO

:oJd PUZ

S'MTLL

inidw

:p

'WflWXUI pu tfl WnW!UTW JO 'I'Z I661 stmdi fl!M

c 9E

g
6E

i i 'tiv ssq '2U3WflOp 3LIqdsoW2 ioj aaisv jULILp 'usITN-Jo1uof 'puj uIq!.Io2IE icquIopJ uoisjndoij'iClo2Eioqfl 'tUPEd fr66i j 'nwis s f 100H i' v 'i'ii swu. suonJusqo JO O3JJflS '11S-XH 3I iuvn wj suagoJs 'lddy 'jjj 'zuij 'A dd '9ZZZ-tZZZ v i '!dsII!O s '2mioa s 'oo-j j 2uns pu r uoemthsuIIwo iaiootp sisq '2U1T10P 9661 uosndoij'icioiioq'-1 'U3PSEd 91 S '!H pu v 'iii u. 'r ppj iussiui 'iwoipds wj wc u.ioq.iy paw4ujp7qsi8u!vwI 'EU3SEd 'VD '81-LI c661 .iwwo.ipadS(snuAv) 'dOyS.LO 7A 'Z if iq 'j-i f fl S POIO3O pijui-p j-) (uirl'O1Osnou2 'Ainqs i s iq pu uds jo
.LOAO

'u
1

'OLIE-cSvN LE 'dd qd t66T ci 'ruoonud o 'JAooH pu

4I ULI j11JUi JIAISSIUI 'uuuods JflOf OW?1 s'uas qsiu (piisq siuo jins-puj oioioqi sisq UUIn3Op

!)

UtIdVf UOTSJA

'(z)iT -o:z

miiduii mq.'o2

UOIS.IA

'

IS'l) '(ULV

'I 3E1:1UoD

UOflOJ.IO3 P01110W

'q

I1IAISSIUI U! S661

suv.LL

UO!SJfldO1j

'S3pOJ

ot'

it

'qPS A& 0 'llD P" S V '12flID LL MVLLQON LIZ 2Iod1 pu naiq PU O d 'uos.Ipuir sdqj 'q'-j si(qdocj 'E1o31!U 'SOdDfld 'ppom Aq WOOSUH 'aiv VW '1EL10 flU03'Z1O-D-T68Z96Id I IWAON 966L A 1'661 2f 'onui'j .iod jo swu. u.ioqi uo1qo :2uowuodx 32S3'W1 'uhiOJ!IED 6 uoisjndoij 'O1Ofl L "dd ot 'w I'66T A 'oinui ,i 'uoH v uuqi piijui uiddui JO 'I2 T y uoqq pu c j '!1!d

I X 'S1cZ3tU

IcAIns I!d-UO uodi '98918 'dd 8861 'LrnqsE'l S 11j'A pu Q 'UyQ J!AISSIUI3 JO 1U?J U?g UOJhtUg 'Zt' '901-ES z:661

''1V I LS NY&LAO'1

SjiW U! fl iI8 iiiil ziqdsout1E'Mopu!M

74 'nqy 0 d

pp

'UOS1pU1T

'j

I H 'puM3q

d 'IU4S V "Ifl '- 'uisuio ci D 'U011I01d

J3qO1D 8861 U1flEdnN

'u 'Pi njr{

'OUOIOA

.{'ojouix;jo1'cc 'WEe Z661

sdnj

3jdS/6 Ion

LL9Z

Downloaded from SPIE Digital Library on 25 Jul 2011 to 129.137.163.64. Terms of Use: http://spiedl.org/terms