NASA Technical Reports Server (NTRS) 20030014735: Simulation and Correction of Triana-Viewed Earth Radiation Budget with ERBE/ISCCP Data

5 Pages·0.37 MB·English

by  NASA Technical Reports Server (NTRS)

Simulation and Correction Of Triana-Viewed Earth Radiation Budget
with ERBE/ISCCP Data
Jianping Huang
AS&M, Inc.
Hampton, VA USA
Patrick Minnis
NASA Langley Research Center
Hampton, Virginia USA
David R. Doelling
AS&M, Inc.
Hampton, VA USA
Francisco P. J. Valero
Atmospheric Research Laboratory
Scripps Institute of Oceanography
La Jolla, CA USA
IGARSS 2002
Toronto, Canada
24-28 June 2002
SIMULATION AND CORRECTION OF TRIANA-VIEWED
EARTH RADIATION BUDGET WITH ERBE/ISCCP DATA
J. Huang 1,P. Minnis 2,D. R. Doelling 1,F. P. J. Valero 3
1AS&M, Inc.
Hampton, Virginia USA
2Atmospheric Sciences
NASA Langley Research Center
Hampton, Virginia USA
3Atmospheric Research Laboratory
Scripps Institute of Oceanography
La Jolla, CA USA
ABSTRACT Triana will be placed in an elliptical Lissajous orbit about
This paper describes the simulation of the earth L1 and will vary from 4° to 15° about the Earth-sun line.
radiation budget (ERB) as viewed by Triana and the It will continuously measure between 92 and 97% of the
development of correction models for converting Triana- sunlit Earth, never viewing the dark side of the planet.
viewed radiances into a complete ERB. A full range of Triana's Earth-viewing instruments consist of the Scripps-
Triana views and global radiation fields are simulated EPIC (Earth Polychromatic Imaging Camera), a 10-
using a combination of datasets from ERBE (Earth channel imager, and the Scripps-NISTAR (National
Radiation Budget Experiment) and ISCCP (International Institute of Standards and Technology Absolute
Satellite Cloud Climatology Project) and analyzed with a Radiometer), a single-pixel 4-channel broadband cavity
set of empirical correction factors specific to the Triana radiometer. Data from the Scripps-EPIC and Scripps-
views. The results show that the accuracy of global NISTAR will be used to monitor the Earth's radiant power
correction factors to estimate ERB from Triana radiances and to analyze weather systems and cloud patterns in an
is a function of the Triana position relative to the entirely new way.
Lagrange-1 (L1) or the Sun location. Spectral analysis of ERB measurements at the top of the atmosphere
the global correction factor indicates that both shortwave (TOA) are fundamental quantities for monitoring the
(SW; 0.2 - 5.0 gm) and longwave (LW; 5 -50 gm) global climate system. These measurements have
parameters undergo seasonal and diurnal cycles that traditionally been obtained from ERB instruments on
dominate the periodic fluctuations. The diurnal cycle, polar-orbiting satellites that typically view a region of the
especially its amplitude, is also strongly dependent on the earth only twice each day. They cannot provide
seasonal cycle. Based on these results, models are continuous spatial coverage of the Earth's entire surface
developed to correct the radiances for unviewed areas and at a specific time or provide continuous temporal
anisotropic emission and reflection. A preliminary coverage for a specific location. NISTAR is designed to
assessment indicates that these correction models can be facilitate determination of the radiation budget for an
applied to Triana radiances to produce the most accurate entire hemisphere every 10 min or less from a single set
global ERB to date. of 3 measurements. However, to derive the ERB from
these radiances, it is necessary to convert each radiance to
I. INTRODUCTION a flux and to account for the radiation field of the dark
half of the earth. The Earth's surface and atmosphere are
Triana is designed to continually monitor the sunlit anisotropic reflectors and emitters resulting in a relatively
side of the earth and promises to offer new insights into complex variation of radiance leaving the Earth as a
how our planet's climate works as an integrated system function of the viewing and illumination conditions.
[1]. It will be the first Earth-observing satellite in a Triana views the Earth from a limited range of angles
Lagrange-1 (L1) position at distance of roughly 1.5 corresponding to scattering angles between 165 ° and
million kilometers from the Earth. The L1 point is the 176 °. Therefore, to convert radiance to flux requires the
location where the Earth's gravitational field equally use of anisotropic directional models (ADM) to account
counters that of the sun. Since the strength of the for the emittance and reflectance anisotropies.
gravitational attraction determines the orbital period, Additionally, a sliver of the sunlit Earth (missing light) is
Triana will orbit the sun at the same rate as the Earth.
outofview(replacebdya darksliver)becaustehe
satellitweillnotbepositioneedxactloyntheEarth-sun
line.Thismissinglightmusbtetakenintoaccounfotra
compleEteRBF. inallyn,oLWmeasuremeanrettsakenat
nightT.huss,omemeanissneedetodaccounfotrtheLW
fluxesatnightT. hispapedrescribethsesimulatioonfa
Triana-viewEedRBandthedevelopmeonftcorrection
modeulssingcombinatioonfsERBEandISCCPdata.
II.ERBE/ISCDCAPTA
ERBEdatawerecollectefdromthreesatellitetsh,e
EarthRadiatioBnudgeStatellitEeRBSN, OAA-9a,nd
NOAA-10[2].EachsatellitemeasurebdroadbanSdW
0.20 0.40 0.60 0.80 1.00
andLWradiancewsithcross-tracskcannearsndwide- REFLECTANCE
field-of-vierwadiometeTrsh.escanndeartaareusedhere Fig.1.Simulated Triana-viewed reflectance from ERBE data
becautsheeyprovidtehespatiarel solutioanndscenteype forMarch 21, 1986at15°east fromL1.
informationneededto converrtadiancetso fluxes.
Becausethe ERBEdataare temporallysparse, III. SIMULATION OF TRIANA-VIEWED ERB
interpolatitoenchniquaerseusedtoestimattheevalueast
eachhourbasedononlya fewsamplepserday.To The simulation of the Triana-viewed radiances is
improveontheERBEsampling3,-hourlgyeostationary used to generate correction factors that can be applied to
dataareusedto fill thegapsbetweentheERBE NISTAR-measured radiances to compute global SW
measureme[n3t]s.ISCCPDXgeostationa3r-yhourly albedo and LW flux. Having the hourly fluxes at a 2.5 °
visible(0.65gin)andinfrare(dIR,11gin)radiance[4s] scale, it is possible to compute the Triana-viewed
weregriddedintothe2.5° equal-angrelegionussedby radiance at any set of viewing and illumination angles by
ERBEN.arrowband-to-broacdobnavnedrsioconefficients specifying the location of the satellite in the Lissajous
werecomputefdoroceanl,andandsnowtousein orbit, computing the solar and viewing zenith angles and
standafrodrmula[e5]foreachgeostationsaarytellitaend relative azimuth angle for each viewed region, and then
monthusingcoincideanntdcollocateEdRBEfluxesand multiplying the SW and LW flux for the region by the
ISCCPfluxesT.helatterwerederivedfromtheISCCP bidirectional reflectance and limb-darkening factors for
radiancesby applicationof ERBEbidirectional the scene [6, 7]. The radiances and fluxes are then
reflectancaendlimb-darkenimngodels[6,7].TheIR integrated over the viewed disc. The fluxes for the entire
coefficientswerenot basedon scenetype.The globe are also computed for the same time.
narrowband-to-broacdobnavnedrsiownesrethenapplied Figure 1 shows an example of Triana-viewed
toalloftheISCCPradiancetoscomputaeSWalbedo regional reflectance assuming an offset position of 15°
East from L1. The dark sliver on the left-hand side of the
andalongwavfeluxforeach2.5° regionT.heISCCP-
derivedbroadbanflduxeswerethennormalizetdothe figure corresponds to the unilluminated portion of the
existinEgRBEfluxeussingtheratiooftheERBEfluxto Triana-viewed half disc. A corresponding sunlit sliver
theISCCPfluxatthenearetsimt eofcoincidenfcoerboth behind the right side of the disc out of the field of view is
satelliteSs.incpeolar-orbitisnagtellitecsrossthepoles14 referred to as the missing light area. It follows that the
timesaday,theISCCPbroadbanflduxeswereonly global albedo at any given time must account for the
applietdoregionbsetwee6n0°Sand60°NT. hedataset anisotropy of the viewed SW radiance such that the
thenconsisotsf3-hourlIySCCP-basfleudxesandscene albedo derived from a radiance for the viewed disc should
informatio(pnercecnltoudcovers,urfacteype)combined match the integrated albedo from a scene like that in Fig.
withERBEfluxesandscenteypesatothehroursF. luxes 1. Additionally, the contribution of the missing light area
fortheremaininhgourwsithoudtataarefilledinusingthe must be taken into account so that the albedo of the
ERBEinterpolatimonethodtosprovidaecomplehteourly measured area is adjusted to match the global albedo. The
datasefotrtheentireglobeT.heinclusioonftheISCCP first SW correction factor is the global reflectance
dataprovidea morerealistichourlydatasetthan anisotropic factor (GRAF) that is simply the ratio of the
previousulysed[8]becaustheeISCCPdatameasutrhee disc albedo to the disc reflectance. The missing light or
actuaclhangeinscloudcovearndtheresultinfgluxetshat albedo correction factor (ACF) is the ratio of the global
wereonlyestimatewdhenusingtheERBEdataalone[8]. albedo to the disc albedo. Thus, the global albedo for a
Thiscombinatioofnmultiplesatelliteissusedtosimulate given view then is the product of the measured
globaral dianctehsatwouldbemeasurbeydTriana. reflectance, the GRAF, and the ACF for the particular
viewed disc.
(a) Albedo Correction Factor (a) 15°East from L1
N
1.010
1.005
1.015
E 1012
1.000
1010 2 3 4 5 6 7 8 9 10 11 12
Time(month)
I 1014
1004
(b) 15°West from L1
s
(b) OLR Correction Factor 1.010
N
1.005
1.015
1.000
2 3 4 5 6 7 8 9 10 11 12
E O994 Time(month)
Fig. 3. Predicted (dashed line) and simulated 1988 missing light
O99O albedo correction factors (solid line) for (a) 15°East from L1;
I O998 (b) 15°West from L1.
0986
S IV. DEVELOPMENT OF CORRECTION MODELS
Fig.2. Monthly mean global Triana missing light albedo and
nighttime correction factors as functions of LO at 0000 UTC for Time series analysis indicates that the long-term
September 1987. trend, diurnal and seasonal variations are also significant
for all of the correction factors. Therefore, the correction
Similarly, the observed disc LW radiance must be model should include the long-term trend, the seasonal
corrected for limb darkening and for the unviewed part of cycle, and the seasonally modulated diurnal cycle. For
the Earth at each hour to compute the global outgoing LW any time and position, the general correction model can
radiation (OLR). The global limb-darkening factor (LDF) be written as:
is the ratio of the disc OLR to the disc radiance, while the
nighttime OLR correction factor (OCF) is the ratio of the Y(r_, re,0,0)=Co(O,o)+cl(o,o)*r_ +G(0,0)*r_ 2 +
global OLR to the disc OLR. c3(o,O)*cos(2xfjTj) +c4(o,O)*cos(2xfgTg)+
Figure 2(a) shows the monthly mean values of the cs(o,O)*cos(2_fjTj )*cos(2_f gTg) (1)
global ACF as a function of L1 offset position, LO, at
0000 UTC, September 1987. The values of ACF increase where ¢ is the angle from North (in degrees), 0 is the
monotonically with distance from L1. The maximum
distance from L1 (in degrees), and fj and fg are the
missing light albedo factor appears near 15° N or 15°S off frequencies of the seasonal and diurnal cycles,
L1 because significant portions of the highly reflective respectively. The last term in equation (1) represents the
polar regions are within the relatively large unviewed
effect of modulation. The coefficients, (ci(o,o), I=0 .... 5)
sliver of the sunlit disc. At the equinox, both poles are
were computed using a least squares multiple regression
equally sunlit. However, observing only one pole is
fit for each position. These models can be used to predict
apparently not sufficient to compensate for the unviewed
the hourly correction factors for any given time (Tj, Tg)
polar region. Values for the GRAF (not shown) can be as
and position (¢, 0). Three years (Jan. 1985 to Dec. 1987)
large as 1.2 as LO approaches L1. Figure 2(b) shows the
of simulated data are used as the historical dataset to
mean values of OCF for the same time and month. The
determine the regression coefficients. The resulting
patterns of the nighttime OLR correction factor are very
coefficients were used to predict the factors for 1988 and
different from those for albedo (Fig. 2a), which is more
then compared with the simulated data. Examples of the
symmetrical around the L1 point. The lowest values of
predicted missing light albedo correction factors at 2
OCF (-0.986), result when LO is northeast of L1 and
maximum LOs are shown in Fig. 3. The predicted (dashed
increase gradually towards the southwest to values of
line) and simulated (solid line) albedo correction factor
-1.0. This plot indicates that the Triana-viewed disc OLR
values are in good agreement. The respective mean and
is 1.64% greater than the global OLR at the southeast standard deviation of the differences are 0.00021 and
point. Values for LDF (not shown) can be as small as
0.0017. Preliminary results indicate that the correction
0.970.
model is capable of resolving the global correction factor.
VCONCLUSAIONNDDSISCUSSIONS
Thispaperprovidesabasicframeworfkorthe
simulationof Triana-vieweEdRB.Thepreliminary
predictiornesultsindicatethatthesecorrectiomnodels
canbeusedtoproductehemostaccuragtelobaElRBto
dateH.owevethr,ecorrectiomnodelasrepurelystatistical
anditisnotpossibtloedistinguisbhetweepnhysicaalnd
randomrelationshiinpsthedataA.lsot,hemodemlsaybe
biasebdythesamplinpgatternosftheERBEsatellites.
CorrectiofnactorsbasedonthecombineEdRBEand
ISCCPdataseptsrovideamoreaccurarteepresentaotiof n
thediurnaclycleA. moreadvanceadpproacuhsingcloud
informatiofrnomtheEPICisalsoundedrevelopmetont
explicitlyaccounfotrphysicavlariationinseachviewed
scenethatarenottakenintoaccounwtiththepurely
statisticaml ethodusedhere.In themeantimeth,e
techniqudeevelopheedreshoulpdrovidaehighlyreliable
methofdormonitorinthgeglobarladiatiobnalancferom
Triana.
REFERENCES
[1]F.P.J.ValeroJ,.HermanP,.MinnisW, .D.Collins,
R.SadournWy,.WiscombDe.,Lubina,ndK.Ogilvie:
"TrianaaDeepSpacEearthandSolaOr bservatory,"
NASAbackgrournedport1,999.
[2]B.R.BarkstroEm.,F.HarrisoGn,.L.Smith",Results
fromtheEarthRadiatioBnudgeEtxperiment
(ERBE),"Adv.
Space Res., 9, pp. 775 782, 1989.
[3] D. F. Young, P. Minnis, D. R. Doelling, G. G. Gibson
and T. Wong: "Temporal interpolation methods for the
Clouds and the Earth's Radiant Energy System
(CERES) Experiment," JAppI. MeteoroI, 37, pp. 572-
590, 1998.
[4] W. B. Rossow and R. A. Schiffer, "Advances in
understanding clouds from ISCCP," Bull. Am.
MeteoroI. Soc., 80, pp. 2261-2287, 1999.
[5] P. Minnis and W. L. Smith, Jr.: "Cloud and radiative
fields derived from GOES-8 during SUCCESS and the
ARM-UAV Spring 1996 Flight Series," Geophys. Res.
Lett., 25, pp. 1113-1116, 1998.
[6] J. T. Suttles, R. N. Green, et. al.: "Angular radiation
models for Earth-atmosphere system: Volume 1-
Shortwave radiation," NASA RP 1184, VoI. I, 1988.
[7] J.T. Suttles, R. N. Green, et al., "Angular radiation
models for the Earth-atmosphere systems: Volume II:
Longwave radiation," NASA RP 1184, Vol. II, 1989.
[8] P. Minnis, J. Huang, D. R. Doelling, and F. P. J.
Valero: "Simulation and correction of Triana-viewed
Earth radiation budget with ERBE data," Proc. SPIE
Conf. on Sensors, Systems, and Next-Generation
Satellites V, Toulouse, France, Sept. 12-17, pp. 391-
401, 2001.