Full report

Original page 1

NASA Technical Memorandum 104566, Vol. 9 SeaWiFS Technical Report Series Stanford B. Hooker, Editor NASA Goddard Space Flight Center Greenbelt, Maryland Elaine R. Firestone and A. W. Indest, Technical Editors General Sciences Corporation Laurel, Maryland Volume 9, The Simulated SeaWiFS Data Set, Version 1 Watson W. Gregg NASA Goddard Space Flight Center Greenbelt, Maryland Frank C. Chen, Ahmed L. Mezaache, Judy D. Chen, and Jeffrey A. Whiting General Sciences Corporation Laurel, Maryland National Aeronautics and Space Administration Goddard Space Flight Center Greenbelt, Maryland 20771 1993

Original page 2

Original page 3

W.W. Gregg, F.C. Chen, A.L. Mezacche, J.D. Chen, and J.A. Whiting ABSTRACT Data system development activities for the Sea-viewing Wide Field-of-view Sensor (SeaWiFS) must begin well before the scheduled 1994 launch. To assist in these activities, it is essential to develop a simulated SeaWiFS data set as soon as possible. Realism is of paramount importance in this data set, including SeaWiFS spectral bands, orbital and scanning characteristics, and known data structures. Development of the simulated data set can assist in identification of problem areas that can be addressed and solved before the actual data are received. This paper describes the creation of the first version of the simulated SeaWiFS data set. The data set includes the spectral band, orbital, and scanning characteristics of the SeaWiFS sensor and SeaStar spacecraft. The information is output in the data structure as it is stored onboard. Thus, it is a level-0 data set which can be taken from start to finish through a prototype data system. The data set is complete and correct at the time of printing, although the values in the telemetry fields are left blank. The structure of the telemetry fields, however, is incorporated. Also, no account for clouds has been included. However, this version facilitates early prototyping activities by the SeaWiFS data system, providing a realistic data set to assess performance. 1. INTRODUCTION The Sea-viewing Wide Field-of-view Sensor (SeaWiFS) is an ocean color sensor scheduled for launch in 1994. A follow-on to the highly successful Coastal Zone Color Scanner (CZCS) mission (1978-1986), SeaV_'iFS is intended to provide improved estimates of chlorophyll concentrations in the world's oceans. The SeaWiFS mission will provide an unprecedented set of global ocean color data on a regular basis (every two days in the absence of clouds, about a week otherwise). The CZCS, a proof-of-concept mission, was able to produce a single global image in its nearly eight year lifetime. The volumes of data produced by an operational global ocean color sensor pose challenges for a data processing system that were not encountered by the CZCS system. Development and rigorous testing activities must begin well before launch to avoid the sorts of delays in data processing and distribution that plagued the CZCS mission. To facilitate the development of the data system for processing SeaWiFS data, it is essential to develop a simulated SeaWiFS data set as soon as possible. This data set should incorporate the characteristics of the SeaWiFS sensor, including spectral bands and orbital and scanning characteristics, and should do so in a realistic manner to allow proper preparation for actual data processing. While many of the data characteristics, contents, and anomalies of geometries. A comparison between SeaWiFS orbital and mav not be known until just before launch, development the simulated data set can assist in identification of problem areas that can be addressed and solved before a fuller understanding of the data is achieved. This paper describes the details of the creation of the first version of the simulated SeaWiFS data set. The data set includes the spectral band, orbital, and scanning characteristics of the SeaWiFS sensor and SeaStar spacecraft. is data recorder area. There is estimated to be about 20 The information is output in the data structure as it stored onboard. Thus, it is a level-0 data set which can be taken from start-to-finish through a prototype data system. The data set is complete and correct at the time of printing, although, the values in the telemetry fields are left blank (the structure of the telemetry fields is incorporated, however). Also, no account for clouds has been included. CZCS pigment data and a variety of radiative transfer models were used to create realistic SeaWiFS total radiances. These total radiances were Earth-located using a SeaWiFS orbital model to provide SeaWiFS-like viewing and solar geometries. 2. BACKGROUND The improvements of SeaWiFS over CZCS are largely expected due to the existence of four new spectral bands: one at 412nm (to discern dissolved organic matter, or Gelbstoff), one at 490 nm (to enable better estimates of chlorophyll at high concentrations), and two in the nearinfrared at 765 and 865 nm (to improve the determination of atmospheric aerosols). Thus, SeaWiFS contains eight. spectral bands compared to the four used in CZCS processing (CZCS actually contained six bands, but two were not used). Table 1 shows a comparison of the SeaWiFS spectral bands to the CZCS bands. In addition to different spectral bands, SeaWiFS is also in a different orbit than the CZCS and has different viewing sensor characteristics and those of the CZCS is summarized in Table 2. Due to limited onboard data recorder space, SeaWiFS will be able to obtain global coverage only at reduced resolution (every fourth pixel along-track and along-scan). These data are called Global Area Coverage (GAC) data. High resolution data may be recorded in the remaining minutes per day of high resolution Local Area Coverage

Original page 4

TheSimulatedSeaWiFSDataSet,Version1 (LAC) data. Theeffort hereis to createboth typesof solar and viewing geometries, a realistic atmosphere was data,in orderto simulatea representativeSeaWiFSdata added by including Rayleigh scattering, aerosol scattering set. Table 1. SeaWiFSspectralbandsand center wavelengths,areshownwith thoseof CZCS,for comparison.Wavelengths(A)arein nm. Sea WiFS CZCS Band No. Band No. 1 412 1 443 2 443 2 520 3 490 3 550 4 510 4 670 5 555 5 7501 6 670 6 13,5002 7 765 8 865 1. Intended for surface vegetation analyses only (Williams et al. 1985). 2. Data suspect after 1979 (\Villiams et al. 1985). 3. METHODS and absorption, water vapor absorption, oxygen absorption, and sun glint reflectance. Radiance contributions of each were summed, and diffusely transmitted waterleaving radiance was added to produce simulated total radiance received by the sensor. Finally, data structures for telemetry were inserted and science data were formatted into 10-bit words to create level-0 data. 3.1 Normalized Water-Leaving Radiances Creating normalized water-leaving radiances Lw_,_ is the first step, because LwN is independent of viewing and solar geometry. The method for creating LWN is taken from Gordon et al. (1988a). According to this model, LwN(A) = (i - pn)(i - p_)F0()R(0-,) n2Q [1 - rR(O-, A)] (2) where Pn is the Fresnel reflectance of the sea surface for normal incidence, PN is a normalized mean value of surface reflectance for direct and diffuse irradiance for a flat sea, In order to maximize the usefulness and realism of the F0 is the extraterrestrial irradiance corrected for Earthsimulated SeaWiFS data set, CZCS pigment data were sun distance, R(0-,A) is the irradiance reflectance just chosen as the data source. The method for simulating below the sea surface, n is the index of refraction, Q is SeaWiFS data involves creating both LAC and GAC data. the irradiance-to-radiance ratio (equals 7r for totally dif- Details of the method are described in the following, but fuse radiance), and r is the water-air reflectance for totally a general overview here is helpful. diffuse irradiance. The basic equation of radiative transfer for ocean color From Fresnel's Law, Pn is known a priori (Jerlov 1976), assumes separability of the radiance contributions from at- PN is estimated from Gordon et al. (1988a) at 0.043, n is mospheric and oceanic components: taken to be 1.341 (Austin 1974), and r is taken to be 0.48 (Gordon et al. 1988a). The ratio R(A)/Q is wavelength Lt(A) = L_(A)+T(A)Lg(A )+L.(A)+T(A)LW(A) (1) dependent, and is taken from Gordon et al. (1988a) as: where Lt is the total radiance at the sensor, Lr is the R(A) bb(A) Rayleigh radiance, La is the sun glint radiance, L_ is the Q - 0.110 K(A) (3) aerosol radiance, Lw is the water-leaving radiance, and the where bb(A) is the spectral backscattering coefficient, and T is the total transmittance (direct plus diffuse) from vari- K(A) is the spectral attenuation coefficient. K(A) is deterocean through the atmosphere to the spacecraft. The Lt, mined from Baker and Smith (1982) as able of interest in the creation of simulated data is which is what the sensor actually detects. To obtain realistic values of Lt, with ranges and means observed the real ocean, it is desirable to create it using representative values of the components. This process begins with K(A) = Kw(A) + Kc(A) + Kg(A) (4) in where Kw (A) is the attenuation coefficient of pure seawater, Kc(A) is that for phytoplankton, and Kg(A) is that for the generation of Lw derived from CZCS pigment values, Gelbstoff, or yellow substances. and adding atmospheric contributions from simulated Sea- WiFS viewing and solar geometries. For both LAC and GAC data, real CZCS images pigments were selected. The pigment fields were used to Use of this model for a SeaWiFS simulation requires values for wavelength-dependent parameters be chosen for of SeaWiFS bands. Values for the mean extraterrestrial irradiance F0 were taken from Neckel and Labs (1984), and create normalized water-leaving radiances, LWN(,,), at the spectrally weighted over all of the SeaWiFS bandwidths eight SeaWiFS bands using a semi-analytic radiance model modified for SeaWiFS (Gordon et al. 1988a). Then, simulated SeaWiFS orbits were propagated through these simulated radiance fields to produce actual SeaWiFS viewing geometries. An arbitrary day was chosen (the vernal equinox) to produce solar geometry. Given these simulated assuming a full-width half-maximum (FWHM), Gaussian spectral response (Table 3). The values were corrected for Earth-sun distance following Gordon et al. (1983): Fo() = _0(A) 1 +0.0107cos 5_g (5)

Original page 5

W.W. Gregg, F.C. Chen, A.L. Mezacche, J.D. Chen, and J.A. Whiting Table 2. Orbit and sensor characteristics for the Sea\ViFS and the CZCS, shown for comparison. Characteristic Altitude (kin) Period (minutes) Inclination Equator Crossing Time (local) CZCS Sea IViFS 955 705 104.0 98.9 99.28 ° 98.25 ° Noon Noon Node Type Ascending Descending Scan Width Instantaneous Field of View (IFOV) Ground IFOV at Nadir (km) Pixels Along Scan Scan Period (seconds) -t-20° in 2 ° steps +20o, oo, 2oo Scan Plane Tilt Scan Ground Coverage (km) Maximum Spacecraft Zenith Angle Digitization (bits) where F0(A) is the mean extraterrestrial irradiance and D is the sequential day of the year. Table 3. Values for mean extraterrestrial irradiance F0(A) (mW cm -2 zm-1), kc(A), k'c(A ), bw(A) (m-l), A(A) (m-l), and B(A) used in the creation of the SeaWiFS simulated data set. A Fo kc k'c bur A B 410 171.92 0.208 1.077 0.0067 0.00313 0.20401 443 189.05 0.175 1.001 0.0048 0.00300 0.21770 490 193.60 0.121 0.963 0.0031 0.00284 0.23674 510 188.41 0.103 1.006 0.0026 0.00346 0.34070 555 185.90 0.076 1.144 0.0019 0.00330 0.35666 670 152.82 0.137 1.463 0.0008 0.00238 0.29579 765 123.27 0.040 1.732 0.0005 0.00221 0.32081 865 100.34 0.010 1.732 0.0003 !0.00206 0.34400 The values for Kw (A) were taken from Baker and Smith 39.34 ° 58.3 ° (LAC); 45 ° (GAC) 0.05 ° 0.09 ° 0.825 1.12 1,968 1,285 (LAC); 248 (GAC) 0.134 0.667 (LAG); 1.5 (GAC) 1,566 2,802 (LAC); 1,502 (GAC) 46.8 ° 70.8 ° (LAC); 51.7 ° (GAC) 8 10 scattering ratio for pure seawater. The backscattering coefficient for plankton, bbc(A), was determined from an empirical relationship developed by Gordon et al. (1988a) bbc(A) = A(A)[chl. a]s(a) (8) where, the coefficients A(A) and B(A) were computed for SeaWiFS bands as shown in Table 3. The only remaining unknown for the model, then, is [chl. a], which was obtained from CZCS data. 3.2 Orbit Model A simulated SeaWiFS orbit is required to obtain realistic viewing and solar geometries, which are then used to produce the atmospheric contribution to the total radiance detected by the sensor. The method used here is to propagate a SeaWiFS orbit over a field of CZCS pigment values, (1982), and again spectrally weighted FWHM assuming a use the values to compute LwN (A) for Sea\ViFS bands, and Gaussian response (Table 3). Kc(A) was computed from Baker and Smith (1982): Kc(A) = kc(A)[chl, a]exp -k;(A) log10 [chLCrerallJ 2 (6) + O.O01[chl. a]2 where [chl. a] is chlorophyll concentration, Cref is a reference chlorophyll value (0.5), and k(A) and k(A) are spectral fit coefficients and were spectrally weighted over the SeaWiFS bands (Table 3). For this simulation, Kg(_ ) was assumed to be zero. The backscattering coefficient bb() may be expressed using components similar to the K(A) formulation (4): bb(A) = 0.5bw'(A) + bbc(A) (7) where bw(A) is the total scattering coefficient for pure seawater and bbc(. ) is the backscattering coefficient of phytoplankton. The coefficient 0.5 is the backscattering-to-total use the simulated scan and tilt positions to determine the location of the pigment value in Earth coordinates (geocentric latitude and longitude). Given this information, the spacecraft and solar zenith and azimuth angles can be calculated and then used to simulate the radiative properties of an atmosphere. This section describes the orbit model employed, the equations to compute latitude and longitude, and the viewing and solar geometries. The orbital model used assumes a circular orbit. Thus, the only parameters required to propagate a simulated orbit were inclination, altitude, and epoch (see Table 1 for SeaWiFS inclination and altitude). The epoch was chosen such that the first orbit crossed the equator on the descending node at the Greenwich Meridian at exactly noon local time. In such a simple orbit model, the location of successive sub-satellite points (the point on the Earth where the projection of the satellite lies) is a simple function of the 3

Original page 6

TheSimulatedSeaWiFSDataSet,Version1 previoussub-satellitepointandthe traveltime between The pixel longitude is determined from Ace by either thetwo points.Fromsphericaltrigonometry, sin[%(t - to)] = cos(5 sin [s(t0)] + sin(5 cos [q2s(t0)] cos(/') where @s is the sub-satellite latitude at times to and t-to, adding to, or subtracting from, the sub-satellite longitude, depending on which side of the scan the pixel is on; if it is west of the sub-satellite point a subtraction is performed. (9) 3.4 Viewing and Solar Geometries is the great circle distance from (t0) to g2(t-to), and Once the pixel has been navigated to Earth coordii' is inclination i - 90 °. The great circle distance is known nates, computation of viewing geometry (spacecraft zenith from At (5 = 360 (10) and azimuth angles) is straightforward. In fact, the spacecraft zenith angle 0 has already been found (14). This leaves the computation of the spacecraft azimuth angle, where P is the nodal period. The longitude difference, which is defined as the angle from the pixel to the sub- A_os, is calculated similarly: [cos(5-sin[@s(t- to)]sin[@s(to)]] (11) -- oos-L J The new longitude, w, equals the old longitude 0J0 mi- Earth where _ is the spacecraft azimuth angle (the other terms nus the longitude difference AaJ plus a correction for rotation (12) The determination of the solar geometry requires addiw = w0 - Aw + 0.004167t satellite point, measured counterclockwise from true north (at the pixel) sin _ - cos8 sin 5 = sin(5 cos (17) are defined above). tional information to the pixel location. Specifically, this where the coefficient 0.004167 is the rate of rotation (° s- 1). additional information includes the solar declination lat- 3.3 Navigation of Pixels Given sub-satellite longitude and latitude, and the scan itude II/d, equation of time Te [both obtained using standard methods, e.g., Iqbal (1983)], and the Greenwich Mean Time (GMT). GMT is determined by and tilt angles, the Earth position of any given pixel is cal- O_Je distance HGMT -- tEc + t, (18) culable. First, one must calculate the great circle from the sub-satellite point to the pixel: 15 where HGM T is GMT in hours, tEC is the equator crossing (5 = 0-0' s (13) time (in local time, i.e., noon) w, is the equator crossing where 0 is the spacecraft zenith angle, and O's is the scan angle Os modified by tilt. Given 05, O's is computed a simple pitch rotation in the direction of the tilt. The spacecraft zenith angle, 0, is computed directly from longitude, and t_ is the time difference to or from the most recent equator crossing to the present position (in hours). by Given this information, plus the pixel location, the solar zenith angle 00 is computed from cos00 = sine2sin_d + COSO2COS_dCOSft (19) 0 = sin-l[ Re+HssinO's]Re (14) where Q is the solar hour angle, computed from where Re is the mean Earth radius (6371.2 kin) and Hs is = (H-12)15 + w + Te. (20) the spacecraft altitude (705 km). The pixel latitude and longitude are determined from The solar azimuth angle q)0 is computed by spherical trigonometry. For the latitude: = sin -1 [sin_s cos(5 + cosq_s sin_ cosrl] (15) sinff/d -sinkO cos 0o] where rI is the bearing from the sub-satellite point to the pixel along the direction of motion of the satellite (either 270 ° + i' for the left-hand side of the scan facing north or 90 ° - i' for the right-hand side facing south). The longitude difference, Aw, from the sub-satellite point to the pixel is determined by A_ -- COS-1 Icos C'-OS--(5 -sinco'--s"qJssin k_] (16) 4 ¢0 = cos-' co;- 2E0 J (21) The viewing geometries 0 and ¢ and solar geometries 00 and q_0 are now known and attention may be focused on the addition of an atmosphere. 3.5 Atmospheric Contributions This section focuses on the computation of realistic atmospheric contributions to the total radiance, based on representative atmospheric conditions and the simulated

Original page 7

W.W.Gregg,F.C.Chen,A.L.Mezacche,J.D.Chen,andJ.A. Whiting SeaWiFSviewingandsolargeometries.This meansdewa- contribution L_ to the total radiance received at the satelterminingthe Rayleighscattering,ozoneabsorption, The next step is to compute the Rayleigh scattering scatter- lite. Since multiple scattering Rayleigh tables are not yet ter vaporabsorption,oxygenabsorption,aerosol Asa first available for SeaWiFS bands, the tables for the CZCS ingandabsorption,andsunglintcontributions. must bands are used (Gordon et al. 1988b). There are two costep,absorptioncoeffÉcientsfortheatmosphericgases overthe SeaWiFSincident bands for SeaWiFS and the CZCS (see Table 1): bedeterminedandspectrallyweighted irradiance.one at 443 nm and the other at 670 nm. The CZCS multibands,aswasdoneforthemeanextraterrestrial andoxygenple scattering Rayleigh tables are used for these two bands. Absorptioncoefficientsforozone,watervapor, of whichallthree Values for non-coincident SeaWiFS bands are derived as- (thespectralabsorptioncharacteristics are suming a A-4 principle. For bands 1-5 (412-555 nm), the fall to someextentwithinsomeSeaWiFSbandwidths) thicknessprocedure is as follows. Let A = 443 nm; then, shownin Table4,alongwiththeRayleighoptical for theSeaWiFSbands. Table 4. Spectrallyweighted(FWHM)Rayleigh opticalthickness(Tr),ozoneabsorptioncoefficient (aoz),watervaporabsorptioncoefficient(awv),and oxygenabsorptioncoefficient(ao×).Wavelengths, A, arein nm. Theunitsfor the absorptioncoefficientsarein cm-1 Tris dimensionless. A T r aoz Owv aox 410 0.3191 0.0000 0.0000 0.0000 443 0.2364 0.0027 0.0000 0.0000 49O 0.1562 0.0205 0.0000 0.0000 510 0.1326 0.0382 0.0000 0.0000 555 0.0938 0.0898 0.0000 0.0000 670 0.0437 0.0463 0.0006 0.0000 765 0.0255 0.0083 0.0000 6.9900 865 0.0155 0.0000 0.0008 0.0000 Following Gordon et al. (1983), the effects of the absorbing gases in the atmosphere were determined by assuming two trips of the extraterrestrial irradiance through the atmosphere. This was accomplished by Fd = Yoro_(00)Toz(O)T,,v(Oo)T,,.v(O)Tox(Oo)Tox(O) (22) where wavelength dependence has been dropped and T represents the transmittance of the gaseous component in subscript. The 00 and 0 dependencies represent the two paths through the atmosphere: 00 is the path from the sun to the Earth and 0 is the path from the Earth to the satellite. The path lengths for each path are computed for water vapor and oxygen by (Kasten 1966): 1 M = (23) cost/ + 0.15(93.885-r/) -1253 where 7/is 00 or 0. The path length for ozone transmittance is slightly different due to the high altitude of ozone concentrations (Paltridge and Platt 1976): 1.0035 Mo_ = (24) cos2r/ + 0.007 .0.5. These formulations avoid the assumption of a fiat Earth. z : I(A)A 4. (25) Now, for A = 412 and 555 z() = za - (26) from which the Rayleigh scattering is derived: L_(A) = I(A)F_(A) (27) where I(A) is the Rayleigh intensity taken from the CZCS multiple scattering tables (Gordon et al. 1988b). The procedure is similar for bands 6-8 (670-865 nm), except 670 is substituted for A in (25), and 670 and 865 replace 412 and 555 in (26). When this method is applied to the CZCS bands at 520 and 550nm, a maximum error of 3.2% is obtained for 00 = 60 and 0 = 50. The error is less at all smaller angles tested. Next, the aerosol radiances, La, must be determined. According to Gordon and Castafio (1989) the aerosol radiance is determined by La - w_-u_- (28) 4_ where o is the single scattering albedo of the aerosol, - is the optical thickness, and Pa is a factor to account for the probability of scattering to the spacecraft for three different paths from the sun (Gordon and Castafio 1989). In (28) M has been substituted for Gordon and Castafio's (1989) 1/cos0 to allow calculations on a curved Earth. The assumption is made that the aerosol obeys the Angstr6m formulation -,-a= ),-' (29) and the Gregg and Carder (1990) model may be used to obtain _ and a, as a function of wind speed. Wind speeds were obtained from six years of data from the Fleet Numerical Oceanography Center (FNOC), made available from the NASA Climate Data System (NCDS). The mean wind speeds over these six years at 2.5 x 2.5 ° spatial resolution were used. 5

Original page 8

TheSimulatedSeaWiFSDataSet,Version1 Theaerosolscatteringalbedo,we, was also computed Some of the scattered Rayleigh and aerosol components using Gregg and Carder (1990). The factor Pa was com- continues forward and contributes to the total irradiance puted from Gordon and Castafio (1989) 1 P" = 4--, [P(0-) + [p(00)+p(O)]P(O+)] (3o) at the surface. These are expressed separately as: ts = toz tw,, to× ta [ 1- 2t°9s-r + tr1.5Ya(1 - t_,) ] (36) where p(Oo) and p(O) are the Fresnel reflectances for solar where taa is the aerosol transmittance after absorption and viewing geometries, respectively, and P is the aerosol scattering phase function for the forward scattering angles taa = e -(1-")'°'v (37) P(O+), i.e., scattering toward the surface and reflected off the ocean or reflected first and then scattered toward the and tas is analogously the transmittance after scattering spacecraft, and backward angles P(0-), i.e., backscattered from the atmosphere. (Justus and Paris 1985) P(O ±) is determined using a Henyey-Greenstein func- t = e(..... M). (38) tion for marine aerosols a(1 - 921) P(O±) = (1 + gl2 - 2gl cos0±) 15 (31) optical thickness. (1 - a)(a - o22) + (1 + g - 292 cosO±) 1.s where a = 0.983, gl = 0.82, and g2 = -0.55 Gordon and Castafio 1989). As defined in Gordon and Castafio (1989): cosO± =±cosOo cosO- sinOo sinO cos(C-(I)0). (32) The next contribution to be accounted for is that of sun glint radiance diffusely transmitted to the spacecraft (TLg). Sun glint at the surface for a given viewing and In (36)-(38), Fa is the forward scattering probability of the aerosol, taken from Bird and Riordan (1986), w_ is the single scattering albedo of the aerosol, and 'a is the aerosol If L 9 (A) is the sun glint radiance at the surface, TLg (£) is that received by the sensor, where T is the total transmittance from the Earth to the satellite. It is defined in (34)-(38) with 0 substituted for 00. Finally, the normalized water-leaving radiance LWN must be converted into the water-leaving radiance measured at the satellite TLw. This is expressed by Gordon (1990) as TLw = LWN(1 - p) To cosOo. (39) solar geometry may be expressed by (Cox and Munk 1954): The weighted direct plus diffuse reflectance, p, is computed as: L9 = p,To(A, Oo)Fo(A)p,(O,¢ Oo,¢o,W) (33) P = pnWd + pNWs (40) 4 cos 0 cos 4 ON glit- irradiance at the surface, and W is the diffuse irradiance where pw(O, (I), 00, (I)0) is the probability of seeing sun ter in the direction 0, (I) given the sun in position 00,(I)0 re- function of the atmospheric constituents and solar zenith a function of wind speed (W), and ON is the angle with a reflec- angle using the model of Gregg and Carder (1990), as are spect to nadir of the sea surface slopes to produce term the direct and diffuse reflectances, which are functions of tion angle to the spacecraft (Viollier et al. 1980). The T0(A, 00) represents the total downward transmittance irradiance as a function of 00 and the absorbing gases the in (1) is calculated by summing (27), (28), (33), and (39). the atmosphere. It may be expressed as the sum of direct and diffuse transmittances (again dropping wavelength dependence) To(Oo) = te(Oo) + ts(Oo) (34) where Wd equals the direct irradiance divided by the total as divided by the total. These components are computed as a of the wind speed. in The simulated total radiance at the spacecraft, given 3.6 Atmospheric Conditions There are five attenuating sources in the atmosphere that lie within SeaWiFS bands: Rayleigh scattering, ozone absorption, water vapor absorption, oxygen absorption, where the subscripts d and s represent direct and diffuse and aerosol scattering and absorption. Each was taken components, respectively. The transmittance td is simply for the simulated data set at mean values. Rayle_gh scatthe transmittance after absorption by the gaseous comtering and oxygen absorption are functions of atmospheric ponents of the atmosphere, scattering and absorption by pressure for which a standard value of 1,013.25mb was aerosols, and scattering by Rayleigh (dropping the angular dependence): (35) be 1.5 cm. tu = tr taa toz tw_ to×. used. Ozone absorption was taken to be 340 Dobson units (DU) and absorption due to water vapor was assumed to

Original page 9

W.W.Gregg,F.C.Chen,A.L.Mezacche,J.D.Chen,andJ.A.Whiting Aerosolsmaybecalculatedasa functionof visibility. a commonlyreportedmeteorologicalparameter.Visibility Table 5. LNER and saturation radiances (Lsat) for ;eaWiFS (units are mW cm -2 #m -1 sr-1). wassetto 15kmfor the simulation,whichcorresponds Band A [nm] LNER L_at to a mildlyturbidatmosphere.Relativehumidity(which 412 0.0182 13.63 of the aerosols) 443 0.0125 13.25 determines,in part,thesizedistribution wassetat 80C7c.The24-hourmeanwindspeedrequiredfor the aerosolmodelusedby GreggandCarder (1990), was estimated at 4.75m s -1, the global mean from 6 years of FNOC data, and the current wind speed, as noted before, was determined as the mean over 2.5 ° x 2.5 ° grid cells from the 6-year FNOC data set. 3.7 Ten-bit Words and Data Structures The final step in developing a realistic SeaWiFS simulated data set is to convert the science data into 10-bit words and to structure the data fields according to the expected format for the downlink data. Not all of the telemetry fields have been decided upon, but the purpose here is to create as reasonable a data structure as possible to allow data system testing. Further developments can be incorporated as they are made. Fig. 1 shows the preliminary stored data structures for 490 0.0098 10.50 510 0.0088 9.08 555 0.0077 7.44 670 0.0056 4.20 765 0.0035 3.00 865 0.0023 2.13 4. APPLICATION OF CZCS DATA The SeaWiFS GAC simulated data set was derived from the global CZCS data set (Feldman et al. 1989). This level-3 (remapped and gridded) data set was in dimensions of 2,048 × 1,024 (0.176 ° longitude by 0.176 ° latitude), and was cloud- and land-masked. The simulated SeaWiFS orbits passed through this CZCS pigment field in descending node (from north to south) for approximately 6.8 orbits to simulate a single data recorder dump of data. Start time for the simulation was the first contact at Wallops Flight LAC and GAC data. Depicted are the structures for a Facility (WFF), the downlink station for SeaWiFS GAC major frame of GAC and a major frame of LAC, each consisting of 3 minor frames. A major frame of GAC thus consists of 15 GAC scan lines, while a major frame of LAC consists of 3 LAC scan lines. The number of bytes in each minor frame is also denoted. A single LAC scan line of SeaWiFS data contains 1,288 pixels, comprising 1,285 science pixels, a start synchronizaa minimum of six years of International Satellite Cloud Clition pixel (also a 10-bit word) preceding the first pixel, stop synchronization pixel at the end of the scan line, followed by a time delay" integration (TDI) pixel. A GAC scan likewise contains 251 pixels, with 248 actual data pixels and three others in the same format as the LAC data. This representation is included in the simulated data set. SeaWiFS simulated total radiance data were packed into 10-bit words according to DC10(A) = Lt(A) - LNER(A) (41) s() where DClo is the digital counts at 10-bit digitization, LNER(A) is the noise equivalent radiance (NER), or the minimum detectable radiance, and s is the slope for the range 0-1,023, which is applicable for 10-bit digitization, given by Lsat(A) - LNER(A) (42) s(a) = 1,023 as a the correct amount of time left on the recorder after GAC where L_at(A) is the saturation radiance for the sensor function of wavelength. LNER(A) and Lsat() are shown in Table 5. This provides a realistic data set corresponding to the format stored on board the data recorder, and thus may be considered a level-0 data set. data. Succeeding start times, as well as all stop times, were derived in previous simulations to maximize GAC coverage. The ultimate limit of the sensor was assumed to be 2 NER values, and the recorder was turned on whenever this limit was reached, subject to land features and ice cover. An 80% ice concentration was assumed as the limit for remote sensing purposes. Ice concentrations used were the matology Project (ISCCP) data, made available by NCDS. The start and stop times, in minutes travelled along the orbital track from equator crossing at ascending node, are given in Table 6. The tilt was left at nadir pointing (0 °) for this simulation to test the response of the data system to excessive sun glint. Normally, the sensor will be tilted aft when approaching the solar declination latitude from the north, and turned forward for the remainder of the descending node track; however, this strategy has yet to be finalized. For the GAC simulation, no attempt was made to inelude clouds, and land features were masked (set to zero) as they occurred in the CZCS global data set. The LAC simulation was initiated with the same global CZCS pigment data, using the same orbit trajectories, as will be the case for the real spacecraft. The sensor LAC recorder was turned on for three separate occasions, comprising one 3-minute scene and two 2-minute scenes, for a total of 7 minutes of LAC record time. This is about data recording and some calibration activities. The start and stop times for the LAC recorder, and the associated simulated data set, are provided in Table 7. The difference between LAC and GAC data is the amount of data in a

Original page 10

The Simulated SeaWiFS Data Set, Version 1 1 7 1 < < -i S ]. q: r C . = e.. T _ _ + 1 T v v _ ¢9 _ 52 1 ¢; T _= .= _= + T = = _=÷ .",1 1 "q O T l 9 1 T . • .£ _. _ - T T © D "4 I

Original page 11

W.W. Gregg, F.C. Chen, A.L. Mezacche, J.D. Chen. and J.A. Whiting Table 6. Start and stop times (measured in minutes travelled along the orbit track from the equator crossing on ascending node), total time, start and stop latitudes, and solar zenith angles for the orbital simulations for the SeaWiFS simulated data set. The first contact with Vv'FF is made on orbit 4 at 33.8743 min. and the second contact is on orbit 11 at night. Thus, the simulated data begins at 33.8743 min. and ends at the stop time of orbit number 10. Day Orbit Times Total Start Stop Start Stop Number Start Stop Latitude Solar Zenith 80 1 27.0571 71.3008 44.243 78.285 -76.530 81.396 79.507 80 2 26.9876 69.2139 42.226 78.460 -69.996 81.648 71.923 80 3 26.9043 69.9826 43.078 78.670 -72.499 81.954 74.716 80 4 26.9154 70.7818 43.866 78.642 -74.995 81.910 77.621 80 5 27.7986 70.5321 42.733 76.235 -74.230 78.701 76.713 80 6 26.8932 71.6449 44.751 78.698 -77.496 81.991 80.757 8O 7 26.9099 71.6449 44.735 78.656 -77.496 81.931 80.757 80 8 26.8932 69.2139 42.320 78.698 -69.996 81.991 71.923 80 9 30.3202 68.4647 38.144 68.207 -67.492 69.537 69.200 80 10 28.7511 68.4647 39.713 73.336 -67.492 75.239 69.200 80 11 27.8374 68.4647 40.627 76.121 -67.492 78.560 69.200 80 12 26.8932 68.4647 41.571 78.698 -67.492 81.991 69.200 80 13 26.9876 69.2139 42.226 78.460 -69.996 81.648 71.923 80 14 27.0571 69.2139 42.156 78.282 -69.996 81.396 71.923 LAC scan (Table 1), which is 1,285 pixels, with no gaps along-scan. Table 7. Start and stop times (in minutes travelled along the orbit track from the equator crossing on ascending node) orbit numbers, and corresponding latitudes for the LAC simulation. Orbit Times Latitude Number Start Stop Total Start Stop 4 54.6 57.6 3.0 -18.35 -29.13 6 64.3 66.3 2.0 -53.03 -60.06 6 68.6 70.6 2.0 -67.95 -74.44 5. RESULTS AND DISCUSSION sorption (see Table 4). This absorption reduces the amount of saturation in the sun glint regions. Little evidence of water-leaving radiances is evident in most of the images. This is because the atmosphere dominates the total radiance over the oceans. Some evidence may be seen in the image of band 2 (443 nm), where the contribution of water-leaving radiance to the total is greatest under low chlorophyll concentrations. The LAC recording areas are also depicted in Figs. 2-4. Note that they correspond with portions of the GAC orbits, but have a wider swath. Note also that the extra data across-track have very large Lt(A) values, due to increased atmospheric contribution as a function of the greater LAC spacecraft zenith angle at the edges. This first version of the SeaWiFS simulated data set Imagery of the simulated SeaWiFS total radiances Lt (A) contains realistic total radiances, in a representative data for all 8 bands is shown in Figure 2 for GAC data. The start of the simulation was at the first visibility at WFF, which can be noted by the truncated orbit at the eastern edge of the image. Recall that for this simulation the sensor remained untilted, resulting in substantial sun glint influence near the equator (which is the solar declination for the vernal equinox, the time of the simulation). Tilting the sensor aft of the velocity vector when approaching the solar declination, and then forward when leaving would decrease the sun glint contribution to the total radiance substantially. Note the decreases in Lt(A) toward the poles, due to a reduction in the flux of solar irradiance into the oceans, and the increases in L_(A) at the scan edges, due to increased atmospheric contribution. The bands show progressively less total radiance from 412-865 nm. However, band 7 (765 nm) is the lowest because of large oxygen abrecord scenario: approximately 6.8 orbits of GAC data beginning at the first visibility of the downlink station at WFF, and 8 minutes of LAC record time, interspersed in the SeaWiFS orbits. Future improvements should incorporate representative values in the telemetry fields, accounting for clouds and land radiances. Also, a tilt strate_" to minimize sun glint should be included. However, this version facilitates early prototyping activities by the Sea- WiFS data system, providing a realistic data set to assess performance. 6. DATA AVAILABILITY Version 1 of the simulated SeaWiFS data set is available from the authors on 9-track tape at 6,250 bits per inch (bpi). The tape is written using the VAX/VMS BACKUP command, on the assumption that most users have access

Original page 12

The Simulated Sea\ViFS Data Set, Version 1 Band 1 Band 2 Band 3 Fig. 2. Simulated total radiances for GAC data for SeaWiFS bands 1-3. 10

Original page 13

W.W.Gregg,F.C.Chen,A.L.Mezacche,J.D.Chen.andJ.A. Whiting Band4 Band5 Band6 Fig. 3. Simulatedtotal radiancesfor GACdataforSeaWiFSbands4-6. 11

Original page 14

TheSimulatedSeaWiFSDataSet,Version1 Band 7 Band 8 Band 2 LAC Fig. 4. Simulated total radiances for GAC data for SeaWiFS bands 7-8, and LAC data for band 2. 12

Original page 15

W.W. Gregg, F.C. Chen, A.L. Mezacche J.D. Chen, and J.A. Whiting to Digital Equipment Corporation (DEC) hardware. FOR- TRAN programs to read the data and strip out telemetry fields and science data are provided in the Appendices. APPENDICES A. Description B. Program RGAC.FOR C. Program RLAC.FOR D. Program EXPAND.FOR E. Program B2W.FOR Appendix A Description The SeaWiFS simulated data sets and programs to read them are provided on a 6,250 bpi 9-track tape. The tape was generated using the VMS (version 5.4-3) BACKUP command and the following command sequence: MOUNT/FOREIGN device BACK-t/P/LOG device: simdat ./label=simdat/block=32256 DI SMDUNT device where device isthe tape drive device name. (Note: 105 MBytes of free disk space are required to read the tape in itsentirety.) C c c integer4 MinorFrameNo,msec_of_day, spacecra/t_id integer2 day_of_year integer4 i,j,k c c Orbit number, xlons, ylats, and 8"mr time(hr) for 5 GAC c scan lines within one minor frame. These fields are c temporary. c integer4 iorb (5) real4 xlons (5) ,ylats(5) ,gmt (5) c c Map the above fields within SCSOH c equivalence (iorb(1),SCSOH(197)), (xlons(1),SCS{3H(217)) equivalence (ylats(1),SCS[}H(23Z)), (_t(1),SCSOH(257)) c c Map fields within minor frame to the three segments c equivalence (FrameSync(1),sl(1)), (SCID(1),sI(7)) equivalence (TimeTag(1),sl(9)), (SCSOH(1),s2(1)) equivalence (InstTLM(1),s3(1)), (CalTable(1),s3(45)) equivalence (Spare(1),s3(89)), (GAC(I,I,I),s3(277)) equivalence (PADI(1),s3(I0357)), (AuxSync(1),s3(10359)) equivalence (PAD2(1) ,s4(1)) c I00 format (i5,3f I0.4) 200 format (8i5) nscan = 0 %Vhen the tape is loaded onto disk, 6 filesare put in the open (4, file = 'gac. dat ' ,status= 'old ' ,form= 'unf ormatt ed' ) current directory.GAC. DAT and LAC. DAT are the simulated data 20 continue sets for GAC and LAC data, respectively. Listings of the c source codes to read these data sets are provided in the follow- c Read in one minor frame P_LAC.FOR c ing. This includes the main programs RGAC.FOR and and the two supporting subroutines, EXPAND. FOI_and B2W. FOR. EXPAND.FOR isused to extract the minor frame data segments. B2W.FOR isused to unpack a 5-byte fieldinto four 2-byte integers. Appendix B Program RGAC.FOR progr_r, rgac C c Program to read and sort simulated SeaWiFS GAC data. c by: Frank Chen, GSC. c implicit none character c(13864) integer,2 si(12) character s2(775) integer.2 s3(10458) character s4(2) c c Minor frame data structures c integer2 FrameSync(6),SCID(2),TimeTag(4) character SCSOH(775) integer2 InstTLM(44),CalTable(44),Spare(188) integer2 GAC(8,252,5) integer2 PADl(2),AuxSync(lO0) character PAD2(2) c c Science data array c integer2 ilt(8,248) c c Total number of scan lines c integer*4 nsca_ c c Minor frame number(l,2,3) read(4, end=999) c c c Convert all lO-bit-word to 16-bit-word c call expand(c, el, s2, s3, s4) c c Calculate s/c id, minor fram number, day of year, c and msec of day c spacecraft_id = MOD(SCID(1)/8,16) MinorFrameNo = MOD(SCID(1)/128,4) day_of_year = TimeTag(1)/2 msec_of_day = TimeTag(2)1024.01024.0 + * TimeTag(3)1024.0 + TimeTag(4) do i=I,5 c c GAC(k, 1,i),k=1,8 80 bits GAC GainTDI c GAC(k, 2,i),k=1,8 80 bits GAC StartSync c GAC(k, 3,i),k=1,8 80 bits GAC DarkRestore c GAC(k,252,i),k=l,8 80 bits GAC StopSync c c Check if the GAC data are valid c if ((GAC(l,3,i).eq.O) .and. (GAC(2,3,i).eq.O) .and. * (GAC(3,3,i).eq.O) .and. (GAC(4,3,i).eq.O) .and. * (GAC(5,3,i).eq.O) .and. (GAC(6,3,i).eq.O) .and. * (GAC(7,3,i).eq.O) .and. (GAC(8,3,i).eq.O)) then write(6,)' no more valid data' goto 999 end if nSCan = nscan + 1 do j=4,251 do k=1,8 ilt (k, j -3) =GAC (k, j, i) end do end do end do goto 20 999 continue 13

Original page 16

The Simulated SeaWiFS Data Set, Version 1 close(4) write(6,)'Total no. of scan lines =',nscan c end Appendix C Program RLAC.FOR program rlac c c Program to read and sort SeaWiFS simulated LAC data. c by: Frank Chert, GSC. c implicit none character c(13864) integer2 sl(12) character s2(775) integer2 s3(I0458) character s4(2) c c Minor frame data structures c integer2 FrameSync(6),SCID(2),TimeTag(4) character SCSOH(775) integer2 InstTLM(44),LAC(8,1289) integer2 PADl(2),AuxSync(lO0) character PAD2(2) c c History SCSOH and 0X25 in minorframe 2,3 c character HistSOH(775),OX25(TZ5) c c Ancilliary TLM and Calibration Table in minor c frame 2,3 c integer2 AncTLM(44),CalTable(44) c c Science data array c integer2 ilt(8,1285) c c Total number of scan lines c integer4 nscan c c Minor frame number(l,2,3) c integer-4 MinorFrameNo,msec_ofday.spacecraft_id integer2 day_of_year integer.4 i,j.k c c Orbit number.xlons,ylats, and gmt time(hr) for LAC c scan line within one minor frame. These fields c are temporary. c integer4 iorb real4 xlons,ylats,gmt c c Map the above fields within SCSOH c equivalence (iorb,SCSOH(197)), (xlons,SCSOH(217)) equivalence (ylats, SCSDH(237)), (gmt,SCSOH(257)) c c Map fields within minor frame to the three segments c equivalence (FrameSync(1),sl(1)), (SCID(1),sl(7)) equivalence (TimeTag(1),sl(9)), (SCSOH(1),s2(1)) equivalence (InstTLM(1),s3(1)), (LAC(I,1),s3(45)) equivalence (PADI(1),s3(10357)), (AuxSync(1),s3(10359)) equivalence (PAD2(1),s4(1)), (HistSOH(1),SCSOH(1)) equivalence (OX25(1),SCSOH(1)), (AncTLM(1),InstTLM(1)) equivalence (CalTable(1),InstTLM(1)) c 14 100 format (i5,3f 10.4) 200 format (8i5) nscan = 0 c open (4, f i le = 'lac. dat , status= ' old ' ,form= 'unf ormatt ed ' ) 20 continue c c Read in one minor frame c read (4, end=999) c c c Convert all lO-bit-word to 16-bit-word c c all expand (c, s 1, s2, s3, s4 ) c c Calculate s/c id, minor fram number, day of year, c and msec of day c spacecraft_id = MOD (SCID (I)/8,16) MinorFrameNo = MOD(SCID(1)/128,4) day_of_year = TimeTag (i)/2 msec_of_day = TimeTag(2)1024.01024.0 + * TimeTag(3)1024.0 + TimeTag(4) c c write(6.)'day:'.day_of_year.' scid:', c * spacecraft_id, c * ' frame: ' ,MinorFrameNo c c LAC(k, 1) ,k=l,8 80 bits LAC GainTDI c LAC(k, 2) ,k=1.8 80 bits LAC StartSync c LAC(k, 3),k=I,8 80 bits LAC DarkRestore c LAC(k,1289),k=l,8 <80 bits LAC StopSync c nscan = nscan + 1 do j=4,1288 do k=1,8 ilt (k, j-3) =LAC (k, j ) end do end do goto 20 999 continue close(4) write(6,*)'Total no. of scan lines =' ,nscan end Append D ProKramEXPAND.FOR c This subroutine converts the input buffer (c) into c four minor frame segments: sl, s2, s3, and s4. c Sl consists of 12 lO-bit words containing the frame c sync, the spacecraft id, and the time tag. $2 c is a the S/C SOH array. $3 is the remainder 10458 c lO-bit words without the last padding (1.5 bytes). c $4 is the last padding. c subroutine expand(c,sl,s2,s3,s4) implicit none character c(13864) integer2 si(12) character s2(775) integer2 s3(I0458) character s4(2) character b(5) integer2 w(4) integer4 i,J,k c c Process first 1.2 segment c do i=0,11,4 k = i.5/4 do j=1,5

Original page 17

W.W. Gregg, F.C. Chen, A.L. Mezacche, J.D. Chen, and J.A. Whiting b(j) : c(k+j) end do call b2w(b,w) do j=l,4 sl(i+j) : w(j) end do end do c c Process second CI segment c GLOSSARY bpi bits per inch CZCS Coastal Zone Color Scanner DEC Digital Equipment Corporation DU Dobson Units FNOC Fleet Numerical Oceanography Center k : 12"5/4 FORTRAN Formula Translation (computer language) do i=i,775 s2(i) = c(k+i) end do c c Process third 1.2 segment c do i=0,10455,4 k = (i+12).5/4 + 775 do 3=1,5 b(3) : c(k+j) end do call b2w(b,w) do j=l,4 s3(i+j) = w(j) end do end do k = 13860 do j:l,3 b(3) = c(k+j) end do b(4) = char(ichar(c(k+4)) .and. 240) b(5) = char(O) call b2w(b,w) do 3=1,3 s3(lO456+j) = w(j) end do s4(1) = char(MOD(ichar(c(13863)),16)16 + * ichar(c(13864))/16) s4(2) = char(MOD(ichar(c(13864)),16)16) return end Append_ E Program B2W.FOR c Convert packed 5 bytes character(b) into four c two-byte-integer(w) each contains 10-bit information. c MOD(n,2.m): modulus 'n' by '2**m'. extract lowest c 'm' bits from the number 'n'. c n * (2-m): shift left the number 'n' by 'm r bits. c n / (2**m): shift right the number 'n' by 'm r bits. c subroutine b2w(b,w) implicit none character b(5) integer2 w(4) integer4 c(5) integer4 i integer4 p0 /I/, p2 /4/, p4 /16/, p6 /64/, integer4 p8 /256/ c do i=I,5 c(i) = ICHAR(b(i)) end do w(1) = MOD(c(1),p8)*p2 + c(2)/p6 w(2) = MOD(c(2),p6)*p4 + c(3)/p4 w(3) = MOD(c(3),p4)*p6 + c(4)/p2 w(4) = MOD(c(4),p2)*p8 + c(5)/pO return end FWHM Full-Width Half-Maximum GAC Global Area Coverage GMT Greenwich Mean Time IFOV Instantaneous Field-of-View ISCCP International Satellite Cloud Climatolog3, Project LAC Local Area Coverage Level-0 Raw data. Level-3 Gridded and averaged derived products. MB Megabytes MF Major Frame mF Minor Frame NCDS NASA Climate Data System NER Noise Equivalent Radiance OSC Orbital Sciences Corporation SeaWiFS Sea-viewing Wide Field-of-view Sensor s/c Spacecraft SOH State of Health TDI Time Delay Integration TLM Telemetry VAX Virtual Address Extension VMS Virtual Memory System WFF Wallops Flight Facility SYMBOLS a A constant equal to 0.983. aox Coefficient for oxygen absorption. aoz Coefficient for ozone absorption. awv Coefficient for water vapor absorption. A(A) Coefficient for calculating bb(A). bb() Spectral backscattering coefficient. bbo() Spectral backscattering coefficient for phytoplankton. b_(A) Total scattering coefficient for pure seawater. B() Coefficient for calculating bb(A). Crcf Reference chlorophyll value (0.5). [chl. a] Chlorophyll concentration. D Sequential day of the year. DCIo Digital counts at 10-bit digitization. Y0() Mean extraterrestrial irradiance. F0 Extraterrestrial irradiance corrected for Earthsun distance. E_ Extraterrestrial irradiance corrected for the atmosphere. Fo Forward scattering probability of the aerosol. 15

Original page 18

The Simulated SeaWiFS Data Set, Version 1 91 A constant equal to 0.82. 92 A constant equal to -0.55. HGMT GMT in hours. H_ Altitude of the spacecraft (for SeaStar 705 km). i Inclination angle. i t Inclination angle minus 90 ° . I Rayleigh intensity. Spectral fit coefficient weighted over the Sea- WiFS bands. k', (A) Spectral fit coefficient weighted over the Sea- WiFS bands. K() Spectral attenuation coefficient. Kc(A) Attenuation coefficients for phytoplankton. KE() Attenuation coefficient downwelled irradiance. K_(Z) Attenuation coefficients for Gelbstoff. KL(z, ,) Attenuation coefficient upwelled radiance. K() Attenuation coefficients for pure seawater. Lo() Aerosol radiance. Lg(A) Sun glint radiance. LNER,X) Noise equivalent radiance. L_() Rayleigh radiance. Saturation radiance for the sensor. L,(:) Total radiance at the sensor. L_(z, A) Upwelled spectral radiance. Lw()) Water-leaving radiance. LwN(,) Normalized water-leaving radiances leaving an ocean of given optical properties assuming no atmosphere and the sun directly overhead. M Path length through the atmosphere. Mo_ Path length for ozone transmittance. n Index of refraction. Q Irradiance-to-radiance ratio (equals rr for totally diffuse radiance). Pa A factor to account for the probability of scattering to the spacecraft for three different paths from the sun. p.w The probability of seeing sun glitter in the direction 0, q_ given the sun in position 80, Co as a function of wind speed (W). P Nodal period. P_ Probability of scattering to the spacecraft. P(O +) Phase function for forward scattering. P(O-) Phase function for backward scattering. Water-air reflectance for totally diffuse irradiance. R(O-, ) Irradiance reflectance just below the sea surface. Re Mean Earth radius (6371.2 km). Slope for the range 0-1,023. t Time variable. to Initial time. taa Aerosol transmittance after absorption. t Aerosol transmittance after scattering. td Direct component of transmittance after absorption by the gaseous components of the atmosphere, scattering and absorption by aerosols, and scattering by Rayleigh. 16 t_ Time difference in hours between present position and most recent equator crossing. tEC Equator crossing time. toz Transmittance after absorption by ozone. t_ Transmittance after Rayleigh scattering. t, Diffuse component of transmittance after absorption by the gaseous components of the atmosphere, scattering and absorption by aerosols, and scattering by Rayleigh. twv Transmittance after absorption by water vapor. T(A,0) Total transmittance (direct plus diffuse) from the ocean through the atmosphere to the spacecraft along the path determined by the spacecraft zenith angle 0. T(A,O)Lw(A) The water-leaving radiance transmitted to the spacecraft. To(,Oo) Total downward transmittance of irradiance. T, Equation of time. Tox Transmittance of oxygen (O2). To, Transmittance of ozone (Oa). Ts() Transmittance through the surface. Twv Transmittance of water vapor (H20). W Wind speed. Wd Direct irradiance divided by the total irradiance at the surface. Ws Diffuse irradiance divided by the total irradiance. c_ The power constant in the _ngstrSm formulation. 13 A constant in the/_ngstr6m formulation. 6 Great circle distance from _(t0) to _(t - to). At Time difference. Aw The longitude difference from the sub-satellite point to the pixel. AtJ8 Longitude difference. r/ Bearing from the sub-satellite point to the pixel along the direction of motion of the satellite. A Wavelength of light. Pixel latitude. O2a Solar declination latitude. 's(t) Sub-satellite latitude as a function of time. Spacecraft azimuth angle. o Solar azimuth angle. P Weighted direct plus diffuse reflectance. p(o) Fresnel reflectance for viewing geometry. p(Oo) Fresnel reflectance for solar geometry. pn Sea surface reflectance for direct irradiance at normal incidence for a flat sea. pN Reflectance for diffuse irradiance. 0 Spacecraft zenith angle. 00 Solar zenith angle. ON The angle with respect to nadir that the sea surface slopes to produce a reflection angle to the spacecraft. 0 Scan angle of sensor. 0 Scan angle of sensor adjusted for tilt.

Original page 19

W.W.Gregg,F.C.Chen,A.L.Mezacche,J.D.Chen,andJ.A. Whiting 7a Aerosol optical thickness. rr Rayleigh optical thickness. rs(a) Spectral solar atmospheric transmission. Longitude variable. cJ0 Old longitude value. 'a Single scattering albedo of the aerosol. Me Equator crossing longitude. cds Longitude variable. fl Solar hour angle. REFERENCES Austin, R.W., 1974: The remote sensing of spectral radiance from below the ocean surface, In: Optical Aspects of Oceanography, N.G. Jerlov and E. Steemann-Nielsen, Eds., Academic Press, 317-344. Baker, K.S., and R.C. Smith, 1982: Bio-optical classification and model of natural waters, 2. Limnol. and Oceanography, 27, 500-509. Bird, R.E., and C. Riordan, 1986: Simple solar spectral model for direct and diffuse irradiance on horizontal and tilted planes at the Earth's surface for cloudless atmospheres, J. of Climate and Applied Meteorology, 25, 87-97. Cox, C., and W. Munk, 1954: Measurement of the roughness of the sea surface from photographs of the sun's glitter, J. Mar. Res., 44, 838-850. Feldman, G.C., N. Kuring, C. Ng, W. Esaias, C.R. McClain, J. Elrod, N. Maynard, D. Endres, R. Evans, J. Brown, S. W'alsh, M. Carte, G. Podesta, 1989. Ocean Color: Availability of the global data set, EOS, Transactions of the American Geophysical Union, 70, 634-635, 640-641. Gordon, H.R., 1990. Radiometric considerations for ocean color remote sensors, Applied Optics, 29, 3,228-3,236. , D.K. Clark, J.W. Brown, O.B. Brown, R.H. Evans, and W.W. Broenkow, 1983: Phytoplankton pigment concentrations in the Middle Atlantic Bight: Comparison of ship determinations and CZCS estimates, Applied Optics, 22, 20-36. --, O.B. Brown, R.H. Evans, J.W. Brown, R.C. Smith, K.S. Baker, and D.K. Clark, 1988a. A semianaly_ic radiance model of ocean color, J. o] Geophys. Res., 93, 10.909- 10,924. --, J.W. Brown, and R.H. Evans, 1988b: Exact Rayleigh scattering calculations for use with the Nimbus-7 Coastal Zone Color Scanner, Applied Optics, 27, 862-871. --, and D.J. Castafio, 1989: Aerosol analysis with Coastal Zone Color Scanner: A simple method for including multiple scattering effects, Applied Optics, 28, 1,320-1,326. Gregg, W.W., and K.L. Carder, 1990: A simple spectral solar irradiance model for cloudless maritime atmospheres, Limnol. and Oceanography, 35, 1,657-1,675. Iqbal, M., 1983: An Introduction to Solar Radiation. Academic Press, 390 pp. Jerlov, N.G., 1976: Marine optics, Elsevier Scientific Publishing Co., New York, 231 pp. Justus, C.G., and M.V. Paris, 1985: A model for solar spectral irradiance and radiance at the bottom and top of a cloudless atmosphere, J. of Climate and Applied Meteorology, 24, 193-205. Kasten, F., 1966: A new table and approximate formula for relative optical air mass, Geophys. Biokimatol., B14, 206- 223. Neckel, H., and D. Labs, 1984: The solar radiation between 3300 and 12500/. Solar Physics, 90, 205-258. Paltridge, G.W., and C.M.R. Platt, 1976: Radiative processes in meteorolo-v and climatology, Developments in Atmospheric Science, Vol. 5. Elsevier Scientific Publishing Co., New York, 318 pp. Viollier, M., D. Tanre, and P.Y. Deschamps, 1980: An algorithm for remote sensing of water color from space,Bound.- Layer Meteorology, 18, 247-267. Williams, S.P., E.F. Szajna, and W.A. Hovis, 1985: Nimbus 7 Coastal Zone Color Scanner (CZCS) Level 1 data product users' guide, NASA Technical Memorandum 86203, 49 pp. 17

Original page 20

REPORT DOCU MENTATION Form Approved PAG E OMBNo.o;o4-o188 PUbtK: reporting burden for this collection of information is estirnaled to average 1 hour per response, including the hme for rev=ewing instructions, searching existing data sources, gathering and maintaimng the data needed, and completing and reviewing the oolteclion of information. inlormation, including suggestions for reducing this burden, to Washington Headquarters 1204. Adin_l[on. VA 22202-4302. and to the Oflee of Management and 8ud_let. Paperwork 1. AGENCY USE ONLY (Leave blank) 2. REPORT DATE May 1993 4. TITLE AND SUBTITLE SeaWiFS Technical Report Series Volume 9-The Simulated SeaWiFS Data Set, Version 6. AUTHOR(S) Watson W. Gregg, Frank C. Chen, Ahmed L. Mezaache, and Jeffrey A. Whiting Send comments regarding this burden estimate or any other aspect of this cellection of Services, Directorate lor Information Operations and Reports, 1215 Jefferson Davis Highway, Suite Reduction Proiect (0704-0188). Weshin_lton. DC 20503. 3. REPORT TYPE AND DATES COVERED Technical Memorandum S. FUNDING NUMBERS 1 970.2 Judy D. Chen, Editors: Stanford B. Hooker, Elaine R. Firestone, and A. W. Indest ADDRESS(ES) 8o PERFORMING ORGANIZATION 7. PERFORMING ORGANIZATION NAME(S) AND Laboratory for Hydrospheric Processes Goddard Space Flight Center Greenbelt, Maryland 20771 ADDRESS(ES) 10. SPONSORING/MONITORING 9. SPONSORING/MONITORING AGENCY NAME(S) AND National Aeronautics and Space Administration Washington, D.C. 205460001 11. SUPPLEMENTARY NOTES REPORT NUMBER 93B00086 AGENCY REPORT NUMBER TM-104566, Vol. 9 Frank C. Chen, Ahmed L. Mezaache, Judy D. Chen, Jeffrey A. Whiting, Elaine R. Firestone, and A. W. Indest: General Sciences Corporation, Laurel, Maryland. 12a. DISTRIBUTION/AVAILABILITY STATEMENT Unclassified-Unlined Subject Category(_.) _£tL-L _ ICI ti 12b. DISTRIBUTION CODE Report is available from the National Technical Information Service, U.S. Dept. of Commerce, 5285 Port Royal Road, Springfield, VA 22151; (703) 557--4650. 13. ABSTRACT(Ma_mum2OOw_ds) Data system development activities for the Sea-viewing the scheduled 1994 launch. To assist in these activities, Wide Field-of-view Sensor (SeaWiFS) must begin well before it is essential to develop a simulated SeaWiFS data set as soon as possible. Realism is of paramount importance in this data set, including SeaWiFS spectral bands, orbital and scanning characteristics, and known data structures. Development of the simulated data set can assist in identification of problem areas that can be addressed and solved before the actual data are received. This paper describes the creation of the first version of the simulated SeaWi.FS data set. The data set includes the spectral band, orbital, and scanning characteristics of the SeaWiFS sensor and SeaStar spacecraft. The information is output in the data structure as it is stored onboard. Thus, it is a level-0 data set which can be taken from start to finish through a prototype data system. The data set is complete and correct at the time of printing, although the values in the telemetry fields are left blank. The structure of the telemetry fields, however, is incorporated. Also, no account for clouds has been included. However, this version facilitates early prototyping activities by the SeaWiFS data system, providing a realistic data set to assess performance. 14. SUBJECT TERMS Orbital Characteristics, 17 SeaWiFS, Oceanography, Data Set, Spectral Band, Scanning Characteristics 15. NUMBER OF PAGES 16. PRICE CODE '17. SECURITY CLASSIFICATION 18. SECURITY CLASSIFICATION 19. SECURITY CLASSIFICATION 20. LIMITATION OF ABSTRACT OF REPORT OF THIS PAGE Unclassified Unclassified NSN 7540-01-280-5500 OF ABSTRACT Unclassified Unlimited Standard Form 298 (Rev. 2-89) Prescribed by ANSI Std. 239-18,298-102