Full report

Original page 1

NASA Technical Memorandum 104566, Vol. 22 SeaWiFS Technical Report Series Stanford B. Hooker, Editor NASA Goddard Space Flight Center Greenbelt, Maryland Elaine R. Firestone, Technical Editor General Sciences Corporation Laurel, Maryland James G. Acker, Technical Editor Hughes STX Lanham, Maryland Volume 22, Prelaunch Acceptance Report for the SeaWiFS Radiometer Robert A. Barnes Man Tech, Inc. Wallops Island, Virginia William L. Barnes Wayne E. Esaias Charles R. McClain NASA Goddard Space Flight Center Greenbelt, Maryland National Aeronautics and Space Administration Goddard Space Flight Center Greenbelt, Maryland 20771 1994

Original page 2

This publication is available from the NASA Center for AeroSpace Information, 800 Elkridge Landing Road, Linthicum Heights, MD 21090-2934, (301) 621-0390.

Original page 3

R.A. Barnes, W.L. Barnes, W.E. Esaias, and C.R. McClain ABSTRACT The final acceptance, or rejection, of the Sea-viewing Wide Field-of-view Sensor (SeaWiFS) will be determined by the instrument's on-orbit operation. There is, however, an extensive set of laboratory measurements describing the operating characteristics of the radiometer. Many of the requirements in the Ocean Color Data Mission (OCDM) specifications can be checked only by laboratory measurements. Here, the calibration review panel (composed of the authors of this technical memorandum) examines the laboratory characterization and calibration of SeaWiFS in the light of the OCDM performance specifications. Overall, the performance of the SeaWiFS instrument meets or exceeds the requirements of the OCDM Contract in all but a few unimportant details. The detailed results of this examination are presented here by following the outline of the specifications, as found in the Contract. The results are presented in the form of requirement and compliance pairs. These results give conclusions on many, but not all, of the performance specifications. The acceptance by this panel of the performance of SeaWiFS must only be considered as an intermediate conclusion. The ultimate acceptance (or rejection) of the SeaWiFS data set will rely on the measurements made by the instrument on orbit. 1. INTRODUCTION to the OCDM performance specifications is the responsibility of OSC. The contract between OSC and SBRC has In addition to its role as an ocean color experiment, the incorporated the OCDM performance specifications almost Sea-viewing Wide Field-of-view Sensor (SeaWiFS) serves completely. There have been a few instances in which the as a satellite procurement experiment for the National deliverables in the subcontract between OSC and SBRC Aeronautics and Space Administration (NASA). For the have not provided the information needed by the Project to SeaWiFS Project, NASA is procuring data, not an inassure compliance with the OCDM specifications. In those strument designed by the agency. NASA has entered into instances, the informal arrangement has given a mechaa contractual agreement, the Ocean Color Data Mission nism for the Project to obtain the necessary information. (OCDM) contract, (hereinafter referred to as the Contract, The major events in the construction and testing of the unless otherwise stated) with Orbital Sciences Corpora- SeaWiFS instrument are given in Table 1. Throughout this tion (OSC) to obtain, at a fixed price, an ocean color data period, the Project has worked as an active partner with set. OSC has, in turn, entered into an agreement with the SBRC. This partnership has been of great advantage to the Hughes Santa Barbara Research Center (SBRC) for which Project. The Project representative has been given access SBRC, as a subcontractor, has built the satellite sensor re- to all of the technical information about SeaWiFS, plus quired to provide these data. In this arrangement, SBRC access to the engineers and technicians working on the inhas had the freedom to design an instrument which meets strument. Technical problems that arose during constructhe predetermined set of specifications. The design of the tion were openly discussed by SBRC and the Project. In testing procedures for the instrument has also been left to addition, test procedures were developed, in part, through SBRC. informal talks between SBRC engineers and the Project Although not written in the specifications, it is the re- representative. The representative was an active particisponsibility of the Project to understand the design, the pant in several of the tests. operation, and the calibration of the satellite sensor. The Again, this active participation has been of great ad- Project also has the responsibility of transferring this un- vantage to the Project. In this review of the SeaWiFS derstanding to the community of scientists who will use the specifications, the review panel can base its conclusions on ocean color data set. Without a specific requirement in the a set of tests and calibrations with results and procedures Contract, SBRC, OSC, and the Project set up an unofficial that the panel understands. More importantly, the acprogram of visits to the instrument builder by a Project tive participation has provided a greater understanding of representative during the construction and calibration of the operation of SeaWiFS. This includes an understand- SeaWiFS. From the outset, these visits developed into a ing of some of the characteristics that are particular to collaboration between SBRC and the Project, rather than the instrument--characteristics that the SBRC engineers strict supervision by NASA, over the instrument develop- have called its personality. These include the instrument's ment. along-track modulation transfer function (MTF), discussed This arrangement has been non-standard in another below, and its stray light characteristics, which will be disway, since the Contract remains an agreement between cussed at length in a future volume within the SeaWiFS NASA and OSC. Care has been taken to ensure no inter- Technical Report Series. Such understanding will be cruference from the Project in the contractual obligations be- cial as the Project works to interpret the data that the tween OSC and SBRC. It should be noted that adherence radiometer will transmit from orbit.

Original page 4

PrelaunchAcceptanceReportfortheSeaWiFSRadiometer [able 1. Majoreventsin theconstructionandtestingof the SeaWiFS instrument. Date Noteworthy Event 16 May 1991 Letter Contract Signed between OSC and SBRC 6 July 1991 Preliminary Design Review 16 December 1991 Critical Design Review 16 September 1992 Engineering Design Unit Completed 27 October 1992 SBRC and GSFCt Integrating Sphere Comparison 8 December 1992 First Field Test (lunar and solar measurements) 15 January 1993 Vibration and Thermal-vacuum Testing 7 March 1993 Second Field Test 27 April 1993 Initial Pre-Ship Review 27 May 1993 Stray Light Paths Review (at GSFC) 3 August 1993 Performance Specification Modification (at OSC) 26 October 1993 Vibration Testing of Modification Workmanship 1 November 1993 Third Field Test 22 November 1993 Completion of Instrument 2 December 1993 Post-Modification Pre-Ship Review Goddard Space Flight Center It is the opinion of the review group that in light of The along-scan and along-track line spread functions the instrument design and testing program, the prelaunch performance of the SeaWiFS radiometer meets, or exceeds, de- values are representative of the instrument scanning from the requirements of the Contract in all but a few small it left to right. In the same manner, the along-track values tails, which are considered to be minor. In addition, the are given in the direction of flight, i.e, for an instrument is the opinion of the review group that the testing of instrument has also been adequate to allow these conclusions. per- normalized to unity. As shown in Fig. 1, the right-most In Sections 2 through 22, individual parts of the sec- data points in both scans have been set to zero. The scans formance specifications are addressed. Some of these tions include requirements that do not involve the radiometer's performance. One such section is Section 20, which concerns satellite pointing data. These additional requirements are discussed in the compliance section for those the the along-scan and along-track measurements. These respecifications. Section 23 gives a short summary of conclusions. 2. FIELD-OF-VIEW 2.1 Requirement and lations found in the SeaWiFS Calibration and Acceptance The instantaneous field-of-view (IFOV) at nadir be Data Package (SCADP). The SCADP was generated by A. 0 ° tilt shall be between 1 and 1.21 km. Sampling shall done once per nominal (square) IFOV. 2.2 Compliance for the eight SeaWiFS bands are shown in Fig. 1. The along-scan values are given in the direction of scan, i.e., the moving left to right. For the purposes of Fig. 1, the offsets have been removed from the data, and the data have been in Fig. 1 also provide the basis for the MTF calculations and for the band-to-band registration calculations. The basic method for determining the field-of-view is to calculate the full-width at half-maximum (FWHM) for sults are presented in Table 2. In addition, the results are presented in terms of the length of an arc that subtends the angle at a distance of 705 km. Such an arc represents the width of the footprint of the SeaWiFS measurement, assuming that the instrument is 705 km above the Earth. The values calculated here are in agreement with the calcu- Holmes of SBRC in the course of constructing, calibrating, and testing the SeaWiFS instrument. Many of the conclusions made by the review panel are based on information found in the SCADP. The field-of-view of the instrument is determined from However, these calculations do not adequately reprenadir sent the two-dimensional nature of the field-of-view. The data measured by scanning a narrow slit across the pixel, both along-scan and along-track. The narrow slit is shape of a SeaWiFS footprint is not a perfect rectangle, the since the line spread functions are not perfect square waves. 0.16 mrad wide (about 0.1 of the width of a pixel), and slit is scanned in 0.1 mrad increments. The slit overfills the To better represent the SeaWiFS footprint, the along-scan pixel in the direction perpendicular to its narrow opening. The term for the results from this type of measurement a line spread function. and along-track values for each band have been combined is into two-dimensional arrays. Each element in the array is the product of the value at the ordinate (the along-scan

Original page 5

R.A.Barnes,W.L.Barnes,W.E.Esaias,andC.R.McClain 1.0 0.9 0.8 0.7 0.6 0.3 0.2 0.1 0.0 1 2 3 4 -0.1 Increment Imtml) 1.0 0.9 0.8 0.7 0.6 0.4 0.3 0.2 0.1 0.0 1 2 3 4 -0.1 Increment (mRad) -- BAND I ........ BAND 2 .... BAND 3 ....... BAND 4 ........ BAND 5 -- BAND6 ........ BAND 7 .... BAND 8 5 6 7 10 -- BAND 1 ........ BAND 2 BAND 3 ....... BAND 4 ..... BAND 5 -- BAND 6 ........ BAND 7 BAND 8 5 6 7 8 9 10 Fig. 1. The along-scan and along-track line spread functions. The top figure shows the along-scan line spread function. In this figure, the zero offsets have been removed from the data, and the results have been scaled to unity. The bottom figure shows the along-track line spread function. As in the top figure, the zero offsets have been removed from the data, and the results have been scaled to unity.

Original page 6

PrelaunchAcceptanceReportfortheSeaWiFSRadiometer Table2. Field-of-viewcalculations.All measurementsgivenaretheFWHMof thelinespreadfunction. Along-Scan Values Along-Track Values Band These Results SCADP These Results SCADP No. [mrad] [km] [mrad] 1 1.84 1.30 2 1.78 1.26 3 1.82 1.28 4 1.78 1.25 5 1.82 1.28 6 1.80 1.27 7 1.76 1.24 8 1.77 1.2 Mean 1.80 1.27 1.8 Std. Dev. 0.03 0.02 along- than 1500 km for tilts of +20 ° to enable two-day global relative response) and the value of the abscissa (the track relative response). This result can be represented [km] [mrad] [km] [mrad] [km] 1.70 1.20 1.69 1.19 1.68 1.19 1.70 1.20 1.69 1.19 1.67 1.18 1.66 1.17 1.65 1.16 1.27 1.68 1.18 1.7 1.20 0.02 0.01 as coverage from the nominal altitude. All scan data shall be a three-dimensional figure (Fig. 2), with the base given as transmitted in the local area coverage (LAC) broadcast. the ordinate and abscissa locations, and the height as the Global area coverage (GAC) data subsampled from the product of the ordinate and abscissa values. instru- than -t-45 ° . Figure 3 shows the 50% cross section of the ment response for band 7. It gives the edge of the threedimensional figure at the half maximum of the response for the band in Fig. 2. Figure 3 also shows the values for the major axes, i.e., for the axis with an along-scan value unity. No. Sectiont Length _ Length § unity and for the axis with an along-track value of Each of the two axes gives the widest possible distance cross-track scan need not include data taken at greater Table 3. Field-of-view, calculated from the foot- )rint area. of Band Cross- Side Side 2.67 1.64 1.15 across the cross section in each direction. When these 2.51 1.58 1.12 widest possible distances are multiplied together, they give 2.55 1.60 1.12 an area that is larger than the actual footprint in Fig. 3. 2.44 1.56 1.10 Figure 4 shows the 50% cross section without the added 2.73 1.65 1.16 axes. This footprint for band 7 is nominally, i.e., roughly, 2.48 1.57 1.11 square. For the purposes of this review, it seems preferable to define the field-of-view of the SeaWiFS measurements terms of the area of the (nominally) square footprint. For the eight SeaWiFS bands, the area within the 50% cross sections was calculated, and the length for the side of a square that would enclose those areas was subsequently determined. Table 3 gives the results of the calculations. rep- 3.2 Compliance The lengths of the sides in Table 3 then give the best resentation of the fields-of-view of the eight bands ill the instrument, as determined by this review. The average side length for SeaWiFS is 1.60mrad, 1.13 km, at an altitude of 705 km. The actual values range from 1.10-1.16 km. All values conform to the requirement 2.53 1.59 1.12 in 2.50 1.58 1.11 Mean 1.60 1.13 Std. Dev. 0.03 0.02 "mrad 2 :mrad §kin The angular portion of the SeaWiFS measurements is determined by the rotation rate of the optics within or the instrument, the sampling frequency of the instrument, and the number of samples in a scan line. The optics within the SeaWiFS instrument rotate 6 times per secof the specifications for an IFOV between 1 and 1.21 km. ond 6x360 ° s-l). The time period between pixels is 42 #s. 3. CROSS-TRACK SCAN 3.1 Requirement The active portion of the cross-track scan shall not be 116.6 ° The specifications call for this angle to be between 90 ° less than 90 ° (+45 ° about nadir) nor greater than less and 116.6 °. It should be noted that the distance from the (:k58.3 ° about nadir). The swath width shall not be 4 There are 1,285 pixels per scan line. The calculation is shown in (1): 42 x 10-6s 1,285pixels 116.58 ° 6 x 360°s - pixel scanline scanline" (1)

Original page 7

R.A.Barnes,W.L.Barnes,W.E.Esaias,andC.R.McClain 1.0 ,_* 0.8 O.C) 0.2 0.0 Fig. 2. The response (field-of-view) for SeaWiFS band 7 in three dimensions. The axes for the base are in the along-scan and along-track directions. The height shows the response of the band normalized to unity.

Original page 8

Prelaunch Acceptance Report for the SeaWiFS Radiometer 35 ' • ' ' I I 3O • sa2 O _" 20 15 I .... I ' 1;1 .Sol -1 10 .... I .... I , , , I , I I I J t I i J ; 45 50 55 NOn 9 Scan [O.lmRad/tick] 50 65 7C Fig. 3. The cross section of the response of SeaWiFS band 7 at the 50% response level. The two axes within the cross section correspond to the FWHM values in Table 1. The 1:1 point marks the maximum response for the band. The other four points are located on the half maximum contour. _5 l I S0 25 E _" 20 _5 ' ' ' I , , , , ) , f t 10 , , , , I , , , , I , , , , I I , , , I 45 50 55 AIoflg Scan [0.1mRod/tick] Fig. 4. The measurement footprint for SeaWiFS 50 55 7c band 7. The best estimate of the field-of-view for this band is calculated as the side of a square with an area equal to this footprint. 6

Original page 9

R.A. Barnes, W.L. Barnes, W.E. Esaias, and C.R. McClain start to the end of the first pixel is one pixel. Thus, the angles include the optical paths through the instrument. distance from the start of the first pixel to the end of the Thus, the values in Table 4 include measurements with 1,285th pixel is 1,285 pixels. The distance from the center both sides of the half angle mirror. All of the values in this of the first pixel to the center of the 1,285th pixel is 1,284 table are given with respect to the alignment mirrors. The pixels. As a result, the angular distance from the center of SCADP gives the uncertainty for the aft tilt as +0.003 °, pixel 1 to the center of pixel 1,284 is 116.48 °. and for the nadir and forward tilt the error is given as A figure in the SCADP gives a clock rate of 1.905 MHz +0.006 °. The specifications call for tilt knowledge to be +1%. It also shows that there are 80 bits per pixel, or within 0.01 °. 41.99 microseconds (+1%) per pixel. The rotation rate The SeaWiFS scanner tilts from -20 to +20 ° in about and the number of pixels per scan line are also described 13 seconds. The speed profile for the tilt is Gaussian, with in the SCADP. a maximum speed of about 6 ° s-1. Subsampling of each scan line for GAC will be performed by the SeaStar bus. This subsampling will give the 90 ° angular range (centered about nadir), which is required in the specification. 4. FORE-AND-AFT POINTING 4.1 Requirement The data shall be taken with a sensor capable of pointing the swath fore and aft (positive and negative with respect to the velocity vector, respectively) to avoid specular solar reflectance from the ocean's surface. The sensor scan must be capable of tilts of 0.0 °, +20.0 °, and -20.0 ° from nadir. Changes in the tilt angle from -20 ° to +20 ° shall take less than 30 seconds. Data describing the tilt angle Table 4. SeaWiFS tilt angles. Tilt Mirror Side An_Je Uncertainty Aft A 19.896 ° +1-0.003° Aft B 19.888 ° +0.003 ° Aft Average 19.892 ° +0.003 ° Nadir A 0.075 ° +0.006 ° Nadir B 0.068 ° +0.006 ° Nadir Average 0.072 ° +0.006 ° Fore A - 19.850 ° +0.006 ° Fore B - 19.857 ° +0.006 ° Fore Average - 19.853 ° +0.006 ° shall be accurate to within 0.01 ° . 5. DARK LEVEL 4.2 Compliance 5.1 Requirement The nadir direction for SeaWiFS is given as the +x A portion of every scan shall contain sensor output data axis. The instrument scans in the (x,z) plane, and the solar while the field-of-view is obscured and the input radiance diffuser points in the +y direction. The diffuser points at is less than the Noise Equivalent Differential Spectral Rathe sun after the instrument has completed its Earth views diance (NEdL). and is passing over the South Pole. The diffuser is at the back of the instrument. SeaWiFS flies in the -y direction. 5.2 Compliance A forward tilt has a -y direction and a backward tilt has a +y direction; the values here are from the SCADP. The dark restore value, for each scan of the instrument. This fore and aft angles have been measured relative to nadir. value is provided in accordance with the dark level mea- The angles for nadir direction were measured with respect surements specification. For details on the design and poto the alignment mirrors mounted on the instrument. sitioning of the dark restore, the reader is referred to Fig. 1 Here is a brief description of the procedure that was of Woodward et al. (1993), which shows dark direct current used to measure the fore and aft tilt angles relative to (DC) restore at the angular range between 140 ° and 220 ° nadir. With the instrument at zero tilt, a collimated light from nadir for each SeaWiFS scan. SeaWiFS provides the source was aligned with pixel 643 of the SeaWiFS scan. dark level measurements required by the specifications. The entire instrument was then rotated in a specially designed cradle fore and aft 20 ° , and the scanner was tilted 6. BAND TOLERANCES in the other direction. The cradle was designed to rotate the instrument to these 20 ° locations with great accuracy. 6.1 Requirement The instrument tilt realigned the nadir pixel of the instrument with the collimated light source. The collimated The SeaWiFS instrument incorporates a zero offset, or The location of the band edges shall be +2 nm (3 a) light source was moved in the y direction to determine the of the values in Table 1, and shall be stable to less than difference in the tilt angle from 20 ° . +1 nm over the duration of the ground test program. The Since the tilt angles used in this procedure were de- edge range shall not exceed 50% of the bandwidth in any termined by SeaWiFS measurements of the source, these spectral band.

Original page 10

PrelaunchAcceptanceReportfortheSeaWiFSRadiometer 6.2 Compliance 6.2.1 Band Edges The band edge calculations are described in Barnes et al. (1994). The results presented here have been taken from that manuscript, and are summarized in Tables 5-7. These tables include the 5%, 50%, and 80% response levels edge 7.2.2 1% Response Point that are required to calculate the band edges and the ranges. The specifications call for the band edges to be within ±2 nm of the values in the specifications. The right (upper) band edge for band 8 falls at the 2.0nm limit, and the remaining band edges are within the specification limits. The band edge and edge range results in Barnes et in the lations in Barnes et al. (1994). The calculations have been al. (1994) are in agreement with the results in found SCADP. 6.2.2 Edge Ranges The specifications require that the edge range for each band shall not exceed 50% of the bandwidth. The edge range is the wavelength interval between 5% of peak response and 80% of peak response. The edge range values are given in Tables 6 and 7. These values were also taken from the calculated results in Barnes et al. (1994). The edge range for band 2, left edge, is 54% of the bandwidth. The edge range for band 6, right edge, is 51%. The remaining ratios fall within the 50% limits. 7. OUT-OF-BAND RESPONSE 7.1 Requirement The out-of-band response shall be less that 5% of the within-band value. Each 1% point shall be within 1.5 times the bandpass from the corresponding band edge. Compliance with this specification shall be determined for a source with spectral shape equivalent to Lcloud (the spectral radiance of a Lambertian surface of 100_ reflectance illuminated by the sun at 22.5 ° zenith angle). Note: The 1994). All of the calculated out-of-band values are well within the 5% value in the specifications. Table 9 contains a comparison of the results from Barnes et al. (1994) with those from the digitized spectral data provided as part of the SCADP. The two sets of calculations show reasonable agreement. The specifications state that each 1% response point shall be within 1.5 times the bandwidth from the corresponding band edge. The calculated results are given in Tables 10 and 11. These results are also derived from calcumade using a 5,900 K blackbody source, in accordance with the specifications. All of the 1% response points are well within the specification limits. In general, the shapes of the SeaWiFS bands are a significant improvement over the shapes required by the specifications. 8. SPECTRAL DIFFERENCES 8.1 Requirement If multiple detector elements are used within a band, the spectral response of all detector elements in a band shall be compared, as to location and shape, by use of normalized spectral response curves. The central wavelength of any element must be within ±0.5 nm of the average central wavelength for all elements of the band. The integrated spectral response between the 10% response points shall not differ by more than 10% for any two elements in the band. 8.2 Compliance Ldoud radiances can be found in Table 18, below. They 8.2.1 Center Wavelength are also called the SeaWiFS maximum cloud radiances. 7.2 Compliance 7.2.1 Out-of-Band Response. The in-band response is defined as the integrated response of each band between the 1% transmission points. The out-of-band response is defined as the integrated response at all other wavelengths. The ratio of out-of-band response to in-band response is used to give the percent out-of-band response. These values have been calculated by Barnes et al. (1994). Tables 8 and 9 have been adapted on the quantum efficiency of the type of silicon photodiode from Table 12 of Barnes et al. (1994) and are based re- used in SeaWiFS is roughly constant over the 10-20 nm measurements using a 5,900K blackbody source, as quired in the specifications. A 5,900K blackbody dupliwave- ness implies a very minor effect from the detectors on the cates the spectral shape of the solar output over the No spectral measurements have been specifically designed for the SeaWiFS instrument to confirm this specification. However, there is evidence to indicate that the central wavelength for each of the four channels in each SeaWiFS band is well within 0.5 nm of the average central wavelength for that band, as called for in the specifications. The design of SeaWiFS has the four channels for each band located below a single interference filter for that band. As shown in Barnes et al. (1994), the shape of the spectral response of the filter dominates the determination of the central wavelength. Barnes et al. (1994) also shows that half-widths of the interference filters. This spectral flatlength range of the SeaWiFS measurements (Barnes et al. central wavelength of the band.

Original page 11

R.A.Barnes,W.L.Barnes,W.E.Esaias,andC.R.McClain Table 5. Reference wavelengths for band edge and edge range calculations. All measurements are in nanometers. Band Left No. 5% Point 50% Point 80% Point 398.4 403.2 407.4 427.8 434.1 438.3 474.6 480.8 483.3 492.8 498.9 501.1 540.4 545.5 547.6 653.4 658.3 660.1 734.5 744.7 748.9 835.8 845.7 849.2 Right 80% Point 50% Point 5% Point 421.1 423.4 428.9 451.0 453.7 459.5 499.4 501.4 507.1 518.9 521.2 527.2 561.4 563.8 570.4 675.6 678.2 685.8 780.8 785.0 798.3 884.1 887.0 896.8 Table 6. Band edge calculations. All measurements and calculated quantities are in units of nanometers. Band Left Band Edge Results No. Specified Measured Difference 1 402 403.2 1.2 2 433 434.1 1.1 3 480 480.8 0.8 4 500 498.9 - 1.1 5 545 545.5 0.5 6 660 658.3 - 1.7 7 745 744.7 -0.3 8 845 845.7 0.7 Right Band Edge Results Specified Measured Difference 422 423.4 1.4 453 453.7 0.7 500 501.4 1.4 520 521.3 1.3 565 563.8 -1.2 680 678.2 - 1.8 785 785.0 0.0 885 887.0 2.0 Table 7. Edge range calculations. All measurements are in nanometers. Band Bandwidth Left Edge Percent of Right Edge Percent of No. Range Bandwidth Range Bandwidth 1 20.2 9.0 2 19.6 10.5 3 20.6 8.7 4 22.4 8.3 5 18.3 7.2 6 19.9 6.7 7 40.3 14.4 8 41.3 13.4 45 7.8 39 54 8.5 43 42 7.7 37 37 8.3 37 39 9.0 49 34 10.2 51 36 17.5 43 32 12.7 31

Original page 12

PrelaunchAcceptanceReportfortheSeaWiFSRadiometer Table 8. Calculated out-of-band responses for the eight SeaWiFS bands. The instrument responses are given as the output of the photodiode in picoamperes (pA). The 5,900 K radiances in the calculations are normalized to the expected saturation radiance for each band at the nominal center wavelength for each band. The upper and lower extended band edges come from Tables 7 and 8 of Barnes et al. (1994). These results are calculated over the wavelength range from 380-1,150nm. Lower Lower In-Band Upper Upper Out-of-Band Band Out-of-Band Extended Response Response No. Response [pA] Band Edge [nm] [pA] Extended Out-of-Band Band Edge [nm] Response [pA] [%] 1 3.38 395.2 2175.34 433.6 11.77 0.70 2 9.59 424.1 3418.80 463.7 1.56 0.33 3 6.48 470.7 4301.14 511.7 28.08 0.80 4 17.32 488.1 4586.23 530.7 8.96 0.58 5 39.14 536.3 3631.84 577.2 46.14 2.35 6 12.66 646.7 2071.19 692.5 7.84 0.99 7 10.17 727.3 2818.97 813.4 29.58 1.41 8 66.36 826.4 2191.97 907.5 15.43 3.73 Table 9. A comparison of the calculated instrument out-of-band response results found in Barnes et al. (1994) with SCADP results for the same parameter. Both sets of calculations show the instrument's out-of-band response to be within the 5% value in the specifications. Band Barnes et al. (1994) SCADP No. [%] [%] 0.70 0.65 0.33 0.40 0.80 0.80 0.58 0.57 2.35 2.35 0.99 0.98 1.41 1.41 3.73 3.73 The detectors in each band are etched from a single 10%, i.e., the range from greatest to least integrated response for the elements can be no more than 10% of the value of the average integrated response. Measurements of individual SeaWiFS channels with the SBRC integrating sphere have shown that the output from the individual detectors in each band corresponds at the 1% level. That is, the integrated response of the detectors agrees at the 1% level. Of course, the output from each channel has been adjusted (or rather, the values of the resistors in the operational amplifiers have been adjusted) to give uniformity at the 1% level. In general, switching from detector to detector in a SeaWiFS band will give a change in output from the instrument on the order of l%--considerably better than the 10% limit in the specification. This 10% specification was included in the SeaWiFS specification, in case the design of the flight instrument included area arrays with many individual detectors. piece of silicon. The overall dimensions for the set of four 9. BAND CO-REGISTRATION detectors are approximately 0.05in × 0.01 in. Assuming reasonable uniformity in the manufacture of the silicon ma- 9.1 Requirement terial, the four detectors in each band should have nearly identical spectral response over the half-width of the interference filter. It is the response of the interference filter, not the detectors, that dominates the determination of the central wavelength for each SeaWiFS band. The estimate of the within-band spectral differences is based primarily The IFOVs from all spectral bands shall be ,:o-registered to within 0.3 pixel (1 a). 9.2 Compliance The specification requires that the IFOVs from all specon an assumption of uniformity in the manufacture of the tral bands be co-registered within 0.3 pixel. From the fieldan of-view calculations in Section 2 above, it was determined interference filter. This estimate of uniformity covers area of 0.05 in x 0.01 in. 8.2.2 Integrated Spectral Response that the IFOV for the instrument is 1.6mrad on a side. The band-to-band registration results presented here are based on the along-scan and along-track line spread functions in Section 2. The line spread measurements were The within-band spectral response specification also re- made by moving a narrow slit across the field-of-view of quires an integrated spectral response, between the 10% the nadir SeaWiFS pixel. The slit was part of a colliresponse points for the individual elements, to be within 10 mated light source, and the slit was moved in increments

Original page 13

R.A.Barnes,W.L.Barnes,W.E.Esaias,andC.R.McClain Thble 10. Reference wavelengths for one percent response Band Left point calculations. All values are in nanometers. Center Right No. 1% Point 50% Point Wavelength 50% Point 1% Point 395.2 403.4 424.1 434.2 470.7 480.8 488.1 498.9 536.3 545.4 646.7 658.3 727.3 744.4 826.4 845.5 413.4 423.4 433.6 444.0 453.8 463.7 491.1 501.4 511.7 510.1 521.3 530.7 554.6 563.8 577.2 668.2 678.1 692.5 764.6 784.9 813.4 866.1 886.7 907.5 Table 11. One percent response point calculations. The specifications call for a ratio to the bandwidth that is less than 150%. The bandwidths were previously given in Table 7. Band Bandwidth Left 1% from Percent of Right 1% from Percent of No. [nm] Left 50% [nm] Bandwidth Right 50% [nm] Bandwidth 20.0 8.2 19.6 10.1 20.6 10.1 22.4 10.8 18.4 9.1 19.8 11.6 40.5 17.1 41.2 20.1 41 10.2 51 52 9.9 5O 49 10.3 5O 48 9.4 42 50 13.4 73 59 14.4 73 42 28.5 7O 49 20.8 5O (ticks) equal to a 0.1 mrad angular displacement as seen 10.2 Compliance by the instrument. The center for each pixel (in ticks) was calculated as the average of the positions of the two half The noise in the SeaWiFS instrument was measured maximum points for each band. The along-scan and along- by viewing the SBRC integrating sphere. The SNR was track centers were calculated independently. The absolute calculated by determining the mean and the standard devalues for the central positions (in ticks) are not important viation in a 21 pixel-long section of the scan of the sphere's in these results--it is the relative locations of the central center. These measured results are listed in Table 14. The points that are relevant here. Table 12 gives the along- measurements were made close to, but not at, the exact scan results. Table 13 gives the along-track results. Fig- typical radiance levels required by the specifications. The ure 5 shows the results relative to the average for the set results have been scaled to the typical levels by changing of band centers, and also shows the locations of the eight the SNRs, assuming that the noise in the measurement band centers in two dimensions. Figure 5 also includes a varies as the square root of the change in signal level over square 0.1 pixel wide, centered on the average band cen- this small range. These measured SNRs exceed the reter. The maximum distance between band centers in the quirements of the specifications. However, the SNR meaalong-scan direction is 0.15 pixels, and the maximum dis- surements include the non-uniformity in the output of the tance along-track is 0.21 pixels. The instrument's spectral sphere. This additional variation makes the measured valband-to-band registration is significantly better than the ues lower than the actual SNRs. requirements of the specifications. the noise in the dark output from the instrument. These SBRC has provided a calculation of the SNRs, based on 10. SENSITIVITY calculated values are also listed in Table 14. For purposes of evaluation, it seems reasonable to assume that the actual SNRs for SeaWiFS fall between the measured and 10.1 Requirement calculated SNRs in Table 14. In all cases, however, the Table 14 provides the signal-to-noise ratio (SNR) speci- measured values (the lower limits for the actual SNRs) are fications for all bands at a gain value of unity. The required better than the requirements in the specifications. Sec- SNR shall be achieved at the typical spectral radiance lev- tion 19 describes the model for instrument noise devised els (Ltypical). NEdL may be calculated from the expression: by the Project. The SNRs from this model fall between the NEdL= Ltypical/SNR. measured and calculated values in Table 14. For the review 11

Original page 14

Prelaunch Acceptance Report 02 0.1 x for the SeaWiFS Radiometer f2_ %/5 ¢- 0.0 O Average _ 6 e- _o -0.1 -0.1 0.0 0.1 0.2 Along Track (Pixels) Fig. 5. Band-to-band registration of SeaWiFS. The locations of the individual band centers are given relative to the average for the eight bands. The distances are given in "pixels," where 1 pixel equals 1.6mrad. The figure also includes a square that average center location for the eight bands. es- Table 13. Along-track band center measurements. panel, the results in Section 19 give the best prelaunch radi- The same procedure used to obtain the measured timates of the SNRs for the instrument at the Ltypical ances. However, those values remain prelaunch estimates. is 0.1 pixel on each side and which is centered on the values in Table 12 was employed here. An extensive set of on-orbit measurements (McClain et al. Center Distance Distance 1992 and Woodward et al. 1993) will be used to obtain an Band Position from center from center improved set of SNRs. Table 12. Along-scan band center measurements. The measurements were made in ticks, with each tick equal to 0.1 mrad. The results are also given in mrad and in pixels, where one pixel is 1.6 mrad. The range gives the distance between the two bands that are farthest apart. Center Distance Distance Band Position from center from center No. [ticks] [mrad] [pixel] 57.56 -0.003 -0.002 56.71 -0.088 -0.055 58.08 0.049 0.031 57.03 -0.056 -0.035 59.00 0.141 0.088 58.26 0.066 0.041 57.52 -0.008 -0.005 56.58 -0.101 -0.063 Range 2.42 0.242 0.151 12 No. [ticks] [mrad] [pixel] 1 22.50 0.050 0.031 2 20.29 -0.170 -0.106 3 21.15 -0.085 -0.053 4 20.78 -0.121 -0.076 5 23.02 0.103 0.064 6 21.83 -0.016 -0.010 7 23.67 0.167 0.104 8 22.73 0.073 0.046 Range 3.37 0.337 0.210 11. POLARIZATION 11.1 Requirement The radiometric data shall be nominally insensitive to linear polarization. The polarization factor (PF) as defined below, shall be no greater than 2% over scan angles from +45 ° to -45 ° for all bands and tilt angles between -20 ° and +20 °. /max and Imin are the recorded maximum

Original page 15

R.A.Barnes,W.L.Barnes,W.E.Esaias,andC.R.McClain andminimumoutput when the plane of incoming 100% linearly polarized light is rotated through 180 ° . show a dominant signal of 1 cycle per 360 ° and show an data for band 1 are shown in Fig. 6. These data are representative of the results for bands 1-6. The input data PF - /max - Imin average value of 100. This allows a direct conversion of < 0.020 (2) /max + Imin the average-to-peak values from the Fourier analyses into percentages. Table 14. SNRs for the SeaWiFS Bands. Band Specified Measured Calculated Table 15. They are given in terms of the average-to-peak No. SNR SNR SNR values for the input to the Fourier transforms, for the one- The results of the Fourier analysis are summarized in 1 499 940 1107 cycle component, and for the two-cycle component. It 2 674 950 1269 is the two-cycle component that contains the information 3 667 1156 1402 about the polarization of the instrument. Since the aver- 4 640 1055 1373 age value for the measurements is 100, the average-to-peak 5 596 690 1242 values translate directly into percentages. 6 442 798 846 As shown in Table 15, the one-cycle Fourier compo- 7 455 860 971 nent accounts for almost half of the average-to-peak signal 8 467 670 796 in the data from bands 1-6. In the judgement of the review panel, the majority of the observed variation in the measurements of these bands does not come from polariza- 11.2 Compliance tion sensitivity in the SeaWiFS instrument. Presumably, Polarization is defined in this requirement in terms of the bulk of the variability derives from a non-uniformity the difference from the average for the output range. This in the polarizer plate used in the measurements. The efdefinition then gives a polarization value that is half of fects of the polarizer plate repeat for each of the 6 bands the difference between the maximum output and the mini- for which the polarizer plate was used. The polarization mum output. The polarization sensitivity of SeaWiFS was in the SeaWiFS instrument for these bands is less than checked using two linear sheet polarizers, at different times, 0.25%. between the light source and the instrument. The first polarizer (Polaroid HR) checked bands 1-6, and the second (Polaroid HN) checked the two near infrared (IR) bands-bands 7 and 8. Changes in the output of the instrument during these tests included variations in the light through the polarizers as well as polarization dependent changes in the output of the instrument itself. The variation in light through the polarizer is an artifact of the measurement procedure, since the rotation axis for the polarizer is not located at the center of the input aperture of the instrument. Crossed polarizers showed good extinction for each band, indicating that the polarizers were creating polarized light. Using knowledge of the expected angular dependence of the polarization effect, an effort was made to separate the instrument's portion of the polarization pattern. The po- Table 15. Summary of polarization results. The values in columns 3-5 are average-to-peak values, expressed as percentages. Band Input One-Cycle Two-Cycle No. to Transform Output Output 1 1.90 0.96 0.15 2 1.10 0.53 0.09 3 0.90 0.30 0.11 4 0.60 0.22 0.08 5 1.10 0.39 0.20 6 1.00 0.43 0.14 7 0.90 0.01 0.26 8 1.10 0.05 0.35 The test results for bands 7 and 8 show a pattern that larization measurements were made by rotating each polar- is similar for the two bands (see Fig. 7 for band 7). Howizer through 360 ° in 22.5 ° increments. Since polarization ever, the pattern is noticeably different from the pattern changes from a minimum to a maximum in 90 °, the re- for bands 1-6. Of course, bands 7 and 8 used the same posponse of the instrument should show two cycles, i.e., two larizer plate, a different polarizer plate than the one used maxima and two minima, in the 360 ° rotation. Fourier for measurements on bands 1-6. The patterns for bands analysis was used to look for the expected two cycle sinu- 7 and 8 show essentially no one-cycle component. In adsoidal function in the polarization results. dition, the two-cycle component in the pattern is some- The data were taken at 16 angles, from an arbitrarily what larger than for bands 1-6, falling between 0.25% and determined zero, through 360 °, back to the original an- 0.35%. No instrument based explanation can explain why gular zero. The two measurements at 0 ° were averaged the polarization in SeaWiFS for bands 7 and 8 should be to give a data set with 15 points from 0 to 337.5 °. From different from that for the other 6 bands. For this reason, a the Fourier transform of these data, sinusoidal waveforms large part of the two-cycle Fourier component for bands 7 with 1-9 cycles per 360 ° were extracted for analysis. The and 8 appears to be an artifact caused by variations in the 13

Original page 16

PrelaunchAcceptanceReportfortheSeaWiFSRadiometer polarizerplateforthesebands.Thisassumptionexplains13. QUANTIZATION thesimilarityin theangularpatternforthemeasurements onbands7 and8,in thesamemannerthat it explainsthe similarityin thepatternsforbands1-6. in Data shall be quantized at 10 bits. The differential Thus,forall eightSeaWiFSbands,thepolarization theinstrumentis estimatedto belessthan0.25%. 12. DYNAMIC RANGE 12.1 Requirement The sensor shall be designed to operate over a dynamic range that extends from the noise floor (NEdL) in each spectral band to the maximum levels (Ldoud) given in Table 18. (There is an amendment to the Contract. With this addition, there are 3 radiance levels in the specifications: saturation radiance, maximum ocean radiance, and maximum cloud radiance. The specification radiances are given in Tables 16, 17, and 18, respectively.) 12.2 Compliance for themselves. The specification, which gives the saturation levels the eight SeaWiFS band, was supplemented by an amendment to the contract in August 1993. The amendment provides the maximum ocean radiances and the maximum cloud radiances for the instrument. These changes to the specifications were required due to the use of bilineax gains in SeaWiFS. However, the saturation values from the original specification remain an important part of the SeaWiFS measurements. The saturation values are the maximum radiances that the SeaWiFS bands would provide if the knees in ground. the point where the radiance vs. counts slope changes) the bilinear gains are ignored. The saturation values define the sensitivity of SeaWiFS, in counts per unit radiance, for the instrument's ocean measurements. In summary, the saturation radiances define the sensitivities of the SeaWiFS bands for ocean measurements, i.e., for measurements below the knees of the bilinear gains. The maximum ocean radiances are the radiances at the knees for the eight bands. The maximum cloud radiances are the greatest values that the SeaWiFS bands will measure. With the addition of the "maximum ocean" and the "maximum cloud" radiances to the specifications in the contract amendment, the term "saturation" radiance has become a misnomer. However, the term has been retained here. Table 16 gives the saturation radiances from the Sea- WiFS calibration (ignoring the bilinear gain knees). Table 17 gives the maximum ocean radiances, and Table 18 gives the maximum cloud radiances. All radiances are calculated in terms of milliwatts per square centimeter per micrometer per steradian (mW cm-2# m-lsr - 1). The differences from the specifications are also calculated. They are small, and present no problem with regard to specification compliance. 14 13.1 Requirement linearity of the quantizer(s) shall be better than one-half a least significant bit. 13.2 Compliance The data from SeaWiFS are quantized at 10 bits. The data from each detector in each band are digitized with 12-bit analog-to-digital (A/D) converters. For 4-to-1 Time Delay and Integration (TDI), the 12-bit values from each detector are summed to give a 14-bit value. The bottom four bits from this output are removed, with the upper 10 bits sent to SeaStar. Measurements by the instrument manufacturer (SBRC), using a voltage ramp, show linearity at better than one-half of the least significant bit of the output 10 bits. This procedure was relatively easy, since the least significant bit for each detector corresponds to four times the least significant bit in the A/D converters, 14. MTF 14.1 Requirement The MTF of the data shall equal or exceed the values tabulated in Table 19 below, in both the along-track and cross-track directions for a sine wave input. The Nyquist frequency? has a spatial period equal to two IFOVs on the The MTF specifications shall be satisfied for modulations between dark and Ltypical and between dark and Lmax, for every detector element in each spectral band. Data describing the MTF shall be provided from prelaunch testing to verify that the specification is met. Data from lunar views and/or internal stabilized sources will also be provided for analysis of the MTF on orbit. 14.2 Compliance The SeaWiFS MTFs are calculated from the line spread functions in Fig. 1. They consist of the Fourier transforms of the line spread functions. The results in Table 20, below, use the MTF calculation program provided by SBRC. The values in the specification give minimum amplitudes for several low frequency sinusoidal waves from the Fourier analysis. The waves are given in terms of their wavelengths relative to the width (FWHM) of the field-of-view, i.e., in cycles per pixel. t The Nyquist frequency is the minimum sampling frequency of a digital system sufficient to reconstruct the original information. For SeaWiFS, the original information is the input sequence of radiances.

Original page 17

R.A.Barnes,W.L.Barnes,W.E.Esaias,andC.R.McClain 102 S I01 bl i I00 99 98 0 I 2 3 4 5 6 7 a) 102 / I01 blk -- I00 99 \ /t \ / 98 x k 337.5 b) 102 _¢- - \ I01 --,oo _ 99 k , , 98 -- 0 ×k 337.5 d) 8 9 I0 11 12 13 14 15 i I 0.5 / \ ! • o t -o- \ /! -0.5 / -i - S 0 x k 337.5 c) 0.2 ! \ / _'Tk o \ / \ / / _).2 0 x k 337.5 e) Fig. 6. Polarization measurements for SeaWiFS band 1. a) The instrument response measured at 22.5 ° increments of polarizer rotation. These values are the input values for Fourier analysis, b) The polarization values with and without the one-cycle component bl (k) equal the measured values, and h(k) equal the measured values without the one-cycle component, c) The one-cycle polarization component generated by Fourier analysis, d) The polarization values with and without the two-cycle component bl(k) equal the measured values and h(k) equal the measured values without the two-cycle component. e) The two-cycle polarization component generated by Fourier analysis. 15

Original page 18

Prelaunch Acceptance Report 102 101 b7 i 1 O0 -0- 99 98 0 I 2 3 4 5 6 7 I02 101 --100 _ \j/ 99 , i 98 ! o x k 337.5 b) 102 I01 --1oo.'- \ × --.- 99 98 0 x k 337.5 d) for the SeaWiFS Radiometer 8 9 I0 I1 12 13 14 15 i 0.1 0.05 F-- It--di ffk 0 __.). _ j,- _- p.-t ]P--4 -0.05 -01 ! o 337.5 x k c) 0.3 ;\ /\ 0.15 '\ A _k o ,\ 1 \ / -0.15 \ -0.3 0 x k 337.5 e) Fig. 7. Polarization measurements for SeaWiFS band 7. a) The instrument response measured at 22.5 ° increments of polarizer rotation. These values are the input values for Fourier analysis, b) The polarization values with and without the one-cycle component b7(k) equal the measured values, and h(k) equal the measured values without the one-cycle component, c) The one-cycle polarization component generated by Fourier analysis, d) The polarization values with and without the two-cycle component b7(k) equal the measured values and h(k) equal the measured values without the two-cycle component. e) The two-cycle polarization component generated 16 by Fourier analysis.

Original page 19

R.A. Barnes, W.L. Barnes, W.E. Esaias, and C.R. McClain Table 16. SeaWiFS saturation radiances (in mWcm-2/_ m-lsr-1). The measured and specified saturation radiances appear in columns 3 and 4, respectively. B and Measured No. Gain Radiance 1 13.76 2 13.44 3 10.52 4 9.22 5 7.47 6 4.25 7 3.02 8 2.15 Specified Percent Radiance Difference 13.63 1.0 13.25 1.4 10.50 0.2 9.08 1.5 7.44 0.4 4.20 1.3 3.00 0.5 2.13 1.0 Average % Difference 0.9 Greatest _ Difference 1.5 Least % Difference 0.2 Table 17. SeaWiFS maximum ocean radiances (in mW cm-2p m-lsr-1). The measured and specified ocean radiances appear in columns 3 and 4, respectively. Band Measured No. Gain Radiance 10.90 10.56 8.18 7.16 5.74 3.25 2.29 1.64 Specified Percent Radiance Difference 10.84 0.5 10.46 0.9 8.19 -0.1 7.05 1.6 5.74 0.0 3.21 1.2 2.29 0.0 1.62 1.0 Average % Difference 0.6 Greatest % Difference 1.6 Least % Difference -0.1 Table 18. SeaWiFS maximum cloud radiances (in mW cm-2# m-lsr-1). The measured and specified cloud radiances appear in columns 3 and 4, respectively. Band Measured No. Gain Radiance 1 60.16 2 67.91 3 68.21 4 66.47 5 64.97 6 54.93 7 42.98 8 34.38 Specified Percent Radiance Difference 60.02 0.2 66.24 2.5 68.17 0.1 65.62 1.3 65.16 -0.3 53.78 2.1 42.95 0.1 34.05 1.0 Average % Difference 0.9 Greatest % Difference 2.5 Least % Difference -0.3 17

Original page 20

PrelaunchAcceptanceReportforthe SeaWiFSRadiometer TheMTF givesanideaof the responseofthe instru- 16. TRANSIENT RESPONSE mentto variationsin theinputradiance,bothalong-scan andalong-track.Forexample,considera casein which 16.1 Requirement theinstrumentscansacrossa scenewitha variationofthe pe- Radiometric data should be relatively free of effects of inputradiancehavinga sinusoidalshapeanda spatial varies overshoot and ringing when the IFOV scans across a steep riodoftwopixelsperperiod.In this case,thescene to thespecification,gradient in radiance, from a maximum radiance of Lcloud at the Nyquistfrequency.According (from to a minimum radiance of Ltypical. For this radiance step theoutputfromthesensorshouldshowa variation of change, the output signal shall settle to within 0.5% of its pixelto pixel)acrossthescenethatis greaterthan30% theamplitudeof inputsinusoidalradiancevariation. Table 19. MTF requirementsfor spatialresolu- ;ion. Frequency/ MTF NyquistFrequency 0.00 1.0 0.25 0.9 0.50 0.7 0.75 0.5 1.00 0.3 15. GAINS 15.1 Requirement Band independent gains shall be provided, which are or During the April 1993 SeaWiFS Pre-Ship Review, it commandable band by band, and which will increase decrease sensitivity according to the following: (3) time, it was decided that the instrument manufacturer S,_ = G,, x S, where S, and Sn are the initial detector signal and the at filters, incorporation of bilinear responses and correspondsignal with gain, respectively, and Gn is the gain factor gain setting n. The nominal G= values for gain settings in and testing. The modifications did not include changes of n equal to 2, 3, and 4 shall be based on the values on- to the instrument's focal planes. A description of the re- Table 21. These G values include those required for be vised BTR specification is given in the following section. board solar and lunar capabilities. The values of G will be The discussion presented here centers on the performance within 5% of the specifications in Table 21, and shall known relative to G1 = 1 with an accuracy of greater than 99.5%. The nominal G,_ value for n = 2 is 2. 15.2 Compliance Gain values for gain settings 3 and 4 were re-evaluated in the summer of 1993 during a meeting at OSC. Gains 3 and 4 are used for solar and lunar measurements. The SCADP contains the predicted on-orbit lunar and solar diffuser radiances. They have been derived from ground final value within 10 IFOVs. 16.2 Compliance 16.2.1 Original Specification The original SeaWiFS specification describes the required response of the instrument when it scans across a steep gradient in radiance. The specification describes a gradient that is expected be found on orbit, i.e., from the radiance level for a cloud (Lcloud) to the typical radiance level for ocean measurements (Ltypical). The original specification requires that the output from the instrument settles to within 0.5% of its final value (ntypical) within 10 pixels. This limit, calculated in counts using values from the specifications, is given in Table 22. Since the values are small, the limit is given to 0.1 count relative to the quantization of the data. was determined that the instrument did not meet the original bright target recovery (BTR) specification. At that would rework SeaWiFS to improve its stray light characteristics. The series of modifications included the tilting of ing changes to gains, refiguring the polarization scrambler, of the reworked SeaWiFS radiometer only with respect to the original specifications. The results in Table 23 were derived from laboratory measurements using the output of the SBRC integrating sphere. The measurements were made in four sets_ne for each focal plane in the instrument. Color glass filters were placed over the output from the sphere in order to give a spectral shape which approximated the sun over the wavelength range of the bands on each of the four focal planes. Cross-talk between bands on a focal plane is a based solar and lunar measurements at SBRC (Biggar et significant factor in the recovery of the instrument from al. 1993). The values presented in Table 21, below, are bright targets. The measurements were made for a three appropriate for those radiances. The gains are given for a pixel wide slit. 4:1 TDI. These relative gain values can be measured on orbit. The results in Table 23 give the distance, in pixels, required for the instrument to settle to less than 0.5% of Current plans will have these values checked twice a day Ltypical, using the counts given in Table 22. For bands 1at the start of the mission. 18 5, the results give the pixels required to settle to 3 counts

Original page 21

R.A.Barnes,W.L.Barnes,W.E.Esaias,andC.R.McClain Table 20. MTF calculations for SeaWiFS. The amplitude 0.500, 0.375, 0.250, and 0.125 cycles per pixel. is shown for each band at four different wavelengths: Band Along-Scan MTF Amplitude Along-Track MTF Amplitude No. 0.500 0.375 0.250 0.35 0.57 0.76 0.36 0.58 0.76 0.36 0.58 0.77 0.36 0.58 0.77 0.35 0.57 0.76 0.37 0.59 0.78 0.39 0.61 0.79 0.34 0.55 0.74 Minimum Amplitude 0.30 0.50 0.70 0.125 0.500 0.375 0.250 0.125 0.93 0.44 0.65 0.84 0.96 0.92 0.41 0.62 0.81 0.95 0.93 0.41 0.62 0.82 0.95 0.93 0.40 0.61 0.81 0.95 0.93 0.48 0.67 0.84 0.96 0.93 0.43 0.63 0.82 0.95 0.94 0.49 0.68 0.85 0.96 0.91 0.47 0.66 0.83 0.96 0.90 0.30 0.50 0.70 0.90 above background. For bands 6-8, the results give the radiances between 0.2 Ltypical and 0.9 Lmax, the absolute pixels required to settle to 2 counts. Table 23 also gives radiometric accuracy shall be within +6%. Measurements the number of pixels required for the instrument to settle to of the accuracy shall be made, as a minimum, at scan anzero counts after illumination by the slit. All eight bands gles centered at 0 °, -40 °, and +40°; tilt angles of 0 °, -20 °, settle to less than the specification limit within 10 pixels. and +20°; special tilt angles, should they be required, used to view the moon; and at all gains. 2hble 21. SeaWiFS gain values. These values are given relative to gain 1. The nominal value for gain setting 2 is 2. 17.2 Compliance Band No. G1 G2 63 64 1 1 1.931 1.302 1.642 2 1 1.940 1.303 1.648 bration at the 5% level. The SeaWiFS instrument was cal- 3 1 1.951 0.900 1.655 17.2.1 Accuracy at Nadir This specification calls for an absolute radiometric cali- 4 1 1.955 0.796 1.658 ibrated radiometrically by the manufacturer (SBRC) using 5 1 1.961 0.652 1.579 an integrating sphere that was calibrated with standards 6 1 1.969 that are traceable to the National Institute of Standards 0.376 0.671 7 1 1.969 0.323 0.583 and Technology (NIST). In addition, the SBRC sphere has 8 1 1.975 0.272 been compared with the GSFC sphere, which has also been 0.507 calibrated using standards traceable to NIST. The comparison of the GSFC and SBRC spheres showed agreement at 16.2.2 Revised BTR Specification the 2% level. In the summer of 1993, there was a meeting at OSC F_ndamentally, the accuracy of the radiometric caliin Chantilly, Virginia. At that meeting, a set of improve- bration of SeaWiFS reduces to the accuracy of the calibraments to ameliorate stray light in SeaWiFS was discussed. tion of the integrating sphere. The absolute uncertainty Each proposed improvement was presented, accompanied in the radiances from the sphere is the largest of the set by a corresponding estimate of the resulting improvement of uncertainties in the instrument calibration. In addito performance of the instrument. A final and accepted tion, many uncertainties in the radiometric calibration of set of instrument improvements from that meeting was in- SeaWiFS, such as the alignment of the sphere and the incorporated into a Contract modification. As demonstrated strument, duplicate uncertainties in the calibration of the by SBRC testing, this set of modifications has created the sphere, such as the alignment of the sphere, the radiance anticipated improvements to the performance of the sen- standard, and the transfer instrument. sor. However, as described above, the reworked SeaWiFS The current understanding of the uncertainties in the radiometer also meets the original BTR specifications. calibration of the GSFC sphere is reported by Walker et al. (1991). Two of the authors specialize in radiometric 17. ABSOLUTE ACCURACY calibrations at NIST, and the third author is the principal investigator for the GSFC sphere. In the abstract, 17.1 Requirement Walker et al. (1991) states: "Recent measurements performed at NIST and NASA Goddard Space Flight Center An absolute radiometric accuracy of 5% (1 a) shall be have demonstrated that the uncertainty of sphere-source achieved at the typical spectral radiance levels. At spectral radiance measurements can be improved from the present 19

Original page 22

PrelaunchAcceptanceReportfortheSeaWiFSRadiometer Table 22. Constantsusedin the calculationof thespecificationcountlimit. TheLcloud and Ltypica I values come from the performance specifications. The slopes (or sensitivities for each band) come from the SBRC calibration data. Band Lcloud Ltypical Lcloud Radiance 0.5% of No. [mW cm-2# m-lsr -1 ] [mW cm-2# m-lsr -1 ] Ltypical 60.0 9.10 66.2 8.41 68.2 6.56 65.6 5.64 65.2 4.57 53.8 2.46 43.0 1.61 34.0 1.09 5-10% level to a 1-2% level." This is a general statement about the technique for calibrating sphere sources. in Recent discussions indicate that error propagation in However, it does not describe the actual uncertainty radiances from the GSFC sphere. Table 23. This table shows the SeaWiFS BTR characteristics. Column 2 gives the value of 0.5% of Ltypical, in counts, for each band. The instrument response must settle below this level of residual counts after scanning a bright target. Column 3 shows the number of pixels after the slit required for the instrument to settle to the level of the specified residual counts. Column 4 provides the number of pixels the instrument requires to settle to a level of zero residual counts. Band Residual Cnts. Pixels to Pixels to No. (0.5% Ltypical) Spec. Zero Cnts. 1 3.3 10 2 3.1 9 3 3.1 10 4 3.1 15 5 3.1 15 6 2.9 9 7 2.7 11 8 2.5 10 per Count Ltypica I [counts] 6.59 0.0137 3.3 7.87 0.0134 3.1 10.4 0.0105 3.1 11.6 0.00920 3.1 14.3 0.00746 3.1 21.9 0.00425 2.9 26.7 0.00301 2.7 31.2 0.00215 2.5 dent check method has not been determined. More work will have to be done if this problem is to be resolved." NASA's independent check method has not changed. The discrepancy between the NASA 5-10% error estimate and the 1-2% estimate in Walker et al. (1991) remains unresolved. This assessment has led to the conclusion that the uncertainty in the radiances from the GSFC sphere is in the range of 2-5%. This conclusion is nothing more than an educated guess. If the 2-5% uncertainty in the GSFC sphere is correct, and if the GSFC and SBRC spheres agree at the 2% level, then the estimate of the uncertainty in the SBRC sphere is 2-5%. This is also the review panel's estimate for the uncertainty in the calibration of SeaWiFS. 17.2.2 Accuracy Over a Scan Line These measurements were made at nadir. The specifications also call for a 5% accuracy at scan angles of +40 ° and -40 °. SBRC has made measurements to check the output of the instrument at scan angles 20 °, 40 °, and 58 ° on both sides of nadir. The results of the measurements at these angles give the instrument output relative to the output at nadir; this effect is called scan modulation. A fitted curve to the scan modulation gives a correction that is good to 0.5% at all scan angles. This correction has At the conclusion of the report, Walker et al. (1991) been incorporated into the data reduction procedures for states: "The normal stated uncertainty for the NASA cal- SeaWiFS. The absolute radiometric calibration of the inibration of their large-area sphere source is presently 5- strument remains within 5% over its entire scan range. 10%. The repeatability of the calibration from month to month is about 1 percent. The major contributors to 17.2.3 Accuracy Over Tilt Angles NASA's large overall uncertainty are the uncertainties asthe The specification calls for the radiometric calibration sociated with the standard lamp, the dimensions of and to be known at all tilt angles. For SeaWiFS, the entire apertures, the distance measurement, the alignment, of scanner tilts, including all components from the optical the lamp current of the sphere source lamps. The results our measurements confirmed the accuracy of the spectral irradiance method employed by NASA, and in the future of angle is not considered a factor in the radiometric accuracy it will permit them to state uncertainties in the range 1-2%. The reason for the difference in NASA's indepen- 2O inlet to the focal planes. The optical path through the instrument does not change as a function of tilt angle. Tilt of SeaWiFS.

Original page 23

R.A.Barnes,W.L.Barnes,W.E.Esaias,andC.R.McClain 18. RELATIVE ACCURACY of the sensitivity at each light level from the average sensitivity for the band was calculated. These differences should be less than 1%. Except for one outlier in each band, all 18.1 Requirement differences are observed to be less than 1%. For light level The design shall be capable of achieving an accuracy 7, each band shows the sensitivity to be between 1.4% and within 2% (1 a) relative to the sun. The calibrated data 1.9% higher than the other levels. The review panel feels shall be linear to within +1% over the full range of input that this discrepancy does not come from the instrument, signals. but from the calibration of the integrating sphere. It occurs at one light level, and it is present in all bands. 18.2 Compliance values for the bands in Table 24 have been calculated with- With the knowledge of this discrepancy, the average 18.2.1 Accuracy Relative to the Sun out the sensitivity for light level 7. The inclusion of this light level skews the results for the remaining measure- The instrument shall be capable of achieving an accuments. In addition, it is interesting to note that the linracy, within 2%, relative to the sun. SBRC has made field entity of the SeaWiFS measurements is sufficiently good measurements of the solar flux in March 1993 and in Octo reveal a 1.5-2.0% inconsistency in the calibration of the tober 1993. Atmospheric transmission measurements for integrating sphere. It should also be noted that the conthese two field measurements were made by the Universistency of the calibration of the sphere is an integral part sity of Arizona (Biggar et al. 1993). The differences in the of the linearity measurement. predicted on-orbit solar measurements from these two field tests averaged 1.5%. The greatest difference was 3.6% in band 8. These field measurements include uncertainties in the measurement of the atmospheric transmission that 18.2.2.2 November 1993 Measurements Although not reported in the SCADP, a set of linearity may amount to 2 or 3%. The field test results also show measurements was made by SBRC for the eight SeaWiFS a consistency in the SeaWiFS solar measurements at the bands in November 1993. The results of these measure- 2-3% level. However, with the error source of atmospheric ments are given in Table 25. The sensitivity of each band transmission, ground based measurements cannot confirm is presented in the table at three radiance levels, approxian accuracy of 2% in the SeaWiFS solar measurements. mating ntypical, two-thirds Ltypical, and one-third Ltypica I. They do, however, give some confidence that the instru- The least linear of the bands in Table 25 is band 2, with ment is capable of that accuracy. differences of 1.2% and 1.3% from the average sensitivity. As discussed above, the consistency in the calibration of 18.2.2 Linearity the sphere is a fundamental part of these results. The review panel feels that imprecision in the repeatability of 18.2.2.1 February 1993 Measurements the SBRC integrating sphere output has created a substantial portion of these differences. The repeatability of The measurements presented here were made in Februband 2 from the February 1993 measurements in Table 24 ary 1993. They were taken during the radiometric calibra- is significantly better than that in Table 25. As a result, tion of the instrument. The radiance measurements were the review panel concludes that the linearities of the eight made at eight light levels to cover the radiance ranges of SeaWiFS bands meet specification. the SeaWiFS bands. The results from those measurements The results in Table 24 also indicate that the response are given as the averages of 25 individual measurements for of the SeaWiFS bands are linear at low radiance levels. each band (see Table 24). Due to the spectral shape of the As shown for bands 1 and 2 in Table 24, the response of light from the integrating sphere, only SeaWiFS band 1, the instrument is linear for radiance levels corresponding at 412 nm, produced output that did not saturate at any to output from 15-50 counts. of the radiance levels. For SeaWiFS band 2, light level 1 Table 24 lists only the results for SeaWiFS bands 1-4. caused the output to saturate. For that band and that The results for the other four bands are consistent with light level, the band produced only its maximum digital those presented here, at least to the extend that they can output. For band 3, two of the radiance levels caused the be checked. For band 8 in particular, it is difficult to check band to saturate. This sequence continued up to band 8, linearity using one point only. which saturated for seven of the eight light levels. However, this final light level was sufficient for the radiometric 19.0 SYSTEM NOISE calibration of band 8. Table 24 gives the sensitivity for SeaWiFS bands 1-4 19.1 Requirement at each light level. It is the consistency of the instrument's sensitivity with light level that is required by the specifica- The SNR shall be determined for all bands at a suffitions. To check this sensitivity, the difference (in percent) cient number of spectral radiance levels between 0.2 Ltypical 21

Original page 24

PrelaunchAcceptanceReportfortheSeaWiFSRadiometer Thble 24. SeaWiFS radiometric measurements given for four SeaWiFS bands at eight light levels. The sensitivity for each band is calculated for each light level. The differences are the percent differences for the average sensitivity for each band. See text for details. Band Light Counts Offset Net No. Level Counts 1 1 842.77 21.76 821.01 2 435.54 21.76 413.78 3 257.50 21.76 235.74 4 194.59 21.76 172.83 5 111.85 21.76 90.09 6 54.03 21.76 32.27 7 39.60 21.76 17.84 8 36.98 21.76 15.22 Average sensitivity 0.014380 2 1 Band output saturates 2 710.43 19.46 690.97 3 413.26 19.46 393.80 4 308.19 19.46 288.73 5 170.34 19.46 150.88 6 73.63 19.46 54.17 7 49.35 19.46 29.89 8 44.90 19.46 25.44 Average sensitivity 0.013814 3 1 Band output saturates 2 Band output saturates 3 872.75 21.05 851.70 4 646.69 21.05 625.64 5 349.98 21.05 328.93 6 139.07 21.05 118.02 7 86.97 21.05 65.92 8 75.92 21.05 54.87 Average sensitivity 0.010874 4 1 Band output saturates 2 Band output saturates 3 Band output saturates 4 897.99 20.80 877.19 5 482.88 20.80 462.08 6 186.47 20.80 165.67 7 113.04 20.80 92.24 8 97.02 20.80 76.22 Average sensitivity 0.009361 Radiance Sensitivity Difference [mWcm-2#m-lsr -1] [radiance/count] [%] 11.80 0.014373 -0.1 5.97 0.014428 0.3 3.40 0.014423 0.3 2.49 0.014407 0.2 1.29 0.014319 -0.4 0.46 0.014255 -0.9 0.26 0.014574 1.4 0.22 0.014455 0.5 9.56 0.013836 0.2 5.45 0.013840 0.2 3.99 0.013819 0.0 2.08 0.013786 -0.2 0.75 0.013845 0.2 0.42 0.014052 1.7 0.35 0.013758 -0.4 9.28 0.010896 0.2 6.80 0.010869 -0.0 3.56 0.010823 -0.5 1.28 0.010846 -0.3 0.73 0.011074 1.8 0.60 0.010935 0.6 8.19 0.009337 -0.3 4.30 0.009306 -0.6 1.55 0.009356 -0.1 0.88 0.009540 1.9 0.72 0.009446 0.9 and 0.9 L,_ to characterize the signal dependence of the over short intervals of time, have caused a small (but sigsystem noise. 19.2 Compliance The SNRs for the eight SeaWiFS bands have been meanificant) decrease in the SNRs from these measurements. SBRC has created a model of their instrument noise, based on dark noise measurements and on the electronic design of the instrument. These model-based SNRs are slightly higher than the measured values at Ltypical, which is in sured near Ltypical. These measurements were made using agreement with the review panel's understanding of the the SBRC integrating sphere. Uncertainties in the output measurements. of the sphere, both over the area of the output aperture and The SeaWiFS SNR model gives the calculated results 22

Original page 25

R.A.Barnes,W.L.Barnes,W.E.Esaias,andC.R.McClain "l_able 25. SeaWiFS radiometric measurements given for all bands at three light levels. The differences are the percent differences from the average sensitivity for each band. These measurements were made in November 1993. Band Counts Offset Net Radiance Sensitivity Difference No. Counts [mW cm-2# m- l sr- 1] [radiance/count] [%] 1 695.60 20.84 674.76 516.70 20.84 495.86 268.60 20.84 247.76 Average sensitivity 0.013768 2 701.30 18.00 683.30 491.60 18.00 473.60 258.30 18.00 240.30 Average sensitivity 0.013532 3 709.00 20.45 688.55 503.80 20.45 483.35 248.50 20.45 228.05 Average sensitivity 0.010522 4 699.50 20.12 679.38 480.30 20.12 460.18 248.60 20.12 228.48 Average sensitivity 0.009220 5 703.30 22.00 681.30 527.60 22.00 505.60 281.40 22.00 259.40 Average sensitivity 0.007474 6 590.60 23.10 567.50 450.90 23.10 427.80 195.70 23.10 172.60 Average sensitivity 0.004259 7 706.50 22.73 683.77 347.10 22.73 324.37 209.60 22.73 186.87 Average sensitivity 0.002985 8 526.40 20.14 506.26 365.50 20.14 345.36 210.30 20.14 190.16 Average sensitivity 0.002131 9.246 0.013703 -0.5 6.817 0.013748 -0.1 3.432 0.013852 0.6 9.122 0.013350 -1.3 6.485 0.013693 1.2 3.257 0.013554 0.2 7.216 0.010480 -0.4 5.076 0.010502 -0.2 2.414 0.010585 0.6 6.212 0.009144 -0.8 4.223 0.009177 -0.5 2.134 0.009340 1.3 5.073 0.007446 -0.4 3.791 0.007498 0.3 1.940 0.007479 0.1 2.427 0.004277 0.4 1.835 0.004289 0.7 0.727 0.004212 -1.1 2.057 0.003008 0.8 0.966 0.002978 -0.2 0.555 0.002970 -0.5 1.075 0.002123 -0.4 0.734 0.002125 -0.3 0.408 0.002146 0.7 from the SBRC generated values for instrument noise. The the counts and from their associated SNRs in the SCADP SBRC results give the SNRs for the eight SeaWiFS bands data. The counts of noise for one-quarter Ltypical in bands at three radiance levels: Ltypical, one-half Ltypica h and one- 3 and 4 in the SBRC data were smaller than the accepted quarter Ltypical- The Ltypical radiance levels are listed in value for digitization noise (0.289 count). As a result, it Barnes and Holmes (1993). They are also given as part was concluded that digitization noise had not been incorof Table 26. The prelaunch calibration coefficients used to porated into the SBRC data. Digitization noise is included convert the counts from the instrument into the measured in the SeaWiFS noise model presented in Table 26. The radiance and vice versa are found in Table 6 of Barnes et addition of digitization noise to the model lowers the SNRs al. (1994). in the model's results. However, the model SNRs remain The model presented here gives the noise from each slightly higher than the measured results. band, in counts, as a function of the number of counts Noise values from the calibration data show a strong in the measurement by that band. The noise values that linear dependence with the measured counts. Based on form the basis for the model have been calculated from this linear dependence, a noise value for each band has 23

Original page 26

PrelaunchAcceptanceReportfortheSeaWiFSRadiometer Table 26. SNRscalculatedfromthe SeaWiFSnoisemodel.Thenoisemodelis a linearfunctionof thecounts fromeachSeaWiFSband.Thismodelisgivenforthestandardgain(gain1)andthe standard detector configuration (TDI 4) for the SeaWiFS ocean measurements. noise. The noise values in the model incorporate digitization Ltypical Ltypical Noise Intercept Noise Slope Noiset SNR Band No. [mW cm-2/z m- Isr-I] [counts] [counts] [counts/count] [counts] at Ltypica I 1 9.10 638.4 2 8.41 618.2 3 6.56 613.0 4 5.64 612.2 5 4.57 607.1 6 2.46 584.9 7 1.61 531.6 8 1.09 513.5 Noise at Ltypica[ been calculated for zero radiance. The noise model for Sea- WiFS band 7 is shown in Fig. 8. The linear noise model for the eight SeaWiFS bands (intercepts and slopes) are given in Table 26. For the review panel, the SNRs in Table 26 give the best prelaunch estimates of the SNRs for the instrument at the Ltypica I levels. As part of the SCADP data, there is a complete scan line of data from the third outdoor field test on 1 November 1993. A description of the components of a solar scan 0.420 0.0003528 0.645 990 0.372 0.0003141 0.566 1,091 0.348 0.0002875 0.524 1,170 0.352 0.0002928 0.531 1,152 0.375 0.0003178 0.568 1,069 0.510 0.0004083 0.749 781 0.424 0.0003674 0.619 859 0.494 0.0004144 0.707 726 ment to meet the specifications for that band. However, the measured results for band 5 in the SCADP data are also substantially lower than those from previous laboratory measurements by SBRC. The review panel feels that the value of 1,069 from Table 26 givesa more proper value forthe SNR forband 5. This value,along with those for the other seven bands, remains a prelaunch estimate. An extensiveseriesof on-orbitmeasurements (McClain et al. 1992 and Woodward et al.1993) willbe used to obtain line are given in Woodward et al. (1993). For three of an improved set of SNRs. A summary of the prelaunch the SeaWiFS bands, the solar diffuser measurements are made using gain 1. The results for these bands from the field test are given in Table 27. They are based on the average counts from 25 consecutive measurements across the diffuser. The noise values in Table 27 represent one standard deviation about each average. The linear model in Table 26 is used to convert those noise counts into the noise values at Ltypical. The SNRs for bands 3, 4, and 5 in Table 27 compare favorably with those in Table 26. Table 27. Measured SNRs from a SeaWiFS solar scan. For bands 3, 4, and 5, the solar diffuser measurements are made using gain 1. The noise model in Table 26 is used to convert the noise values from the measured count levels to those at Ltypical. These SNRs compare favorably with those in Table 26. Band Average Noise Noiset SNRt No. Counts [counts] [counts] 3 282.0 0.400 0.495 1,238 4 331.2 0.431 0.513 1,193 5 421.5 0.571 0.630 963 At Ltypical For the review panel, the results from the solar scan indicate that there has been a problem with SBRC SNR measurements for band 5 (Section 10), the cause of which is undetermined. The SBRC measurements for the calibration data book (SNR of 690 for band 5) show the instru- 24 SNR model for SeaWiFS bands 1-4 is presented in Table 28. A similarsummary for bands 5-8 is presented in Table 29. 20. POINTING KNOWLEDGE Pointing knowledge is a system level requirement. It includes knowledge of the nadir vector and scan plane of the instrument relative to the spacecraft; of the location of the spacecraft on orbit; and of the yaw, pitch, and roll angles of the spacecraft relative to the Earth. 20.1 Requirement The contractor shall provide in the downlinked data stream, data describing the spacecraft attitude and location and sensor pointing angles required for calculation of the location (in latitude and longitude) of each ocean IFOV to within one IFOV at all scan and tilt angles. 20.2 Compliance SBRC has provided OSC with the pointing coordinates for the SeaWiFS radiometer, information also found in the SCADP. These coordinates are given with respect to a set of transfer mirrors mounted on the radiometer. These values, from the instrument, are only part of the information required for pointing knowledge, since OSC must transfer these coordinates into their own system for the spacecraft

Original page 27

R.A.Barnes,W.L.Barnes,W.E.Esaias,andC.R.McClain 08 SeoWiFS - Band 7 Modeled Noise 0.7 06 0.5 v •_ 0.4 O Z -_0,3 Q; O 0.2 0.1 Q 0.0 IIIllllllllllllllll|lllllllllIlllllllll | 200 400 600 800 Signal (counts) 1200 SeoWiFS - Bond 7 Signal-to-Noise Ratio b 200 400 600 800 Signal (counts) Fig. 8. Results of the noise model for SeaWiFS band 7. The input values for the model come from the SBRC calibration data book. The top figure shows the noise model for SeaWiFS. The curve gives the results of the noise model. The symbols give the input noise values at Ltypical, one-half Ltypical, and one-quarter ntypica I. These three values include digitization noise. The bottom figure displays the calculated SNRs for SeaWiFS. The curve gives the calculated results from the model. The symbols give the input values from SBRC, with digitization noise added. 25

Original page 28

PrelaunchAcceptance Report for the SeaWiFS Radiometer T_ble 28. SNR model for SeaWiFS bands 1-4. The counts at each reference level are calculated using Table 6 of Barnes et al. (1994). The noise values are calculated using Table 20. Radiance Band 1 Band 2 Band 3 Band 4 [mW cm-2# m- lsr- 1] Counts SNR Counts 11.30 793 1,133 11.10 779 1,121 10.90 765 1,109 10.70 751 1,096 787 10.50 737 1,083 772 10.30 723 1,071 757 10.10 709 1,058 742 9.90 695 1,044 728 9.70 680 1,031 713 9.50 666 1,017 698 9.30 652 1,003 684 9.10 638 989 669 8.90 624 975 654 8.70 610 961 640 8.50 596 946 625 8.30 582 931 610 8.10 568 916 595 7.90 554 900 581 7.70 540 885 566 7.50 526 869 551 7.30 512 853 537 7.10 498 836 522 6.90 484 819 507 6.70 470 802 493 6.50 456 785 478 6.30 442 767 463 6.10 428 749 448 5.90 414 731 434 5.70 400 713 419 5.50 386 694 404 5.30 372 675 390 5.10 358 655 375 4.90 344 635 360 4.70 330 615 346 4.50 316 594 331 4.30 302 573 316 4.10 288 552 301 3.90 274 530 287 3.70 260 507 272 3.50 246 485 257 3.30 232 461 243 3.10 217 438 228 2.90 203 414 213 2.70 189 389 198 2.50 175 364 184 2.30 161 338 169 2.10 147 312 154 26 SNR Counts SNR Counts SNR 1,271 1,256 1,242 1,227 1,212 1,196 1,181 1,165 1,149 1,133 1,116 1,100 1,083 776 1,358 1,065 757 1,338 1,048 738 1,318 1,030 719 1,297 1,011 701 1,275 993 682 1,254 974 663 1,231 771 1,334 955 645 1,209 749 1,311 935 626 1,186 727 1,287 915 607 1,162 706 1,263 895 589 1,138 684 1,238 874 570 1,114 662 1,213 853 551 1,088 640 1,187 832 533 1,063 619 1,160 810 514 1,037 597 1,133 788 495 1,010 575 1,105 766 477 983 554 1,077 742 458 955 532 1,048 719 439 926 510 1,018 695 420 897 488 987 671 402 867 467 955 646 383 836 445 923 621 364 805 423 889 595 346 773 402 855 568 327 740 380 820 541 308 706 358 784 514 290 672 336 747 486 271 636 315 709 457 252 600 293 669 428 234 563 271 629 398 215 524 250 587 367 196 485 228 544

Original page 29

R.A.Barnes,W.L.Barnes,W.E.Esaias,andC.R. McClain Table 28. (cont.) SNR model for SeaWiFS bands 1-4. The counts at each reference level are calculated using Table 6 of Barnes et al. (1994). The noise values are calculated using Table 20. Radiance Band 1 Band 2 Band 3 Band 4 [mW cm-2p m- lsr-*} Counts SNR Counts 1.90 133 285 140 1.70 119 258 125 1.50 105 230 110 1.30 91 202 96 1.10 77 173 81 0.90 63 143 66 0.70 49 112 51 0.50 35 81 37 0.30 21 49 22 0.10 7 17 7 Table 29. SNR model for SeaWiFS bands 5-8. The counts SNR Counts SNR Counts SNR 336 178 445 206 500 304 159 403 185 454 271 140 361 163 407 238 121 317 141 359 203 103 272 119 309 168 84 226 98 257 133 65 178 76 203 96 47 129 54 148 58 28 79 33 90 20 9 27 11 31 at each reference level are calculated using Table 6 of Barnes et al. (1994). The noise values are calculated using Table 20. Radiance Band 5 Band 6 Band 7 Band 8 [mW cm-_# m- lsr- 1] Counts SNR Counts 5.75 764 1,237 5.65 751 1,223 5.55 737 1,210 5.45 724 1,197 5.35 711 1,183 5.25 697 1,169 5.15 684 1,155 5.05 671 1,141 4.95 658 1,126 4.85 644 1,111 4.75 631 1,096 4.65 618 1,081 4.55 604 1,066 4.45 591 1,050 4.35 578 1,034 4.25 565 1,018 4.15 551 1,002 4.05 538 985 3.95 525 969 3.85 511 951 3.75 498 934 3.65 485 916 3.55 472 899 3.45 458 880 3.35 445 862 3.25 432 843 3.15 418 824 749 3.05 405 804 725 2.95 392 785 701 2.85 379 764 678 2.75 365 744 654 2.65 352 723 630 SNR Counts SNR Counts SNR 918 90O 881 861 842 821 27

Original page 30

PrelaunchAcceptanceReportfortheSeaWiFSRadiometer Table 29. (cont.) SNR model for SeaWiFS bands 5-8. The counts at each reference level are calculated using Table 6 of Barnes et al. (1994). The noise values Lre calculated using Table 20. Radiance Band5 Band 6 Band 7 Band 8 [mW cm-2_ m-lsr -1 ] Counts SNR Counts SNR Counts SNR Counts SNR 2.55 339 702 606 800 2.45 325 680 582 779 2.35 312 658 559 757 2.25 299 636 535 734 743 1,066 2.15 286 613 511 711 710 1,037 2.05 272 590 487 687 677 1,006 1.95 259 566 464 663 644 975 1.85 246 542 440 638 611 942 1.75 232 518 416 612 578 908 1.65 219 493 392 585 545 873 1.55 369 558 512 836 730 917 206 468 1.45 193 442 345 530 479 798 683 879 1.35 179 415 321 501 446 758 636 840 166 388 297 471 413 717 589 798 1.25 153 361 273 440 380 674 542 754 1.15 1.05 250 408 347 629 495 708 139 333 0.95 126 304 226 375 314 582 448 659 0.85 113 275 202 341 281 532 400 607 0.75 I00 245 178 306 248 481 353 552 0.65 86 215 155 270 215 427 306 493 0.55 131 232 182 370 259 431 73 183 0.45 60 152 107 193 149 310 212 364 0.35 46 119 0.25 33 86 0.15 20 52 0.05 7 18 Sea- at equal radiance levels, but separated by any time period to provide on-orbit pointing vectors for SeaWiFS and Veri- up to two weeks. This includes the effects of perturbations Star. Pointing knowledge is a system level function. fication of pointing knowledge will be made using measure- 83 153 116 248 165 293 59 III 83 182 118 217 36 68 50 112 71 135 12 23 17 38 24 47 at the orbital period. This stability requirement shall also a sufficient be met for short-term temperature excursions that may ments of land targets on orbit. Verification of Review be expected to occur during sunlit portions of the orbit. set of downlinked data must wait for the Pre-Ship of the SeaStar spacecraft. 21. STABILITY &: REPEATABILITY Data from lunar views, corrected for secular changes in lunar radiance exitance, collected on several consecutive orbits while the moon is near full phase, shall be provided with sufficient frequency to assess short term and long term Bias errors will be removed from the data during ground stability. processing in order to improve radiometric accuracy. To accomplish this, the sensor data must be stable over time, 21.2 Compliance as defined below. 21.1 Short-Term Stability Requirement Due to the short period between the completion of instrument modification and the Pre-Ship Review, it was not possible to test this specification before launch. The Short-term stability applies to time intervals less than requirements will require examination of measurements on re- orbit. However, the radiometric calibration equations for two weeks. This stability also applies to radiometric data. the SeaWiFS bands (Barnes et al. 1994) contain factors, sponses corrected on the ground using calibration shall such as the temperature dependences for their radiometric The mean radiometric response of each spectral band, mea- sensitivities, which will be applied on orbit. Lunar meanot differ by more than +1% from another response operating surements (Woodward et al. 1993) are also planned for surement made while viewing the same source 28

Original page 31

R.A. Barnes, W.L. Barnes, W.E. Esaias, and C.R. McClain each measurement opportunity. These measurements are eliminated the period of time for extended testing by the planned for twice each month, as the moon reaches 96% manufacturer. However, the anecdotal information in Secof full before and after each full phase. The correction tion 21.4 indicates that the performance of the instrument for temperature dependence will also be applied to lunar is at the level required by this specification. measurements. 22. IN-FLIGHT DATA 21.3 Long-Term Stability Requirement 22.1 Requirement Long-term stability applies to time intervals between two weeks and 5 years. The mean corrected and calibrated Data for calibration and stability monitoring shall be radiometric response of each spectral band shall not change obtained from direct lunar views when the moon is greater by more than =t=2% over these time intervals. Compliance than 80% full phase, and either an onboard stabilized will be demonstrated by an estimate based upon short- source viewed by all optical elements or a solar diffuser. term tests plus analysis. The sources shall fill the optical aperture of the sensor. These data shall be obtained for all channels with a SNR 21.4 Compliance no less than 10% of the SNR values specified in Section 10, and shall measure changes in gain or throughput of the op- As shown in Table 1, the SeaWiFS instrument was comtical, focal plane, and electronic subsystems, using either pleted (after stray light modifications) on 22 November onboard, lunar, or solar sources. In-flight radiometric char- 1993. The post modification Pre-Ship Review was held acterization, i.e., output digital value versus input spectral ten days later on 2 December, which was an insufficient radiance, shall be made with sufficient accuracy to assure interval of time to test long-term stability. However, meathat the calibration and stability requirements delineated surements by the instrument before its modification to in this specification are achieved. ameliorate the effects of stray light indicated radiometric stability from the instrument at the 1% level over a few months. In addition, solar measurements during "field tests" of the instrument (Section 18) have shown a consis- 22.1.1 Lunar Calibration Provision shall be made to use the moon at near full tency in the instrument's measurements at the 2-3% level phase as a target source for monitoring stability. The lowfrom March to November 1993. These tests suggest, but est (least sensitive) gain shall accommodate direct viewing do not guarantee, long term stability on orbit. Long-term at near full lunar phase without saturation in any band. stability during the SeaWiFS mission will be monitored, and instrument changes will be corrected, through a series of lunar, solar, and ground based measurements (McClain et al. 1992 and Woodward et al. 1993) diffuser characteristics shall be provided which, when combined with data from other calibration systems, will be ad- 22.1.2 Solar Diffuser If a solar illuminated diffuser is selected, data on the 21.5 Band-to-Band Stability Requirement equate to maintain knowledge of the calibration and stability of the radiometric data to within stated specifications The relative amplitude stability between all pairs of throughout the five-year mission lifetime. spectral bands shall be better than =t=0.5% measured at full-scale, and +1% at half-scale. Each band shall be exposed to a source, and the mean calibrated responses determined. To compare outputs between bands, the ratio of 22.1.3 Internal Source In-flight data on characteristics of onboard sources, to the means shall be calculated for each band with respect show performance within the specifications, are required to a common band. In addition, ratios shall be calculated (should that approach be taken). for selected pairs of bands, which will be used in common retrieval algorithms. These ratios shall remain constant, 22.2 Compliance within ±0.5_ at full-scale and -i-l_ at half scale, over times separated by any interval up to two weeks. to make both lunar and solar diffuser measurements. There are no internal sources within the instrument. As discussed 21.6 Compliance in Section 15.2, there are gains for each SeaWiFS band The SeaWiFS radiometer has been specifically designed As discussed in Sections 21.2 and 21.4 above, the time specifically set for the expected on-orbit solar and lunar period between the completion of the SeaWiFS instrument radiances. Also, as discussed extensively in the SeaWiFS and the Pre-Ship Review was less than two weeks. This Technical Report Series (McClain et al. 1992, Woodward short time period was instituted in an effort to help ensure et al. 1993, and Barnes et al. 1994) and Biggar et al. 1993, the launch of SeaWiFS and SeaStar at the earliest possible lunar and solar diffuser measurements form a fundamental date. The practical requirements for an early launch have part of SeaWiFS on-orbit calibrations. 29

Original page 32

Prelaunch Acceptance Report for the SeaWiFS Radiometer 23. SUMMARY This technical memorandum contains only a portion and SNR Signal-to-NoiseRatio the information from the prelaunch characterization calibration of the SeaWiFS radiometer. The SCADP much more extensive. However, as discussed in this acceptance report, the requirements of the SeaWiFS specifications have determined the design of the instrument. fun- bl(k) Input data for polarizationcalculationsfor SeaWiFS Compliance with these specifications has defined the damental operation of SeaWiFS. SBRC Santa Barbara Research Center SCADP SeaWiFS Calibrationand Acceptance Data Package of SeaWiFS Sea-viewing Wide Field-of-viewSensor is TDI Time Delay and Integration SYMBOLS band 1. Input data forpolarizationcalculationsforSeaWiFS This report summarizes prelaunch analyses that have b7(k) been made by the review group, who are the co-authors this technical memorandum. The review group concludes that, in the period between the delivery and the launch exceeds h(k) Residual values without the calculated sinusoidalrethe instrument, the SeaWiFS radiometer meets or band 7. of G Gain factor. of G, Gain factor at gain setting n. all applicable specifications. Within the restriction that sponse. the instrument has not yet flown, SeaWiFS is found to Recorded maximum instrument output in response be acceptable. However, the complete set of the informa- to linearlypolarized light. tion necessary for the acceptance, or rejection, of SeaWiFS exten- to linearlypolarized light. and SeaStar is not yet available. There must be an sive analysis of the on-orbit operational characteristics SeaWiFS before a final judgement about the acceptability of the ocean color data set obtained by SeaWiFS can informa- bit. made. The completion of this work will require lmin Recorded minimum instrument output in response of Lcioud Maximum radiance from reflectedlightoffof clouds. Lm_ Maximum saturation radiance. be Ltypical Expected radiance from the ocean measured on ortion from 60 days of on-orbit operation by the satellite and n Gain setting. its instrument. GLOSSARY A/D Anaiog-to-Digital BTR Bright Target Recovery DC Direct Current FWHM FhJll-Width at Half-Maximum GAC Global Area Coverage GSFC Goddard Space Flight Center HN (Polaroid) Not an acronym; a linear sheet polarizer used to check the polarization sensitivity of bands 7 and 8. HR (Polaroid) Not an acronym; a linear sheet polarizer used to check the polarization sensitivity of bands 1-6. IFOV Instantaneous Field-Of-View IR Infrared. LAC Local Area Coverage MTF Modulation Transfer Function NASA National Aeronautics and Space Administration NEdL Noise Equivalent Differential Spectral Radiance NIST National Institute of Standards and Technology OCDM Ocean Color Data Mission OSC Orbital Sciences Corporation 3O PF Polarizationfactor. S, Initialdetector signal. Sn Detector signal with gain. x Abscissa or longitudinal coordinate, or the pixel number within a scan linedepending on usage. y Ordinate or meridional coordinate. z Mantissa coordinate. a One standard deviation. REFERENCES Barnes, R.A., and A. Holmes, 1993: Overview of the SeaWiFS Ocean Sensor. Proc. SPIE, 1,939, 224-232. Barnes, R.A., A. Holmes, W.L. Barnes, W.E. Esaias, and C.R. McClain, 1994: The SeaWiFS Prelaunch Radiometric Calibration and Spectral Characterization, NASA Tech. Memo. 104566, Vol. 23 S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, (in press). Biggar, S.F., P. Slater, K. Thome, A. Holmes, and R. Barnes, 1993: Preflight solar-based calibration of SeaWiFS, Proc. SPIE, 1,939, 233-242. McClain, C.R., W.E. Esaias, W.L. Barnes, B. Guenther, D. Endres, S.B. Hooker, B. Mitchell, and R.A Barnes, 1992: SeaWiFS Calibration and Validation Plan, NASA Tech. Memo. 104566, Vol. 3, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 41 pp.

Original page 33

R.A.Barnes,W.L.Barnes,W.E.Esaias,andC.R.McClain Walker,J.H., C. Cromer, and J. McLean, 1991: A technique Vol. 9 for improving the calibration of large-area sphere sources. Gregg, W.W., F.C. Chen, A.L. Mezaache, J.D. Chen, J.A. Pro(:. SPIE, 1,493, 224-230. Woodward, R.H., R.A. Barnes, C.R. McClain, W.E. Esaias, W.L. Barnes, and A.T. Mecherikunnel, 1993: Modeling of the SeaWiFS Solar and Lunar Observations. NASA Tech. Memo. 104566, Vol. 10, S.B. Hooker and E.R. Fire- Vol. 10 stone, Eds., NASA Goddard Space Flight Center, Green- Woodward, R.H., R.A. Barnes, C.R. McClaln, W.E. Esaias, belt, Maryland, 26 pp. THE SEAWIFS TECHNICAL REPORT SERIES VoI. 1 Hooker, S.B., W.E. Esaias, G.C. Feldman, W.W. Gregg, and VoL 11 Whiting, 1993: The Simulated SeaWiFS Data Set, Version 1. NASA Tech. Memo. 104566, Vol. 9, S.B. Hooker, E.R. Firestone, and A.W. Indest, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 17pp. W.L. Barnes, and A.T. Mecherikunnel, 1993: Modeling of the SeaWiFS Solar and Lunar Observations. NASA Tech. Memo. 104566, VoL 10, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 26 pp. C.R. McClain, 1992: An Overview of SeaWiFS and Ocean Part, F.S., C.M. Hoisington, W.W. Gregg, and P.L. Coronaclo, Color. NASA Tech. Memo. 104566, VoL 1, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 24 pp., plus color plates. Vol. 2 Gregg, W.W., 1992: Analysis of Orbit Selection for SeaWiFS: Ascending vs. Descending Node. NASA Tech. Memo. Vol. 12 1993: Analysis of Selected Orbit Propagation Models for the SeaWiFS Mission. NASA Teeh. Memo. 104566, Vol. 11, S.B. Hooker, E.R. Firestone, and A.W. Indest, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 16 pp. 104566, VoL 2, S.B. Hooker and E.R. Firestone, Eds., Firestone, E.R., and S.B. Hooker, 1993: SeaWiFS Technical Re- NASA Goddard Space Flight Center, Greenbelt, Maryland, 16 pp. Vol. 3 McClain, C.R., W.E. Esaias, W. Barnes, B. Guenther, D. Endres, S. Hooker, G. Mitchell, and R. Barnes, 1992: Cal- Vol. 13 ibration and Validation Plan for SeaWiFS. NASA Tech. McClain, C.R., K.R. Arrigo, J. Comiso, R. Fraser, M. Darzi, Memo. 104566, Vol. 3, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 41 pp. Vol. 4 McClain, C.R., E. Yeh, and G. Fu, 1992: An Analysis of GAC Sampling Algorithms: A Case Study. NASA Tech. Memo. 104566, Vol. 4, S.B. Hooker and E.R. Firestone, Eds., Vol. 14 port Series Summary Index: Volumes 1-11. NASA Tech. Memo. 104566, Vol. 12, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 28 pp. J.K. Firestone, B. Schieber, E-n. Yeh, and C.W. Sullivan, 1994: Case Studies for SeaWiFS Calibration and Validation, Part 1. NASA Tech. Memo. 104566, Vol. 13, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 52pp., plus color plates. NASA Goddard Space Flight Center, Greenbelt, Mary- Mueller, J.L., 1993: The First SeaWiFS Intercalibration Roundland, 22 pp., plus color plates. Vol. 5 Mueller, J.L., and R.W. Austin, 1992: Ocean Optics Protocols for SeaWiFS Validation. NASA Tech. Memo. 104566, Vol. 5, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard VoL 15 Robin Experiment, SIRREX-1, July 1992. NASA Tech. Memo. 104566, Vol. 14, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 60 pp. Space Flight Center, Greenbelt, Maryland, 43 pp. Gregg, W.W., F.S. Patt, and R.H. Woodward, 1994: The Sim- VoL 6 Firestone, E.R., and S.B. Hooker, 1992: SeaWiFS Technical Report Series Summary Index: Volumes 1-5. NASA Tech. Memo. 104566, Vol. 6, S.B. Hooker and E.R. Firestone, Vol. 16 Eels., NASA Goddard Space Flight Center, Greenbelt, ulated SeaWiFS Data Set, Version 2. NASA Tech. Memo. 104566, Vol. 15, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 42 pp., plus color plates. Maryland, 9 pp. Mueller, J.L., B.C. Johnson, C.L. Cromer, J.W. Cooper, J.T. Vol. 7 Darzi, M., 1992: Cloud Screening for Polar Orbiting Visible and IR Satellite Sensors. NASA Tech. Memo. 104566, Vol. 7, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 7pp. Vol. 17 Vol. 8 Abbott, M.R., O.B. Brown, H.R. Gordon, K.L. Carder, R.E. Hooker, S.B., W.E. Esaias, and L.A. Rexrode, 1993: Proceedings of the First SeaWiFS Science Team Meeting. NASA Tech. Memo. 104566, Vol. 8, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 61 pp. McLean, S.B. Hooker, and T.L. Westphal, 1994: The Second SeaWiFS Intercalibration Round-Robin Experiment, SIRREX-2, June 1993. NASA Tech. Memo. 104566, VoL 16, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 121 pp. Evans, F.E. Muller-Karger, and W.E. Esaias, 1994: Ocean Color in the 21st Century: A Strategy for a 20-Year Time Series. NASA Tech. Memo. 104566, Vol. 17, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 20 pp. 31

Original page 34

PrelaunchAcceptanceReportfortheSeaWiFSRadiometer Vol. 18 yol. 21. Firestone,E.R., and S.B. Hooker, 1994: SeaWiFS Technical Re- Acker, J.G., 1994: The Heritage of SeaWiFS: A Retrospecport Series Summary Index: Volumes 1-17. NASA Tech. Memo. 104566, Vol. 18, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, (in press). Vol. 1.__9 McClain, C.R., R.S. Fraser, J.T. McLean, M. Darzi, J.K. Firestone, F.S. Patt, B.D. Schieber, R.H. Woodward, E-n. Yeh, S. Mattoo, S.F. Biggar, P.N. Slater, K.J. Thome, A.W. Holmes, R.A. Barnes, and K.J. Voss, 1994: Case Studies for SeaWiFS Calibration and Validation, Part 2. NASA Tech. Memo. 104566, Vol. 19, S.B. Hooker, E.R. Firestone, and J.G. Acker, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 73 pp. Vol. 2.____0 Hooker, S.B., C.R. McClain, J.K. Firestone, T.L. Westphal, E-n. Yeh, and Y. Ge, 1994: The SeaWiFS Bio-Optical Archive and Storage System (SeaBASS), Part 1. NASA Tech. Memo. 10J566, Vol. 20, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 40 pp. 32 tive on the CZCS NIMBUS Experiment Team (NET) Program. NASA Tech. Memo. 104566, Vol. 21. S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, (in press). yo. 22 Barnes, R.A., W.L. Barnes, W.E. Esaias, and C.R. McClain, 1994: Prelaunch Acceptance Report for the SeaWiFS Radiometer. NASA Tech. Memo. 10_566, Vol. 22, S.B. Hooker, E.R. Firestone, and J.G. Acker, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 32 pp.

Original page 35

Original page 36

Form Approved REPORT DOCUMENTATION PAGE OMB No. 0704-0188 Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of infon'nation. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington, VA 2220_-43C_, and to the Office of Mana_mant and Budget, Paperwork Reduction Project 10704-01881, Washington, DC 20503. 1. AGENCY USE ONLY (Leave blank) 2. REPORT DATE September 1994 4o TITLE AND SUBTITLE SeaWiFS Technical Report Series SeaWiFS Radiometer 970.2 Volume 22--Prelaunch Acceptance Report for the 6. AUTHOR(S) 3. REPORT TYPE AND DATES COVERED Technical Memorandum 5. FUNDING NUMBERS Robert A. Barnes, William L. Barnes, Wayne E. Esaias and Charles R. McClain Series Editors: Stanford B. Hooker and Elaine R. Firestone Technical Editor. James G. Acker ADDRESS(ES) 8, PERFORMING ORGANIZATION 7. PERFORMING ORGANIZATION NAME(S) AND Laboratory for Hydrospheric Processes Goddard Space Flight Center Greenbelt, Maryland 20771 REPORT NUMBER 94B00126 AGENCY NAME(S) AND ADDRESSOES) 10. SPONSORING/MONITORING 9. SPONSORING/MONITORING National Aeronautics and Space Administration Washington, D.C. 20546-0001 11. SUPPLEMENTARY NOTES Robert A. Barnes: ManTech, Inc., Wallops Island, Virginia; Maryland; and James G. Acker: Hughes STX, Lanham, Maryland 12a. DISTRIBUTION/AVAILABlUTY STATEMENT Unclassified-Unlimited Subject Category 48 Report is available from the NASA Center for AeroSpace Landing Road, Linthicum Heights, MD 21090; (301) 621-0390. 13. ABSTRACT 200 t_o'ds) AGENCY REPORT NUMBER TM- 104566, Vol. 22 Elaine R. Firestone: General Sciences Corporation, Laurel, 12b. DISTRIBUTION CODE Information, 800 Elkridge The final acceptance, or rejection, of the Sea-viewing Wide Field-of-view Sensor (SeaWiFS) will be determined by the instrument's on-orbit operation. There is, however, an extensive set of laboratory measurements describing the operating characteristics of the radiometer. Many of the requirements in the Ocean Color Data Mission (OCDM) specifications can be checked only by laboratory measurements. Here, the calibration review panel (composed of the authors of this technical memorandum) examines the laboratory characterization and calibration of SeaWiFS in the light of the OCDM performance specification. Overall, the performance of the SeaWiFS instrument meets or exceeds the requirements of the (-'DM contract in all but a few unimportant details. "Ihe detailed results of this examination are presented here by following the outline of the specifications, as found in the Contract. The results are presented in the form of requirements and compliance pairs. These results give conclusions on many, but not all, oftbe performance specifications. The acceptance by this panel of the performance of SeaWiFS must only be considered as an intermediate conclusion. The ultimate acceptance (or rejection) of the SeaWiFS data set will rely on the measurements made by the instrument on orbit. 14. SUBJECT TERMS SeaWiFS, Oceanography, Acceptance, Radiometer, Dark Level, Compliance, Polarization, Pointing Knowledge 15. NUMBER OF PAGES Fore-and-Aft Pointing Requirement, 32 16. PRICE CODE CLASSIRCATION 19. SECURITY CLASSIRCATION 20. UMITATION OF ABSTRACT 17. SECURITY CLASSlRCATION 18. SECURITY OF REPORT OF THIS PAGE Unclassified Unclassified NSN 7540-01-280-5500 OF ABSTRACT Unclassified Unlimited Standard Form 298 (Rev. 2-89) Preacrlbed by ANSI Std. 239-18, 298-102