Full report

Original page 1

NASA Technical Memorandum SeaWiFS Technical Report Stanford B. Hooker, Editor NASA Goddard Space Flight Center Greenbelt, Maryland Elaine R. Firestone, Technical Editor General Sciences Corporation Laurel, Maryland 104566, Vol. 41 Series Volume 41, Case Studies for SeaWiFS Calibration and Validation, Part 4 Eueng-nan Yeh Edward A. Early Robert A. Barnes B. Carol Johnson Michael Darzi National Institute of Standards and Technology General Sciences Corporation Gaithersburg, Maryland Laurel, Maryland Lakshmi Kumar James L. Mueller Hughes STX Charles C. Trees Landover, Maryland Center for Hydro.Optics and Remote Sensing/SDS U San Diego, California National Aeronautics and Space Administration Goddard Space Flight Center Greenbelt, Maryland 20771 1997

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

Case Studies for SeaWiFS Calibration and Validation, Part 4 PREFACE he scope of the Sea-viewing Wide Field-of-viewSensor (SeaWiFS) Calibrationand Validation Program encompasses a broad varietyof topics,as evidenced by the contentsof three previouscase studiesvolumes in the SeaWiFS TechnicalReport Series---Volumes13, 19, and 27. Each case studiesvolume contains several chapters discussingtopicsgermane to the Calibrationand ValidationProgram. Volume 41,the fourth collection of case studies,further demonstrates both the breadth and complexity of the issuesthat the Program must address,and provides furtherjustificationfora comprehensive calibrationand validationeffort. The chapters in this volume present discussions of the: a) Calibration and characterization of the GSFC sphere; b) Revised SeaWiFS algorithm for the diffuse attenuation coefficient K(490); c) Simplified out-of-band correction algorithm for SeaWiFS; and d) Stray light correction algorithm implementation. --C. R. McClain Greenbelt, Maryland June 1997

Original page 4

p • _ • , ........ i_o_1-rl i _ ........ j h •

Original page 5

F.,-n.Yeh, R.A. Barnes, M. Dm'zi,L. Kumar, E.A. Early,B.C. Johnson, J.L. MueUer, and C.C. Trees Table of Contents Prok_ ...oo........,...............o0.. ............... 0 ................................................ 1 1. Calibration sad Characterization of the GSFC Sphere .............................................. 3 1.1 Introduction ....................................................................................... 3 1.2 Aperture Uniformity ............................................................................... 4 1.3 SpectralRadiance ................................................................................11 I.S.I Experimental Procedures .........................................................................II 1.8.2 Rmult_ ........................................................................................... 13 1.4 AncillarySXR Measurements .....................................................................15 1.5 Concltmiom .......................................................................................17 2, Revised SeaWiFS Algorithm for the Diffuse Attenuation Coefficient K(490) ........................ 18 2.1 Introduction ...................................................................................... 18 2.2 Data sad Methods ................................................................................ 19 2.8 RmultJ ...........................................................................................19 2.4 Discussion ........................................................................................ 21 8. Simplified Out-of-Band Correction Algorithm for SeaWiFS ........................................ 22 3.1 Introduction ...................................................................................... 22 8.2 The out.bandsub Algorithm ...................................................................... 22 4. SeaWiFS Stray Light Correction Algorithm ........................................................ 24 4.1 Introduction ...................................................................................... 24 4.2 LAC Correction .................................................................................. 24 4.8 GAC Correction .................................................................................. 30 4.4 Stray Light Mssking .............................................................................. 30 GLOSSARY ............................................................................................... SYMBOLS ................................................................................................ REPUtBN(_BS ............................................................................................ 31 31 32 THE SEAWIFS TECHNICAL REPORT SERIES ............................................................ 33 iii

Original page 6

Original page 7

E-n.Yeh,R.A.Barnes,M. Darzi,L. Kumar,E.A.Early,B.C.Johnson,J.L.Mueller,andC.C.Trees ABSTRACT This document provides brief reports, or case studies, on a number of investigations sponsored by the Calibration and Validation Team (CVT) within the Sea-viewing Wide Field-of-view Sensor (SeaWiFS) Project. Chapter 1 describes the calibration and characterization of the GSFC sphere, which was used in the recent recalibration of the SeaWiFS instrument. Chapter 2 presents a revision of the diffuse attenuation coefficient, K(490), algorithm based on the SeaWiFS wavelengths. Chapter 3 provides an implementation scheme for an algorithm to remove out-of-band radiance when using a sensor calibration based on a finite width (truncated) spectral response function, e.g., between the 1% transmission points. Chapter 4 describes the implementation schemes for the stray light quality flag (local area coverage [LAC] and global area coverage [GAC]) and the LAC stray light correction. Prologue The purposes of the Sea-viewing Wide Field-of-view Sensor (SeaWiFS) Project is to obtain valid ocean color data of the world ocean for a five-year period, to process that data in conjunction with ancillary data to meaningful biological parameters, and to make that data readily available to researchers. The National Aeronautics and Space Administration (NASA) Goddard Space Flight Center (GSFC) will develop a data processing and archiving system in conjunction with the Earth Observing System Data and Information System (EOSDIS), which includes a ground receiving system. In addition, the Sea- WiFS Project will oversee a calibration and validation effort which is designed to ensure the integrity of the final products. The Calibration and Validation Team (CVT) has three main tasks: 1) Calibration and characterization of the SeaWiFS instrument; 2) Development and validation of the operational atmospheric correction algorithm; and 3) Development and validation of the derived product algorithms, such as chlorophyll a concentration. Some of this work will be done internally at GSFC, while the remainder will be done externally at other institutions. NASA and the Project place the highest priority on ensuring the accuracy of derived water-leaving radiances globally, and over the duration of the entire mission. If these criteria are met, the development of global and regional biogeochemical algorithms can proceed on many primarily of contributions related to atmospheric correction methodologies, ancillary data sets required for level-2 processing of Coastal Zone Color Scanner (CZCS) and Sea- WiFS data, laboratory techniques for instrument calibration relevant to calibration round-robins, and field observations designed for transferring the prelaunch calibration to orbit, and in interpreting the on-orbit lunar calibration data. The second case studies volume, Volume 19, contains chapters on atmospheric and glint corrections; solar-, lunar-, and integrating sphere optical measurements; data format considerations; and the use of ancillary data (including surface wind velocities) in SeaWiFS processing. Volume 27 is the third in the set of such volumes and contains chapters on measuring immersion coefficients, oxygen absorption, solar calibration experiments, ship shadow effects on radiance and irradiance measurements, and the definition of the SeaWiFS data day for level-3 data binning. Volume 41 is the fourth of the case studies volumes. A short synopsis of each chapter in this volume is given below. 1. Ca/ibration and Characterization of the GSFC Sphere A large integrating sphere source, which is owned and maintained by the Sensor Development and Calibration Branch at GSFC, was calibrated and characterized by the Optical Technology Division at the National Institute of Standards and Technology (NIST). This effort, in support of the GSFC SeaWiFS Project, is part of an interagency agreement between NASA and NIST. The spectral radiance was measured for four different lamp settings fronts. These various activities are discussed in detail in of the sphere source from 370-1,100nm every 10nm us- The SeaWiFS Calibration and Validation Plan (McClaln et al. 1992). Because many of the studies and other works undertaken with the Calibration and Validation Program are not extensive enough to require dedicated volumes of the SeaWiFS Technical Report Series, the CVT has decided to publish volumes composed of brief, but topically specific, chapters. Volume 13 is the first volume, and consists ing a NIST standard tungsten strip lamp and a prismgrating monochromator which was equipped with a silicon photodiode. The results are presented as a function of wavelength with a relative standard uncertainty of 0.29- 2.6% (k=l), depending on the measurement conditions. During these measurements, the SeaWiFS 2_ransfer Radiometer (SXR) was used to determine the spectral radiance of the sphere source at six fixed wavelengths. In

Original page 8

Case Studiesfor SeaWiFS Calibrationand Validation,Part 4 a separate study, the SXR was used to measure the spa- thiswavelength combination are significantlydifferent,at tial uniformity of the spectral radiance at the exit port the 95% confidencelevel,from those of the CZCS K(490) of the sphere source. The measurements were made at algorithmbased on the ratioLw(443)/Lw(550). It isrecthree wavelengths and at the same four lamp settings for ommended, therefore,that thecoeffÉcientsfrom the present the sphere. The radiance in the aperture was found to analysis,ratherthan those of the CZCS algorithm,be used be uniform to within 0.6-2.2% (peak-to-valley variability), depending on the measurement conditions. This study was forthe prelaunch SeaWiFS K(490) algorithm. motivated by the requirement that the sphere source per- 3. A Simplified Out-of-Band form post-environmental calibrations of the SeaWiFS in- Correction Algorithm for SeaWiFS strument, as well as to compare the various radiometric scales in use by the ocean color community which are all NIST traceable. The measurements in this report were made just before the fourth SeaWiFS Intercalibration Round-Robin Experiment (SIRREX-4). 2. Revised Sea WiFS Prelaunch Aigorithm for the Diffuse Attenuation Coefficient K(490) The SeaWiFS instrument willscan through a broad area of known radiance, and the measurement for each band may contain signalsfrom outside the desiredbandwidth. A calculationthat willquickly remove the out-ofband contamination ispresented in thischapter. 4. Sea WiFS Stray Light Correction Algorithm The algorithm relating the diffuse attenuation coeffi- SeaWiFS will scan through not only the dark ocean surcient K(490) at 490 nm to the ratio of water-leaving radi- face, but also the bright land, clouds, and ice objects. Beances Lw(443)/Lw(555) is evaluated through regression analysis of radiometric profiles from oceanographic cruises in the Arabian Sea, Sargasso Sea, the California Current System, the Gulf of California, and the North and South cause of stray light in the SeaWiFS instrument, light from these bright sources can contaminate ocean measurements several pixels away fTom a bright source. A mathematical formula that can be used to correct this contamination is Atlantic Oceans. The resulting algorithm coefficients for discussedin thissection.

Original page 9

E-n. Yeh, R.A. Barnes, M. Darzi, L. Kumar, E.A. Early, B.C. Johnson, J.L. Mueller, and C.C. Trees Chapter Calibration and Characterization EDWARD A. EARLY AND i of the GSFC Sphere B. CAROL JOHNSON National Institute of Standards and Technology Gaithersburg, ABSTRACT Maryland A large integrating sphere source, which is owned and maintained by the Sensor Development and Calibration Branch at GSFC, was calibrated and characterized by the Optical Technology Division at NIST. This effort, in support of the GSFC SeaWiFS Project, is part of an interagency agreement between NASA and NIST. The spectral radiance was measured for four different lamp settings of the sphere source from 370-1,100nm every 10 nm using a NIST standard tungsten strip lamp and a prism-grating monochromator which was equipped with a silicon photodiode. The results are presented as a function of wavelength with a relative standard uncertainty of 0.29-2.6% (k--l), depending on the measurement conditions. During these measurements, the SXR was used to determine the spectral radiance of the sphere source at six fixed wavelengths. In a separate study, the SXR was used to measure the spatial uniformity of the spectral radiance at the exit port of the sphere source. The measurements were made at three wavelengths and at the same four lamp settings for the sphere. The radiance in the aperture was found to be uniform to within 0.6-2.2% (peak-to-valley variability), depending on the measurement conditions. This study was motivated post-environmental calibrations of the SeaWiFS instrument, by the requirement that the sphere source perform as well as to compare the various radiometric scales in use by the ocean color community which are all NIST traceable. The measurements in this report were made just before SIRREX-4. 1.1 INTRODUCTION lute measurements. The work was performed in the Optical Technology Division of NIST as part of an interagency Broadband, uniform, large area sources can be used agreement with NASA. for absolute radiometric calibration of radiometric sensors if the relative spectral response of the sensor is known. exit aperture is 39.5cm (15.5in) in diameter. The sphere Integrating sphere sources like the NASA GSFC sphere, source was originally built for the National Oceanic and which can be shipped to various sites, play a critical role in Atmospheric Administration (NOAA) and was acquired The GSFC sphere is 1.07m (3.5 ft) in diameter; the the calibration of field equipment, intercomparison of spec- by GSFC in 1989; it is similar to that described in Hotral radiance scales, and even in the calibration of satellite vis and Knoll (1983). As part of the NOAA programs, sensors themselves. The GSFC sphere was used at four the sphere source was calibrated by Optronic Laborato- SIRREXs, where intercomparisons were made with various ties, Inc., and used to calibrate spectroradiometers that sources and radiometers. It was used in October 1992 at were flown on high altitude aircraft in a satellite underpass Hughes Santa Barbara Remote Sensing (SBRS) in a com- configuration over White Sands, New Mexico, in order to parison with the SBRS integrating sphere source that was calibrate the fifth Land Resources Satellite (LANDSATused for the initial calibration of the SeaWiFS instrument 5) Thematic Mapper (Smith et al. 1990), the Advanced shortly after it was built (MueUer 1993a). Very High Resolution Radiometer (AVHRR) on NOAA-9 The GSFC sphere is scheduled to be used for the post- (Smith et al. 1988 and Smith et al. 1989a), and the Visibleenvironmental test calibration of SeaWiFS. The applica- Infrared Spin-Scan Radiometer (VISSR) on the sixth Geotion motivated a careful study of the uniformity of the stationary Operational Environmental Satellite (GOES-6) spectral radiance across the exit aperture using the SXR, spacecraft (Smith el al. 1989b). a muitichannel filter radiometer (Johnson et al. 1997), and The GSFC sphere is configured with 16 45W quartz a measurement of the spectral radiance of the sphere source halogen lamps, which are mounted on the inside wall and from 370-1,100nm. A prism-grating monochromator and surround the exit aperture. Small diffuse screens block the a gas-filled tungsten strip lamp were used for these abso- direct optical path between each lamp and the exit aper- 3

Original page 10

Case Studies for SeaWiFS Calibration and Validation, Part 4 ture. The interior surface is coated with a paint containing barium sulfate. The source is cooled using a fan mounted SIRREX-2, the GSFC 746/ISIC and the SXR agreed to within 5% for measurements of the spectral radiance of on a baffled aperture on top of the sphere; air flows in the GSFC sphere. At SIRREX-3, the corresponding result the sphere through a filtered fan port at the top and then through a baffled aperture at the bottom. Four power was about 3% (Mueller et al. 1995). The measurements reported here were done in April supplies are used to operate the fan and the 16 lamps in 1995 just prior to SIRREX-4 (Johnson et al. 1996). They a constant current mode of operation--one power supply for each set of four lamps; the four lamps are connected in series. Various radiance levels are achieved by turning off lamps associated with a particular power supply, followed by turning off the supplies (e.g., to transition from 13 lamps to 12 lamps, the fourth power supply is turned off). The sphere was recoated by GSFC in April 1994. Prior characterizations and calibrations of the GSFC sphere were done at GSFC and at the SIRREX activities (Mueller 1993a, Mueller et al. 1994, Mueller et al. 1995, and Johnson et al. 1996). At GSFC, a NASA filter radiometer and the SXR were used to map the spectral radiance in the exit aperture (Mueller et al. 1994). GSFC also investigated the degree to which water vapor absorption affects the spectral radiance of the sphere using a single-grating monochromator, model 746 from Optronic Laboratoriest with foreoptics (McLean, pers. comm.). The were motivated by the opportunity to measure the response of the SeaWiFS instrument to the radiance from the GSFC sphere as part of post-environmental test calibrations at Orbital Sciences Corporation (OSC) (Barnes et al. 1994b). It was also a good opportunity to assign a spectral radiance scale directly to the GSFC sphere at NIST, so that the SXR and 746/ISIC measurements of the GSFC sphere at SIRREX-4 could be compared to a NIST radiance scale assigned to a gas-filled tungsten standard strip lamp. The GSFC sphere was operated at four radiance levels in this work by operating 16, 8, 4, and 1 lamp, which gave adequate overlap with the desired calibration levels for SeaWiFS (Barnes et al. 1994b). 1.2 APERTURE UNIFORMITY To calibrate a radiometer, a source with known rabulk of the measurements by GSFC, however, consists of diance (e.g., the exit aperture of a sphere) is aligned so determinations of the average spectral radiance of the exit aperture using a standard irradiance lamp and a doubleaperture method for the irradiance-to-radiance transfer (e.g., Walker et al. 1987b, or Appendix A in Johnson et that the entrance window of the radiometer is centered and coplanar with the source; the entrance pupil must be filled by the source as well. This part of the measurement equation (ignoring any spectral aspects) involves integraal. 1996). The collection optic for the monochromator is a ting the spectral radiance as a function of position and small integrating sphere with a known aperture; this configuration is termed the 746/ISIC and it denotes the Optronic Laboratories model 746 monochromator equipped with an integrating sphere irradiance collector (ISIC). At the annual SIRREX activities, the data were obtained on the sphere's temporal stability, radiance levels, and the degree to which Lambert's law is followed. At SIRREX-2, -3, and -4, information was obtained on the consistency of independent, NIST-traceable, spectral radiance scales by measuring the spectral radiance of the GSFC sphere using the 746/ISIC system and the SXR. The former is traceable to NIST via the irradiance scale assigned to an FEL 1,000W tungsten-halogen standard lamp; the latter is traceable to NIST via the radiance scale assigned to a small integrating sphere source. Both of the NIST scales are derived using the same facility and starting point for the scale realization (Walker et al. 1987a and Walker et al. 1987b). The main results indicated that the radiance of the GSFC sphere was variable across the exit aperture, with peak-to-valley differences of about 3% (Mueller et al. 1994) and that the variability was reduced to about 0.5% after recoating the sphere walls. At t Identification of commercial equipment to adequately specify the experimental problem, does not imply recommendation or endorsement by NASA or NIST, nor does it imply that for SXR on an x,y translation stage 100 cm from the exit aperthe equipment identified is necessarily the best available the purpose. direction over the source area and all relevant directions, which depend on the location and size of the entrance pupil of the radiometer (Kostkowski and Nicodemus 1978). If the spectral radiance is uniform across the aperture and independent of viewing direction (lambertian), the measurement equation simplifies to a product of the spectral radiance and a geometric factor that can be measured (or calculated) and associated with the radiometer. If the spectral radiance is nonuniform, the relative spatial response function of the radiometer must be measured. This relative spatial response corresponds to radiation from a point source at x,y when the radiometer is focused at 0,0. In turn, this quantity, the point-spread response (PSR), must be multiplied by the actual spatial distribution of the spectral radiance across the aperture. The product is then integrated, as described above, to estimate a correction factor for the source nonuniformity. In practice, it is more common to assign a value corresponding to a level of nonuniformity that will not affect the final results by more that some stated uncertainty, and then to measure the source in question to determine if it can be used in the calibration procedure. If the source does not meet the requirement, it is usually easier to improve the source than to calculate correction factors. The measurement procedure consisted of mounting the ture of the GSFC sphere and recording the signal at one

Original page 11

E-n. Yeh, R.A. Barnes, M. Darzi, L. Kumar, E.A. Early,B.C. Johnson, J.L.Mueller,and C.C. Trees 1.07 m _1 Fig. I. The locationof the 16 lamps in the GSFC sphere as observed when facingthe sphere aperture. The lamp identificationnumber isshown insidethe small circles,and the power supply number (1-4) is given around the exit aperture (inner-mostcircle)in slantedtype. measurement wavelength, or SXR channel, for a grid of having two lamps offin both power supply I and 2. GSFC points coveringthe exitaperture.A 50 x 50 cm 2 area was has never triedthisconfiguration,and sincethey could not measured with a step sizeof 5 cm, which was chosen be- supply a wiring diagram of the system with detailson how cause the target area viewed by the SXR at thisdistance shunt resistorsare used in place of a lamp when that lamp isabout 4.5cm and the sphere aperture is 39.5cm in di- isturned off,only the normal procedure was used. Itisnot ameter. A background signalwas recorded by coveringthe possibleto operate the sphere with a singlelamp and still objectivelensof the SXR with the lens cap. produce a symmetric illuminationfield;the resultsindicate Four radiance levelsof the sphere were measured, cor- thisconfigurationisthe leastuniform. Another disadvanresponding to having all16 lamps on, then 8, 4, and 1. tage of the electricalsystem forthe sphere isthat the fan For each lamp configuration,the sphere was mapped at isoperated on the power supply that operates lamps 1-4, three SXR channels (wavelengths). The current,drawn so power supply 1 must always be on. Since GSFC does by the lamp power suppliesand the voltage across each not advisethe operationof a power supply without having lamp, was recorded priorto each spatialscan. The spatial at leastone lamp on, lamps 1-4 get more use. scans and data acquisitionwere fullyautomated using a personal computer (PC), two Aerotech linearpositioning stageswith directcurrent(DC) servo motors, an Aerotech controller,and a Hewlett Packard (HP) 3457A digitalvoltmeter. The InstituteofElectricaland ElectronicEngineers (IEEE) 488 interfacebus was used to establishcommunications,and the stage positionand data acquisitionsoftware was written in Visual Basic. Figure I shows the lamp arrangement for the GSFC sphere and the power suppliesthatcontrolthe lamps. Since the lamps require20 minutes to stabilize,the most efficient measurement procedure was to startwith all16 lamps on, the uniformity studies. The choice of measurement waveand then to turn offthe selectedlamps or power supplies lengths for the SXR was determined by: for the other levels.The 16 lamp configurationresultsin symmetric illuminationof the sphere,since the lamps are evenly spaced every 22.5°. As can be seen from Fig.I,the eight-lamp configurationisalsosymmetric (power supplies I and 2),but the four-and one-lamp configurationsare not symmetric (sincelamps I-4 and lamp 1 on power supply 1 were operated forthese configurations). Itwould be possibleto operate the sphere in a symmetricmanner with lamps 1,2, 5,and 7,but thiswould require Table 1. Measurement configurations for the uniformity studies of the GSFC sphere. Lamps SXR Wavelengths [nm] 1-16 411.5 441.6 661.8 1-8 411.5 487.1 661.8 1-4 411.5 487.1 661.8 1 411.5 548.0 661.8 Table 1 summarizes the measurement configurations for 1) Considering the amount of time available for the study, 2) The spectral coverage required, 3) The fact that the SXR has a more ideal PSR at particular channels, and 4) The relationship between the sphere radiance and the spectral bands of the SeaWiFS instrument that are appropriate for that radiance. 5

Original page 12

CaseStudiesforSeaWiFSCalibrationandValidation,Part4 Table 2. Lamp parameters during uniformity measurements of the GSFC sphere. Lamp Lamp Hours Number 11 April 12 April [A] [V] 1 1.0 7.0 6.501 7.144 2 6.0 6.501 7.148 3 6.0 6.501 6.984 4 6.0 6.501 7.196 5 4.5 6.500 7.065 6 4.5 6.500 7.274 7 4.5 6.500 7.144 8 4.5 6.500 7.016 9 2.5 6.501 7.173 10 2.5 6.501 7.359 11 2.5 6.501 7.200 12 2.5 6.501 7.138 13 2.5 6.500 7.237 14 2.5 6.500 7.192 15 2.5 6.500 7.159 16 2.5 6.500 7.455 Lamp Current [A] and Voltage [V_ [A] [V] [A] IV] [A] IV] 6.501 7.138 6.501 7.136 6.500 7.134 6.501 7.142 6.501 7.141 6.501 6.977 6.501 6.976 6.501 7.187 6.501 7.186 6.499 7.057 6.499 7.270 6.499 7.140 6.499 7.001 The SXR was operated at an overall gain of unity and and changed by as much as 15mV when going from the the digital voltmeter was programmed to average over 10 16-1amp to the 8-1amp configuration. The repeat meapower line cycles, with the autozero mode enabled, and surements at the center of the scan indicate that the comthe number of samples set to 10. At each position, five bination of the GSFC sphere radiance and the SXR was readings of the digital voltmeter were recorded, and the stable to between 0.005-0.05%, except in the case of the five results along with the average and standard deviation were stored in an American Standard Code for Information Interchange (ASCII) file. measurement at 411.5 ran in the single lamp configuration, when this value was 0.14%. These variations correspond to the standard deviation of the measurements at the center The standard deviations for the measurements in the of the scan position. exit aperture corresponded to a relative standard uncer- The results are shown in Figs. 2-5 in the form of containty of approximately 0.003%. The procedures for each tour plots. The net average signal at each scan position measurement consisted of recording the lamp .settings in was normalized by the average of the net signal recorded a laboratory notebook, measuring the SXR background for the center scan position. Scan positions that corresignal, and entering this value in the data acquisition pro- sponded to a view of the edge of the sphere aperture were gram. Then, under computer control, the signal at the excluded. The 16-1amp configuration is shown in Fig. 2, center coordinates for the scan was measured, and a hor- the 8-1amp in Fig. 3, the 4-1amp in Fig. 4, and the 1-lamp izontal scan from left to right (facing the sphere) at the in Fig. 5. The minimum contour corresponds to a square lowest vertical position was made. The horizontal scans about 30 cm on a side. Within the square, the maximum were repeated for all the vertical positions, and two read- and the minimum value for the normalized net signals are ings at the center coordinates were made between each also indicated for each of the three wavelengths and the horizontal scan. The data files consist of a header section four radiance levels. All of the data indicate a local mincontaining the relevant information followed by the stage imum in the upper left-hand quadrant when facing the position and signal data. Each complete mapping took sphere, and there seems to be little dependence on waveabout 25 minutes. length for a given lamp configuration. The average current and voltages for the lamps dur- Table 3 summarizes all of the results in terms of the ing the three measurements for each radiance level of the peak-to-valley variability using the minimum and maxi- GSFC sphere are given in Table 2, as well as the number mum values. The 16-1amp and 8-1amp configurations show of hours that the lamps were operated for these tests (11 similar variability, between 0.6-0.9%. The variability of and 12 April 1995). For all of the measurements, the cur- the 4-1amp configuration is greater, between 1.1-1.2%, and rent for each set of four lamps was stable to within 1 mA, the pattern of variability is consistent with the lamp arwhich corresponded to the precision of the indicators on rangement (Fig. 1). The greatest variability is observed for the power supplies. The voltage for a particular lamp was the 1-1amp configuration, between 2.1-2.3%, and again the stable to within 1 mV during a given lamp configuration, 6 pattern is consistent with the lamp arrangement.

Original page 13

E-n. Yeh, R.A. Barnes, M. Darzi, L. Kumar, E.A. Early, B.C. Johnson, J.L. Mueller, and C.C. Trees 20 a) 15" I0- 5" 0" -5 -I0 -15 20 . o_ 15 _ b) 1.0027"r-'_ 1"0c2 T101 _ 0.99 o I I I 2O c) 15 10- 5- 0- -5-- 1. -10- -15 - -20 , , , -20 -15 -I0 -5 _2 '1 _':i 34 )4 I I I I 1.004 , , , , 0 5 I0 15 20 HorizontalPosition[cm] Fig. 2. Contour plot of the radiance in the exit aperture normalized to the average of the measurements at the center position of the GSFC sphere as measured with the SXR at a) 411.5 nm, b) 441.6 nm, and c) 661.8 nm with lamps 1-16 at 6.5 A. The values in the plots indicate the ratio of the radiance along the contour to the radiance at the center. The short lines perpendicular to the contours indicate that the radiance is decreasing in that direction.

Original page 14

Case Studies for SeaWiFS Calibration and Validation, Part 4 20 a) 15- 10- 5- 0 -5 -10 -15 20 b) 15 E _-. 10 o 5 r/3 o 0 ¢a, "--* -5 -10 0.984 > -15 I I 20 15- 10- 5- 0- -5- -10- 0.984 -15- -20-15-10-5 1.002 1.004 0.998 I I I I I 1.002 II 1.004 l 0.998 0 5 10 15 20 Horizontal Position [cm] Fig. 3. Contour plot of the radiance in the exit aperture normalized to the average of the measurements at the center position of the GSFC sphere as measured with the SXR at a) 411.5 nm, b) 487.1 nm, and e) 661.8 nm with lamps 1--8 at 6.5 A. The values in the plots indicate the ratio of the radiance along the contour to the radiance at the center. The short lines perpendicular to the contours indicate that the radiance is decreasing in that direction. 8

Original page 15

E-n. Yeh, R.A. Barnes, M. Darzi, L. Kumar, E.A. Early, B.C. Johnson, J.L. Mueller, and C.C. Trees 20 a) 15- I0- 5- 0- -5- -I0- -15- I 2O b) ,'_ 15- _ lO- ..=o5t i"J ,],lJJ o_ -5- > -15- 1.002 I I I I I 1.002 I I I I I I I 20 c) 15- I0- 5- 0- -5 0.! -I0 0.982 -15 -20 1.004 1.006 I I I I I I I -20 -15-10-5 Horizontal 0 5 10 15 2O Position [cm] Fig. 4. Contour plot of the radiance in the exit aperture normalized to the average of the measurements at the center position of the GSFC sphere as measured with the SXR at a) 411.5 nm, b) 487.1 nm, and e) 661.8 nm with lamps 1-4 at 6.5 A. The values in the plots indicate the ratio of the radiance along the contour to the radiance at the center. The short lines perpendicular to the contours indicate that the radiance is decreasing in that direction.

Original page 16

Case Studiesfor SeaWiFS Calibrationand Validation,Part 4 20 a) 15 10 5 0.! -15 2O I i ! b) 15o 10 = 0 5_ 0 O- "d "5o -10o > -15- 0.998 I I I 20 c) 15- 10- 5- 0- -5l° -10- -15 - -20 I I I ! ! I ! 1.012 .014 I I I I I 0.998 1.006 -----1.014 I I I I -20 -15 -10 -5 0 5 10 15 2O Horizontal Position [cm] Fig. 5. Contour plot of the radiance in the exit aperture normalized to the average of the measurements at the center position of the GSFC sphere as measured with the SXR at a) 411.5 nm, b) 548.0 nm, and c) 661.8 nm with lamp 1 at 6.5 A. The values in the plots indicate the ratio of the radiance along the contour to the radiance at the center. The short lines perpendicular to the contours indicate that the radiance is decreasing in that direction. 10

Original page 17

E-n. Yeh, R.A. Barnes, M. Darzi, L. Kumar, E.A. Early, B.C. Johnson, J.L. Mueller, and C.C. Trees Table 3. Results of the uniformity studies for the GSFC sphere presented as the peak-to-valley variability in }ercent. Lamps SXR Wavelengths [nm] On 411.5 441.6 487.1 548.0 661.8 1-16 0.76 0.61 0.84 1-8 0.67 0.78 0.87 1-4 1.10 1.14 1.24 1 2.27 2.14 2.15 1.3 SPECTRAL RADIANCE The spectral radiance at the center of the exit port of the GSFC integrating sphere source was determined at wavelengths from 380-1,100nm with 16, 8, 4, and 1 lamp operating. Signals as a function of wavelength from both a calibrated radiance source and from the sphere were measured, and the spectral radiance of the sphere was determined by the following procedure. Assuming S and S are the signals from sources with known and unknown spectral radiance, respectively, the simplest measurement equation for these signals is given by (1) length scan over a given range, in which the signal from the = LR and = LR (2) where L and L are the known and unknown spectral radiances of the sources, respectively, and R is the spectral radiance responsivity (SRR) of the detector system. Measuring the signal from a source with a known spectral radiance, therefore, determines the SRR of the detector, from (1), which is then used in (2) along with the measured signal from the source with the unknown radiance to determine the spectral radiance of the unknown source. of determined by viewing, with the detector system, the exit The effects of radiant flux from wavelengths outside the nominal spectral bandpass of the detector have been neglected in (1) and (2), as well as effects from a nonlinear response of the detector. To minimize these effects as much as possible, the known and unknown spectral radiances of the sources should be as similar as po6sible. For the measurements described in this effort, this was achieved by determining the spectral radiance of the sphere with 16 lamps operating from the known spectral radiance of a gas filament lamp. The spectral radiance of the sphere with 16 lamps operating was then used to determine the spectral radiance of the sphere with 8, 4, and 1 lamp operating. 1.3.1 Experimental Procedures 1.3.1.1 Equipment The GSFC integrating sphere source and its accompanying power supply were described in Sect. 1.1. The source of known spectral radiance was a gas-filled tungsten ribbon filament lamp with the designation 92-04. This lamp was calibrated for spectral radiance by the Facility for Automated Spectroradiometric Calibrations (FAS- CAL) at NIST. The current through the lamp was supplied by an HP 6032A power supply and monitored by an HP 3457A digital voltmeter which measured the voltage across a calibrated shunt resistor. The detector system consisted of imaging optics, a prism-grating double monochromator with adjustable slit widths, and a silicon photodiode with an adjustable gain. A metal mask with a 0.6 x 0.8 mm 2 hole in front of the entrance slit of the monochromator defined the imaged area of the source. Details of this system can be found in Mielenz et al. (1990). The voltage output of the silicon photodiode was measured by an HP 3457A digital voltmeter. The entire detector system was mounted in a box on a carriage which was translated horizontally to view either the filament lamp or the integrating sphere. The wavelength of the monochromator and the position of the carriage were under computer control; the computer recorded the monochromator wavelength as determined by an absolute encoder, and the voltage from the silicon photodiode was measured by the voltmeter. A measurement generally consisted of a wavephotodiode was recorded at discrete wavelengths within the range. 1.3.1.2 Auxiliary Calibrations Two calibrations were performed prior to measuring the spectral radiance of the integrating sphere. The first determined the actual wavelength setting of the monochromator in terms of the wavelength measured by the encoder, while the second determined the ratio of the gains of the silicon photodiode. The wavelength calibration of the monochromator was port of a small integrating sphere which had a line source at the entrance port. The line source was either a HeNe laser or a gas emission lamp. Wavelength scans were performed over ranges that included the line source wavelengths. The HeNe laser line at 632.816 nm was scanned with monochromator slit widths ranging from 0.125-1.000mm. From these scans, the bandwidth of the monochromator was determined from the difference in wavelengths at which the signals were half that of the maximum signal. The dispersion of the monochromator is approximately 4.25 nm mm- 1 at 632.8 nm and increases slightly with a decrease in slit width. The centroids of the HeNe laser line were also determined from these scans, and were all within 0.015 nm of each other; thus, the slit width has a minimal effect on the centroid of a line source. Gas emission lamps were chosen which had intense singlet lines, within the wavelength range 350-1,100nm, and were separated from other lines by at least 2 nm. Wavelength scans of the lines from At, Hg, Ne, and Xe lamps 11

Original page 18

Case Studies forSeaWiFS Calibrationand Validation,Part 4 were usually performed with 0.25 mm slit widths at 0.05 nm of detector system, as was the gas filament lamp. The center intervals over 4 nm ranges centered on the wavelengths the lines. A few of the scans of lines from the Ne lamp were performed with 0.125mm slit widths in order to resolve the lines that are close together in wavelength. The centroids of the emission lines, in terms of the encoder wavelength, were calculated from the scans, and the differences between the actual wavelengths of the lines and the centroid wavelengths as a function of centroid wavelength were fit with a third-order polynomial. The actual wavelength of the monochromator, therefore, is given by the encoder wavelength plus the fit of the difference. Using this sphere was placed the same distance--29.0 cm--from the of the cover of the exit aperture was marked previously with a cross, and a glass slide was placed over it. Both the rotation of the integrating sphere about a vertical axis and the height of the rear sphere frame were adjusted to reflect the laser beam back onto itself, thereby making the exit aperture perpendicular to the optic axis. The height of the integrating sphere and the position of the carriage were adjusted to center the exit aperture on the optic axis. These rotation and height adjustments were iterated several times to center the laser beam on the cross and reflect calibration, the standard uncertainty in the wavelength of it back off itself. The exit aperture was centered to within the monochromator was 0.2 nm. While the magnitude of the spectral radiance of the gas filament lamp was several times that of the sphere with 16 lamps operating, the spectral distribution was similar. Therefore, effects from stray light in the detector system were minimized, but different gains for the photodiode had to be used when measuring the two sources. The gain ratio is simply the ratio of the higher gain to the lower gain, and is equal to the ratio of the signals at those two gains. The gain ratio was determined by viewing the gas filament lamp, fixing the wavelength of the monochromator at 700 nm, and recording the signals from the silicon photodiode at consecutive gain settings. The slit width was adjusted so that a nearly full-scale signal was obtained at the higher gain setting. Average signals were determined 2mm, but the cover over the exit aperture was slightly bowed, so the optic axis was perpendicular to the exit aperture to within 2 ° . The SXR was placed on a tripod behind the carriage containing the detector system and oriented to view the center of the exit aperture of the integrating sphere. The distance from the exit aperture to the front plate of the SXR was 136 cm. 1.3.1.4 With 16 Lamps The gas filament lamp was operated at 41.6 A, the same current at which it was calibrated for spectral radiance at FASCAL, with an uncertainty of 0.2 mA. All 16 of the lamps of the integrating sphere were operated at 6.5 A. The four times at gains of 108, 109, and 101° with slit widths of slit width of the monochromator was chosen by setting the 0.3 mm and 1.2 ram. The gain ratio for gains of 101° and l0 s was 99.975751 ± 0.011240. 1.3.1.3 Alignment Proper alignment of both the gas filament lamp and the exit port of the integrating sphere with the optic axis of the detector system is critical for correctly determining the spectral radiance of the integrating sphere. The wavelength to 670 nm, viewing the gas lamp, and adjusting the slit width to yield a signal close to the full scale for that gain setting. This resulted in a slit width of 0.9 mm, corresponding to a bandpass of 4.2 nm at a wavelength of 632.8 nm. The voltmeter was configured to average over 100 power line cycles and to autozero. Wavelength scans of both the gas filament lamp and the integrating sphere were performed from 360-1,100 nm beam from a HeNe laser mounted behind the exit slit of at 10nm increments, and took 20 minutes to complete. the monochromator, and passing through the monochro- Three voltage signals were recorded at each wavelength, mator and the imaging optics, defined the optic axis for and gains of l0 s and 101° on the silicon photodiode were the detector system. The gas filament lamp was mounted on a bipost base attached to tilt and translation stages on an optical table. The lamp was operated at half its normal current for the purpose of optical alignment. Tilt and translation stages were adjusted so that the laser beam defining the optic used with the gas filament lamp and the integrating sphere, respectively. The gas filament lamp was scanned first, and then the carriage with the detector system was translated to view the integrating sphere. Another wavelength scan was performed, and the process was completed by translating the carriage back to the gas filament lamp for a final axis passed through the notch in the lamp filament, as wavelength scan. The radiance of the integrating sphere well as the tip of an arrow etched into the back of the lamp envelope. This oriented the filament perpendicular to, and centered vertically on, the optic axis. The lamp was then translated to center the laser beam horizontally on the filament. Finally, the lamp was translated along the optic axis to focus the image of the filament onto the mask, which is in front of the entrance slit of the monochromator. The integrating sphere was placed on the floor next to the optical table. The front of the exit aperture of the 12 was measured with the SXR before and after each wavelength scan with the detector system. 1.3.1.5 With 8, 4, and 1 Lamp The procedure for measuring the radiance of the integrating sphere with 8, 4, and 1 lamp was similar to that given above for 16 lamps. The slit width was increased to 3.0 mm (12 nm bandpass) in order to maximize the signals

Original page 19

E-n.Yeh,R.A.Barnes,M. Darzi,L. Kumar,E.A.Early, B.C. Johnson, J.L. MueUer, and C.C. Trees oper- preceding paragraph, and by the gain ratio determined in fromthe integratingspherewhenall 16lampswere wasfixedat Sect. 1.3.1.2. The spectral radiance at 760nm was obating,andthegainof thesiliconphotodiode per- viously low when compared to the spectral radiances at 101°. Wavelengthscansoftheintegratingspherewere andtheradi- neighboring wavelengths. This was due to the local miniformedwith 16,8,4, and1 lampoperating, scan. mum in the SRR at that wavelength, as shown in Fig. 6a. ancewasmeasuredbythe SXRbeforeandaftereach sig- Consequently, the spectral radiance from 660-860 nm, ex- After all the lampswereturnedoff, the background gain. cluding the value at 760 nm, was fit with a second-order nalofthesiliconphotodiodewasdeterminedateach × polynomial, and this fit was used to calculate the spectral This backgroundsignalwas-6.733x 10-5 V+ 0.846 x radiance at 760 nm. The resulting radiance, as a function 10-sv at a gainof l0s and6.0771x 10-4V±3.1762 10-a V at a gainof 10l°. of wavelength, is plotted in Fig. 6b. time for the lamps Sources of uncertainty in the spectral radiance of the in- The total accumulatedoperating measurementswas tegrating sphere arise from the gas filament lamp radiance of the integratingsphereduringthese 9 hoursfor lamp1, 8.5hoursforlamps2-4, 8 hoursfor and current, the detector system signals and wavelength, lamps5-8, and7.5hoursfor lamps9-16. The voltages acrossthe lampswerecomparableto thoselistedin Table2, with variationsin the voltageof eachlamponthe orderof severaltensof millivolts. 1.3.2 Results For each wavelength scan, the data were reduced to an average corrected signal as a function of wavelength. First, the wavelength of the monochromator, given by the encoder, was converted to the actual wavelength using the results from the calibration process (Sect. 1.3.1.2). Second, the three voltage measurements at each wavelength were averaged with the standard uncertainty given by the standard deviation of the mean. Third, the background signal of the photodiode for that gain was subtracted from the average signal, and the standard uncertainty of the background signal was propagated. The resulting signal as a function of wavelength was fit with a natural cubic spline to wavelengths from 380-1,100nm at 10 nm increments. The standard uncertainty of the fit signal was taken to be the standard uncertainty of the signal at the actual wavelength closest to the fit wavelength. 1.3.2.1 With 16 Lamps The spectral radiance of the gas filament lamp had been determined by FASCAL at 34 wavelengths from 225- 2,400 nm, and the standard uncertainty was given at seven of those wavelengths. Both the spectral radiance and the fit ties, from each of the sources detailed here in determinstandard uncertainty as a function of wavelength were with a natural cubic spline to the wavelength range 380- 1,100nm at 10 nm increments. The corrected, average signals from both scans of the gas filament lamp were averaged and the uncertainties of the signals were propagated. These final signals were divided by the spectral radiance to obtain the SRR of the detector system with slit widths of 0.9 mm and a gain of l0 s. The results are shown in Fig. 6a, where the SRR is plotted as a function of wavelength. The SRR has abrupt changes near 580, 780, and 980 nm, which are caused by the optics of the detector system. The spectral radiance of the integrating sphere with all 16 lamps operating was obtained by dividing the averaged, corrected, and fitted signal by the SRR, determined in the and the sphere alignment. The standard uncertainty in the spectral radiance of the gas filament lamp is given from the calibration performed by FASCAL. The 2 mA uncertainty in the current through the lamp yields a relative standard uncertainty in the spectral radiance dL/L given by L where ,_ is the wavelength in nanometers and dI is the current standard uncertainty in milliamps (Walker et al. 1987b). The standard uncertainty in the signal is the propagation of the uncertainties from the average signals, the background signal, and the gain ratio. The relative standard uncertainty in radiance due to an uncertainty in the wavelength is given by the following: dL dL dA T = d), L" (4) An analytical expression for the derivative of the radiance with respect to wavelength was obtained by differentiating the Wien radiation law. Fits of the spectral radiance of each source to the Wien radiation law were used to obtain the parameters needed for the analytic expression. The results of the SIRREX-3 activity (Mueller et al. 1995) show that the maximum change in radiance as a function of viewing angle is 1.2% at 5°; therefore, assuming a maximum uncertainty of 2 ° along either the horizontal or the vertical axis, the relative standard uncertainty is approximately 0.0020. The relative standard uncertaining the spectral radiance of the integrating sphere with 16 lamps operating, are given in Table 4 (at four selected wavelengths). Both the detector signal and the wavelength become significant sources of uncertainty at the shorter wavelengths. 1.3.2.2 With 8, 4_ and 1 Lamp The SRR of the detector system with the larger slit widths and higher gain used with 8, 4, and 1 lamp operating was obtained from the wavelength scan of the integrating sphere with 16 lamps operating. The average, corrected, fit signals from this scan were divided by the spectral radiance, calculated in the preceding section, to yield 13

Original page 20

Case StudiesforSeaWiFS Calibrationand Validation,Part4 0.008 a) E e= t.m 0.006 E 0 '7 0.004o_ < 0.002- .=. clef OI i l I I IIII Ilia llll 350 450 550 650 Wavelength [nm] ,._ 60- '_ 45" .+f ¢_ 30". 8 15- I I l l I I J I I i I I 750 850 950 1050 1150 I I I I IIII IIII w I w I IIIi IIII IIII IIII 35O 450 550 650 Wavelength [nm] 0.6" c) llllIlll lllIlllllllllIlllll IIIIIIIII .O 0.45- .g n- Q o 0.3- ¢= ¢= o-'-' ,,+0",+,'-,-"' +'+'+'+-'+-'-'-" "0 ¢g 0.15" 750 850 950 1050 1150 lllIIllllllllllllllllllllllllllllI_ Key.. D8Lamps 0 4 Lamps 1 Lamp " "-"'+.':' '-'++'¥.. .+d....+.+.+.+..+.+...+.+,+.,-,,;0 /VV_MA AAAAAAAA A/WWW wwww I I I I I I I I I I I l l I I I I I I I I i I I l 350 450 550 650 Wavelength [nm] 750 850 950 1050 1150 Fig. 6. Radiance analyses for the calibration of the GSFC sphere: a) SRR as a function of wavelength as determined from the gas filament lamp at a gain of 10s and a slit width of 0.gram; b) radiance as a function of wavelength at the center of the exit aperture of the GSFC sphere with 16 lamps operating; and c) the ratio of radiance of the integrating sphere with the indicated number of lamps operating to the radiance with 16 lamps operating as a function of wavelength. 14

Original page 21

E-n. Yeh, R.A. Barnes, M. Darzi, L. Kumar, E.A. Early, B.C. Johnson, J.L. Mueller, and C.C. Trees 1.03 o 1.02 t- = .__ lo i 1.01 ,.r ¢} i t- (2. T CO 1.00 0 C [] .- _ 0.99 X 0.98 O_ m 0.97 | w ........ ! w ! I 4_ 0 600 Wavelength [nml E] 1 Lamp 0 4 Lamps I] A 8 Lamps 0 16 Lamps | I 7 8_ Fig. 7. Spectral radiance comparison of the GSFC sphere calibration with the standard lamp and spectroradiometer compared to the SXR measurements that were made during the calibration procedure. The different symbols correspond to the lamp configuration; the relative uncertainties in the SXR measurements are indicated by the vertical bars for the 16 lamp configuration. the SRR. This responsivity has the same spectral shape as decreasing number of operating lamps, as expected. The that shown in Fig. 6a, although it is larger in magnitude because of the larger slit widths and higher gain. Also, this SRR was determined only once since no lamps were operating after the series of measurements, detailed in this section, were complete. Table 4. The relative standard uncertainties (in percent) from the indicated sources at the given wavelengths for the spectral radiance of the GSFC sphere with 16 lamps operating. Source of Wavele_h [nm] Uncertainty 400 600 800 1,000 Lamp Radiance 0.44 0.33 0.29 0.27 Lamp Current 0.02 0.02 0.01 0.01 Detector Signal 0.14 0.03 0.03 0.03 Detector Wavelength 0.26 0.08 0.02 0.00 Sphere Alignment 0.10 0.10 0.10 0.10 RSStTotal 0.54 0.35 0.31 0.29 t Root sum squared. The spectral radiance of the integrating sphere with 8, 4, and 1 lamp operating was obtained by dividing the average, corrected, fit signal from the wavelength scan with each lamp combination by the SRR. Again, while the shape of the spectral radiance for each lamp combination is similar to that shown in Fig. 6b, the magnitude decreases with ratio of the spectral radiance with a given number of lamps operating to the spectral radiance with 16 lamps operating, is given in Fig. 6c as a function of wavelength. The ratios are all less than the nominal values of 0.5, 0.25, and 0.0625 for 8, 4, and 1 lamp, respectively, but approach these values at longer wavelengths. The spectral radiance of the integrating sphere at each lamp combination, along with the relative standard uncertainty, is given in Table 5. The number of sources of uncertainty in the spectral radiance are considerably reduced from those of the preceding section. The uncertainties in the spectral radiance with 16 lamps operating are given by the results from the preceding section, and those from the detector signals are still present. However, the uncertainty in radiance from the current through the lamps of the integrating sphere is unknown, and the encoder wavelengths and sphere alignment are unchanged for all the wavelength scans, so there is no uncertainty associated with them. 1.4 Ancillary SXR Measurements In order to track the performance of the GSFC sphere during SIRREX-4, as well as the post-environmental Sea- WiFS calibration, the SXR was used to measure the spectral radiance of the GSFC sphere for the four radiance levels in between the measurements with the spectroradiometer. These measurements were compared by fitting 15

Original page 22

CaseStudiesforSeaWiFSCalibrationandValidation,Part4 Table 5. The radiance, L [#Wcm -2 nm -1 sr-1], and relative standard uncertainty, 6 [%], as a function of wavelength, A, of the GSFC sphere with the indicated number of lamps operating. A 16 Lamps 8 Lamps 4 Lamps 1 Lamp [nm] L ,_ L _ L _ L 380 1.943 0.61 0.926 0.68 0.457 0.74 0.113 1.51 390 2.449 0.57 1.170 0.61 0.579 0.64 0.145 1.13 400 3.047 0.54 1.459 0.56 0.723 0.58 0.182 0.90 410 3.753 0.51 1.801 0.53 0.894 0.54 0.226 0.75 420 4.536 0.50 2.179 0.51 1.082 0.52 0.275 0.66 430 5.357 0.49 2.576 0.50 1.280 0.50 0.326 0.60 440 6.239 0.47 3.005 0.49 1.494 0.49 0.380 0.56 450 7.181 0.46 3.462 0.46 1.722 0.46 0.439 0.52 460 8.188 0.45 3.952 0.45 1.967 0.45 0.502 0.49 470 9.251 0.44 4.470 0.44 2.226 0.44 0.568 0.47 480 10.36 0.43 5.014 0.44 2.497 0.44 0.637 0.47 490 11.52 0.42 5.580 0.42 2.781 0.42 0.709 0.44 500 12.73 0.41 6.170 0.42 3.075 0.42 0.784 0.43 510 13.96 0.40 6.772 0.41 3.377 0.41 0.861 0.42 520 15.22 0.40 7.392 0.40 3.686 0.40 0.940 0.41 530 16.51 0.39 8.023 0.39 4.002 0.39 1.021 0.40 540 17.80 0.38 8.659 0.39 4.321 0.39 1.102 0.40 550 19.12 0.38 9.306 0.38 4.645 0.38 1.184 0.39 560 20.43 0.37 9.955 0.37 4.970 0.37 1.267 0.38 570 21.76 0.37 10.61 0.37 5.297 0.37 1.350 0.38 580 23.05 0.36 11.24 0.37 5.615 0.37 1.431 0.37 590 24.34 0.36 11.88 0.36 5.936 0.36 1.514 0.36 600 25.61 0.35 12.51 0.35 6.250 0.35 1.593 0.36 610 26.86 0.35 13.13 0.35 6.559 0.35 1.672 0.36 620 28.10 0.34 13.74 0.34 6.865 0.34 1.750 0.35 A 16 Lamps 8 Lamps 4 Lamps 1 Lamp [nm] L _ L _ L 6 L 6 750 40.64 0.31 19.95 0.32 9.987 0.32 2.534 0.32 760 41.27 0.31 20.27 0.31 10.15 0.31 2.573 0.32 770 41.90 0.31 20.59 0.31 10.31 0.31 2.614 0.32 780 42.45 0.31 20.86 0.31 10.45 0.31 2.647 0.32 790 42.94 0.31 21.11 0.31 10.57 0.31 2.677 0.32 800 43.37 0.31 21.33 0.31 10.68 0.31 2.704 0.32 810 43.75 0.31 21.53 0.31 10.79 0.31 2.729 0.32 820 44.05 0.31 21.67 0.31 10.85 0.31 2.745 0.32 830 44.35 0.30 21.83 0.30 10.93 0.31 2.764 0.32 840 44.63 0.30 21.97 0.31 11.01 0.31 2.782 0.32 850 44.82 0.30 22.07 0.30 11.06 0.30 2.794 0.32 860 44.99 0.30 22.16 0.30 11.10 0.30 2.805 0.32 870 45.10 0.30 22.22 0.30 11.14 0.30 2.813 0.32 880 45.17 0.30 22.26 0.30 11.16 0.30 2.818 0.32 890 45.21 0.30 22.28 0.30 11.17 0.30 2.821 0.32 900 45.10 0.30 22.24 0.30 11.15 0.30 2.815 0.32 910 45.04 0.30 22.21 0.30 11.14 0.30 2.811 0.32 920 45.01 0.30 22.20 0.30 11.13 0.30 2.809 0.32 930 44.72 0.29 22.06 0.30 11.06 0.30 2.792 0.32 940 44.56 0.29 21.98 0.29 11.03 0.30 2.781 0.32 950 44.26 0.29 21.84 0.29 10.96 0.30 2.763 0.33 960 44.23 0.29 21.83 0.30 10.95 0.30 2.760 0.34 970 44.14 0.29 21.80 0.29 10.94 0.29 2.757 0.33 980 43.90 0.29 21.69 0.29 10.88 0.29 2.744 0.32 990 43.71 0.29 21.59 0.29 10.83 0.29 2.729 0.32 630 29.29 0.34 14.33 0.34 7.161 0.34 1.824 0.35 1,000 43.41 0.29 2.708 0.34 21.45 0.29 10.76 0.29 640 30.45 0.33 14.90 0.34 7.451 0.34 1.898 0.34 1,010 43.12 0.29 21.33 0.29 10.70 0.30 2.690 0.37 650 31.62 0.33 15.48 0.33 7.742 0.33 1.971 0.34 1,020 42.82 0.29 21.19 0.30 10.62 0.31 2.669 0.42 660 32.70 0.33 16.02 0.33 8.011 0.33 2.039 0.34 1,030 42.48 0.29 21.03 0.31 10.54 0.32 2.643 0.49 670 33.76 0.33 16.55 0.33 8.278 0.33 2.106 0.33 1,040 42.14 0.30 20.86 0.32 10.45 0.34 2.618 0.60 680 34.77 0.32 17.05 0.32 8.528 0.32 2.169 0.33 1,050 41.78 0.31 20.70 0.34 10.36 0.37 2.591 0.75 690 35.71 0.32 17.52 0.32 8.766 0.32 2.229 0.33 1,060 41.36 0.33 10.24 0.43 2.559 0.97 20.48 0.39 700 36.67 0.32 18.00 0.32 9.006 0.32 2.289 0.33 1,070 40.93 0.35 10.12 0.51 2.530 1.24 710 37.55 0.32 18.44 0.32 9.227 0.32 2.344 0.32 20.25 0.45 1,080 40.48 0.40 20.04 0.53 10.00 0.62 2.504 1.58 720 38.35 0.32 18.84 0.32 9.430 0.32 2.395 0.32 1,090 40.04 0.46 2.486 2.01 19.83 0.64 9.883 0.77 730 39.16 0.31 19.24 0.32 9.632 0.32 2.446 0.32 1,100 39.62 0.55 740 39.93 0.31 19.62 0.31 9.824 0.31 2.493 0.32 16 19.65 0.81 9.759 0.97 2.483 2.58

Original page 23

F_:-n.Yeh,R.A.Barnes,M. Darzi,L. Kumar,E.A.Early,B.C.Johnson,J.L. Mueller,andC.C.Trees Table6. A comparisonofSXRspectralradiancemeasurements(Ls) to thevaluesmeasuredwith thespectroradiometerandthegas-filledstrip lampstandard(LG),bothgivenin unitsof#Wcm-2 nm-1 sr-1. A 16 Lamps 8 Lamps {nm] Ls LG Ls/LG Ls LG Ls/LG 411.5 3.853 3.869 0.9959 1.848 1.857 0.9951 441.6 6.317 6.387 0.9890 3.040 3.077 0.9881 487.1 11.22 11.18 1.003 5.422 5.413 1.002 548.0 19.08 18.85 1.012 9.269 9.173 1.010 661.8 32.81 32.89 0.9975 16.05 16.11 0.9959 774.8 42.36 42.21 1.003 20.82 20.74 1.003 the spectroradiometer results to a cubic spline and interpolating them to the effective wavelengths of the SXR. Correction was made for a size-of-source effect in the SXR. The results are given in Table 6 as spectral radiance, and the SXR values are divided by the spectroradiometer values. The agreement is within the combined uncertainties of the techniques; however, a degree of consistency in these ratios is evident, as shown in Fig. 7. This indicates a possible error in the absolute calibration because of systematic effects, such as the calibration coefficients for the SXR or the calibration data for the strip lamp. 1.5 CONCLUSIONS The peak-to-valley variability of the spectral radiance in the exit aperture of the GSFC sphere is between 0.6- 0.9% when the sphere is illuminated in a symmetric fashion. When only a single lamp is used, resulting in an asymmetric illumination geometry, these values increase by more than a factor of two. There appears to be no 4 Lamps 1 Lamp Ls LG Ls/LG Ls LG Ls/LG 0.9188 0.9220 0.9965 0.2349 0.2335 1.006 1.512 1.530 0.9886 0.3865 0.3895 0.9925 2.702 2.697 1.002 0.6899 0.6880 1.003 4.626 4.578 1.010 1.180 1.167 1.011 8.026 8.059 0.9958 2.040 2.051 0.9948 10.42 10.39 1.003 2.642 2.633 1.004 spectral dependence to the variability, and there is a consistent region of lower radiance when compared to the average of the central values. The spectral radiance of the sphere was determined at four different lamp settings using both a prism-grating monochromator from 370-1,100nm every 10 nm and the SXR at its six fixed wavelengths. The spectral radiance was a smoothly varying function of wavelength, with a peak of 45.21 #W cm -2 st- x nm- 1 at 890 nm with all 16 lamps operating. The magnitude of the spectral radiance scaled approximately with the number of lamps. The spectral radiances determined by the SXR agreed with those determined by the monochromator and the standard lamp within their combined standard uncertainties. ACKNOWLEDGMENTS The authors are grateful to Ambler Thompson and Robert Saunders of the Optical Technology Division at NIST for their advice and assistance, and to Charles Gibson who provided a calibrated strip lamp on short notice. This work was supported in part by NASA Interagency Agreement S-64096-E. 17

Original page 24

Case Studies for SeaWiFS Calibrationand Validation,Part 4 Chapter 2 Revised SeaWiFS Prelaunch Algorithm for the Diffuse Attenuation Coefficient K(490) JAMES L. MUELLER AND CHARLES C. TREES Center .for Hydro-Optics and Remote Sensing/SDSU San Diego, California ABSTRACT The algorithm relating the diffuse attenuation coefficient K(490) at 490 nm to the ratio of water-leaving radiances Lw (443)/Lw (555) is evaluated through regression analysis of radiometric profiles from oceanographic cruises in the Arabian Sea, Sargasso Sea, the California Current System, the Gulf of California, and the North and South Atlantic Oceans. The resulting algorithm coefficients for this wavelength combination are significantly different, at the 95% confidence level, from those of the CZCS K(490) algorithm based on the ratio Lw(443)/Lw(550). It is recommended, therefore, that the coefficients from the present analysis, rather than those of the CZCS algorithm, be used for the prelaunch SeaWiFS K(490) 2.1 INTRODUCTION The diffuseattenuation coefficientat 490 nm, K(490), was one of the standard ocean data products calculated from Nimbus-7 CZCS imagery. Austin and Petzold (1981) derived the Nimbus-7 CZCS algorithm relatingK(490), in per unit meters, to the ratio of water-leavingradiances Lw(443)/Lw(550), at wavelengths of 443 and 550nm, as "Lw(443)" -1'491m_1. K(490) = 0.022 ÷ 0.088 Lw(550) (5) Root mean square (rms) uncertainties in K(490) estimated from CZCS data are less than 20% (la) based on direct comparisons with in situ radiometric profiles (e.g., Mueller 1993b). The SeaWiFS Science Team formally recommended the as confidence limits of the coefficients of (6). Mueller (1995) Austin and Petzold 1981 algorithm, (5), be adopted the prelaunch K(490) algorithm for SeaWiFS data processing. This algorithm replaces the CZCS water-leaving radiances Lw(443) and Lw(550) with the SeaWiFS norma/ized water-leaving radiances LWN(443) and LwN(555). Some members of the SeaWiFS Science Team, however, questioned whether the shift from 550 to 555 nm might lead to significant systematic errors in SeaWiFS K(490) estimates if the coefficients in (5) are used. Mueller (1995) compared K(490) and LwN(443)/LwN(555) derived from 45 radiometric profiles made during recent cruises using instruments configured with SeaWiFS wavelengths. The profiles that made up this small sample were contributed by: 18 algorithm. 1) C. Trees (San Diego State University [SDSU] Center for Hydro-Optics and Remote Sensing [CHOIRS]), 16 profiles from the Arabian Sea; 2) G. Mitchell (University of California San Diego [UCSD] Scripps Institution of Oceanography [SIO]), 18 profiles from the California Current System; and 3) D. Siegel (University of California Santa Barbara [UCSB]), 11 profiles from the Sargasso Sea. The logarithmic regression analysis of these data resulted in the algorithm LwN(443) /((490) = 0.022 + 0.0984 _ m -1, (6) with a standard error of 0.018m -i . The coefficients of (5) (Austin and Petzold 1981) fell well within the 90% concluded, therefore, that there was insufficient evidence to reject the hypothesis that the two algorithms are equivalent. During the past year, 242 data triplets [of K(490), LWN (443), and LWN (555)] were accrued using radiometric profile data from: a) Three additional cruises in the Arabian Sea (C. Trees); b) One cruise in the Gulf of California (E. Valdez, H. Maske, and their colleagues, of Centro de Investigaci6n Cient_fica y de Educaci6n Superior de Ensenada, [CICESE] Baja, California, and J. Mueller of CHORS); and

Original page 25

E-n.Yeh,R.A.Barnes,M. Darzi,L. Kumar,E.A.Early,B.C.Johnson,J.L. Mueller,andC.C.Trees c) ThefirstAtlanticMeridionalTransect(AMT-1)of where L=(0-,A) and E,(A) are taken from the analyzed the North and South Atlantic Oceans in September/October 1995 (G. Moore of Plymouth Marine Laboratory [PML], United Kingdom [UK]). These cruises were carried out, in part, under the Sea- WiFS Science Team investigations of the participants. The regression analysis of this larger sample yields algorithm coefficients with much narrower (95%) confidence limits than those associated with (6) (Mueller 1995). In contrast to that earlier result, the new algorithm is significantly different from (5) (Austin and Petzold 1981). 2.2 DATA AND METHODS Profiles of spectral downwelling irradiance Ed(z, A) and upwelling radiance Lu(z, A) at the SeaWiFS wavelengths (within 2 nm) were obtained from cruises in the: 1) Arabian Sea (C. Trees); 2) California Current System (G. Mitchell); 3) Sargasso Sea near Bermuda and the Joint Global Ocean Flux Study (JGOFS) Bermuda Atlantic Time Series (BATS) site (D. Siegel); 4) Central Gulf of California (J. Mueller, E. Valdez and H. Maske); and profiles, tf ($) is the upward Fresnel transmittance through the air-sea interface for radiance, and F0(A) is the mean extraterrestrial solar spectral flux. Within less than 1%, tl(443)/tl(555 ) - 1.0, so that from (9) the ratio of normalized water-leaving radiances at 443 and 555 nm may be expressed as LwN( 443) L(O-, 443)Fo(443)Es(555) ___ LWN(555) Lu(0-, 555)F0(555)Es(443) (10) From the work of Neckel and Labs (1984), F0(443) = 198.5 and Fo(555) = 190.0 #W cm -: nm -1. The logarithmic regression model relating K(490) to the ratio LWN(443)/LwN(555) may be expressed as [LwN(443)] In[K(490) - 0.022] = In(A) +Bln LLwN(555)], (11) where using (5), A = 0.088 and B = -1.491, or from (6), A = 0.0984 and B = -1.403. The attenuation coefficient for pure water, Kw (490) = 0.022 m-1, is the minimum possible value for K(490). The present sample of 242 profile measurements were used to determine best-fit coefficients for (11) using simple linear regression. 5) North and South Atlantic Oceans, AMT-1 cruise of 2.3 RESULTS September-October 1995 (G. Moore). Table 7 lists the instruments used in conjunction with the profiles taken, in addition to where the instruments were calibrated. Each calibrated set of profiles of Ed(z, A), L_(z, A), and Es(A), as measured with each instrument's deck radiometer, was analyzed using the integral least-squares solution of Mueller (1991) to determine profiles of Ka(z, A) and KL(Z, A), and the vertical attenuation coefficients for Ea(z, A) and L_(z, A), respectively. The AMT-1 profiles, however, were analyzed by G. Moore at PML. From these profiles, the authors extracted irradiance above the sea surface, Es(A), and upwelling radiances just below the sea surface, Lu(O-, A), for wavelengths (A) of 443 and 555 nm. This was, together with remote sensing K(490), calculated by averaging Kd(z, 490) over the first optical attenuation length, i.e., 1 f,_O0 (7) K(490) = _9o ]J0 Ka(z, 490)dz, where z9o is the depth and where E(490, z) - fo _° Kd(z,ago)az e-1 E(490,0) = e = = 0.37. (8) Normalized water-leaving radiances LWN(A) for A -_443 and 555 nm are calculated as F0( ) (9) LwN(A) = Lu(O-,A)tf(A) Data from the sample described above were combined, using (10) and (11), to assemble a regression sample of size N - 242 The linear least-squares fit to this data is In [K(490) - 0.022] = -2.30261 [LwN(443)] (12) - 1.29966 In L J' with the multiple correlation coefficient, R 2 = 0.90, and a residual standard deviation of 0.293 (in log space). The scatter of these data are illustrated in Fig. 8, together with the best-fit regression line defined by (12). Also shown in Fig. 8, as a dashed line, is (5) (Austin and Petzold 1981). Equation (12) may be transformed to the form of (5) as "LwN( 443)" -1.30o m -I. (13) K(490) = 0.022 + 0.1000 LWN(555) The measured data pairs are compared to (13) in Fig. 9 [LwN(443)/LwN(555) on a logarithmic scale compared to linear K(490)]. The linear residual standard deviation of/<(490) (standard error of the estimate) associated with (13) is 1/2 SKi ----- (14) O.O17m -1, 19

Original page 26

Case Studies for SeaWiFS Calibration and Validation, Part 4 O I 0 [] I I 2 InILwN(44a)/L SSS)l • Trees (Arab-l) A Si©gcl (Sarg. Sea) Tr_s (Arab-p6) •= MucllcffMask¢ (GoCal) [3 Mitchell (CCS) X Trees (Arab-p2) O Trees (Arab-p7) ...... Austin-Pctzold Moore Q_I.& S. Atl_mticr AMT-I Fig. 8. Scattergram comparing K(490) and the normalized ratio LWN(443)/LwN(555) from the Arabian Sea, California Current System (CCS), Sargasso Sea, Gulf of California (GoCal), and the North and South Atlantic (AMT-1). The solid line is the least-squares fitto the data, and the dashed line illustratesthe CZCS K(490) algorithm of Austin and Petzold (1981). The resulting regression equation is In [K(490)- 0.022] = -2.30261 -1.29966 In [Lw_v(443)/Lwjv(555)] with R 2 = 0.90; and the standard error of the estimate isa = 0.017m -I. 0,4 0.35 0,3 O 'm 0.25 m 0.2 _ 0.15 0.1 0.05 0 I 0.1 ! I 10 lO0 Lv 443) SSS) • Trees (Arab-l) A Siegel (Sarg Sea) X Trees (Arab-p6) _- MucUu/Maske(OoCal) Fig. 9. Linear K(490) versus a logarithmic scaling gression fit (solid curve) from Fig. 8. 2O [] Mitchell (CCS) X Trce$ (Arab-p2) 0 Trees (Arab-p7) 0 Moorc (N & S Aflantic, AMT-l ) LWN(443)/LwN(555) display of the data and re-

Original page 27

E-n. Yeh, R.A. Barnes, M. Darzi, L. Kumar, E.A. Early, B.C. Johnson, J.L. Mueller, and C.C. Trees Table 7. This table lists the instruments used, as well as where the instruments were calibrated, for their respective set of profiles. Profiles Arabian Sea California Current System Sargasso Sea Gulf of California North and South Atlantic Oceans (AMT-1) Instruments Used CaJibration Site MER-1032t CHORS MER-2040t MER-2040t CHORS SIIt MER-2040t UCSB PRR-600 CHORS OCI-200 PML All four MER instruments were manufactured by Biospherical Instruments, Inc. (BII), of San Diego, California. and the fit is unbiased. When (5) is applied to this data set, the linear residual standard deviation (in this case the standard error of prediction) SKx = 0.018m -1, and the mean linear bias of the Austin and Petzold (1981) estimates of/(490), is -0.007m -1. The (upper, lower) 95% confidence limits of the intercept in (12) are (-2.27452, -2.36003), and the 95% confidence limits of the slope are (-1.24596, -1.35335). Both the intercept (-2.4303) and slope (-1.491) of the natural log-transform of (5) fall outside these limits. On this basis, there is sufficient evidence at the 95% confidence level to reject the hypothesis that (5) is equivalent to (13). 2.4 DISCUSSION The analysis outlined in this paper addressed the question of whether the Austin and Petzold (1981) K(490) algorithm, (5), which is based on the Lw(443)/Lw(550) ratio, will produce accurate K(490) estimates when it is used with the LwN(443)/LwN(555) ratio. The log-transformed coefficients of (5) fall outside the 95% confidence limits of the coefficients of (12), the log-linear least-squares fit to the present sample of 242 [K(490), LWN(443)/LwN(555)] pairs. Therefore, there is sufficient evidence at the 95% confidence level to reject the hypothesis that (5) is equivalent to (13). The difference between the two sets of predictions, while small, is nevertheless statistically significant at the 95% confidence level. Changing this algorithm's coefficients, therefore, will improve the statistical uncertainty associated with the SeaWiFS K(490) at-launch products. Figures 8 and 9 emphasize scatter at different levels of This change should be simple to implement, and the au- K(490). In the log-log display of Fig. 8, the largest deviations from the regression fit occur at low values of K(490), within less than 0.01 m -1 of pure water. When the K(490) data are displayed on a linear axis (Fig. 9), it is immediately apparent that these deviations at low K(490), which appear to be very large in Fig. 8, are actually very small discrepancies. In fact, the linear K(490) scale (Fig. 9) clearly shows that the largest contributors to the linear standard error of the estimate, SKx, (14), are at the high values [K(490) > 0.1]. thors recommend that it be done. In closing, note that the normalized water-leaving radiances used here were not corrected for either instrument self shading, or the ocean bidirectional reflectance distribution function (Morel and Gentili 1996). Furthermore, the Ed(490, z) profiles from which K(490) values were determined were not corrected for Raman scattering. These corrections, which are assumed to be small, are deferred for possible use in a postlaunch refinement of this SeaWiFS algorithm. 21

Original page 28

CaseStudiesforSeaWiFSCalibrationand Validation, Part 4 Chapter 3 A Simplified Out-of-Band Correction Algorithm for SeaWiFS EUENG-NAN YEH ROBERT A. BARNES General Sciences Laurel, Maryland ABSTRACT Corporation The SeaWiFS instrument will scan through a broad area of known radiance, and the measurement for each band may contain signals from outside of the desired bandwidth. A calculation that will quickly remove the out-of-band contamination is presented here. 3.1 INTRODUCTION SeaWiFS is an eight-band filter radiometer. For Sea- WiFS, as well as for other satellite instruments managed at GSFC, the spectral response of the instrument is divided into in-band and out-of-band responses. The in-band response includes the wavelength region where the response of the band is greater than 1% of the maximum response; at all other wavelengths, the response is out-of-band. is ing array of correctedlevel-lbdata. This subroutine is For each SeaWiFS band, the out-of-band response within the specification for the instrument (Barnes et al. 1994a). These out-of-band responses, however, must be accommodated in the data reduction algorithms for on-orbit SeaWiFS measurements. The original out-of-band correction algorithm was derived for measurements by SeaWiFS from a radiant source with the spectral shape of a 5,900 K blackbody (Barnes et al. 1994b). This is the reference source spectral shape in the SeaWiFS performance specifications (Barnes et al. 1994a). Subsequently, an out-ofband correction scheme was developed that uses SeaWiFS measurements to determine the spectral shape of the upwelling Earth radiances which SeaWiFS measures (Barnes et al. 1995). Although this correction technique gives accurate out-of-band responses, the technique requires an algorithm that uses a great deal of computation time. As a result, Barnes et al. (1996) derived an alternate approach that produces out-of-band calculations much faster in terms of computer time. The equations in Barnes et al. (1996) are the basis for the computer code presented here. A benchmark test of 200 SeaWiFS scan lines has shown that the new algorithm operates about 120 times faster than its predecessor. In addition, the benchmark test showed the calculated results from the two algorithms to be equivalent at the 0.1% level. 22 3.2 THE out_bandsub ALGORITHM Prior to the applicationof atmospheric correctionto derive geophysicalvalues,the level-1radiancesfrom Sea- WiFS need an out-of-band correction.The correctionsfor out-of-band response range from about 0.5_ to more than 5%. The level-1data correctionroutine,out_bandsub, takes an array of calibratedlevel-lbradiancesand, after applyingthe out-of-bandcorrection,returnsa corresponddesignedto work one scan linewith eight bands at a time. This procedure iswritten in C. The input arguments forthisalgorithm are: 1) rad (realnumber, array size8× 1285): the level-lb radiances(bands I-8) per line; 2) oxygen (realnumber): the oxygen absorption correctionfactor;equals 1 iftad(76 nm) isalready corrected,otherwiseitequals 1.12; 3) xsample (integer):number of pixelsper line(maximum of 1,285). There isonly one output argument, which is: tad (realnumber, array size8x 1285): the level-lb out-of-bandcorrectedradiances(bands I-8). This simplifiedout-of-band method appliesthe ratioof in-band response to the totalband response as an out-ofband correctionfactor(kb,Table 11, Barnes et al.[1996]) foreach band. This process willgo through each pixel of a line: for (m - 0; m < xsample; m++) ( original_input_8 = rad[7] [m] ; orlglnal_input_6 = rad[5] [m] ; orlginal_input_5 - rad[4] [m] ; for (n ffiO; n < 8; n++) tad[n] [In] *ffiKb[n] ;.

Original page 29

E-n. Yeh, R.A. Barnes, M. Darzi, L. Kumar, E.A. Early, B.C. Johnson, J.L. Mueller, and C.C. Trees Before bands 8, 6, and 5 can be further revised, the The correction calculations for band 6 is a simplified oxygen absorption effect and radiance at 765 nm should be version of that for band 8. The spectral response, S, is established: xoffr = rad[6] [m] * oxygen; OUT9 = 2 m rad[7] [m] - XOUT;. The simplified out-of-band calculation of SeaWiFS band 8 divides the spectral response, based on a spectrally fiat source, into six components (Fig. 20, Barnes et al. [1996]). These response values, S, are listed in Table 13 of Barnes et al. (1996); thus, for a given pixel the new k b of band 8 is XN = rad[7][m] m S[4]; XD = radiO] [m]mS[O] + rad[3] _]S[I] + rad[4] [m]mS[2] + x0ffr.sc3] + XN + otrr9.s[5]; NEW Kb - XN / XD; and the final result for band 8 is tad [7] [m] = NEW_Kb original_input_8 ;. divided into three components (Table 14, Barnes et al. [1996]). The revised k b of band 6 is XN = radfS][m] * S[l]; XD = rad [2] [m] mS [O] + XN + rad [7] [nO mS [2] ; NEW_Kb = XN / XD; and the final result for band 6 is radCS] [m] - NEW_Kb * original_input_6;. Similarly, the three components of the band 5 spectral response, S, are listed in Table 15 of Barnes et al. (1996). The revised k b of band 5 is XN = rad[4][m] m S[1]; XD = rad[l][m]*S[O] + XN + rad[5] [m]*S[2]; NEW_Kb = XN / XD; and the final result for band 5 is rad[4] [m] = NEW_Kb * origlnal_input_5; ). 23

Original page 30

CaseStudiesfor SeaWiFSCalibrationandValidation,Part4 Chapter SeaWiFS Stray Light 4 Correction Algorithm EUENG-NAN YEH AND MICHAEL DARZI General Sciences Corporation Laurel, Maryland LAKSHMI KUMAR Hughes STX Lanham, ABSTRACT Maryland SeaWiFS will scan through not only the dark ocean surface, but also the bright land, clouds, and ice objects. Because of stray light in the SeaWiFS instrument, light from these bright sources can contaminate ocean measurements several pixels away from a bright source. this contamination is discussed here. 4.1 INTRODUCTION Two routines were developed, stray_light_lac and stray_light_gac, to apply a stray light correction to Sea- WiFS level-1 LAC and GAC data, respectively. The routines are integrated as part of the operational level-la read routine; their algorithms are based on the methodology described by Barnes et al. (1995). The routines perform radiance manipulations to label pixels that are bright sources or are in the vicinity of bright sources. In addition, the A mathematical formula that can be used to correct Both routines are designed to work on individual scan lines even though along-track processing is required. This is accomplished by using a buffer that retains a rolling window of consecutive scan lines. 4.2 LAC CORRECTION There are two major steps involving the SeaWiFS LAC image stray light correction for level-lb data (Fig. 10). The first step is to identify the left and right bright source radiances of pixels that are along scan to, and near, a edges. Two conditions, radiance value and gradient value, bright source are corrected for stray light contamination from that bright source. Because the input data to the routines are radiances, the calibrated levet-1 data, i.e., level-lb data, must be used. The labelling identifies pixels that are part of a bright source, near a bright source in the along-track direction, diagonal to the ends of a bright source, or near a bright source in the along-scan direction. These along-scan pixels are correctable for stray light contamination as a function of their distance from the bright source. For this reason, they are labelled by their pixel distance from the nearest edge of a bright source to the right of that pixel and from the nearest edge of a bright source to the left of that pixel. are applied to SeaWiFS level-1 band 8 (865 nm) radiance (L8) to identify these edges. The radiance of a pixel has to exceed a certain threshold, Stray_thresh, which is defined as a fraction of the value of L8 at knee 1. The value of the typical stray light fraction, Styp frac, is given as 0.9. The radiance gradient, AL, is defined as the L8 difference between two adjacent pixels n + 1 and n. The left edge is found if the radiance value is greater than Stray_thresh and AL is greater than the typical gradient threshold value. This value is defined as a fraction of the bigger one of the following two quantities: Ltypica 1(8), or the difference between L8 of pixel n ÷ 1, i.e., L8[n ÷ 1], and Ltypical(8 ). The Ltypical(8) is the typical quan- For LAC, 14 pixels to the left and 12 pixels to the right of tity of sea surface radiance measured by band 8 and the a bright source edge are considered to be affected by stray light from that bright source. For GAC, 3 pixels on each side of a bright source are so considered. If a stray light pixel is within this range between two bright sources, the pixel distance to the bright source on the right is multiplied by 1,000; the product is then added to the pixel distance to the bright source on the left. 24 value is 1.09mWcm-2#m -lsr -1. The radiance fraction Ltyp_frac is given as 0.25. The right edge is found if the radiance value is greater than Stray thresh and A L is greater than the typical gradient threshold value, which is defined as a fraction of the bigger one of the following two quantities: Ltypical(8), or the difference between L8 of pixel n, i.e., LS[n], and Ltypical(8).

Original page 31

E-n. Yeh, R.A. Barnes, M. Darzi, L. Kumar, E.A. Early, B.C. Johnson, J.L. Mueller, and C.C. Trees _all the routine for each scan line in the scene. If along-track processing is requested: ,) Will work with a rotating buffer of 5 scans. The buffer's center scan, s, will be processed for bright sources, along-scan flagging, and corrections. S will be returned by the routing after it rotates it to the first (earliest) scan in the buffer. , Otherwise: Will work with one scan, S, at a time. Do for entire scan S: Flag as BT those pixels whose band 8 radiances are greater than the band 8 calibration knee 1 value. I Pixel index--0 I I Increment pixel index ] _'es _.. ]-yes 0 has been detected Mark up to 14 pixels (that are not flagged as BT, AT, or DIAG pixels) to the left of LE b), their distance Keep incrementing n and testing for RE until RE is found or there are no more pixels left in S. from LE. I DIAG pixels) to the right of RE b)' their distance from RE. I If no LE had I Markbeen found,upto assume12 pixels that(thatthe arefirstnotpixelflaggedin scanas BT,is theAT,LE.or I Assume that the last pixel in scan is the RE. I _J [ Flag all pixels between and includin$ LE and RE as BT pixels. For each BT pixel, flag corresponding pixels in earlier scans S- 2 and S- 1 (if there are such scans in the buffer and the pixels are not flagged asBT pixels), and later scans S+l and s+2 (if there are any such scans in the buffer)r as being along-track contaminated (AT). Flag pixels diagonally to the left (if any) of LE and diagonally to the right (if any) of RE in S- 1 and S+ 1 (if there are any such scans in the buffer and if the pixels in S- 1 and S+l are not marked as BT) as being diagonally contaminated (DIAG). no _ Applybright sourcecorrectionedgeto pixelsbutare notmarkedotherwiseby theirflagged.distance from a $ (Return earliest scan (S, S-1, or S-2) with corrected radiances and flags.) + Identify all pixels flagged as BT, AT, DIAG, or whose radiances have been corrected by I using the level-2 quality control stray light flag. If that flag has been designated as a mask, pixels will not be generated. I then all level-2 products for the identified Fig. 10. A logic flow diagram for the detection and correction of stray light in SeaWiFS LAC level-1 data is shown. 25

Original page 32

CaseStudiesforSeaWiFSCalibrationandValidation,Part4 -LAC A: Along track from BT D: Diagonal to BT BT SCAN S: Pixel adjacent to BT +: Corrected I I -GAC TRACK o,iiiiii,o AAAAAA Fig. 11. This is a schematic illustration of how the stray light correction is done for LAC and GAC scenes. The arrow of SCAN indicates the instrument scanning direction. The arrow of TRACK indicates the direction of satellite movement. The shaded area represents the bright source region. Symbols A, D, and S represent flags for along-track from, diagonal to, and the scan adjacent to the bright source, respectively. The + sign represents a stray light corrected pixel. These pixels are also flagged as the distance from the bright source edge. The second step is to mask and correct the bright source affected pixels (Fig. 11). All pixels between the left and right bright source edges will be flagged as a bright target (BT) for subsequent masking by the SeaWiFS processing program. In addition, 14 LI_NGE (extra pixels to the left), and 12 RRANGE (extra pixels to the right of bright source edges) will be flagged by their distances from their respective bright source edge as stray-light contaminated pixels if they have not been flagged as BT, along-track (AT), or diagonal (DIAG) pixels before. These pixels will later be corrected by the along-scan direction correction algorithm for the contamination. If along-track masking is selected, all pixels not identified as BTs between the left and right bright source edges on the previous two lines and next two lines will be flagged as along-track pixels. Only one extra pixel to the left and one extra pixel to the right of bright source edges of one previous line and one subsequent line will be flagged as diagonal pixels. The along-scan direction stray light correction (Lcorr) 26 for pixel 0 proceeds in the following sequence: Sequence 1, set the correction term to zero: Lcorr = 0. (15) Sequence 2, replace pixel 0 radiance, L[0], as: Lcorr = Lcor_ + L[0](1 - g[0]), (16) where K[0] is the stray light correction constant for pixel 0. The values of K are listed in Table 8. Sequence 3, remove the stray light contributions from the adjacent pixels: Lcor_ = Lcorr - L[-1]K[+I], (17) L¢o_ = Leo, - L[-2]K[+2], (18) Lcorr = Lco_ - L[+I]K[-1], (19)

Original page 33

E-n. Yeh, R.A. Barnes, M. Darzi, L. Kumar, E.A. Early, B.C. Johnson, J.L. Mueller, and C.C. Trees Table 8. Along-scan responses for SeaWiFS bands 1-8 with their respective wavelengths, in nanometers, in )arentheses. These are the constants, K, used in the LAC stray light correction procedure. Pixel Band Numbers and Wavelengths [nm] Offset 1 (412) 2 (443) 3 (490) -12 0.00000 0.00005 0.00000 -11 0.00000 0.00014 0.00000 -10 0.00000 0.00029 0.00000 -9 0.00000 0.00048 0.00000 -8 0.00004 0.00071 0.00004 -7 0.00010 0.00086 0.00010 -6 0.00030 0.00105 0.00025 -5 0.00080 0.00124 0.00073 -4 0.00240 0.00141 0.00195 -3 0.00639 0.00494 0.00534 -2 0.01757 0.01271 0.01531 -1 0.13172 0.09747 0.21432 0 0.73254 0.83142 0.73217 4 (510) 5 (555) 6 (670) 7 (765) 8 (865) 0.00006 0.00080 0.00005 0.00000 0.00003 0.00020 0.00000 0.00016 0.00000 0.08010 0.00034 0.00000 0.00027 0.00000 0.00016 0.00052 0.00000 0.00048 0.00000 0.00026 0.00070 0.00002 0.00076 0.00000 0.00039 0.00078 0.00005 0.00098 0.00003 0.00048 0.00090 0.00015 0.00129 0.00010 0.00064 0.00104 0.00045 0.00140 0.00031 0.00086 0.00157 0.00135 0.00168 0.00096 0.00118 0.00444 0.00343 0.00292 0.00232 0.00239 0.01044 0.00925 0.00599 0.00758 0.00710 0.10994 0.06548 0.05577 0.18735 0.23668 0.86085 0.67127 0.70802 0.67749 0.66297 1 0.06387 -0.00424 -0.00267 -0.08104 0.23406 0.19771 0.09616 0.05512 2 0.01998 0.03237 0.01956 -0.00418 0.00737 0.01381 0.01971 0.01586 3 0.01038 0.01271 0.00607 4 0.00350 0.00419 0.00176 5 0.00300 0.00141 0.00146 6 0.00230 0.00052 0.00108 7 0.00200 0.00019 0.00083 8 0.00140 0.00007 0.00065 9 0.00110 0.00002 0.00049 10 0.00050 0.00080 0.00031 11 0.00013 0.00000 0.00020 12 0.00000 0.00000 0.00007 13 0.00080 0.00000 0.00000 14 0.00000 0.00000 0.00000 and Lcorr = Loot, - L[+2]K[-2], (20) until pixels LRANGE (equals 14) and RRANGE(equals 12) are reached. Sequence 4, correct the radiance: L[0] = L[0] + Lcorr. (21) The above sequences are applied to all eight SeaWiFS bands. This subroutine is designed to work one scan line with eight bands at a time; this procedure is written in C. The variables, both for input and output, needed for the function stray_light lac are summarized in Table 9. This subroutine, stray_light_lac, calls for each scan line in the scene: status = stray_light_lac(initial, Ltyp_frac, Styp_frac, nscans, nsamples, scan_no, gn, rads, lib_data, sl_scan, sl_flag). 0.00914 0.00298 0.00550 0.00372 0.00796 0.00301 0.00150 0.00194 0.00133 0.00382 0.00090 0.00059 0.00081 0.00079 0.00191 0.00026 0.00052 0.00030 0.00061 0.00089 0.00008 0.00045 0.00011 0.00044 0.00048 0.00003 0.00035 0.00004 0.00031 0.00029 0.00000 0.00030 0.00002 0.00028 0.00016 0.00000 0.00022 0.00000 0.00021 0.00010 0.00008 0.00015 0.00000 0.00014 0.00010 0.00000 0.00007 0.00000 0.00010 0.00006 0.00000 0.00000 0.00008 0.00007 0.00003 0.00080 0.00000 0.00000 0.00003 0.00000 If along-track processing is requested, this subroutine will use a rotating buffer of five scans to mark as BT and alongscan flagging and corrections on the current scan line, S. Other scans in the buffer will store information of pixels flagged with AT and DIAG. Otherwise, it will work with one scan, S, at a time. As discussed in previous sections, the stray light algorithm will try to identify stray light affected pixels and correct them if possible. The first step to find stray light pixels is to screen through each pixel, from scan S, and verify if it is a bright source pixel. The test is if L8 is greater than the knee value, then set the flag buffer as BT. Table 10 is the first knee value of band 8, which will be used for Stray_thresh value calculation. This screening can speed up the determination of bright source edges, which will skip consecutive bright source pixels. The bright source edges can be identified by tests of two quantities: radiance value and gradient value. The edge is a left edge if the gradient shows a positive value. A negative gradient means a right bright source edge is detected. The flag buffer will mark up to LRANGEor RRANGE pixels to the left or right of bright source edges by their 27

Original page 34

Case Studies forSeaWiFS Calibrationand Validation,Part 4 Table 9. Listed below are the LAC input and output parameters. Variable Data Type VariableDescription Input Parameters initial int32 Should be setto 1 forthe firstcallof a scene. Ltyp__rac l_oat32 Fraction(=0.25) factorof Ltypica!to determine the bright targetedge. StypJrac Soat32 Fraction(=0.90) factorof band 8 knee 1 radiance to determine the brighttargetedge. nscans int32 This isthe processingindicator.Ifnscans-%, itwillperform only along-scandirectionstray lightcorrectionprocessing.If nscans_maximum of the filerecords,itwilldo both along-scan and along-trackstray lightprocessing. nsamples int32 Number of pixelsper line. scan_no int32 Scan linenumber (0-relative)of input data. gn int16 Band 8 gain of scan linescan_no. rads Boat32 Knee radiance values [8 x 4 x 5]. 11b_data I]oat32 An array of nsamples long containingllb_data (8 bands) of scan line scan_no. Output Paramete.rs llb_data Aost32 An array of stray light corrected level-lb data. sl_scan int32 Scan line number for which the returned 11b_data is applied. slllag int32 An array of nsamplss long containing codes identifying bright target pixels and the proximity of others to bright targets. The meaning of sl_flag value is: 0 A brighttargetflag(BT); -1 An along-trackflag(AT); -2 A diagonal flag(DIAG); >0 The number represents the distance in pixels away from the bright target edges--if a pixel is within range of two bright targets, sl_flag will equal the pixel distance from the left bright target edge plus 1,000 times the pixel distance from the right bright target edge (a sl_flag value of 3001 means three pixels away from the bright target edge on the right and one pixel away from left bright target edge); and -10 None of the above (BLANK). Return value int32 Returns as 0, not done, to indicate that the routine did not return any useful information for the last call, but was filling its buffers with scan line data; and 1, done, to indicate that the llb_data and sl_flag arrays for scan line sl_scan have been set. Table 10. These axe the band 8 knee 1 values for the SeaWiFS bilineax gains. The radiances are in units of cm-2 st-1 #m -1. Gain 1 Knee Radiance 1.64928 Knee Count 762.30 distance from the respective edge, assuming these pixels have not been flagged as BT, AT, or DIAG pixels. For a pixel within the range of two bright sources, the flag buffer will mark the distance as the sum of the pixel distance from the left bright source edge plus 1,000 times the pixel distance from the right bright source edge. Pixels with flag buffers marked as distance will eventually perform the along-scan stray light correction. If a left and 28 2 3 4 0.830554 6.32555 3.30217 758.12 794.66 773.58 right edge pair fails to identify one of the two edges, the first pixel in scan S will be assigned as the beginning of the left edge for the missing left edge, or the last pixel in S will be assigned as the ending of the right edge for the missing right edge. The flag buffer will mark all pixels between, and including, the left and right edges as BT pixels for scan S. If along-track masking is the option, for each BT pixel, the

Original page 35

F_,-n. Yeh, R.A. Barnes, M. Darzi, L. Kumar, E.A. Early, B.C. Johnson, J.L. Mueller, and C.C. Trees Gall the routine for each scan line in the scene. ) Will work with a rotating buffer of 3 scans. The buffer's center scan, S, will be processed for bright sources and along-scan flagging and corrections. s will be returnedIf along-trackprocessingby the routingis requested:after it rotates it to the first (earliest) scan in the buffer. Otherwise: Will work with one scan, S, at a time. Do for entire scan S: Flag as BT those pixels whose band 8 radiances ] are greater than the band 8 calibration knee 1 value. [ Pixel index=0 ] [ Increment pixei index [ ,T- _ 1 yes _A left edge (LE) has been detected. I 119 has been detected A right edge (RE) i Mark up to 3 pixels (that are not flagged as BT or AT ) to the left of LE by their ] pixel (if an),) to the left of LE. I distance from LE and apply correction to the 3rd Keep incrementing n and testing for RE until RE is found or there are no more pixels left in S. ),es ] I ]Vl"-"kup to 3 pixels (that are not flagged as BT or AT) to the rig-]ht of RE by eir distance from RE and apply correction to 3rd [ixel (if any) to right of RE. [ If no LE had been found, assume that the first pixel in scan is the LE. IAssume that the last pixel in scan is the RE. ] / - 11 ] Fla[ all pixels between and includin LE and RE as BT pixels. For each BT pixel, flag corresponding pixels in the earlier scan S- 1 (if there is such a scan in the buffer and the pixels are not flagged asBT pixels), and later scan S+l (if there is such a ] scan in the buffer)r as being along-track contaminated (AT). QRetum earliest scan (s or S- 1) with corrected radiances and fla[s. ) by using the level-2 quality control stray light flag. If that flag has been designated as a Identify all pixels flagged as BT, AT, or have been marked by distance from a bright source [ pixels will not be generated. I mask, then all level-2 products for the identified Fig. 12. Logic flow diagram for the detection and correction of stray light in SeaWiFS GAC level-1 data. 29

Original page 36

CaseStudiesfor SeaWiFSCalibrationand Validation, Part 4 Thble 11. Along-scan GAC stray light correction factors, K. Band Pixels Before Bright Source Number -1 -2 -3 -0.00000 0 0 -0.00060 0 0 -0.00000 0 0 -0.00074 0 0 -0.00000 0 0 -0.00055 0 0 -0.00000 0 0 -0.00037 0 0 corresponding pixel in earlier scans (S-1 and S-2) and later scans (S+l and S+2) will be flagged as AT for stray light contaminated pixels, assuming the scan lines do exist and the pixels have not been marked as BT. The flag buffer will also mark one extra pixel to the left or right of the left or right edge on scans S-t and S+I as DIA0 for stray light contaminated pixels, assuming the scan lines do exist and the pixels have not been marked as BT. The bright source edge detection and flag buffer flagging will run through the whole scan S. The along-scan direction stray light correction algorithm will apply to the pixel, having the distance as the flag indicator. This program will output the earliest scan (8-2, S-l, or S) with stray light corrected radiances and flags. The information contained in the flag record can be used to identify all pixels flagged as BT, AT, DIA0, or the distance from a BT edge. If that flag has been designated as a mask, then all level-2 products for the identified pixels will not be generated. 4.3 GAC CORRECTION The algorithm for GAC stray light correction, identifying bright sources and edges, is similar to that for the LAC correction algorithm (Fig. 12). The GAC along-scan correction, however, is applied only to the third GAC pixel from the bright source edge: L[3] = L[3] + Ledg.K[3], (22) where L[3] represents the level-1 radiance at three pixels away from the bright source edge. K is the GAC stray light correction constant and its value is listed in Table 11. The above equation is also applied to all eight SeaWiFS bands. The GAC stray light correction algorithm will flag three pixels before the left edge and three pixels after the right 30 Pixels After Bright Source 1 2 3 0 0 -0.00090 0 0 -0.00001 0 0 -0.00055 0 0 -0.00000 0 0 -0.00053 0 0 -0.00000 0 0 -0.00055 0 0 -0.00033 edge (by their distances from respective bright source edges) as stray light contaminated pixels if they have not been previously flagged as BT or AT pixels. The parameters used for the stray_light_gac function are similar to that of the stray_light_lac function (Table 9). The only exception is that 0AC does not have DIAG flags. Unlike LAC, the GAC rotating buffer contains only three scan lines if along-track processing is requested and the output will be the earliest scan (S-I, or S). The flag buffers will not contain DIA0 flags. The value of LRANOE and RRANGEis 3 with only the third GAC pixel from the bright source edge corrected for stray light contamination. Other than that, the GAC software is the same as the LAC software. 4.4 STRAY LIGHT MASKING As mentioned before, the stray light affected pixels will be flagged as bright source (BT ----0), along track (AT = -1), diagonal (DIA0 = -2), or distance (greater than or equal to 1, in units of pixels) away from the bright source edges. The distance range for a LAC scene can be 14 pixels to the left of a bright source left edge and 12 pixels to the right of a bright source right edge. All pixels with positive flag settings will be corrected for stray light contaminations. As a GAC scene, however, the distance range will go up to only three pixels on both sides of bright source edges and only the third pixel will be corrected. All pixels not affected by stray light will be flagged as blank (BLANK = -10). Further analysis of the flag output, sl_flag, will reset the flag values between -10 (exclude) and the value of the maximum distance range of 256, i.e., the SeaWiFS level-2 stray light masking. In this way, the pixels with corrections will also be masked out.

Original page 37

E-n. Yeh, R.A. Barnes, M. Darzi, L. Kumar, E.A. Early, B.C. Johnson, J.L. Mueller, and C.C. Trees GLOSSARY AMT Atlantic Meridional Transect ASCII American Standard Code for Information Interchange AVHRR Advanced Very High Resolution Radiometer BATS Bermuda Atlantic Time Series BII Biospherical Instruments, Inc. CCS California Current System CHORS Center for Hydro-Optics and Remote Sensing CICESE Centro de lnvestigacidn Cient(fica y de Educacidn Superior de Ensenada (Baja, California) CVT Calibration and Validation Team CZCS Coastal Zone Color Scanner DC Direct Current EOSDIS Earth Observing System Data and Information System FASCAL Facilityfor Automated Spectroradiometric Calibra_ tions GAC Global Area Coverage GoCal Gulf of California GOES Geostationary Operational Environmental Satellite GSFC Goddard Space Flight Center HP Hewlett-Packard IEEE Institute of Electrical and Electronic Engineers ISIC Integrating Sphere Irradiance Collector JGOFS Joint Global Ocean Flux Study LAC Local Area Coverage LANDSAT Land Resources Satellite LE Left Edge MER Marine Environmental Radiometer NASA National Aeronautics and Space Administration NIST National Institute of Standards and Technology NOAA National Oceanic and Atmospheric Administration OSC Orbital Sciences Corporation PC Personal Computer (IBM) PML Plymouth Marine Laboratory PSR Point Spread Response RE Right Edge rms root mean square SBRS (Hughes) Santa Barbara Remote Sensing SDSU San Diego State University SeaWiFS Sea-viewing Wide Field-of-viewSensor SIO Scripps Institutionof Oceanography SIRREX SeaWiFS Intercalibration Round-Robin Experiment SIRREX-1 The First SIRREX (July 1992) SIRREX-2 The Second SIRREX (June 1993) SIRREX-3 The Third SIRREX (September 1994) SIRREX-4 The Fourth SIRREX (May 1995) SRR Spectral Radiance Responsivity SXR SeaWiFS Transfer Radiometer UCSD University of California, San Diego UCSB University of California,Santa Barbara UK United Kingdom VISSR Visible-Infrared Spin-Scan Radiometer SYMBOLS A Logarithmic regression model coefficient. B Logarithmic regression model coefficient. Ed(z, ) Downwelling irradiance. E0(A) Downwelling irradiance incident at the sea surface. F0(A) Mean extraterrestrial solar flux. I Current. kb Out-of-band correction factor. K Stray light correction constant. K(490) Remote sensing diffuse attenuation coefficient, in per unit meters. The average of Kd(z, 490) over the first attenuation length. /(490) Regression model estimate of K(490) in per unit meters. Kd(z, A) Vertical attenuation coefficient for Ed(z, A), in per unit meters. KL(z, A) Vertical attenuation coefficient for Lu(z, A), in per unit meters. K_ Attenuation coefficient for pure water. L Spectral radiance. L The spectral radiance from a calibrated (known) source. L The spectral radiance from a measured (unknown) source. L8 Level-1 band 8 radiance. Lcorr Along-scan radiance correction factor. Lo Spectral radiance measured with a gas filled striplamp standard. Ls Spectral radiance measured with a spectroradiometer. Ltypical(8) Typical quantity of sea surface radiance measured by band 8. L_(z,A) Upwelling spectral radiance at depth z. Lw Water-leaving radiance. LwN (A) Normalized water-leaving radiance at the top of the sea surface (z = 0+). N Sample size. R Spectral radiance responsivity. R 2 Multiple correlation coefficient. S Response value. ,_ The signal from a calibrated (known) source. The signal from a measured (unknown) source. Linear residual standard deviation of modeled radi- SKx ance ratios from water-leaving radiance ratios. if(A) Upward Fresnel transmittance through the air-sea interface for radiance. x Horizontal spatial coordinate. Y Vertical spatial coordinate. z Depth, in meters, below the air-sea interface. zgo Depth, in meters, equal to the first attenuation length. 5 Relative standard uncertainty. AL Radiance gradient. A Wavelength. a Standard deviation. 31

Original page 38

CaseStudiesforSeaWiFSCalibrationandValidation,Part4 REFERENCES Austin, R.W., and T.J. Petzold, 1981: "The Determination of the Diffuse Attenuation Coefficient of Sea Water Using the Coastal Zone Color Scanner." In: J.F.R. Gower, Ed., Oceanography from Space, Plenum Press, New York, 239- 256. Barnes, R.A., W.L. Barnes, W.E. Esaias, and C.R. McClain, 1994a: Prelaunch Acceptance Report for the SeaWiFS Radiometer. NASA Tech. Memo. 104566, Vol. ££, S.B. Hooker, E.R. Firestone, and J.G. Acker, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 32 pp. ,A.W. Holmes, W.L. Barnes, W.E. Eeaias, C.R. McClain, and T. Svitek, 1994b: SeaWiFS Prelaunch Radiometric Calibration and Spectral Characterization. NASA Tech. Memo. 104566, Vol. £3, S.B. Hooker, E.R. Firestone, and J.G. Acker, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 55 pp. the --, 1995: SeaWiFS Pre-Launch Algorithm for the Diffuse At- , A.W. Holmes, and W.E. Ezaias, 1995: Stray Light in Morel, A., and B. Gentili, 1996: Diffuse reflectance of oceanic waters. III. Implication of bidirectionality for the remote sensing problem. Appl. Opt., 35, 4,850-4,862. Mueller, J.L., 1991: Integral Method for Analyzing Irradiance and Radiance Attenuation Profiles. CHORS Tech. Memo. 007-91, San Diego State University Center for Hydro- Optics and Remote Sensing, San Diego, California, 10 pp. --, 1993a: The First SeaWiFS Intercalibration Round-Robin Experiment, SIRREX-1, July 1992. NASA Tech. Memo. 104566, Vol. I4, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 60 pp. --, 1993b: An Optical Climatology of the Northeast Pacific Ocean. CHORS Tech. Memo. 001-93, San Diego State University Center for Hydro-Optics and Remote Sensing, San Diego, California, 46 pp. plus appendices. SeaWiFS Radiometer. NASA Tech. Memo. 104566, Vol. tenuation Coefficient K(490). CHORS Tech. Memo. 003- 31, S.B. Hooker, E.R. Firestone, and J.G. Acker, Eds., 95, San Diego State University Center for Hydro- NASA Goddard Space Flight Center, Greenbelt, Maryland, 43 pp. _, E-n. Yeh, and R.E. Eplee, 1996: SeaWiFS Calibration Topics, Part 1. NASA Tech. Memo. 104566, Vol. 39, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 18--38. Hovis, W.A., and J.S. Knoll, 1983: Characteristics of an internally illuminated calibration sphere. Appl. Opt., 22, 4,004-4,007. Johnson, B.C., S.S. Bruce, E.A. Early, J.M. Houston, T.R. O'Brian, A. Thompson, S.B. Hooker, and J.L. Mueller, 1996: The Fourth SeaWiFS Intercalibration Round-Robin Experiment (SIRREX-4), May 1995. NASA Tech. Memo. 104566, Vol. 37, S.B. Hooker and E.R. Firestone, EdJ., NASA Goddard Space Flight Center, Greenbelt, Maryland, 65 pp. , J.B. Fowler, and C.L. Cromer, 1997: The SeaWiFS Transfer Radiometer. Sea WiFS Postlaunch Technical Report Series, S.B. Hooker and E.R. Firestone, EAs., Greenbelt, Maryland (submitted). Kostkowski, H.J., and F.E. Nicodemns, 1978: An Introduction to the Measurement Equation. NBS Tech. Note 910-£, Chapter 5, F.E. Nicodemus, Ed., U.S. Dept. of Commerce, National Bureau of Standards, Washington, D.C., 58-104. McOlain, C.R., W.E. Esaias, W. Barnes, B. Guenther, D. Endres, S. Hooker, G. Mitchell, and R. Barnes, 1992: Calibration and Validation Plan for SeaWiFS. NASA Tech. Memo. 104566, Vol. 3, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 41 pp. Spec- ing High-Altitude Aircraft Measurements. NASA Tech. Mielenz, K.D., R.D. Saunders, and J.B. Shumaker, 1990: Optics and Remote Sensing, San Diego, California, San Diego, California, 11 pp. _, B.C. Johnson, C.L. Cromer, J.W. Cooper, J.T. McLean, S.B. Hooker, and T.L. Westphal, 1994: The Second Sea- WiFS Intercalibration Round-Robin Experiment (SIR- REX-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. --, B.C. Johnson, S.B. Hooker, J.T. McLean, and S. Biggar, 1995: The Third SeaWiFS Intercalibration Round-Robin Experiment (SIRREX-3), 19-30 September 1994. NASA Tech. Memo. 104566, Vol. 34, S.B. Hooker, E.R. Firestone, and J.G. Acker, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 78 pp. Neckel, H., and D. Labs, 1984: The solar radiation between 3300 and 12500A. Solar Phys., 90, 205-258. Smith, G.R., R.H. Levin, P. Abel, and H. Jacobowitz, 1988: Calibration of the solar channels of the NOAA-9 AVHRR using high altitude aircraft measurements. J. Atmos. Ocean. Tech., 5, 631-639. , , R.S. Koyanagi, and R.C. Wrigley, 1989a: Calibration of the Visible and Near-Infrared Channels of the NOAA-9 AVHRR Using High-Altitude Aircraft Measurements from August 1985 and October 1986. NASA Tech. Memo. 101063, NASA Ames Research Center, Motfett Field, California, 18pp. , , , and --, 1989b: Calibration of the Visible and Near-Infzared Channels of the GOES-6 VISSR Ustemperature Memo. 101064, NASA Ames Research Center, Moffett troradiometric determination of the freezing of gold. J. Res. Natl. In.st.Stand. Technol.,95, 49-67. 32 Field, California, 14 pp.

Original page 39

E-n. Yeh, R.A. Barnes, M. Darzi, L. Kumar, E.A. Early, B.C. Johnson, J.L. Mueller, and C.C. Trees , , , and --, 1990: Calibration of the Visible and Near-Infrared Channels of the LANDSAT-5 Thematic Mapper Using High-Altitude Aircraft Measurements. NASA Tech. Memo. 102185, NASA Ames Research Center, Moffett Field, California, 15 pp. Walker, J.H., R.D. Saunders, J.K. Jackson, and D.A. McSparron, 1987a: Spectral Irradiance Calibrations. NBS Special Publication 250-20, U.S. Dept. of Commerce, National Bureau of Standards, Washington, D.C., 37 pp. plus appendices. , , and A.T. Hattenburg, 1987b: Spectral Radiance Calibrations. NBS Special Publication 250-1, U.S. Dept. of Commerce, National Bureau of Standards, Washington, D.C., 26pp. plus appendices. THE SEAWIFS TECHNICAL REPORT SERIES Vol. 1 Hooker, S.B., W.E. Esaias, G.C. Feldman, W.W. Gregg, and C.R. McClaln, 1992: An Overview of SeaWiFS and Ocean Color. NASA Teeh. Memo. 104566, Vol. 1, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 24 pp., plus color plates. Vo]. 2 Gregg, W.W., 1992: Analysis of Orbit Selection for SeaWiFS: Ascending vs. Descending Node. NASA Tech. Memo. 104566, Vol. _, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 16 pp. Vol. 3 McClaln, C.R., W.E. Esalas, W. Barnes, B. Guenther, D. Endres, S. Hooker, G. Mitchell, and R. Barnes, 1992: Calibration and Validation Plan for SeaWiFS. NASA Tech. 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., NASA Goddard Space Flight Center, Greenbelt, Maryland, 22 pp., plus color plates. yoJ. 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 Space Flight Center, Greenbelt, Maryland, 43 pp. yol. Firestone, E.R., and S.B. Hooker, 1992: SeaWiFS Technical Report Series Summary Index: Volumes 1-5. NASA Teeh. Memo. 104566, Vol. 6, S.B. Hooker and E.R. Firestone, Eels., NASA Goddard Space Flight Center, Greenbelt, Maryland, 9 pp. 1Iol. 7 Darzi, M., 1992: Cloud Screening for Polar Orbiting Visible and IR Satellite Sensors. NASA Teeh. Memo. 104566, Vol. 7, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 7 pp. Vol. 8 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. Vol. 9 Gregg, W.W., F.C. Chen, A.L. Mezaache, J.D. Chen, J.A. 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, 17 pp. Vol. 10 Woodward, R.H., R.A. Barnes, C.R. McClaln, 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. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 26 pp. Vol. 11 Patt, F.S., C.M. Hoisington, W.W. Gregg, and P.L. Coronado, 1993: Analysis of Selected Orbit Propagation Models for the SeaWiFS Mission. NASA Tech. Memo. 104566, Vol. 11, S.B. Hooker, E.R. Firestone, and A.W. Indest, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 16 pp. Vol. 12 Firestone, E.R.., and S.B. Hooker, 1993: SeaWiFS Technical Report Series Summary Index: Volumes 1-11. NASA Tech. Memo. 104566, Vol. Ig, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 28 pp. Vol. 13 McClaln, C.R., K.R. Arrigo, J. Comiso, R. Fraser, M. Darzi, 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. l& S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 52pp., plus color plates. Vol. 14 Mueller, J.L., 1993: The First SeaWiFS Intercalibration Round- Robin Experiment, SIRREX-1, July 1992. NASA Teeh, Memo. 104566, Vol. 14, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 60 pp. Vol. 15 Gregg, W.W., F.S. Patt, and R.H. Woodward, 1994: The Simulated 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. Vol. 16 Mueller, J.L., B.C. Johnson, C.L. Cromer, J.W. Cooper, J.T. 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. 33

Original page 40

CaseStudiesfor SeaWiFS Calibration and Validation, Part 4 1Iol. 17 Abbott, M.R., O.B. Brown, H.R. Gordon, K.L. Carder, R.E. 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, 20pp. Vol. 18 Firestone, E.R., and S.B. Hooker, 1994: SeaWiFS Technical Report 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, 47pp. Vol. 19 McClain, C.R., R.S. Fraser, J.T. McLean, M. Darzi, J.K. Fire_ stone, 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. Vow, 1994: Case Studies for SeaWiFS Calibration and Validation, Part 2. NASA Tech. Memo. 104566, Vol. I9, S.B. Hooker, E.R. Firestone, and J.G. Acker, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 73 pp. Vol. 20 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. 10,_566, Vol. 20, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 40pp. Vol. 21 Acker, J.G., 1994: The Heritage of SeaWiFS: A Retrospective on the CZCS NIMBUS Experiment Team (NET) Program. NASA Teeh. Memo. i0.566, Vol. 1, S.B. Hooker and E.R. Firestone, F_As., NASA Goddard Space Flight Center, Greenbelt, Maryland, 43 pp. Vol. 22 Barnes, R.A., W.L. Barnes, W.E. Eeaias, and C.R. McClain, 1994: Prelaunch Acceptance Report for the SeaWiFS Radiometer. NASA Tech. Memo. i0.566, Vol. , S.B. Hooker, E.R. Firestone, and J.G. Acker, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 32 pp. Vol. 23 Barnes, R.A., A.W. Holmes, W.L. Barnes, W.E. Esaias, C.R. McClain, and T. Svitek, 1994: SeaWiFS Prelatmch Pd_diometric Calibration and Spectral Characterization. NASA Teeh. Memo. 104566, Vol. 3, S.B. Hooker, E.R. Firestone, and J.G. Acker, Eels., NASA Goddard Space Flight Center, Greenbelt, Maryland, 55pp. Vol. 24 Firestone, E.R., and S.B. Hooker, 1995: SeaWiFS Technical Report Series Summary Index: Volumes 1-23. NASA Teeh. Memo. 10.566, Vol. 2., S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 36 pp. Vol. 25 Mueller, J.L., and R.W. Austin, 1995: Ocean Optics Protocols for SeaWiFS Validation, Revision 1. NASA Tech. Memo. 104566, Vol. 25, S.B. Hooker and E.R. Firestone, F., NASA Goddard Space Flight Center, Greenbelt, Maryland, 67 pp. 34 Vol. 26 Siegel, D.A., M.C. O'Brien, J.C. Soreusen, D.A. Konnoit, E.A. Brody, J.L. Mueller, C.O. Davis, W.J. Rhea, and S.B. Hooker, 1995: Results of the SeaWiFS Data Analysis Round-Robin (DARR), July 1994. NASA Tech. Memo. 104566, Vol. 26, S.B. Hooker and E.R. Firestone, F., NASA Goddard Space Flight Center, Greenbelt, Maryland, 58 pp. Vol. 27 MueUer, J.L., R.S. Fraser, S.F. Biggar, K.J. Thome, P.N. Slater, A.W. Holmes, Ft.A. Barnes, C.T. Weir, D.A. Siegel, D.W. Menzies, A.F. Michaels, and G. Podesta 1995: Case Studies for SeaWiFS Calibration and Validation, Part 3. NASA Tech. Memo. 104566, Vol. 27, S.B. Hooker, E.R. Firestone, and J.G. Acker, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 46 pp. Vol. 28 McClain, C.R., K.R. Arrigo, W.E. Esaias, M. Darzi, F.S. Part, R.H. Evans, J.W. Brown, C.W. Brown, R.A. Barnes, and L. Kumar, 1995: SeaWiFS Algorithms, Part 1. NASA Teeh. Memo. 10._566, Vol. 28, S.B. Hooker, E.R. Firestone, and J.G. Acker, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 38pp., plus color plates. Vol. 29 Aiken, J., G.F. Moore, C.C. Trees, S.B. Hooker, and D.K. Clark, 1995: The SeaWiFS CZCS-Type Pigment Algorithm. NASA Tech. Memo. I0._566, Vol. 9, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 34 pp. Vol. 30 Firestone, E.R., and S.B. Hooker, 1996: SeaWiFS Technical Report Series Cumulative Index: Volumes 1-29. NASA Tech. Memo. I0.566, Vol. 30, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 43 pp. Barnes, R.A., A.W. Holmes, and W.E. Esaias, 1995: Stray Light in the SeaWiFS Radiometer. NASA Tech. Memo. 104566, Vol. 31, S.B. Hooker, E.R. Firestone, and J.G. Acker, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 76pp. Vol. 32 Campbell, J.W., J.M. BlaisdeU, and M. Darzi, 1995:Level-3 SeaWiFS Data Products: Spatial and Temporal Binning Algorithms. NASA Tech. Memo. 10_566, Vol. 3., S.B. Hooker, E.R. Firestone, and J.G. Acker, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 73pp., plus color plates. Vol. 33 Moore, G.F., and S.B. Hooker, 1996: Proceedings of the First SeaWiFS Exploitation Initiative (SEI) Team Meeting. NASA Tech. Memo. 104566, Vol. 33, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 53 pp. voi. 34 Mueller, J.L., B.C. Johnson, C.L. Cromer, S.B. Hooker, J.T. McLean, and S.F. Biggar, 1996: The Third SeaWiFS Intercalibration Round-Robin Experiment (SIRREX-3), 19-- 30 September 1994. NASA Tech. Memo. 104566, Vol. 8, S.B. Hooker, E.R. Firestone, and J.G. Acker, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, "/8pp.

Original page 41

E-n.Yeh,R.A. Barnes, M. Darzi, L. Kumar, E.A. Early, B.C. Johnson, J.L. Mueller, and C.C. Trees Robins, D.B., A.J. Bale, G.F. Moore, N.W. Rees, S.B. Hooker, C.P. Gallienne, A.G. Westbrook, E. Marafi6n, W.H. Spooner, and S.R. Laney, 1996:AMT-1 Cruise Report and Preliminary Results. NASA Tech. Memo. 104566, Vol. 35, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 87pp. Firestone, E.R., and S.B. Hooker, 1996: SeaWiFS Technical Report Series Cumulative Index: Volumes 1-35. NASA Te.ch. Memo. 104566, Vol. 36, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 55 pp. Johnson, B.C., S.S. Bruce, E.A. Early, J.M. Houston, T.R. O'Brian, A. Thompson, S.B. Hooker, and J.L. Mueller, 1996: The Fourth SeaWiFS Intercalibration Round-Robin Experiment (SIRREX-4), May 1995. NASA Tech. Memo. 104566, Vol. 37, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 65 pp. McClain, C.R., M. Darzi, R.A. Barnes, R.E. Eplee, J.K. Firestone, F.S. Patt, W.D. Robinson, B.D. Schieber, R.H. Woodward, and F_,-n. Yeh, 1996: eaWiFS Calibration and Validation Quality Control Procedures. NASA Tech. Memo. 104566, Vol. 38, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 68 pp. Barnes, R.A., E-n. Yeh, and R.E. Eplee, 1996: SeaWiFS Calibration Topics, Part 1. NASA Tech. Memo. 104566, Vol. 39, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 66 pp. Vol. 40 Barnes, R.A., R.E. Eplee, Jr., F,-n. Yeh, and W.E. Esaias, 1997: SeaWiFS Calibration Topics, Part 2. NASA Tech. Memo. 104566, Vol. 40, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 67pp. Yeh, F_,-n., R.A. Barnes, M. Darzi, L. Kumar, E.A. Early, B.C. Johnson, and J.L. Mueller, 1997: Case Studies for Sea- WiFS Calibration and Validation, Part 4. NASA Tech. Memo. 104566, Vol. 41, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 35pp. 35

Original page 42

Original page 43

Original page 44

Publicreportingburdenfor thiscollectiono! Information is timstad to average I hour per response,includingthe time fix reviewinginstructions,searchingexistingdata sources,gathering and maintainingthe data needed, and completingand reviewingthe collectionof information.Send commentsrol_rd_g this burdenestimateor any otheraspectof thiscollectionof infoffnation,includingsuggestions for reducingthisburdan,toWaahingtonHeadquartersServtceo,Dirsctomtefor Infotmatk)nOperationaand Reports, 1215 JeffersonDavisHighway, Suite 1204, Arlington,VA 2220-4302, and tothe Office of Manm_ementand Bud_, Papem_:_kReductionPro_lct{0704-0188}t Wludlin_n I DC 20503. 1. AGENCY USE ONLY (/.crave blank) 2. REPORT DATE August 1997 4. TITLE AND SUBTITLE SeaWiFS Technical Report Series 3. REPORT TYPE AND DATES COVERED Technical Memorandum 5. FUNDING NUMBERS Volume 41: Case Studies for SeaWiFS Calibration and Validation, Part 4 AUTHOP_S) Code 970.2 Eueng-nan Yeh, Robert A. Barnes, Michael Darzi, Lakshmi Kumar, Edward A. Early, B. Carol Johnson, James L. Mueller, and Charles C. Trees Series Editors: Stanford B. Hooker and Elaine R. Firestone 7. PERFORMINGORGANIZATIONNAME(S)ANDADORESS(ES) Laboratory for Hydrospheric Processes Goddard Space Flight Center Greenbelt, Maryland 20771 8. PERFORMING ORGANIZATION REPORT NUMBER 97B00080 AGENCYNAME(S)ANDADDRESS(ES) I0. SPONSORING/MONITORING 9. SPONSORING/MONITORING National Aeronautics and Space Administration Washington, D.C. 20546-0001 AGENCY REPORT NUMBER TM-104566, Vol. 41 11. SUPPLEMENTARYNOTES Elaine R. Firestone, Eueng-nan Yeh, Robert A. Barnes _and Michael Darzi: General Sciences Corporation, Laurel, Maryland; Lakshmi Kumar: Hughes STX, Landover, Maryland; Edward A. Early and B. Carol Johnson: National Institute of Standards and Technology, Gaithersburg, Maryland; and James L. Mueller and Charles C. Trees: Center for Hydro-Optics and Remote Sensing, San Diego State University, San Diego, California 12a. DISTRIBUTION/AVAILABILITY STATEMENT Unclassified-Unlimited Subject Category 48 112b. DISTRIBUTION CODE Report is available from the Center for AeroSpace Information (CASI), 1 800 Elkridge Landing Road, Linthicum Heights, MD 21090; (301)621-0390 13. ABSTRACT (Ma.x_rrcrn200 words) This document provides brief reports, or case studies, on a number of investigations sponsored by the Calibration and Validation Team (CVT) within the Sea-viewing Wide Field-of-view Sensor (SeaWiFS) Project. Chapter 1 describes the calibration and characterization of the GSFC sphere, which was used in the recent recalibration of the SeaWiFS instrument. Chapter 2 presents a revision of the diffuse attenuation coefficient, K(490), algorithm based on the SeaWiFS wavelengths. Chapter 3 provides an implementation scheme for an algorithm to remove out-of-band radiance when using a sensor calibration based on a finite width (truncated) spectral response function, e.g., between the 1% transmission points. Chapter 4 describes the implementation schemes for the stray light quality flag (local area coverage [LAC] and global area coverage [GAC]) and the LAC stray light correction. 14, SUBJECTTERMS 15. NUMBEROF PAGES SeaWiFS, Oceanography, Calibration, Validation, Case Studies, GSFC Sphere, 35 Diffuse Attenuation Coefficient, K(490), Out-of-Band Correction Algorithm, 16. PRICE CODE Stray Light Correction Algorithm 17. SECURITY Ct.ASSIFICATION 18. SECURITY Ct.ASSIFICATION OF REPORT OF THIS PAGE Unclassified Unclassified 19. SECURITY Ct.ASSFICATION 20. UI_rI'ATION OF ABSTRACT OF ABSTRACT Unclassified Unlimited