The first National Aeronautics and Space Adminis- tration (NASA)/National Institute of Standards and Technology (NIST) Earth Observing System (EOS)- sponsored spectral radiometric measurement comparison experiment was conducted at Hughes Santa Barbara Remote Sensing (SBRS) in Goleta, California, and at the Jet Propulsion Laboratory (JPL) in Pasadena, California, from 12 August to 20 August, 1996. Radiance measurements were made by several participants on three integrating sphere sources. These sphere sources are used in the pre-flight radiance calibration of the EOS Moderate Resolution Spectroradiometer (MODIS) and Multi-angle Imaging SpectroRadiometer (MISR) and the Landsat-7 Enhanced Thematic Mapper+ (ETM+). The Optical Technology Division of NIST was the lead laboratory, coordinating the measurement plan and activities with the EOS Calibration Scientist. The measurement sequence was designed to address the issues of repeatability, evaluation of unknown systematic effects, and stability. Due to the limited time available at Hughes SBRS and at JPL, simplification in terms of the number of instruments and sphere output levels was necessary. The goals were, in order of importance, to: 1) compare the spectral radiance of the sphere sources as calibrated by the EOS instrument providers with that determined by NIST using NIST- calibrated radiometers; 2) compare the spectral radiance determined by the participants from the outside laboratories using the sphere sources as common targets; and 3) evaluate the findings in terms of measurement procedure and basic metrology.
At Hughes SBRS, the Spherical Integrating Source 100 (SIS100) used in the radiometric calibration of MODIS was measured at four different levels by five teams of researchers over a three-day interval. On the fourth day, the SIS 122, which is used in the radiometric calibration of ETM+, was measured by the same participants. At JPL, five teams spent three measurement days recording the output of the SIS 165 source. The SIS 165 was recently used to calibrate the nine MISR cameras.
At Hughes SBRS, the participants were: 1) NASA/GSFC-EOS (John Cooper and Jim Butler), with a scanning single grating monochromator that measured from 400 nm to 2500 nm; 2) NASA/GSFC-ETM+ (Brian Markham and Ken Brown), with the recently acquired
Landsat Transfer Radiometer (LXR); 3) NIST (Carol Johnson), with the SeaWiFS Transfer Radiometer (SXR) and the EOS Visible Transfer Radiometer (VXR); 4) the University of Arizona (UA) (Stuart Biggar and Paul Spyak), with the UA visible/near infrared (VNIR) and UA shortwave infared (SWIR) transfer radiometers; and 5) National Research Laboratory of Metrology (NRLM) (F. Sakuma and J. Ishii) with three Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) VNIR transfer radiometers and two ASTER SWIR transfer radiometers. The LXR, SXR, and VXR have six image locations and separate interference filter/detectors at each location. The interference filters are narrow band (~10 nm) with the exception of four of the filters in the LXR, which are similar to the ETM+ flight filters. The UA VNIR and UA SWIR use a rotating filter wheel to alternately measure at select wavelengths that correspond to MODIS and ASTER bands. The NRLM radiometers are separate units, each making measurements at or near particular ASTER bands. Hughes SBRS calibrated the Landsat SIS 122 sphere prior to and during the comparison measurements on the SIS 100. They planned to calibrate the SIS 100 immediately following the comparison starting on 16 August.
Each day consisted of a series of measurements followed by reporting of preliminary results. The typical measurement procedure was to turn the SIS 100 to a given radiant level determined by the specific lamps illuminated, measure using the SXR, then measure using the other participants' radiometers, and then repeat the SXR measurement. In this manner, over the complete course of the comparison, the SIS 100 was measured at the same level at least twice by NIST and in most cases twice by other participants. During the time the other participants were measuring the sphere, the VXR was used off- axis in order to monitor the stability of the sphere. Linda Fulton of Hughes SBRS adjusted the sphere lamp currents according to Hughes SBRS procedures, and records were made of the lamp currents and voltages during the exercise. Two levels were measured in this fashion each day, except for 12 August, where the morning was devoted to organization and laboratory preparation. In addition to the above measurements, a particular level was selected for study with the SXR every day so that the sphere repeatability could be assessed. A single day (i.e., 15 August) was devoted to measurements of the SIS 122. In order to assess the SIS 122 repeatability, the same level was measured by all participants in the morning and afternoon, with the sphere turned off in between. A second level was measured using only the SXR, the LXR, and then the SXR. At the end of each measurement day, the participants reported and discussed preliminary results. Since SBRS had not provided their SIS 100 calibration data beforehand, Hughes representatives were excluded from presentation and discussions of any preliminary results. These preliminary results indicated a very reasonable level of internal consistency in absolute spectral radiance among the outside laboratories (i.e., 1% to 2% scatter). The Landsat 7 ETM+ team was unable to report absolute results at the time of the comparison because the LXR was not characterized or calibrated. The SIS 100 appeared to be stable and, except in one measurement where a 1% shift was observed in the blue, repeatable.
The comparison participants packed their equipment late on 15 August and traveled to JPL on 16 August. The afternoon of 16 August was devoted to unpacking and cleaning equipment at JPL and deploying instruments in the MISR cleanroom. At JPL, the participants were: 1) NASA/GSFC (John Cooper and Jim Butler), with a scanning single grating monochromator operating from 400 nm to 1100 nm; 2) NIST (Carol Johnson), with the SXR and the VXR; 3) the University of Arizona (Stuart Biggar), with the UA VNIR; 4) NRLM (F. Sakuma and J Ishii) and NEC (K. Suzuki) with three ASTER VNIR transfer radiometers and a commercial single grating monochromator that utilizes a diode array; and 5) JPL (Carol Bruegge and Dan Preston), with the MISR laboratory standard radiometer utilizing 4 interchangeable visible filters.
Radiance measurements were made on the SIS 165 on 17, 19, and 20 August with results being reported on 20 August. In accordance with JPL calibration and operation procedures for the SIS 165, the measurement technique was changed from that employed at Hughes SBRS. Instead of operating the sphere at a single level for several hours while participants made measurements, the sphere was turned on to the brightest level of a set of four designated levels to be measured, allowed to warm up for 20 minutes, measured by one participant, turned to the next brightest level of the set, and so on. On 20 August the comparison participants decided to operate and measure the SIS 165 at a single level in order to assess the long-term stability of the sphere. This approach also provides a more accurate comparison of the participating radiometers. As at Hughes SBRS, during the time the other participants were measuring the sphere, the VXR was used off- axis in order to monitor the stability of the sphere. However, on the day at JPL that a single level was selected for study, the SXR was used off- axis as the monitor. At NIST's request, records were made of the lamp currents and voltages during the exercise and were provided to comparison participants. The preliminary results on the SIS 165 indicate up to an 8% spread in absolute spectral radiance among the outside laboratories. The SIS 165, which was measured at four levels, also exhibited drifts of up to 0.08%/min. The preliminary data also indicate that the SIS 165 is non- lambertian at the 1% to 3% level, as measured by the VXR or the SXR at angles up to 50 degrees from normal incidence. This, coupled with the sphere drift, complicates the analysis of these data.
NIST is currently coordinating the data analysis from this radiometric measurement comparison through its Statistical Engineering Division. All comparison participants were given a list of items needed by NIST (e.g., description of radiometers, raw data files, detailed measurement log sheets, etc.) for accurate analysis and reporting of comparison results. Most of this information has been obtained and mounted on a server at NIST. NIST plans to re-examine the raw data, and compare the results to those determined by the participants based upon their algorithms for analyzing the radiometer data. A draft report including the comparison results and findings is planned by the end of 1996.