Radiometric Measurement Comparison on the Integrating Sphere Source Used to Calibrate the Moderate Resolution Imaging Spectroradiometer (MODIS) and the Landsat 7 Enhanced Thematic Mapper Plus (ETM+)

Abstract

As part of a continuing effort to validate the radiometric scales assigned to integrating sphere sources used in the calibration of Earth Observing System (EOS) instruments, a radiometric measurement comparison was held in May 1998 at Raytheon/Santa Barbara Remote Sensing (SBRS). This comparison was conducted in support of the calibration of the Moderate Resolution Imaging Spectroradiometer (MODIS) and the Landsat 7 Enhanced Thematic Mapper Plus (ETM+) instruments. The radiometric scale assigned to the Spherical Integrating Source (SIS100) by SBRS was validated through a comparison with radiometric measurements made by a number of stable, well-characterized transfer radiometers from the National Institute of Standards and Technology (NIST), the National Aeronautics and Space Administration's Goddard Space Flight Center (NASA's GSFC), and the University of Arizona Optical Sciences Center (UA). The measured radiances from the radiometers differed by ±3 % in the visible to near infrared when compared to the SBRS calibration of the sphere, and the overall agreement was within the combined uncertainties of the individual measurements. In general, the transfer radiometers gave higher values than the SBRS calibration in the near infrared and lower values in the blue. The measurements of the radiometers differed by ±4 % from 800 nm to 1800 nm compared to the SBRS calibration of the sphere, and the overall agreement was within the combined uncertainties of the individual measurements for wavelengths less than 2200 nm. The results of the radiometric measurement comparison presented here supplement the results of previous measurement comparisons on the integrating sphere sources used to calibrate the Multi-angle Imaging SpectroRadiometer (MISR) at NASA's Jet Propulsion Laboratory (JPL), Pasadena, CA and the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) at NEC Corporation, Yokohama, Japan.

Keywords: Earth Observing System (EOS); Landsat 7 Enhanced Thematic Mapper Plus (ETM+); Moderate Resolution Imaging Spectroradiometer (MODIS); calibration; integrating sphere; remote sensing; spectral radiometry; transfer radiometers.

Fig. 1

Spectral radiances for the SIS100 as determined by SBRS in the visible/near infrared for the principal radiance levels in the May 1998 comparison for (a) the three brightest levels and (b) the three dimmest levels.

Fig. 2

Spectral radiances for the SIS100 as determined by SBRS in the shortwave infrared for the principal radiance levels in the May 1998 comparison for (a) the two brightest levels (configuration (10-9-11) was not measured) and (b) the three dimmest levels.

Fig. 3

Schematic showing the positioning of the EOS VXR and the UA VNIR on May 14 and 15 for the off-axis monitoring of the SIS100.

Fig. 4

Results of the visible and near infrared radiometer measurements of the SIS100, shown as the percent difference of the band-average radiances measured by the transfer radiometers LB, XR with the band-averaged SIS100 radiances LB, SIS as the reference, for the six principal SIS100 levels: (a) (10-9-11); (b) (10-9-4); (c) (10-9-0); (d) (10-5-0); (e) (10-0-0); and (f) (4-0-0).

Fig. 5

Results from Fig. 4 plotted versus sphere radiance for two spectral regions: (a) near 440 nm, and (b) near 550 nm. For each transfer radiometer, the abscissa gives the calculated radiance, based on the SBRS values (note the scale break). These radiances correspond to the six principal SIS100 lamp configurations, from (10-9-11), the brightest, to (4-0-0), the dimmest.

Fig. 6

Results from Fig. 4 plotted versus sphere radiance for two spectral regions: (a) near 660 nm, and (b) near 870 nm. For each transfer radiometer, the abscissa gives the calculated radiance, based on the SBRS values (note the scale break). These radiances correspond to the six principal SIS100 lamp configurations, from (10-9-11), the brightest, to (4-0-0), the dimmest. For the GSFC LXR, there is no comparison for the 827 nm channel (see text for details).

Fig. 7

Results of the shortwave infrared radiometer measurements of the SIS100, shown as the percent difference of the band-average radiances measured by the transfer radiometers LB, XR with the band-averaged SIS100 radiances LB, SIS as the reference, for the six principal SIS100 levels: (a) (10-9-4); (b) (10-9-0); (c) (10-5-0); (d) (10-000); and (e) (4-0-0).

Fig. 8

The results of Fig. 7a overlaid with the atmospheric transmittance for a 2 m path through the atmosphere, with conditions of 1 × 10⁵ Pa, 25 °C, and 30 % relative humidity. In the model, the absorption is due to water and carbon dioxide.

Fig. 9

Results of the radiometer measurements of the SIS100 for the six principal levels: (a) (10-9-4); (b) (10-9-0); (c) (10-5-0); (d) (10-0-0); and (e) (4-0-0). The results are as in Fig. 7, except that measurements in the atmospheric absorption intervals—1350 nm to 1500 nm and 1850 nm to 2000 nm—have been removed.

Fig. 10

Results from Fig. 9 plotted versus sphere radiance for two spectral regions: (a) near 950 nm, and (b) near 1250 nm. For each transfer radiometer, the abscissa gives the calculated radiance, based on the SBRS values (note the scale break). These radiances correspond to the five principal SIS100 lamp configurations, from (10-9-4), the brightest, to (4-0-0), the dimmest.

Fig. 11

Results from Fig. 9 plotted versus sphere radiance for two spectral regions: (a) near 1650 nm, and (b) near 2200 nm. For each transfer radiometer, the abscissa gives the calculated radiance, based on the SBRS values (note the scale break). These radiances correspond to the five principal SIS100 lamp configurations, from (10-9-4), the brightest, to (4-0-0), the dimmest.

Fig. 12

EOS VXR results of the off-axis measurements of the SIS100 for configurations where lamps of the same wattage were turned off sequentially. The measurements began on May 14 with 14 of the 200 W lamps, all of the 8 W lamps, and all of the 45 W illuminated, see (a); then individual 45 W lamps were extinguished, see (c); and then individual 8 W lamps were extinguished, see (d). In (b), we show the results for the four bright levels, (10-9-18) to (10-9-14), that were acquired on May 15 along with the May 14 results of the 200 W configurations.

Fig. 13

The results of Fig. 12 plotted as a function of number of lamps in the SIS100 that were illuminated for the measurements on May 14 and 15. The lines are to guide the eye. Configurations with one to nine 8 W lamps illuminated have abscissa values from 1 to 9; configurations with all 8 W and one to ten 45 W lamps illuminated have abscissa values from 10 to 19; and configurations with all 8 W, all 45 W and one to eighteen 200 W lamps illuminated have abscissa values from 20 to 37.

Fig. 14

UA SWIR results of the off-axis measurements of the SIS100 for configurations where lamps of the same wattage were turned off sequentially. The measurements began on May 14 with 4 of the 200 W lamps, all of the 8 W lamps, and all of the 45 W illuminated, see (a); then individual 45 W lamps were extinguished, see (b); and then individual 8 W lamps were extinguished, see (c).

Fig. 15

The results of Fig. 14 plotted as a function of number of lamps in the SIS100 that were illuminated for the measurements on May 14. The lines are to guide the eye. Configurations with one to nine 8 W lamps illuminated have abscissa values from 1 to 9; configurations with all 8 W and one to ten 45 W lamps illuminated have abscissa values from 10 to 19; and configurations with all 8 W, all 45 W and one to four 200 W lamps illuminated have abscissa values from 20 to 23.

Fig. 16

Repeatability of the SIS100 and EOS VXR measurement system for the six principal levels: (a) (10-9-11), (b) (10-9-4), (c) (10-9-0), (d) (10-5-0), (e) (10-0-11), and (f) (4-0-0). The EOS VXR results are normalized to the first measurement and plotted as a function of measurement time.

Fig. 17

Typical results of the VXR as an off-axis monitor during the measurement of the six principal levels by the other radiometers for (a) lamp configuration (10-9-4) on May 12 and (b) lamp configuration (4-0-0) on May 13; note the scale change in the ordinate. All six channels of the VXR are shown as a function of the measurement date and time. For each channel, the results were normalized to the initial value and as plotted represent the change in percent.

Fig. 18

Comparisons, using the EOS VXR, of the direct measurements of lamp configurations with more than one lamp wattage type and the sum of separate measurements of the corresponding configurations with only lamps of a similar wattage illuminated. The results are shown as percent difference versus VXR wavelength for (a) configurations (10-9-14) to (10-9-5), (b) configurations (10-9-4) to (10-9-1), and (c) configurations (10-9-0) to (10-1-0).

Fig. 19

Comparisons, using the UA SWIR, of the direct measurements of lamp configurations with more than one lamp wattage type and the sum of separate measurements of the corresponding configurations with only lamps of a similar wattage illuminated. The results are shown as percent difference versus UA SWIR wavelength for (a) configurations (10-9-4) to (10-9-1), and (b) configurations (10-9-0) to (10-1-0).