Analysis of MODIS–MISR calibration differences using surface albedo around AERONET sites and cloud reflectance

Abstract

MODIS and MISR are two Earth Observing System instruments flown onboard the Terra satellite. Their synergistic use could greatly benefit the broad user community by ensuring a global view of the Earth with high-quality products. A necessary condition for data fusion is radiometric calibration agreement between the two instruments. Earlier studies showed about 3% absolute radiometric difference between MISR and respective MODIS land bands in the visible and near-IR spectrum, which are also used in aerosol and cloud research. This study compared two surface albedo products derived from MODIS and MISR L1B data using the AERONET-based Surface Reflectance Validation Network (ASRVN). The ASRVN shows a positive MISR–MODIS albedo bias of + (0.01–0.03). Cross-sensor calibration inconsistencies were identified as a primary cause of the albedo biases. To establish an independent MODIS–MISR calibration link, top-of-atmosphere MODIS and MISR reflectances were regressed against each other over liquid water clouds. The empirical regression results have been adjusted for the differences in the respective MISR and MODIS spectral responses using radiative transfer simulations. The MISR–MODIS band gain differences for the top-of-atmosphere reflectance estimated with this technique are + 6.0% in the Blue, + 3.3% in the Green, + 2.7% in the Red, and + 0.8% in the NIR band. Applying the derived values to rescale the MODIS or MISR L1B data is shown to significantly reduce the cross-sensor ASRVN surface albedo biases. An absolute calibration scale for both sensors could be established based on independent ground-based measurements of the surface albedo at selected AERONET sites.