Abstract
The radiometric resolution of a satellite sensor refers to the smallest increment in the spectral radiance that can be detected by the imaging sensor. The fewer bits that are used for signal discretization, the larger the quantization error in the measured radiance. In satellite inter-calibration, a difference in radiometric resolution between a reference and a target sensor can induce a calibration bias, if not properly accounted for. The effect is greater for satellites with a quadratic count response, such as the Geostationary Meteorological Satellite-5 (GMS-5) visible imager, where the quantization difference can introduce non-linearity in the inter-comparison datasets, thereby affecting the cross-calibration slope and offset. This paper describes a simulation approach to highlight the importance of considering the radiometric quantization in cross-calibration and presents a correction method for mitigating its impact. The method, when applied to the cross-calibration of GMS-5 and Terra Moderate Resolution Imaging Spectroradiometer (MODIS) sensors, improved the absolute calibration accuracy of the GMS-5 imager. This was validated via radiometric inter-comparison of GMS-5 and Multifunction Transport Satellite-2 (MTSAT-2) imager top-of-atmosphere (TOA) measurements over deep convective clouds (DCC) and Badain Desert invariant targets. The radiometric bias between GMS-5 and MTSAT-2 was reduced from 1.9% to 0.5% for DCC, and from 7.7% to 2.3% for Badain using the proposed correction method.
Highlights
The purpose of radiometric calibration is to establish the quantitative relationship between the digital measurements, or raw counts, recorded by a satellite imaging system and the spectral radiance arriving at the imaging sensor
In order to retrieve uniform geostationary Earth orbiting (GEO) cloud properties and surface fluxes, the Clouds and Earth’s Radiant Energy System (CERES) project inter-calibrates all GEO imagers using Moderate Resolution Imaging Spectroradiometer (MODIS) as a common reference based on coincident ray-matched pairs
Because the ray-matched pairs are obtained from pixel count averages over a 0.5◦ × 0.5◦ latitude by longitude grid, the cross-calibration between MODIS and Geostationary Meteorological Satellite-5 (GMS-5) is impaired by the large quantization difference between the two instruments
Summary
The purpose of radiometric calibration is to establish the quantitative relationship between the digital measurements, or raw counts, recorded by a satellite imaging system and the spectral radiance arriving at the imaging sensor. By comparing the raw counts from a target instrument (whose calibration is to be determined) with the calibrated radiances (provided by the reference instrument) that are coincident in time and identical in viewing and solar geometry, the radiometric calibration of the target instrument can be established. Such identical measurements from two satellite instruments are not feasible in reality as the two instruments always exhibit sampling differences that introduce an uncertainty in cross-calibration. The instrumental differences (spectral, spatial, radiometric resolution, etc.) between the reference and target sensors can affect the cross-calibration results [2,3]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
