Evaluating the radiometric consistency of INSAT-3DR and Meteosat-8 TIR observations

Geostationary (GEO) satellites are paramount for continuously monitoring our planet from a fixed position because of their high temporal resolution. It is essential to monitor the quality of these observations before they are used in retrieval or data assimilation. The Global Space-based Intercalibration System calibrates GEO satellite observations against low-Earth-orbit (LEO) observations; however, very few studies directly compare GEO-GEO observations . GEO-GEO intercomparisons also offer valuable insight into inter-sensor consistency on a diurnal scale and, more robustly, on a seasonal scale. In this work, we perform a detailed inter-comparison of the Brightness Temperatures (BTs) of the thermal infrared window channels of two GEO sensors, INSAT-3DR Imager and SEVIRI onboard Meteosat-8, using more than 1.5 billion collocated clear-sky observations. We also compute simulated observations for both these satellites using a radiative transfer model to quantify the differences between observed and simulated (O − M) BTs and their variability. We found INSAT-3DR observations are 1.21 K (at 10.8 µm) and 0.83 K (at 12 µm) warmer than the Meteosat-8 observations, with corresponding mean model BT differences of 0.96 K and 0.84 K, respectively. A sensitivity test revealed that these systematic BT differences are mainly due to satellite viewing geometry and atmospheric total column water vapour. The bias-corrected INSAT-3DR observations showed a residual mean diurnal bias of ∼ 1 K, whereas it was ∼ 0.1 K for Meteosat-8 in both channels. We also implemented a double-difference technique to investigate the cross-sensor agreement. We found a calibration anomaly in INSAT-3DR observations during Oct-Nov 2021 and Mar-Apr 2022. However, Meteosat-8 observations were radiometrically stable, highlighting their utility for correcting anomalies/biases in INSAT-3DR observations. In summary, this study is valuable for determining the accuracy of observations in the window channels, which further affects the retrieval of geophysical parameters, such as SST and its diurnal variation.