Validation of INSAT-3DR sounder retrieved temperature profile with GPS radiosonde and AIRS observations

Abstract

Abstract Profiles of Temperature (1000 hPa to 0.1 hPa at 40 pressure levels) retrieved by physical approach algorithm in cloud-free areas from Indian National Satellite-3D repeat (INSAT-3DR) satellite has been validated with ground-based in situ measurements from GPS based radiosonde measurements and then the same is compared with Atmospheric Infrared Sounder (onboard on NASA Aqua Satellite) derived profiles of temperature. The accuracy of these retrievals is a challenge and backbone of improvements in forecast accuracy. The statistical analysis is done by using INSAT-3DR and Atmospheric Infrared Sounder (onboard on NASA Aqua Satellite) data (September –December-2020). The retrieval performance and accuracy are examined in terms of biases, correlation coefficients (CC), and root means square error (RMSE) which is carried out over the land, hilly and coastal areas and found slight variations in the lower middle and upper levels of the troposphere. The key findings of this study INSAT −3DR vs AIRS are as follows: The correlation coefficient (CC) is greater than 0.7 above 15° N at all pressure levels except 150 hPa (CC ∼ 0.2 between 22° N to 35° N). The CC values are random (0.2 to 0.7) below 15° N to 7° N at all pressure levels. The warm bias of the order of (∼0.0 K to 2.5 K) from 1000 to 600 hPa over Indian land and ocean region except for some areas/, 25° N to 35° N have cold biases of the order of −2.0 K to −4.0 K. Himalayan region have a warm bias - ranges 0.0 to 4.0 K at all pressure levels. Similarly, the findings of INSAT-3R & GPS radiosonde are as follows: All the coastal stations show the random bias of the order of ∼ 3.0 K to + 3.0 K hPa above 100 hPa. RMSE value observed of ∼ 4.0 K for the stations (land, hilly, coastal) at each pressure level. The correlation coefficient (CC) of land and hilly regions is higher (0.7 to 1.0) as compared to coastal stations (0.5 to 0.6). INSAT-3DR derived temperature profiles from surface to 10.0 hPa have large bias overland (plain and desert) ranges from −3.0 K to + 3.0 K as compared to coastal and hilly regions −2.5 K to + 2.5 K, which may be attributed to the large surface emissivity values over land. In general, a positive bias of the order of (+1 to + 3 °K) from 1000 to 800 hPa, ± (1 to 2 °K) at 800 to 200 hPa & ± (1 to 3 °K) above 200 hPa have been observed with INSAT-3DR and radiosonde data. This study is found very useful in daily operational weather forecasting and analysis of weather systems at different temporal and spatial scales with more confidence.