Comparative analysis of Sub division wise rainfall INSAT-3D vs Ground based observations

Abstract

Rainfall monitoring during south west monsoon season both are very important ad crucial activity. It is important mainly because it is a boon for agriculture, a mirror for future for both social and economic activities and crucial for its measurements (ground as well as remote). In this current works authors made an attempt to know the performance of recently Space Application Centre (SAC), ISRO developed INSAT-3D improved rainfall algorithms (Hydro Estimator and corrected IMSARA)  with actual  ground based rainfall data by calculating the bias (Actual –Satellite) for each sub-division. The analysis is done for the southwest monsoon season -2021 in by calculating weekly, monthly and seasonal bias for each subdivisions of Indian domain. It is seen that both the algorithms behave similar fashion (both show increase or decrease, simultaneously) with actual data and mostly satellite overestimate with actual data ranges from ~ 20-40 mm. In some subdivisions bias reached within the range 40 -70 mm (except Konkan & Goa ). Almost 40 % of the subdivisions have bias within 0 to 20 mm range, however the variation on weekly, monthly or seasonal differs subdivision and magnitude-wise. Overall, both the algorithms captures and performance well the trends in weekly, monthly and seasonal accumulated rainfall values. Corrected IMSRA (IMC) algorithm perform slightly better (15-20 %) except heavy rainfall episodes during the monsoon season -2021. In both the heavy and very heavy rainfall cases Hydro Estimators pick up well and performs better (~ 10 -12 %) than IMC algorithm especially over orographic areas. In extremely heavy rainfall cases both the algorithms behave in the same manner and captures the events although it is differing magnitude wise. Seasonal analysis of monsoon 2021 rainfall shows that 8 subdivisions have negative biases in the range of 50-60 mm) and 24 subdivisions have negative biases in the range of 0-20 mm, except Konkan Goa, Coastal Karnataka & A & N Islands have positive biases. Therefore, there is need to strengthen the actual observation rainfall measuring network and re-examine the performance of algorithms with larger data sets so that current algorithms retuned as per changing scenario.