Assessment Of SDSM 4.2 For Generating Regional Climate Projections Under Multiple Emission Scenarios: A Case Study Of Gangapur, Maharashtra India

Abstract

Climate change threatens natural resource sustainability in semi-arid regions such as Ganagapur, Nashik, Maharashtra, where agriculture and water availability are highly sensitive to monsoon variability. Reliable regional climate projections are essential for adaptation planning, but the coarse resolution of General Circulation Models (GCMs) limits local applicability. Statistical downscaling methods like the Statistical DownScaling Model (SDSM) 4.2 help generate finer-scale data suitable for regional analysis. This study evaluates the performance of SDSM 4.2 in downscaling daily temperature and rainfall using historical data (1961–2000) and projections from HadCM3 (CMIP3), CanESM2 (CMIP5), and CanESM5 (CMIP6) under SRES, RCP, and SSP scenarios. Results show high accuracy for temperature, especially with HadCM3 (R² ≥ 0.99). CanESM2 performed well for Tmin (R² ≥ 0.92) but was less effective for Tmax and rainfall (R² ≤ 0.59). CanESM5 projected extreme warming under SSP5-8.5 (Tmax +6.1°C, Tmin +7.77°C by 2080s), though its historical fit was weak, particularly for rainfall (R² ≤ 0.5).

Moderate pathways, including RCP 4.5 and SSP2-4.5, emerged as balanced scenarios. Rainfall uncertainties highlight the need for ensemble or hybrid approaches. The findings underscore urgent adaptation measures for water and agriculture in Gangapur, consistent with IPCC AR6 projections for South Asia.