Regional scale climate change impact assessment efforts to reduce and quantify the predictive uncertainty

Abstract

The thesis focuses on addressing the crucial need for quantifying uncertainties in newlineregional-scale climate change impact studies and devising effective strategies to reduce uncertainties arising from the modelling stages. Impact estimates that fail to adequately communicate the uncertainties can result in inadequate planning and negligent decisions, leading to significant loss of life and property. The Bharathapuzha river basin, situated along the Western Coast of India, has been identified as a climate newlinechange hotspot in Kerala, India. Over the past century, the annual rainfall over the newlinecatchment has seen a reduction of 2.9 mm per year, indicating an overall drying up of newlinethe catchment. The present study endeavours to analyse the potential impacts of climate change on the Bharathapuzha catchment, while keeping in consideration the uncertainties involved in the assessment procedure. The impact assessment studies are based on the projections of General Circulation Models (GCMs) which represent the physical processes on land, air and the ocean that drive the climate using mathematical formulations. GCMs that have significant deficiencies in their newlinerepresentation of regional climate physics may simulate unreasonable climate for the newlinefuture period leading to large uncertainties in the output. Hence, the selection of newlineclimate models for regional scale impact assessment is important. Current study newlinedeveloped a climate model selection procedure based on their ability to simulate the newlineregional climate phenomena, historical climate, and the mutual interdependence of newlineclimate models. The proposed method is validated on the Western Coast of India newline(where Bharathapuzha River Basin is located) and the GCMs CESM1-BGC, CMCC_ESM2, FGOALS-G2, FIO_ESM_2_0 and MIROC4H are identified to be the better performing climate models for analysing climate change.