Bottom-up assessment of climate change vulnerability of a large and complex river basin using emulator models

The Murray-Darling Basin (MDB) is Australia’s most significant river system. Its’ scale and complexity are such that climate change impact assessments using traditional water resource modelling have been restricted to a small number of scenarios, which limits the understanding of climate uncertainty and hinders the identification of robust adaptation responses. This study implements a bottom-up climate vulnerability assessment for the MDB. The approach is designed to overcome the computational and data constraints of traditional top-down methods and offers insights into system robustness to climate uncertainty. It uses computationally efficient machine learning-based emulators, trained on outputs from complex water resource models, to conduct the extensive simulations required. The emulators, driven by conceptual rainfall-runoff models, enable the rapid simulation of the regulated river system to explore a wide range of climate uncertainties, which retaining high accuracy. The bottom-up assessment reveals significant system sensitivities, and non-linearities and thresholds in how ecological metrics respond to climate change. Results contrast differences in hydrological response across the north and south MDB. A key insight is the importance of precipitation reductions of 15 %, which represents a threshold beyond which the long-term performance of environmental targets is significantly compromised across the basin. Such outcomes may be missed in traditional top-down assessments but are crucial for future planning to develop robust water management practices. • Developed emulators to efficiently model complex river system behaviour. • Used 1089 scenarios to stress-test Murray-Darling Basin climate vulnerabilities. • Highlighted differing sensitivities and spatial variation across the basin. • Identified thresholds where flow reductions compromise environmental objectives. • Demonstrated emulators’ utility in supporting bottom-up climate assessment.