Benefits of Calibrating a Global Hydrological Model for Regional Analyses of Flood and Drought Projections: A Case Study of the Yangtze River Basin

Abstract

Uncalibrated global hydrological models are primarily used to inform projections of flood and drought changes under global warming and their impacts, but it remains unclear how model calibration might benefit these projections. Using the Yangtze River Basin as a case study, we compare projected changes in flood and drought frequencies and their impacts—area, population, and gross domestic product affected—at various warming levels, from uncalibrated and calibrated simulations with the Community Water Model. These projections are driven by 10 General Circulation Models (GCMs) from Coupled Model Intercomparison Project Phase 6, within the Inter-Sectoral Impact Model Intercomparison Project framework. Calibration significantly improves simulated discharge, yet the impact of calibration under climate change on projected increases in flood frequency and their associated impacts is minor, in contrast to its notable role in drought projections. We further quantify the relative contribution of GCMs, emission scenarios, and calibration approaches to the projected impacts, finding that GCMs primarily drive projected flood changes, while emission scenarios and calibration contribute more significantly to the variance in drought projections after 2050. The differing sensitivities to calibration are attributed to the dominance of extreme precipitation in flood generation and the influence of long-term evapotranspiration trends on drought occurrence. The findings imply that future projections of relative changes in flood frequency and risks based on uncalibrated hydrological models are likely still quite reliable for warm and humid regions. However, careful calibration and model improvement is crucial for enhancing the reliability of future drought impact assessments.