How Climate Extremes Influence Conceptual Rainfall-Runoff Model Performance and Uncertainty

Rainfall-runoff models are frequently used for assessing climate risks by predicting changes in streamflow and other hydrological processes due to anticipated anthropogenic change, variability, land management. Historical observations commonly calibrate empirically the performance of conceptual mechanisms. As a result, calibration procedures limited when extrapolated novel conditions under future scenarios. In this paper, rainfall-runoff model simulated catchment were explored using JAMS/J2000 Berg River South Africa evaluate tails current distribution climatic conditions. An evolutionary multi-objective search algorithm was develop sets parameters which best simulate “wet” “dry” periods, providing upper lower bounds temporal uncertainty analysis approach identify variables affected these extremes. Variables most included soil-water storage timing interflow groundwater flow, emerging as overall dampening hydrograph. Previous modeling showed that provided “good” simulation periods where yearly long-term mean precipitation shortfall <28%. Above threshold, autumn reduced 50%, paper shows use set is recommended improve performance. These better account change concentration peak flows, occur drier winter years, improving Nash-Sutcliffe Efficiency (NSE) from 0.26 0.60 validation period 2015–2018, although availability data still potential factor. (NSE > 0.7) during calibration, not catchment, but could play large role tropical climates. The results study likely transferrable rainfall/runoff models, may differ various greater variability drives around world, empirically-based projections need assumptions regarding processes, enhanced risk