Item

Abstract

Mediterranean regions are widely considered to be one of the most vulnerable areas to the impacts of climate change, especially in terms of experiencing longer and more severe droughts. In recent years, there has been significant investment in generating more accurate climate projections, with the goal of providing hydrological model users with future climate scenarios. However, hydrological models are often driven by a set of climate forcings, without subjecting the model itself to stress testing its parameters and the simulated processes. We present a new perspective by focusing on stress testing the adaptable rainfall-runoff model JAMS/J2000 for Mediterranean South Africa. Resulting from 7 years of research and development, our systematic approach provides valuable insights into data requirements, simulated hydrological process variability, factors impacting model uncertainty, and methods to improve hydrological process simulation and model robustness. We evaluated: 1) the impact of precipitation station density on model performance and recommendations of minimum number of precipitation station, 2) factors which impact model uncertainty (catchment size, streamflow signal, anthropogenic factors such as dams), 3) driving forces behind reduced model performance and the potential influence of hydrological process changes and 4) the capability of the model to detect different drought types and the most reoccurring seasonal precipitation shortfalls. Finally, we introduce a new hydrological model using stable isotopes and End-Member-Mixing and aim to further develop robust isotope-enabled modelling tools. We conclude that stress testing the hydrological model is as important as the climate inputs and more effort should be placed on developing robust models.