Comparison of perturbation methods for rainfall and temperature data: case of a Belgian catchment

Abstract

Analyses of hydrological impacts of climate change require appropriate methods for perturbing meteoric time-series to represent future climate conditions. Two readily available tools for perturbing rainfalls and temperatures are tested for a medium-sized catchment in Belgium. CCI-HYDR provides three scenarios, tailored for Belgium every decade until 2100. In contrast, KNMI-ADC tool provides 191 scenarios, at a regional level and for two horizons (near and far future). With its three contrasted scenarios of possible future climate conditions, CCI-HYDR is found suitable for forcing computationally intensive detailed hydrological models. With its broader spectrum of climate scenarios, KNMI-ADC tool is suitable for forcing multiple runs of fast conceptual hydrological models. As the two perturbation tools deliver stationary time-series, they are also compared to an alternate method producing transient time-series. Transient stochastic tools are particularly computationally demanding due to their stochastic nature, which is not optimal when combined with detailed distributed hydrological models.