Are we using the right fuel to drive hydrological models? A climate
impact study in the Upper Blue Nile
Stefan Liersch1, Julia Tecklenburg1, Henning Rust2, Andreas Dobler2, Madlen Fischer2, Tim Kruschke3, Hagen Koch1, and Fred Hattermann11Potsdam Institute for Climate Impact Research (PIK), Telegraphenberg A31, 14473 Potsdam, Germany 2Free University of Berlin (FUB), Institute of Meteorology, Carl-Heinrich-Becker-Weg 6-10, 12165 Berlin, Germany 3GEOMAR Helmholtz Centre for Ocean Research Kiel, Germany
Received: 18 Aug 2016 – Accepted for review: 23 Sep 2016 – Discussion started: 26 Sep 2016
Abstract. Climate simulations are the fuel to drive hydrological models that are used to assess the impacts of climate change and variability on hydrological parameters, such as river discharges, soil moisture, and evapotranspiration. Unlike with cars, where we know which fuel the engine requires, we never know in advance what unexpected side-effects might be caused by the fuel we feed our models with. Sometimes we increase the fuel's octane number (bias-correction) to achieve better performance and find out that the model behaves differently but not always as was expected or desired. This study investigates the impacts of projected climate change on the hydrology of the Upper Blue Nile catchment using two model ensembles consisting of five global CMIP5 Earth System Models and ten Regional Climate Models (CORDEX Africa). WATCH forcing data were used to calibrate an eco-hydrological model and to bias-correct both model ensembles using slightly differing approaches. On the one hand it was found that the bias-correction methods considerably improved the performance of average rainfall characteristics in the reference period (1970–1999) in most of the cases. This also holds true for non-extreme discharge conditions between Q20 and Q80. On the other hand, bias-corrected simulations tend to overemphasise magnitudes of projected change signals and extremes. A general weakness of both uncorrected and bias-corrected simulations is the rather poor representation of high and low flows and their extremes, which were often deteriorated by bias-correction. This inaccuracy is a crucial deficiency for regional impact studies dealing with water management issues and it is therefore important to analyse model performance and characteristics, the effect of bias-correction, and eventually to exclude some climate models from the ensemble. However, the multi-model means of all ensembles project increasing average annual discharges in the Upper Blue Nile catchment and a shift in seasonal patterns, with decreasing discharges in June and July and increasing discharges from August to November.
Liersch, S., Tecklenburg, J., Rust, H., Dobler, A., Fischer, M., Kruschke, T., Koch, H., and Hattermann, F.: Are we using the right fuel to drive hydrological models? A climate
impact study in the Upper Blue Nile, Hydrol. Earth Syst. Sci. Discuss., doi:10.5194/hess-2016-422, in review, 2016.