High flows in the 21st Century: analysis with a simple conceptual hydrological models using the input of 3 GCMs (A2 scenario)

Abstract

The study on high flows with a conceptual hydrological model leads to the following conclusions for about 1500 randomly selected land points across the world that have an intermediate soil water supply capacity and an intermediary responding groundwater system: · the probability distributions of Q10 (flow that is equalled or exceeded 10% of the time) for the three GCMs (ECHAM. CNRM, IPSL) are very similar for the control period (1971-2000), irrespective of the Köppen-Geiger climate type, which means that the downscaled, bias-corrected climate data from the climate models for the last part of the 20th century only slightly deviate; · median Q10s derived from the GCMs for the CTRL period differ at least tens of per cent from the Q10 obtained when using WFD as input for the hydrological model (i.e. about 30-60%). The differences for the arid and polar (B- and E-) climates are higher because of the low discharge, and for the B-climate also due to overestimation of the high Q10s (about 100% or more); · the impact of climate change (median of Q10 as metric) is projected to increase over the 21st century according to all three GCMs (A2 scenario) for the C-, D- and E-climates. This is also expected for the A- and B-climates according to ECHAM and CNRM. IPSL provides for these two Köppen-Geiger major climate regions a more diffuse projection; · the projected increase of the median Q10 because of climate change (A2 scenario) varies between about 5 and 20% for the near future (2021-2050), irrespective of the Köppen-Geiger major climate region and the GCM. The only exception is IPSL that expects a decrease of the Q10 by about 5% for the equatorial climate; · the projected increase of the median Q10 by the end of the 21st century (2071-2100) varies between about 15 and 60%, ), irrespective of the Köppen-Geiger major climate region and the GCM. The only exception again is IPSL. This model expects a decrease of the median Q10 by about 2-10% for the equatorial and arid climates (A- and B-climates; · difference between the median Q10 derived from the hydrological model forced with re-analysis data (WFD) and the same model forced with the down-scaled, bias corrected climate output from three climate models for the control period (1971-2000) clearly exceeds the differences found by climate change in the 21st century. The analysis of high flows as obtained with the rather simple hydrological model in this study should be extended with: · the outcome from the GCMs for the B1 scenario; · an analysis at the global scale of the outcome from the suite of large-scale models that have been run within the WATCH project similar to the study of Miller et al. (2011) for Europe.