GC011-02
Asymmetric shifts in dry and wet phases of the terrestrial water cycle attributed to anthropogenic climate warming

Monday, 7 December 2020: 05:34
Virtual
Kedar Otta, The University of Tokyo, Bunkyo-ku, Japan and Hyungjun Kim, The University of Tokyo, Institute of Industrial Science, Tokyo, Japan
Abstract:
With the warming climate, the mean state of the discharge, as well as its variability, is projected to change significantly in many regions globally. In this study, we investigate the asymmetric changes in discharge for wet and dry anomalous deviations in terms of magnitude, duration and severity, and attribute them to human-induced warming. Runoff simulations from 11 CMIP6 multi-model experiments (BCC-CSM2-MR, CanESM5, IPSL-CM6A-LR, MRI-ESM2-0, NorESM2-LM, ACCESS-CM2, EC-EARTH3, INM-CM4-8, INM-CM5-0, MIROC6, MPI-ESM1-2-HR) force a global river routing and inundation model, CaMa flood, to estimate river discharge for the historical all-forcing simulation (historical), historical natural-only run (hist-nat; for the first five models listed above) and two future warming conditions (SSP1-RCP2.6 and SSP5-RCP8.5). Our results show a decline in multi-model mean discharge for Amazon, North-East Brazil, Central Asia, Central North America, Central Europe, South Europe and Mediterranean, Southern Africa, in contrast to an increase in streamflow for Alaska and N.W. Canada, Eastern and Western Africa, South Asia.

Further, we discovered that, in future worlds, the magnitude in dry phase is smaller than in the wet phase for most river basins except for some drying basins like Amazon, Mississippi, Orinoco and St. Lawrence. The dry duration is in general greater than wet duration, however wet duration increases in the future for most basins, except for the drying basins.

The changes get intensified with aggravating conditions for different emission and mitigation scenarios. However, some regions show highly nonlinear shifts; opposite tendencies can also be seen in the two counterfactual worlds, for example, drying under SSP1-RCP2.6 and wetting under SSP5-RCP8.5 in the Zambezi. We conclude that there exist some nonlinearities in the changes of mean streamflow and magnitude, duration and severity of its transience. The drying (wetting) of the basins is caused most likely due to a decrease (increase) in the wet duration, which is driven by changes in rainfall and evapotranspiration patterns. Fractional risk changes of flooding and severity of droughts due to human-induced warming have tipping points between historical to SSP1-RCP2.6 scenario over large areas of mid- and high-latitudes.