Time series analysis of hydrological drought under climate change with anthropogenic water management

Monday, 15 December 2014
Yusuke Satoh, The University of Tokyo, Tokyo, Japan, Kei Yoshimura, Atmosphere and Ocean Research Institute University of Tokyo, Tokyo, Japan, Yadu N Pokhrel, Michigan State University, Department of Civil & Environmental Engineering, East Lansing, MI, United States, Hyungjun KIM, The University of Tokyo, Tokyo, Japan and Taikan Oki, University of Tokyo, Bunkyo-ku, Japan
Human society have altered terrestrial hydrological cycles by water management infrastructure, such as reservoirs and weirs for irrigation, in order to enable stable water use against natural variability. On the other hand, anthropogenic climate change is projected to alter the hydro-meteorological cycles, and it is projected that drought frequency and/or intensity will increase in some regions. Thus reliable projection is a critical issue for our society in order to adapt for the change. However, only few studies have investigated the effect of anthropogenic intervention on drought under climate change.

This study focuses on hydrological drought, particularly on stream flow, as stream flow is one of the most easy-to-access water resource. HiGW-MAT, a state of arts land surface model capable to reproduce energy and water cycle considering the anthropogenic water management, is used to simulate the historical and future terrestrial water cycles. The model includes reservoir operation, water withdrawal and irrigation process. Five CMIP5 GCM outputs with bias-correction provided by ISI-MIP for 1980-2099 are used to force a set of simulations. Time series data of global hydrological drought for 120 years, with and without human activity, is analyzed in order to estimate the impact of climate change and the adaptation capacity of anthropogenic water management. It is identified that Europe, Central and Eastern Asia, East and West part of USA, Chile, Amazon basin and Congo basin will have large increases of drought more than 90 days. According to uncertainty check particular increases in Central USA and Southern and Eastern South America have high robustness. Dividing global land into 26 regions, we characterized the variation of drought time series for each region. Drought does not show abrupt change and show almost linear increase in many regions. Also, it is found that human activity effectively reduces the increasing rate and suppresses the natural variability under projected warming climate. Nevertheless, the increasing trends are significant under RCP8.5 scenario. In the regions where large increase of drought is projected, drought will depart from historical condition to unexperienced phase in each region until 2050. It alarms human society has to consider countermeasures for the coming change promptly.