Runoff of arid and semi-arid regions simulated and projected by CLM-DTVGM and its multi-scale fluctuations as revealed by EEMD analysis

Runoff is a major component of the water cycle,and its multi-scale fluctuations are important to water resources management across arid and semi-arid regions.This paper coupled the Distributed Time Variant Gain Model(DTVGM)into the Community Land Model(CLM 3.5),replacing the TOPMODEL-based method to simulate runoff in the arid and semi-arid regions of China.The coupled model was calibrated at five gauging stations for the period 1980–2005 and validated for the period 2006–2010.Then,future runoff(2010–2100)was simulated for different Representative Concentration Pathways(RCP)emission scenarios.After that,the spatial distributions of the future runoff for these scenarios were discussed,and the multi-scale fluctuation characteristics of the future annual runoff for the RCP scenarios were explored using the Ensemble Empirical Mode Decomposition(EEMD)analysis method.Finally,the decadal variabilities of the future annual runoff for the entire study area and the five catchments in it were investigated.The results showed that the future annual runoff had slowly decreasing trends for scenarios RCP 2.6 and RCP 8.5 during the period 2010–2100,whereas it had a non-monotonic trend for the RCP 4.5 scenario,with a slow increase after the 2050 s.Additionally,the future annual runoff clearly varied over a decadal time scale,indicating that it had clear divisions between dry and wet periods.The longest dry period was approximately 15 years(2040–2055)for the RCP 2.6 scenario and 25 years(2045–2070)for the RCP 4.5 scenario.However,the RCP 8.5 scenario was predicted to have a long dry period starting from 2045.Under these scenarios,the water resources situation of the study area will be extremely severe.Therefore,adaptive water management measures addressing climate change should be adopted to proactively confront the risks of water resources.