干旱区生态环境知识资源中心

An integrated multi-GCM-based stochastic weather generator and stepwise cluster analysis (MGCM-SWG-SCA) method is developed, through incorporating multiple global climate models (MGCM), stochastic weather generator (SWG), and stepwise-clustered hydrological model (SCHM) within a general framework. MGCM-SWG-SCA can investigate uncertainties of projected climate changes as well as create watershed-scale climate projections from large-scale variables. It can also assess climate change impacts on hydrological processes and capture nonlinear relationship between input variables and outputs in watershed systems. MGCM-SWG-SCA is then applied to the Kaidu watershed with cold-arid characteristics in the Xinjiang Uyghur Autonomous Region of northwest China, for demonstrating its efficiency. Results reveal that the variability of streamflow is mainly affected by (1) temperature change during spring, (2) precipitation change during winter, and (3) both temperature and precipitation changes in summer and autumn. Results also disclose that: (1) the projected minimum and maximum temperatures and precipitation from MGCM change with seasons in different ways; (2) various climate change projections can reproduce the seasonal variability of watershed-scale climate series; (3) SCHM can simulate daily streamflow with a satisfactory degree, and a significant increasing trend of streamflow is indicated from future (2015-2035) to validation (2006-2011) periods; (4) the streamflow can vary under different climate change projections. The findings can be explained that, for the Kaidu watershed located in the cold-arid region, glacier melt is mainly related to temperature changes and precipitation changes can directly cause the variability of streamflow.