基于陆面水文耦合模式CLHMS的淮河流域水文过程的模拟评估及其不确定性分析

摘要

The performance of the Coupled Land surface and Hydrologic Model System (CLHMS) in simulating the hydrological processes over the Huaihe River basin is evaluated using a 24-year numerical simulation of its hydrological cycle between 1980 and 2003. The CLHMS model system is driven by CFSR (Climate Forecast System Reanalysis) reanalysis data and observed precipitation and surface air temperature datasets over China. Generally, the CLHMS shows good performance in reproducing observed hydrological processes, and the model’s skill is pretty higher for wet years in simulating the water balance, and in reproducing the seasonal and interannual variation of the observed streamflow over the Huaihe River basin. For dry years with less precipitation, the model discrepancies in overestimating the observed streamflow can be found in both Wangjiaba and Benbu hydrological stations. Significant differences in the model’s performance are also found between the 1980s and the 1990s, and is largely ascribed to the differences in the model’s skill for wet and dry years. The uncertainties regarding these hydrological simulation results were further examined by three sets of numerical simulations using different precipitation forcing. It’s found that the streamflow simulation using CFSR precipitation forcing exhibited a larger bias than simulations using EAG (East Asia Grid data) precipitation forcing. These results demonstrate the importance of the precipitation forcing chosen for hydrological simulation. Further comparative analysis suggests that the temporal disaggregation method for precipitation forcing preserves the strong diurnal variation, and is therefore also important when conducting hydrological simulations over the Huaihe River basin.%利用最新的CFSR（Climate Forecast System Reanalysis）再分析及观测的降水和地表气温资料驱动陆面水文耦合模式CLHMS（Coupled Land surface and Hydrologic Model System），对淮河流域1980～2003年共24年的水文水循环过程进行了模拟，系统评估了CLHMS对淮河流域水文过程的模拟能力及其不确定性。分析结果表明， CLHMS 模式对淮河流域水文过程具有良好的模拟能力，模式尤其对湿润年份流域的水量平衡以及河道流量的季节、年际变化具有很强的模拟能力，而对降水偏少的干旱年份，模式模拟的河道流量通常会高于观测实况，与实况间存在着一定的偏差，而这也是导致CLHMS对流域水文过程模拟能力存在显著年代际差异的主要原因。基于三组不同降水强迫的流域水文过程模拟结果比较表明，降水驱动资料准确与否是陆面水文模拟最主要的不确定性来源之一，正是由于CFSR再分析降水与观测降水之间存在较大的差异，从而导致CFSR降水驱动下模式模拟的淮河流域河道流量与观测存在较大的偏差，其模拟性能相对较差。进一步分析还表明，可以保持较强降水日变化的时间解集方法，也是保证合理模拟流域水文过程的重要因素。