大气、陆面与水文耦合模式在中国西北典型流域径流模拟中的新应用

水文模式在较长的发展阶段由简单的概念模型逐步演变复杂的分布式物理模型,大气学科的各类气候模式在近年来迅猛发展同时也逐步带动了水文学的发展。从大气、水文两个学科发展角度纵向开展研究,通过分析以往研究成果认为,虽然大气、水文模式在其各自的发展已经到达了较为完善的地步,然而其相互耦合并取长补短的优势并未发挥。在探讨大气、陆面及水文模式发展的基础上,选取XJLDAS(Xinjiang Land Data Assimilate Datasets)大气陆面同化系统,CLM3.5(Community Land Model,Version 3.5)公用陆面模式及快速汇流模式RAPID(Routing Application for Parallel computation of Discharge)作为关键耦合对象,利用以上耦合系统对新疆精博河流域径流过程进行模拟。通过研究分析,认为:XJLDAS+CLM3.5+RAPID模式可较好地重现流域地表径流年内分布状况,然而,由于陆面参数化方案选取等原因,研究区径流量出现一定偏差。此外,在本研究中发现,陆面模式径流汇流模拟结果与实际结果存在一定偏差,将这种偏差进行分析后发现:在进行大尺度水文模拟时,需要在考虑在研究区相关地理特性(如地质构造、地表覆被)基础上进行模式相应改进,以最大限度的重现大尺度径流真实过程。

A New Application of Atmosphere, Land Surface and Hydrological Coupling Model in Runoff Simulation of Typical Basins in Northwest China

The hydrological model has developed from simple conceptual model into the complex distributed physical model in the long evolution stage. The climate model of meteorological subject has developed fast in recent years, meanwhile, it has improved the development of hydrology. This paper carried out researches from the perspective of meteorology and hydrology development. After analyzing previous studies, we found that although meteorological and hydrological models have reached to a perfect stage in their respective development, the advantage of coupling and learning from each other has not been exploited. On the basis of ex - ploring the development of atmosphere, land surface and hydrology coupling model, this paper chose XJLDAS atmosphere-land surface coupling system, CLM3.5 public land surface model and Rapid confluence model as the key coupling objects respectively. With the above coupling systems, we simulated the runoff process of Xinjiang Jingbo River Basin. Through analysis, it is deemed that XJLDAS+CLM3.5+RAPID model can reappear the annual distribution of land surface runoff well. However, due to the selection of land surface parameterization schemes, there are some biases of runoff in the study area. In the research, furthermore, we found that there exists some differences between runoff simulation results and observed values. As a result, when taking large scale hy - drological simulations, it is necessary to consider improvements on model based on geographic features (such as geological structure, land surface cover and so on), so as to reappear the utmost large scale runoff process