大尺度生态水文模型的构建及其与GIS集成

Large scale eco-hydrological model and its integration with GIS

基于过程模拟手段揭示森林植被的生态水文功能和变化机制 ,已经成为生态水文学研究的重要手段。由于陆地生态水文过程的非线性和尺度问题的广泛性 ,基于过程的坡面或小流域尺度的分布式水文模型不适合大流域的水文过程的分析和预测 ;另外 ,传统的水文模型主要侧重水文物理过程 ,只有充分耦合植被生态过程 ,才能从机制上揭示森林植被参与水文循环的调控作用。依据大流域的水文过程特点 ,从 5个方面阐述了大尺度生态水文模型构建过程中的主要问题 :1大尺度生态水文模型的概念和结构 ;2尺度的界定与匹配 ;3离散化数据集的建立 ,着重分析了植被覆盖、土壤质地、山地气候等主要数据集的建立方法 ;4分布式与集总式模型 ,这两类模型可以从他们的基本空间单元上进行区分 ,数字流域的建立和空间分析手段使得集总式模型和分布式模型得到了很好的结合 ;5生态水文模型与 GIS的集成 ,分析了 4种不同的集成方式 ,“松散型”的集成方式因其编程工作量小而被广泛采用。集成的目的不仅是要提高模拟的技术水平 。

Process-based modeling is an effective way in the current eco-hydrological stu dy to explore the interacting mechanism of ecological and hydrological processes in relation to forest vegetation. Because of the nonlinearity of terrestrial ec o-hydrological processes and diversity of scales, the process-based distribute d hydrological models based on slope or small catchments can not be used to make an explicit prediction for large scale watersheds. More over, the eco-hydrolog ical function of forest vegetation could not be quantitatively evaluated without coupling the interaction of ecological and hydrological processes because eco-hydrological processes are not merely a physical process of water movement.Acc o rding to the characteristics of large scale eco-hydrological processes, in this paper the following five major issues were discussed in the process-based eco-hydrological models: (1) Conception and structure of large scale eco-hydrolog i cal models. Large scale eco-hydrological models are usually applied to the scal e of more than 10,000km\+2 valleys with relatively low spatial resolution and lo ng temporal interval. The kernels of the model are water balance and the couplin g of ecological and hydrological processes on the basis of spatial elements. (2) The notions of process scale, observational scale and modeling scale are discus sed to determine the temporal and spatial scale of the model. Because the modeli ng scale is usually different with the observation scale, scale matching is need ed to acquire datasets with the same resolution. Spatial scale and temporal scal e, hydrological scale and ecological scale, hydrological scale and general circu lation model (GCM) scale are key pairs to be matched. (3) Development of discret ization datasets. Land use and vegetation cover, soil texture, topographic and c limatic data need to be matched and specified to each grid point of discrete sur faces. Indices of vegetation property are important to simulate eco-hydrologica l processes because they are used to be a joint of the two processes. Several in dices including LAI and NDVI have been compared and tested. To build up discrete surfaces of meteorological factors in grids, surface-fitting software includin g ANUSPLIN, MTCLIM-3D and PRISM are briefly introduced in this article. (4) Dis tributed model and lumped model can be identified by the modeling elements or un its in space such as square grid vs. “catchment-shaped” cell. We suggest an a pproach to combining distributed model with lumped model by means of delineation of digitizing watershed and GIS techniques. (5) Process-based models integrate d with geographic information system. Four integration approaches have been wide ly used, they are: a. Embedding GIS-like functionalities into eco-hydrological modeling packages;b. Embedding eco-hydrological modeling into GIS packages; c . Loose coupling, eco-hydrological modeling and GIS are integrated via data exc hange using either ASCII or binary data format, there is no common user interfac e among different software. d. Tight coupling, integrate eco-hydrological model s with GIS via either GIS macro or conventional programming, this approach needs a well defined interface to the data structure held by GIS. The advantage of “ loose coupling” is that redundant programming can be avoided and it is a realis tic method to be adopted in modeling work. After all, it is conceptualization co mpatibility rather than a technology-driven problem that should be adequately a ddressed in the integration.