洪河沼泽湿地水文过程模型构建及水文功能分析

Construction of Hydrological Processes Model and Analysis of Hydrological Functions of Marsh Wetlands in Honghe

洪河沼泽湿地地势低平、河道多变,洪水季节性泛滥,缺乏水文监测数据,而模拟的水文数据不确定性较大;而且,用地表径流数据验证生态水文模型的模拟结果,效果并不理想。以洪河国家级自然保护区中部的沼泽湿地为研究对象,整合数字河网与高精度数字高程数据,改进分布式生态水文模型SWAT的子流域概化方式,构建湿地水文过程模型;在此基础上,分析研究区沼泽湿地水文功能。对于湿地水文过程模型的构建,除了利用径流观测数据进行验证外,还应用不同深度的多点土壤湿度数据进行验证。研究结果表明,1改进的SWAT模型取得了较高的面向沼泽湿地模型精度,模型决定系数R2达到0.7,相对误差小于21%,表明模拟效果良好;2在研究区内的沼泽湿地中,以草甸黑土—沼泽化草甸、草甸黑土—草本沼泽、沼泽土—沼泽化草甸、沼泽土—草本沼泽4种土壤—植被的组合模式为主,土壤湿度较大,湿地的调蓄水功能较强;3季节性和常年的洪泛水文过程与土壤湿度的梯度特征,控制着研究区的沼泽植物的空间格局;随着土壤湿度梯度的变化,植物群落呈现出明显的条带状梯度分布。

The marsh wetlands in are featured by low-lying, changing river network and seasonal flooding in Honghe National Nature Reserve, China. There is a great challenge to simulate eco-hydrological processes using a hydrological model in the reserve. In this study, we investigated eco-hydrological processes in the marsh wetlands and proposed a new model calibration-validation framework for a hydrological model—Soil and Water Assessment Tool. In addition to the runoff from the outlet of a basin in the reserve, soil moisture from different sub-basins and various depths was measured. These runoff and soil moisture data were combined to calibrate and validate the model based on sensitivity analysis. During the calibration period, the determination coefficient（R2） for the runoff simulation was up to 0.7, and the relative error（RE） was about 21%. For the model validation, the simulation for soil moisture was also acceptable since R2 reached 0.7 and RE was less than20%. The daily soil moisture reflects the status of internal humidity and the relationship between the soil type and the vegetation. Therefore, Soil and Water Assessment Tool model performed well during the calibration and the validation period and could be used to analyze the eco-hydrological processes in the reserve. Moreover, the soil moisture of four soil-vegetation types, meadow black soil-marshy meadow, meadow black soil-marsh, marsh soil-marshy meadow, and marsh soil-marsh, remained at high levels. So the eco-hydrological function of water regulation in the marsh wetlands was very significant. The four soil-vegetation types were within area of the seasonal water-logged or the perennial water-logged zones. The soil moisture gradient was the dominant factor, which controls the vegetation spatial patterns. It showed that the distribution of vegetation communities have accordingly shrunk into a banding distribution along a humidity gradient. So the marsh wetlands were in a well status and it was favorable for biodiversity conservation. The calibration-validation framework is potential to reduce the modeling uncertainties in such wetland area especially when the data is rather sparse.