Using hydrologic equivalent wetland concept within SWAT to estimate streamflow in watersheds with numerous wetlands.

摘要

Using hydrologic models such as the Soil and Water Assessment Tool (SWAT) to quantify the conveyance, storage, and retention functions of wetlands is a cost-effective means of generating the scientific information required for watershed management and wetland protection/restoration. However, it has been a challenge to appropriately represent wetlands in models, and few SWAT applications reported in the literature have considered wetlands. The objectives of this study were to: (1) demonstrate how to incorporate wetlands into a SWAT model using a hydrologic equivalent wetland (HEW) concept; and (2) use the SWAT model to simulate the stream flows in the 4506 km2 upper portion of the Otter Tail River watershed in northwestern Minnesota, USA. The HEWs were defined in terms of six calibrated parameters: the fraction of the subbasin area that drains into wetlands (WET_FR), the volume of water stored in the wetlands when filled to their normal water level (WET_NVOL), the volume of water stored in the wetlands when filled to their maximum water level (WET_MXVOL), the longest tributary channel length in the subbasin, Manning's n value for the tributary channels, and Manning's n value for the main channel. The model performance was judged in terms of three statistics: the Nash-Sutcliffe coefficient (Ej2), the coefficient of determination (R2), and the performance virtue (PVk). For the study area, the HEWs were determined to have a value for WET_FR ranging from 50 to 96%, i.e., draining more than 50% of the subbasins, depending on the spatial distribution and types of the wetlands in each subbasin. A wetland with minimum alteration tended not to contribute water to the streams until it reached its maximum capacity, as indicated by identical calibrated values for WET_NVOL and WET_MXVOL for all the subbasins except for the three that included altered wetlands. The results indicated a strong linear relationship between WET_NVOL and the corresponding surface area (R2=0.98) and an overall average depth of 0.85 m for the HEWs. The parameters WET_NVOL and WET_MXVOL may primarily represent the storage function of the wetlands, whereas the conveyance and retention functions would be reflected by the latter five parameters. In addition, the SWAT model incorporating the HEW assumption had an acceptable or satisfactory performance in simulating the streamflows for an evaluation period from 1 December 1969 to 31 May 1975 at daily, monthly, seasonal, and annual time steps (Ej2 >=0.36, PVk >=0.75), although its performance varied from one evaluation station to another and might be weak for spring and fall in some evaluation years. Nevertheless, additional model runs with the no wetland and synthetic wetland assumptions, which are two approaches typically used to deal with wetlands for modelling purposes, indicated that the HEW concept is superior in incorporating wetlands into SWAT for the study area.