基于SWAT-MODFLOW耦合模型的巴音河中游泽令沟盆地地表水?地下水转换关系研究

为准确刻画资料缺乏、地表水–地下水转换频繁的泽令沟盆地地表水?地下水不同时空尺度上的转换关系，利用河道径流数据、遥感蒸散发数据、地下水位观测数据等对SWAT-MODFLOW耦合模型进行率定和验证，在此基础上对泽令沟盆地水循环过程进行模拟分析。结果显示：① SWAT-MODFLOW耦合模型模拟效果较好，率定期和验证期月径流量的R²≥0.73，NSE≥0.67，PBIAS在?20%~20%。其各子流域实际蒸散发的R²均大于0.91，NSE均大于0.85，PBIAS在?10%~10%。此外，模拟地下水位与实测值误差在0.6 m以内，R²为0.94。② 地表水补给地下水的河道占整个河道长度90.31%，年平均补给量占年总交换水量的50.20%；季节尺度上，补给的最大值出现在7月，为331043.31 m³/d，最小值出现在12月份，为41208.33 m³/d。地下水对地表水的补给量较为稳定，季节性变幅在?2.5%~2.5%，年际变幅处于?6.5%~0.6%之间。

Modelling of the Surface-ground Water Exchange Yield in Zelinggou Basin,Middle Reaches of the Bayin River Based on SWAT-MODFLOW Coupled Model

To reveal the surface-ground water exchange yield in different temporal and spatial scale in Zelinggou Basin, middle reaches of the Bayin River in Qaidam Basin. The SWAT-MODFLOW coupled model was used to simulate the water cycle of the Zelinggou Basin, the monthly stream flow, remotely sensed evapotranspiration data, ground water level data were used to calibrate and validate the model. The results showed that: 1) For the monthly streamflow simulating during calibration and validation period, R² of the SWAT-MODFLOW coupled model are 0.74 and 0.73, NSE are 0.67 and 0.7, and PBIAS are between ?20% and 20%. While for the actual Evapotranspiration in each sub-basin, all R² and NSE are greater than 0.91 and 0.85, and PBIAS are between ?20% and 20%. The bias values are within 0.6 m and R² value is 0.94 between the simulated groundwater level and the measured. 2) The river where surface water to groundwater accounts for 90.31% of the whole length of the river, and the average annual replenishment amount accounts for 50.20% of the total annual exchanged water. Seasonally, the maximum value of recharge is 331043.31 m³/d in July, while the minimum value is 41208.33 m³/d in December. The groundwater discharged yield is relatively stable, it seasonal variation is between ?2.5% and 2.5%, and the annual variation is between ?6.5% and 0.6%. This research can further support the hydrological processes modelling and water resource allocation in data scarce area.