高级检索+

基于多源蒸散发的参数率定对日径流模拟的影响

The Impact of the Calibration Based on Multisource Evapotranspiration Products on the Daily Streamflow Simulation

  • 摘要: 蒸散发作为水文循环中的重要变量,也可以成为水文模型参数率定中替代实测径流的重要变量。研究基于SWAT模型,以湘江流域中上游为例,通过对比基于实测径流和基于蒸散发两种率定方案所得到的模拟日径流,探究了MODIS(Moderate Resolution Imaging Spectroradiometer)中分辨率成像光谱仪蒸散发产品(MOD16A2)、GLEAM(Global Land Evaporation Amsterdam Model)和GLDAS(Global Land Assimilation Data System)蒸散发产品在模型参数率定中的适用性及其对径流模拟的影响,并分别对比湘江中游和上游两个水文站的模拟效果以分析不同流域所造成的差异。研究结果表明基于GLEAM、GLDAS和MOD16的参数率定能够获得比较好的径流模拟效果,尤其是在流域中游的衡阳站。在率定期,其在双牌站的NSE均能达到0.50以上,在衡阳站的NSE均能达到0.70以上,与基于实测径流率定的日径流模拟效果相当。在验证期,其在双牌站得到的径流NSE皆高于0.58,与基于实测径流率定所得到的NSE(0.60)相当;而在衡阳站,基于蒸散发参数率定所得到的径流模拟效果略好于基于实际径流的,其中基于GLDAS和MOD16的径流模拟NSE可达到0.72,明显高于基于实际径流率定得到的NSE(0.60)。研究验证了遥感蒸散发数据在水文模型参数率定中具有的巨大潜力,且在面积更大的流域可能具有更好的适用性。研究可推广应用至其流域,可为解决无资料地区的水文模拟提供思路。

     

    Abstract: Evapotranspiration(ET)is an important variable in the hydrological cycle and is becoming an alternative variable to replace the observed streamflow for parameter calibration in the hydrological models. In this study,the streamflow simulations produced by the SWAT(Soil and Water Assessment Tool)model based on two different calibration strategies,which use observed streamflow and ET as the calibration variable respectively,are compared to investigate the impact of parameter calibration based on ET on the streamflow simulation in the upper and middle part of the Xiang River Basin. Besides,the differences brought by the basin are analyzed. Multisource ET is used in the study,including the ET products from Moderate Resolution Imaging Spectroradiometer(MODIS)(named MOD16A2),Global Land Amsterdam Model(GLEAM)and Global Land Assimilation Data System(GLDAS). The results show that the SWAT model based on the parameter calibration using GLEAM,GLDAS and MOD16 can generate satisfactory streamflow simulation performances. For the calibration based on the three ET products,the NSE values of simulated streamflow are close to that calibrated based on the observed streamflow in the calibration period. In particular,the NSE values of the simulated streamflow are more than 0.50 at the Shuangpai Station and more than 0.70 at the Hengyang Station. Additionally,in the validation period,the NSE values of simulated streamflow obtained at the Shuangpai Station are more than0.58,which are comparable to that calibrated based on the observed streamflow(0.60). At Hengyang Station,the streamflow simulation performance generated from the parameter calibration based on ET is slightly better than that based on observed streamflow. It is worth mentioning that the SWAT model calibrated based on GLDAS and MOD16 can generate streamflow simulation with NSE of 0.72,which are significantly more than that calibrated based on the observed streamflow(NSE=0.6). This study shows that remote sensing ET product has great potential for parameter calibration of hydrological models,and is more suitable for basins with larger areas. The strategy in this study can provide an insight into hydrological modeling in those ungauged regions.

     

/

返回文章
返回