An Alternative Approach for Improving Prediction of Integrated Hydrologic‐Hydraulic Models by Assessing the Impact of Intrinsic Spatial Scales
An Alternative Approach for Improving Prediction of Integrated Hydrologic‐Hydraulic Models by Assessing the Impact of Intrinsic Spatial Scales
复制标题
DOI：
10.1029/2020wr027702
复制
发表时间：
2021-09
影响因子：
5.4
通讯作者：
S. Saksena;V. Merwade;P. Singhofen
中科院分区：
文献类型：
--
作者：
S. Saksena;V. Merwade;P. Singhofen
空间尺度和分辨率的影响已分别量化为不同的水文和水力过程。然而，模型的结构和固有分辨率很少被修改，以准确地捕捉尺度相关的物理过程。虽然存在计算昂贵的集成模型的自动校准方法，在这里提出了一种替代方法依赖于改进模型结构。本研究主张通过量化不同类型的空间尺度对整体流域响应的影响，更好地代表物理过程及其子模型的内在空间尺度。首先，通过评估盆地响应的变化来量化空间范围缩放的效果（例如，河流流量和淹没程度）跨越一个小的和大的子流域为同一地区。其次，修改地表-地下水（SW-GW）子模型的相对内在空间尺度的影响进行了量化。最后，结果被用来实现一个更好的模型结构，以提高两个流域不同的物理特性的预测。研究结果表明，SW-GW子模型的相对内在尺度可能是不同的水文地质系统，这取决于水文-水力过程的特征长度尺度的比例。进行缩放分析可以帮助确定如何在一系列气候和自然地理条件下在集成模型中最好地表示不同的物理过程，这可能会成为分布式模型中广泛校准的替代方案。因此，建议将此分析作为大规模集成模型广泛参数校准的先决条件。
The effect of spatial scale and resolution has been quantified individually for different hydrologic and hydraulic processes. However, the model structure and intrinsic resolution are seldom modified to accurately capture scale‐dependent physical processes. Although automated calibration methods exist for computationally expensive integrated models, an alternate approach reliant on improving the model structure is proposed here. This study advocates for a better representation of the intrinsic spatial scales of physical processes and their submodels by quantifying the impact of different types of spatial scaling on the overall watershed response. First, the effect of spatial extent scaling is quantified by evaluating the change in the basin response (e.g., streamflow and inundation extent) across a small and large subwatershed for the same region. Second, the effect of modifying the relative intrinsic spatial scales of surface‐groundwater (SW‐GW) submodels is quantified. Finally, the results are used to implement a better model structure for improving prediction across two watersheds with distinct physical characteristics. The findings suggest that the relative intrinsic scales of SW‐GW submodels may be different for different hydrogeological systems depending on the ratio of the characteristic length scales of hydrologic‐hydraulic processes. Conducting a scaling analysis can help identify how different physical processes can be best represented in integrated models for a range of climatological and physiographic conditions which can potentially serve as an alternative to extensive calibration in distributed models. Therefore, it is recommended that this analysis should be included as a prerequisite to extensive parameter calibration for large‐scale‐integrated models.