Department of Environmental Engineering, Dokuz Eylül University, 35160 Buca-Izmir, Turkey; +90-232-4127112; fax: +90-232-4531143;
Ground Water. 2014 May-Jun;52(3):434-47. doi: 10.1111/gwat.12077. Epub 2013 Jun 7.
The quantification of groundwater recharge is an important but challenging task in groundwater flow modeling because recharge varies spatially and temporally. The goal of this study is to present an innovative methodology to estimate groundwater recharge rates and zone structures for regional groundwater flow models. Here, the unknown recharge field is partitioned into a number of zones using Voronoi Tessellation (VT). The identified zone structure with the recharge rates is associated through a simulation-optimization model that couples MODFLOW-2000 and the hybrid PSOLVER optimization algorithm. Applicability of this procedure is tested on a previously developed groundwater flow model of the Tahtalı Watershed. Successive zone structure solutions are obtained in an additive manner and penalty functions are used in the procedure to obtain realistic and plausible solutions. One of these functions constrains the optimization by forcing the sum of recharge rates for the grid cells that coincide with the Tahtalı Watershed area to be equal to the areal recharge rate determined in the previous modeling by a separate precipitation-runoff model. As a result, a six-zone structure is selected as the best zone structure that represents the areal recharge distribution. Comparison to results of a previous model for the same study area reveals that the proposed procedure significantly improves model performance with respect to calibration statistics. The proposed identification procedure can be thought of as an effective way to determine the recharge zone structure for groundwater flow models, in particular for situations where tangible information about groundwater recharge distribution does not exist.
地下水补给的量化是地下水流动建模中的一项重要但具有挑战性的任务，因为补给在空间和时间上都存在变化。本研究的目的是提出一种创新的方法，用于估算区域地下水流动模型的地下水补给率和带结构。在这里，使用 Voronoi 图将未知的补给场划分为多个带。通过将 MODFLOW-2000 与混合 PSOLVER 优化算法耦合的模拟-优化模型，将具有补给率的识别带结构相关联。该程序的适用性在先前开发的 Tahtalı 流域地下水流动模型上进行了测试。通过附加方式获得连续的带结构解决方案，并在该过程中使用惩罚函数来获得现实和合理的解决方案。其中一个函数通过强制与 Tahtalı 流域区域相重合的网格单元的补给率之和等于通过单独的降水-径流模型在前一次建模中确定的面积补给率，从而对优化进行约束。结果，选择了一个六带结构作为代表面积补给分布的最佳带结构。与同一研究区域的先前模型的结果进行比较表明，与校准统计数据相比，所提出的程序显著提高了模型性能。所提出的识别程序可以被视为确定地下水流动模型补给带结构的有效方法，特别是在不存在有关地下水补给分布的具体信息的情况下。
