Well drawdown predictions8/13/2023 ![]() ![]() The implicit assumption of linearity between pumping rate and drawdown in the RM approach is valid in simpler cases, but as model non-linearities increase the assumption of this linear relation diminishes ( Niswonger et al., 2011) and the linear assumption becomes less accurate. Once the RM is constructed, it can replace the simulation code and perform fast forward calculations. The approach is based on the assumption of space superposition and uses single simulations of wells to calculate coefficient values that linearly relate a pumping rate at one point to drawdown at different locations in the model ( Psilovikos, 2006). Here, 1,530 model runs took approximately 10 h, which is impractical for rapid decision support.īecause of the long computation time, management models can be approximated using the response matrix (RM) approach ( Gorelick, 1983). (2016) perform a capture fraction analysis of a groundwater model using the flow simulation code MODFLOW ( Harbaugh, 2005). Setting up accurate numerical models requires high numerical grid resolution which comes at the price of long computation times for flow calculations and high memory demand ( Cheng et al., 2014). They can act as decision support tools for tasks such as finding the optimal location for a new well or estimating the effects on wetland and stream depletion. Numerical models can simulate flow dynamics and forecast the effects of management strategies. Therefore, practitioners often use numerical models as decision support tools for planning and managing groundwater ( Hadded et al., 2013). Proper groundwater management prevents depletion or pollution of the resource and saves societies millions of dollars in the remediation of contaminated areas ( Sun and Zheng, 1999). Societies utilize groundwater for drinking water supply, agriculture, and environmental control on a large scale. Resolving the complexity of flow in a groundwater reservoir is a challenging but important task for securing water supply around the world. We discuss the application of the neural network in a decision support framework and describe how the uncertainty estimate accurately describes the uncertainty related to the construction of the training data set. ![]() The network can adapt to non-linearities in the numerical model that the response matrix method fails at resolving. The network has fast predictions with results similar to the full numerical solution. The accuracy and speed of the neural network are compared to results using MODFLOW and a constructed response matrix of the model. A generalized method of constructing and training such a network is demonstrated and applied to a groundwater model case of the San Pedro River Basin. This paper presents a method based on a probabilistic neural network that predicts hydraulic head changes from groundwater abstraction with uncertainty estimates, that is both fast and useful for non-linear problems. Numerical flow simulations are accurate but slow, while response matrix methods are fast but only accurate in near-linear problems. Groundwater resource management is an increasingly complicated task that is expected to only get harder and more important with future climate change and increasing water demands resulting in an increasing need for fast and accurate decision support systems. 3Department of Near Surface Land and Marine Geology, The Geological Survey of Denmark and Greenland (GEUS) Aarhus, Aarhus, Denmark.1Department of Geoscience, Aarhus University, Aarhus, Denmark.Mathias Busk Dahl 1,2 *, Troels Norvin Vilhelmsen 2, Torben Bach 3 and Thomas Mejer Hansen 1
0 Comments
Leave a Reply.AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |