For real-world problems in typically complex hydrogeologic environments, numerical models “… are the only approach to compute capture from different features” (Barlow and Leake, 2012). Models are widely used tools in groundwater analysis. The beauty of the model is that it can be used to project how a particular system might respond to different stresses in the future. There are too many unknowns and uncertainties in the problem to predict accurately and uniquely a future result. On the other hand, one can project how a system might respond, or a range of responses, and at the same time, given potential errors in input data, place some confidence interval about a future projection. This is useful in attempting to understand and manage the system. In other words, one can ask: if we do this, what is the resulting projected future state of the system.
Barlow and Leake (2012) provide an example of such an analysis for the Upper San Pedro Basin aquifer system in southern Arizona, USA, which was studied by Leake et al. (2008). Using this model, which included a representation of the evapotranspiration process, they assessed the response of the system to pumping a hypothetical well at various locations. The results for one such well are plotted in Figure 16 and show the shifting tradeoff over time between groundwater storage change and capture as sources of water to balance the pumpage. Furthermore, it illustrates that salvaged evapotranspiration can be a substantial component of the total capture. The streamflow depletion includes both induced infiltration (increased recharge to the aquifer) and decreased discharge of groundwater to the stream, though these are not shown separately in the plot (though the typical model output will include sufficient information to allow the user to do this).