Cumulative frequency distributions (CFDs) of transit times have been generated for many watersheds using regional scale MODFLOW-MODPATH models. While it is impractical to develop such models for all unstudied watersheds, a large scale analysis is necessary to assess the long term impact of nitrate delivery via groundwater to streams and ultimately coastal environments. Alternatively, the exponential lumped parameter model (ELPM) only requires three spatially averaged, physically based values (recharge, saturated thickness, porosity) and hence can be used to create CFDs with minimal resources. We evaluated the ELPM against MODFLOW-MODPATH for regional watersheds in New Jersey, Delaware, and New Zealand. The ELPM systematically underestimates the percentage of very short and very long transit times, although in many cases the discrepancies are minor. One reason for the discrepancy between the ELPM and MODPATH is that only the latter modeling approach can simulate sinks that do not draw water from the entire aquifer thickness, commonly referred to as weak sinks. However, the MODPATH simulations indicate that the impact of weak sinks on the shape of the CFD is minimal for large watersheds. Hence the ELPM generally performs well at the regional scale, although it will not always yield satisfactory results at subwatershed scales where weak sink effects can be more prominent. A second reason for the discrepancy between the ELPM and MODPATH is that, for aquifers with high resistance to vertical flow, the Dupuit-Forchheimer approximation that underlies the ELPM can potentially (but not necessarily) be invalidated. When it is invalidated, an alternative modeling approach to the ELPM is needed to conduct a rapid assessment of multiple watersheds. It appears that a simple cross-sectional MODFLOW-MODPATH model can be used to simulate flow in only the x and z directions, hence accounting for resistance to vertical flow. The cross-sectional model requires the same inputs as the ELPM plus two: average distance between streams and anisotropy ratio. |