Simulation of Regional Karst Aquifer System and Assessment of Groundwater Resources in Manatí-Vega Baja, Puerto Rico.

The North Coast karst aquifer system of Puerto Rico, the most productive aquifer of the island, is a vital water source for drinking water and local ecosystems. High freshwater demands alter the coastal groundwater system that impacts both human populations and coastal ecosystems of the island. To predict how this system might respond to rainfall events and high pumping demands, we used the equivalent porous medium (EPM) technique to develop a three-dimensional ground-water flow model to estimate hydrogeological parameters and assess groundwater resources in the Manatí-Vega Baja karst aquifer. The approach is based on the hypothesis that the simplified EPM approach will reproduce groundwater hydrodynamics in this complex karst environment. The steady-state model was calibrated with trial and error and parameter estimation methods using an observed groundwater table of 1995 (r = 0.86, p < 0.0001, n = 39). The large-scale simulation suggested that groundwater flow roughly follows the elevation slope [i.e. south to north). Calibrated hydraulic conductivities range from 0.5 to 86 m/d, whereas the hydro-geologic data strongly suggest higher permeability in the middle karst section of the study area. The transient model adequately estimates the observed groundwater fluctuations in response to rainfall events from 1980 until 2014. The transient results indicate that the conceptual model accuracy is more acceptable with a mean error (ME) of -0.132 m, mean absolute error (MAE) of 0.542 m and root mean square (RMSE) error of 0.365 m. The results of water budget simulation show that the total recharge satisfies the total groundwater withdrawal rate in the past, but continuous closure of more contaminated wells causes groundwater levels to increase in the future. The results indicate that the assumption of applicability of EPM approach is sustained and supported by measured data in the study area. Taking future water demands into account, this model could be applied further to predict the changes of groundwater levels and mass balance under different exploitation scenarios.