Modeling Karstic Controls on Watershed-Scale Groundwater Flow in the Floridan Aquifer of North Florida

Abstract

Groundwater flow in the Santa Fe River Basin (SFRB) and the Woodville Karst Plain (WKP) in northern Florida is controlled by large volume karst conduits that transport groundwater and surface water to large discharge points such as Wakulla and Hornsby Springs. Tracer tests have confirmed that these conduits receive surface water input from sinking streams throughout both regions and that water can move through these conduits at velocities greater that 1.5 km/day. Cave surveys have shown that the size and shape of these conduits repeatedly morph between large open passages more than 100 feet in diameter and highly bifurcated networks of small tubes. Groundwater / surface water exchange is reflected by tea-colored spring discharge during wet periods and clear water discharge during drought periods. The combination of these conditions presents a significant challenge to understanding and simulating regional groundwater flow patterns and their susceptibility to human-induced or natural environmental changes. We have had great success in simulating groundwater flow in the SFRB and throughout northern Florida by constructing a finite-element model with conduits represented as embedded discrete element features. This method allows for the assignment of varying conduit area and friction parameters along the flow path. By explicitly defining the conduit network, we have been able to develop flow models that are simultaneously calibrated to head, spring discharge, stream loss, and tracer-defined conduit velocities. In doing this the model has both successfully simulated observed surface water / groundwater interactions and provided insights on the controlling mechanisms. Simulations show that under wet conditions, surface water entering through sinkholes dominates the conduit flow and suppresses the transport of groundwater whereas under dry conditions, surface water input decreases allowing for rapid transport of groundwater. Though it is often dismissed in karst aquifers, these results demonstrate the utility of modeling if an estimated pattern of conduits can be defined through cave surveys, dye tracing, or detailed head mapping. Once a realistic pattern is established, the model itself provides an excellent tool for estimating non-observable conditions such as the effects of internal gradients on conduit-matrix interactions through calibration to heads, spring discharges, and velocities.