GEOSPATIAL ANALYSIS OF DEPRESSIONAL WETLANDS: A POTENTIAL TOOL FOR ASSESSING HYDROPERIOD IMPACTS FROM MINING IN A KARST AQUIFER SYSTEM

Mining in the southeastern coastal plain Floridan aquifer system can result in subsurface fluid movement from: 1) volumetric displacement of ground water, 2) nonmechanical dewatering via increased evaporative losses and 3) mechanical dewatering to facilitate excavation, processing or transport. Determining magnitude and extent of subsurface fluid movement from mining in this regional karst aquifer system is difficult due to characteristic preferential flow-ways, mine blasting and aquifer injections. Such determinations do not reveal impacts of subsurface alterations on the biotic environment. Our pilot project used remote sensing and a geographic information system (GIS) to evaluate mean near infrared (NIR) reflectance for natural herbaceous depressional (wet prairie) wetlands in an 8x16-km area surrounding selected phosphate mines in central Florida. These wetlands historically were dominated by graminoids such as maidencane (Panicum hemitomon Schult.), which reflect less NIR radiation than invasive broad-leaf woody species associated with hydroperiod alterations.

Study-area mines, surrounded primarily by unirrigated pasture, were permitted to initiate maximum groundwater withdrawals of 76,457 m³day^-1 in November 1977 and remain active. Digital color infrared (1-m) aerial imagery acquired in winter 2003-04 and available from the USDA was used to extract NIR reflectance values within depressional wetlands. Digital historic (mid-1950s) 1:24,000-scale USGS topographic quadrangle maps of the study area were used to delineate baseline wetland boundaries using ESRI ArcGIS 9.2. The study area included 567 wet prairie wetlands totaling 1411 hectares. Half (287) had mean NIR digital number (DN) values >100, indicative of hydroperiod-altered wetlands dominated by invasive species. Some of these wetlands were >5 km from study-area mines. Only 20% (111) of the wetlands had low mean NIR DN values (<80), indicative of wet prairies dominated by maidencane. Results suggest that shallow ditches in some wetlands were not the controlling factor influencing mean NIR reflection in these wetlands. Spatial distribution of wetlands with high reflectance was inconsistent with conical drawdown typical of ModFlow-model predictions, but may reflect fluid movement via subsurface preferential flow-ways.