Spatiotemporal Surface-Groundwater Interaction Simulation in South Florida

摘要

South Florida ecosystem is dictated by a large wetland, karst hydrogeology and extended coastal boundary with the Atlantic Ocean. The risks related to the ecosystem include: disruption of groundwater flow as a result of frequent sinkhole formation; flooding in urban areas as a result of the shallow water table; saltwater intrusion from the ocean; and excessive nutrient load to surficial water bodies and subsequently eutrophication because of the intensive utilization of wetlands for nutrient removal. Attempts to understand eco-hydrological processes primarily focus on extensive monitoring and use of distributed hydrological models. However, the relatively flat nature of the region and also the extended coastal boundary with the ocean, makes watershed-based approaches less realistic. A regional spatiotemporal groundwater level modeling approach was attempted using a Dynamic Factor Analysis (DFA) method. The daily water levels of 13 monitoring well sites from major hydrogeologic regions and different land uses were used to conduct the DFA analysis, and six dynamic factors were identified using minimum Akaike Information Criterion (AIC). Further exploratory analysis to relate the dynamic factors with physically attributable explanatory variables has helped to identify five of the major factors that govern the groundwater dynamics in south Florida. Three of the factors were attributable to the Lake Kissimmee water level in the north, Caloosahatcb.ee River water level in the west, and Hillsboro canal in the east. The other two factors identified were the regional averaged rainfall and soil moisture. The spatiotemporal simulation involved interpolation of the loadings of the dynamic factors using an inverse distance weighted method and convoluting with the dynamic factors. The result has shown a good fit with the maximum RMSE of 0.12 m. Retrieval of rainfall, soil moisture, and surface water level from satellite imagery makes spatiotemporal modeling of the groundwater level achievable.