Development of predictive models of soil phosphorus in Water Conservation Area-2A (Everglades), integrating remote sensing, GIS and geostatistics

Integrating GIS and remote sensing technologies with multivariate geostatistical methods for the mapping of spatial soil-vegetative interrelationships has the potential to improve the predictions of soil properties at the landscape-scale. This study presents a hybrid geospatial modeling approach to predict soil total phosphorus (TP) using remotely-sensed data and ancillary landscape properties as supporting variables. Phosphorus is a key indicator variable in subtropical, naturally oligotrophic wetlands to indicate nutrient enrichment. The main objective of this research was to develop spatially-explicit models that predict the distribution and variability of soil TP and other physico-chemical properties (bulk density, total nitrogen, total carbon, and total calcium) throughout Water Conservation Area-2A (WCA-2A), in the Everglades. A set of soil data collected at 111 sampling points in 2003 was analyzed by the Wetland Biogeochemistry Laboratory. We compared two remote sensors, Landsat 7 Enhanced Thematic Mapper (ETM) + and Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER), using various univariate and hybrid multivariate geostatistical techniques (ordinary kriging, regression kriging and co-kriging) to predict TP across WCA-2A. Findings of this study have broad impact to improve predictive geospatial modeling of below-ground properties in WCA-2A and similar wetland ecosystems that are sensitive to internal and external forcing functions. Spectral data and indices provide dense, high-resolution signatures that combined with sparse floc/soil data allow to monitor wetland ecosystems and document restoration success across larger landscapes.