Mapping Flood Agents in the Northern Pantanal Wetland Using Multiple Spatio-Temporal GIS and Remote Sensing Techniques

Abstract The mapping of functional hydrologic units is crucial for enhancing our understanding of flooding and hydroecological processes in large wetlands. These units are typically defined by flooding frequency, duration, and magnitude, but another important hydrologic characteristic is the origin of flooding, known as the flood agent. This study presents an empirical framework utilizing remote sensing and GIS procedures for modeling flood agents in the northern Pantanal wetland. Eleven spatial data layers were derived from multi-year MODIS flood maps, daily rainfall estimates from the TRMM 3B42 grid, in situ discharge data, the BEST Digital Terrain Model, and a hydrographic network layer. These layers were tested for their predictive power in a multinomial logistic regression mapping model. Model performance metrics, along with qualitative validation of mapping outcomes using in situ flooding measurements, and vegetation and soil data from field test sites, support the plausibility of the proposed mapping scheme. However, they also highlight the challenges of flood agent mapping in large tropical wetlands.