Development of a reproducible method for determining quantity of water and its configuration in a marsh landscape

Coastal Louisiana is a dynamic and ever-changing landscape. From 1956 to 2010, over 3,734 km² of Louisiana's coastal wetlands have been lost due to a combination of natural and human-induced activities. The resulting landscape constitutes a mosaic of conditions from highly deteriorated to relatively stable with intact landmasses. Understanding how and why coastal landscapes change over time is critical to restoration and rehabilitation efforts. Historically, changes in marsh pattern (i.e., size and spatial distribution of marsh landmasses and water bodies) have been distinguished using visual identification by individual researchers. Difficulties associated with this approach include subjective interpretation, uncertain reproducibility, and laborious techniques. In order to minimize these limitations, this study aims to expand existing tools and techniques via a computer-based method, which uses geospatial technologies for determining shifts in landscape patterns. Our method is based on a raster framework and uses landscape statistics to develop conditions and thresholds for a marsh classification scheme. The classification scheme incorporates land and water classified imagery and a two-part classification system: (1) ratio of water to land, and (2) configuration and connectivity of water within wetland landscapes to evaluate changes in marsh patterns. This analysis system can also be used to trace trajectories in landscape patterns through space and time. Overall, our method provides a more automated means of quantifying landscape patterns and may serve as a reliable landscape evaluation tool for future investigations of wetland ecosystem processes in the northern Gulf of Mexico.