Projecting Fine Resolution Land-Cover Dynamics for a Rapidly Changing Terrestrial–Aquatic Transition in Terrebonne Basin, Louisiana, U.S.A.

Abstract

Abstract Coastal landscapes are in a constant state of flux and continue to change with sea level rise. Past trends in land cover may be useful for predicting future landscapes under different scenarios of change. In this study, models representing land-cover change were created for a coastal forest–open water transition in a rapidly subsiding region in the Mississippi Delta, U.S.A. Land-cover images manually delineated from aerial photography for 1940, 1953, 1978, 1998, and 2004 served as the basis for the models. A combination of Markov chain analysis, a cellular automaton, and suitability images were used to model past trends and to create future land-cover scenarios. Model validation revealed that one of two model runs more closely matched reference images than null (no change) models. Models were generally better at predicting the location of land-cover classes on the landscape than the quantity of each class. Prediction accuracy varied among individual land-cover classes, with forest being the most ...