Does planned retreat matter? Investigating land use change under the impacts of flooding induced by sea level rise

Abstract

Coastal regions worldwide are during the process of rapid urban expansion. However, expanded urban settlements in land-sea interfaces have been faced with unprecedented threats from climate change related hazards. Adaptation to coastal hazards has received increasing attention from city managers and planners. Adaptation and land management practices are largely informed by remote sensing and land change modeling. This paper establishes a framework that integrates land change analysis, coastal flooding, and sea level rise adaptation. Multilayer perceptron neural network, similarity learning, and binary logistic regression were applied to analyze spatiotemporal changes of residential, commercial, and other built-up areas in Bay County, Florida, USA. The prediction maps of 2030 were produced by three models under four policy scenarios that included the population relocation strategy. Validation results reveal that three models return overall acceptable accuracies but generate distinct landscape patterns. Predictions indicate that planned retreat of residents can greatly reduce urban vulnerability to sea level rise induced flooding. While managed realignment of the coast brings large benefits, the paper recommends different mixes of adaptation strategies for different parts of the globe, and advocates the application of reflective land use planning to foster a more disaster resilient coastal community.