LandCurate Alabama Shiny App

LANDCURATE is an interactive Shiny-based geospatial application designed to support adaptive habitat management on Alabama’s Wildlife Management Areas (WMAs). The application enables users to choose or upload a study area shapefile, and generate spatially explicit land cover classifications using a hybrid modeling framework. The classification approach integrates a Random Forest (RF) model with the National Land Cover Database (NLCD). The RF model predicts forest structural classes, including even-aged and uneven-aged pine and hardwood systems, as well as mixed and young stands. To ensure reliable landscape-wide classification, NLCD is used to define the spatial extent of forest and non-forest areas. The RF model is then applied only within NLCD-defined forest and shrubland classes, while NLCD is used to classify non-forest land cover types. Non-forest areas are grouped into broader classes, including water, developed, barren, herbaceous, planted (agriculture), and wetlands. The final output is a land cover map that combines RF-derived forest structural classifications with NLCD-derived non-forest classes. Additional environmental predictors used in the model include vegetation indices, texture metrics derived from gray-level co-occurrence matrices (GLCM), elevation data, and tree canopy cover. These variables enhance the model’s ability to distinguish forest structural characteristics across heterogeneous landscapes. The application provides visualization through an interactive map interface and allows users to download the final classified raster as a GeoTIFF for use in GIS software.