Remote sensing for adaptive management: Leveraging Landsat-based tools to support spatially explicit decision-making in Alabama’s Wildlife Management Areas

Abstract

Adaptive management emphasizes continuous improvement by learning from outcomes and adjusting strategies. It requires real-time data that static models cannot provide. To address this, we developed a land cover classification model using freely available data sources, multi-temporal Landsat imagery, digital elevation models, and field observations, combined using the Random Forest algorithm. We applied this model to lands managed by the Alabama Department of Conservation and Natural Resources (ADCNR) and classified landscapes into 11 management-relevant categories with ~89% accuracy. To support adaptive management, we turned this model into an interactive R Shiny web application, LANDCURATE. LANDCURATE allows users to upload a shapefile and receive land cover classifications for specified areas and years using our pre-trained model. This tool transforms complex workflows into an accessible, real-time decision-support platform, helping ADCNR overcome limitations of costly field surveys. LANDCURATE empowers land managers to track habitat changes and adapt strategies aligned with long-term conservation goals.