Integrated high-resolution, continental-scale land change forecasting

• DOI: 10.1016/j.envsoft.2023.105749
• Publication Date: 2023-05-26T09:32:04Z
• Authors: Marco Calderón-Loor; Michalis Hadjikakou; Richard Hewitt; Raymundo Marcos-M...; Brett A. Bryan
• Keywords: Cartography; Scale (ratio); Management, Monitoring, Policy and Law; Oceanography; Environmental science; land-use change; Meteorology; Engineering; Soil Evaluation; Machine learning; Climate change; Crop Suitability; Civil engineering; Environmental resource management; Global change; Biology; Land use, land-use change and forestry; Climatology; Global and Planetary Change; Global Analysis of Ecosystem Services and Land Use; Ecology; Geography; Land Suitability; Land-Use Suitability Assessment Using GIS; Geology; Remote Sensing in Vegetation Monitoring and Phenology; FOS: Earth and related environmental sciences; Remote sensing; Grassland; Computer science; FOS: Biological sciences; Environmental Science; Physical Sciences; Land use; FOS: Civil engineering; Random forest

Predicting future land change is crucial in anticipating societal and environmental impacts informing responses at different scales. We designed an integrated, high-resolution, land-change model forecasted Australia's for the years 2020, 2025 2030 Cropland, Forest, Grassland, Built-up land-uses using cloud-based high-performance computing. A spatially explicit set of drivers was fed into a random forest classifier to generate 30-m per-class suitability layers country, which were then used allocating land-use. The validated against 2015 data, land-use projected until 2030. Accuracy national level ∼94%. Forecasts showed increases Grassland areas decreases Forest Cropland. Our modelling framework expands current capabilities large-scale models provides first-of-its-kind multiclass forecast Australia that can inform policy multiple scales Australia.