Abstract
Land use change leads to shifts in species ranges and declines in biodiversity across the world. By mapping likely future land use under projections of socio-economic change, these ecological changes can be predicted to inform conservation decision-making.
We present a land use modelling approach that enables ecologists to map changes in land use under various socio-economic scenarios at fine spatial resolutions. Its predictions can be used as a direct input to virtually all existing spatially-explicit ecological models.
The most commonly used land use modelling approaches provide binary predictions of land use. However, continuous representations of land use have been shown to improve ecological models. Our approach maps the fractional cover of land use within each grid cell, providing higher information content than discrete classes at the same spatial resolution.
When parametrized using data from 1990, the method accurately reproduced land use patterns observed in the Amazon from 1990 until 2018. Predictions were accurate in terms of the fractional amounts allocated across the landscape and the correct identification of areas with declines and increases in different land uses. A small case study showcases the successful application of our model to reproduce patterns of agricultural expansion and habitat decline.
The model source code is provided as an open-source R package, making this new, open method available to ecologists to bridge the gap between socio-economic, land use and biodiversity modelling.
Competing Interest Statement
The authors have declared no competing interest.
