Data from: Modelling the area of occupancy of habitat types with remote sensing

1. A current challenge of biodiversity and conservation is the estimation
of the spatial extent of habitat types across broad territories. In the
absence of fine-resolution maps, predictive modelling helps in assessing
the spatial distribution of vegetation cover. However, such approaches are
still uncommon in regional planning and management. Here, we present a
framework for mapping the area of occupancy (AOO) of habitat types that
allows highly suitable estimates at different scales. 2. We model the
potential AOO with abiotic variables related to topography and climate,
resulting in broad AOO estimates that are subsequently downscaled to the
local AOO with remote sensing. The combination of individual local AOO
estimates allows the defining of the realized AOO, comprising locations
with a high likelihood of occurrence and low uncertainty for each habitat.
We applied this framework to mapping 24 protected habitat types of Natura
2000 sites in northern Spain. 3. Local and realized AOO were highly
accurate, with a 70% overall accuracy for the realized AOO. Remote sensing
data, and especially LiDAR, were the most important predictors in habitat
types related to forests and shrubs, followed by rock outcrops and
pastures. Environmental variables were also relevant for specific habitats
subject to abiotic constraints. 4. The combination of ecological modelling
with remote sensing offers multiple advantages over traditional field
surveys and image interpretation, allowing the harmonization of habitat
maps across large regions and through time. This is particularly useful
for implementing conservation actions under Natura 2000 principles or
assessing IUCN criteria for ecosystems.