Choosing predictors and complexity for ecosystem distribution models: effects on performance and transferability

There is an increasing need for ecosystem-level distribution models (EDMs) and a better understanding of which factors affect their quality. We investigated how the performance and transferability of EDMs are influenced by (1) the choice of predictors, and (2) model complexity. We modelled the distribution of 15 pre-classified ecosystem types in Norway using 252 predictors gridded to 100 m × 100 m resolution. The ecosystem types are major types in the \"Nature in Norway\" system mainly defined by rule-based criteria such as whether soil or specific functional groups (e.g., trees) are present. The predictors were categorised into four groups, of which three represented proxies for natural, anthropogenic, or terrain processes (‘ecological predictors’) and one represented spectral and structural characteristics of the surface observable from above (’surface predictors’). Models were generated for five levels of model complexity. Model performance and transferability were evaluated with data ..., , , # Choosing predictors and complexity for ecosystem distribution models: effects on performance and transferability

[https://doi.org/10.5061/dryad.vq83bk40j](https://doi.org/10.5061/dryad.vq83bk40j)

The dataset contains raster layers (GeoTIFF files) with the probability of presence of 15 ecosystem types in 100 m x 100 m resolution (Coordinate Reference System: ETRS89 / UTM zone 32N). Each ecosystem type corresponds to one major type in the Nature in Norway (NiN) ecosystem typology. The modelled ecosystem types are:

* T1 Bare rock
* T3 Arctic-alpine heath and lee side
* T4 Forest
* T7 Snowbed
* T14 Exposed ridge
* T19 Patterned ground
* T22 Arctic-alpine dry-grass heath
* T27 Boulder field
* T30 Alluvial forest
* T31 Boreal heath
* T32 Semi-natural grassland
* T34 Coastal heath
* V1 Open fen
* V2 Mire and swamp forest
* V3 Bog