Résumé
Description
This dataset aims to represent, in geographic space, the potential distribution of biological tree richness in New Caledonian forests according to a 1 ha grid based on the observed distribution of 148,085 occurrences for 1112 tree species.
For each species, we constructed the environmental niche based on 7 abiotic variables (rainfall, slope, elevation, compound topographic index, substrate, sunshine index, distance to the east coast; cf. Pouteau et al., 2015, 2019 for details). We used species distribution models (SDM) and stacked species distribution models (S-SDM) through the R-package SSDM (Schmitt et al., 2017).
According to the S-SDM model, the potential richness ranges between 18 and 355 tree species per hectare in New Caledonia. We adjusted this range to the richness observed on 24 1 ha plots from the Permanent Plant Inventory Network of New Caledonia (NC-PIPPN), which ranges between 35 and 121 tree species per hectare. Finally, we clipped the resulting raster with the forest map of New Caledonia (version 2024, Birnbaum et al., 2024) to produce the raster of potential distribution of biological tree richness in the New Caledonian forests at 1 ha resolution.
Content
This dataset was produced, analyzed, and verified using a combination of open-source software, including QGIS, PostgreSQL, PostGIS, Python, R, and the GDAL library, all running on Linux.
amap_raster_forest_richness.tif is a GeoTIFF utilizing the WGS84 international coordinate system and consists of a single band with graduated values ranging from 35 to 121 potential tree species per hectare. The NoData value was set to 0.
amap_raster_forest_richness.png is a image illustrating the spatial distribution of the data and values
Limitations
This dataset is strictly based on the relationship between a few environmental variables and a limited set of tree species occurrences. While it provides a valuable overview of potential tree species richness, it represents only a part of the complex biotic and abiotic interactions that lead to the effective presence or absence of a species in the environment. Consequently, the projection of these probabilities onto the geographical space provides only an overview of the potential richness of forest fragments, which should not be considered as the observed diversity.
Additionally, due to a lack of occurrence data, only the Grande-Terre forest is covered in this raster.