Gap Analysis and Specicies Distribution Models

All information presented on this site is in draft format and will be updated.

2024 11 15 update

The following model run include the details from the 11-05 update plus an additional spatial filting process.

Spatial filter was performed by the spthin package. The method was applied different for different types of input data. As the spatial thining is meant to influence the modeling effort, all maps and conservation measures are presented with all data. The spatially thinned data is only applied to the model runs.

1. All Germplasm records we kept for the modeling process as these represent the high confidence observations
1. All herbarium records where spatial thinned to a distance of 5km.
1. All H type records based on inaturalist reserach grade observations were spatial thinned to a distance of 10km.

There is not a rigid standard for what distance should be applied when utilixing the spthin package. Source recommend 10km, a distance related more toward you environmental variables or some distanct from your knowledge of the species. For our case, we are using two spatial thining weights (5km and 10km). 5km represents the cell size of the environmental data. 10km matches the generic, though not agreeable standard distance. By assigin the lower value to the assume higher quality dataset source of herberium records we will across all species be keeping more of the herbarium records for the modeling process then the inaturalist records which are thinned to 10km.

2024 11 05 update

We getting very close to finalizing results for the project. The following examples showcase the difference in results after applying the following data filters.

A state level filter is applied based on Flora of North America desiginated habitat. Any points outside of this area are excluded. This filter only applies to the United States. All data from Mexico and Canada is included.
Any occurrence data that was flag by the county review as “Exclude” or “Questionable” was removed from the analysis.

current run : Spatial Thinning, FNA filter and community evaluation applied

previous run : data filter by geographic variables only (has lat lon and within CA,US,MX )