1) When comparing bioclimatic variables for doing species distribution modelling, which is more relevant when working on narrowly endemic tree species. ( regional study)
2) What is the suitability of BIOCLIM variables for a regional level study?
3) Are there other methodologies for comparing the variables when trying to do distibution modelling ?
Stepwise backward elimination of variables is an alternative to a priori selection by PCA. When we compared both methods, the outcome was near identical. One issue to keep in mind is that PCA is not suitable for categorical variables.
Bioclimatic variables are spatial interpolations of meteorological point data. The quality of the interpolation depends on the number data points and representativeness of these meteorological data. Where we tested the WorldClim interpolations with independent local meteostations, the difference was too large for comfort. Further, we noticed in more than one study, that classical meteorological variables were outperforming the bioclimatic derivatives.
I agree with Yuri from our experience that (bio-)climatic variables maybe suitable at larger areas, say at subcontinental level, but not so much for smaller ones. In mountains elevation plus vegetation variables may outperform climatic variables.
My statements above can be studied in detail in our SDM papers available under my RG profile. Keywords: bear, endemic plants, elephant, wild boar, oak, beech, mountain pine, land use conversion ("deforestation"), GLM and maxent.
Enjoy the read
- Badges
- Science topic
- Similar topics
- Geoscience
- Asia
- India