Do bioclimate variables improve performance of climate envelope models?
Ecological Modelling
By: James I. Watling, Stephanie S. Romañach, David N. Bucklin, Carolina Speroterra, Laura A. Brandt, Leonard G. Pearlstine, and Frank J. Mazzotti
Links
- More information: Publisher Index Page (via DOI)
- Open Access Version: External Repository
- Download citation as: RIS | Dublin Core
Abstract
Climate envelope models are widely used to forecast potential effects of climate change on species distributions. A key issue in climate envelope modeling is the selection of predictor variables that most directly influence species. To determine whether model performance and spatial predictions were related to the selection of predictor variables, we compared models using bioclimate variables with models constructed from monthly climate data for twelve terrestrial vertebrate species in the southeastern USA using two different algorithms (random forests or generalized linear models), and two model selection techniques (using uncorrelated predictors or a subset of user-defined biologically relevant predictor variables). There were no differences in performance between models created with bioclimate or monthly variables, but one metric of model performance was significantly greater using the random forest algorithm compared with generalized linear models. Spatial predictions between maps using bioclimate and monthly variables were very consistent using the random forest algorithm with uncorrelated predictors, whereas we observed greater variability in predictions using generalized linear models.
| Publication type | Article |
|---|---|
| Publication Subtype | Journal Article |
| Title | Do bioclimate variables improve performance of climate envelope models? |
| Series title | Ecological Modelling |
| DOI | 10.1016/j.ecolmodel.2012.07.018 |
| Volume | 246 |
| Year Published | 2012 |
| Language | English |
| Publisher | Elsevier |
| Publisher location | Amsterdam, Netherlands |
| Contributing office(s) | Southeast Ecological Science Center |
| Description | 7 p. |
| Larger Work Type | Article |
| Larger Work Subtype | Journal Article |
| Larger Work Title | Ecological Modelling |
| First page | 79 |
| Last page | 85 |
| Google Analytic Metrics | Metrics page |