Interface to the Global Biodiversity Information Facility API
Developers of rgbif: Scott Chamberlain, Karthik Ram, Vijay Barve, Dan Mcglinn
Other packages used: ggplot2
CRAN homepage
rgbif documentation @ GBIF

The first step is to download occurrence data (e.g. from GBIF; duplicate records are removed). After acquiring these points, it is useful to examine them on a map.

Such datasets can contain errors; as a preliminary method of data-cleaning, here the user can specify records to be removed. Additionally, the user can download the records as a CSV file.

Future versions will allow the user to download occurrence records from other databases, as well as upload their own occurrence records as an alternate option.

Download Occurrence CSV

Spatial Thinning of Species Occurrence Records
Links to software note and CRAN
Citation: Aiello-Lammens, M. E., Boria, R. A., Radosavljevic, A., Vilela, B. and Anderson, R. P. (2015; Early View), spThin: an R package for spatial thinning of species occurrence records for use in ecological niche models.
Other packages used: ggplot2

Datasets of occurrence records typically suffer from the effects of biased sampling across geography. spThin implements one way to reduce the effects of such biases, by spatial thinning that removes occurrence records less than a user-specified distance from other records. The user can download the thinned records as a CSV file. This step is optional.

Download Thinned Occurrence CSV

Packages used: ggplot2, sp, rgeos

The user then chooses which environmental variables to use as predictors. These data are in raster form. For this demonstration, WorldClim bioclimatic variables are made available at 3 resolutions. For Maxent and many other niche/distribution modeling approaches, selection of a study region is critical because it defines the pixels whose environmental values are compared with those of the pixels holding occurrence records of the species (Anderson & Raza 2010; Barve et al. 2011). As one way to do so, the user can choose a bounding box or minimum convex polygon around the occurrence records, as well as buffer distance for either.

Future versions will include other sets of environmental variables, as well as allow users to upload their own sets of environmental variables and designate a shapefile indicating a custom study region.

Worldclim homepage

Anderson, R.P. & A. Raza. (2010). The effect of the extent of the study region on GIS models of species geographic distributions and estimates of niche evolution: preliminary tests with montane rodents (genus Nephelomys) in Venezuela. Journal of Biogeography, 37: 1378–1393.
Barve, N., V. Barve, A. Jiménez-Valverde, A. Lira-Noriega, S.P. Maher, A.T. Peterson, J. Soberón & F. Villalobos. (2011), The crucial role of the accessible area in ecological niche modeling and species distribution modeling. Ecological Modeling, 222: 1810–1819.

Automated Runs and Evaluations of Ecological Niche Models
Links to software note and CRAN
Citation: Muscarella, R., Galante, P. J., Soley-Guardia, M., Boria, R. A., Kass, J. M., Uriarte, M., Anderson, R. P. (2014), ENMeval: An R package for conducting spatially independent evaluations and estimating optimal model complexity for Maxent ecological niche models. Methods in Ecology and Evolution, 5: 1198–1205.
Other packages used: ggplot2, raster, dismo

The output of niche/distribution models varies greatly depending on model settings, in particular those affecting the level of model complexity. This step automates the tedious building of a suite of candidate models with differing limitations on complexity. Furthermore, it quantifies their performance on test records. These metrics can aid the user in selecting optimal settings.

Future versions will include other options for partitioning occurrence data. Furthermore, future development of ENMeval or similar packages could provide similar implementations with other algorithms.

Download ENMeval Results CSV
Packages used: raster

View the prediction rasters. You can download them individually and import into a GIS for further analysis.

Download Current Prediction Raster

Wallace was created by an international team of ecologists:

Jamie M. Kass is a coauthor of ENMeval and a PhD student at CUNY Graduate Center and City College of New York. He is currently supported by a CUNY Science Scholarship.

Matthew Aiello-Lammens is the lead author of spThin and a post-doctoral researcher at University of Connecticut. He is currently supported by NSF DEB-1046328.

Bruno Vilela is a coauthor of spThin and a PhD student at the Federal University of Goiás (Brazil) and in Ecology, Conservation and Restoration of Ecosystems at the University of Alcalá (Spain). He is currently supported by a CAPES grant for doctoral studies.

Robert Muscarella is the lead author of ENMeval and a post-doctoral researcher at Aarhus University (Denmark). He is currently supported by NSF DEB-1311367 and NSF DBI-1401312.

Robert P. Anderson is a coauthor of spThin and ENMeval, and a Professor of Biology at City College of New York CUNY. This work was supported by NSF DEB-1119915.