Download Data: Biological Conservation
Methods Summary
We used a machine learning, “maximum entropy” modeling method called Maxent, and its logistic output format (v.3.3.1; Phillips et al 2006, Phillips & Dudík 2008), to estimate environmental suitability in each grid cell within the study area given the modeled relationship between a given species and environmental covariates. We calculated an index of species richness equal to the summed pixel-level projected suitability (from the Maxent models) across all species. To reduce the influence of sampling bias, we constrained the model background data to the environmental space that was sampled by the surveys (Phillips et al. 2009).
We identified areas of conservation importance using the environmental suitability models for all species within a conservation prioritization framework (Zonation; Moilanen et al. 2005) using the core area analysis. All species were given equal conservation priority and all cells were assumed to have equal potential conservation costs.
Available for download below are ascii format grids of the environmental variables used in the models and of the Maxent models representing environmental suitability for each species. Each grid has an associated xml metadata file containing more information about the specific grid as well as projection information.
Environmental Data
| Name | Definition | Grid (ascii) |
|---|---|---|
| Bathymetry | Depth in meters | Download |
| Prevalence of Circumpolar Deep Water | Temperature and salinity defined water mass | Download |
| Chlorophyll | Mg x m3 averaged over 10 years (Nov–Jan; 1997 - 2006) | Download |
| Distance to Shelfbreak Front | Euclidean distance (m) to the 800-m isobath | Download |
| Summer Sea Ice | Mean percent cover (Dec-Jan; 1998 - 2008) | Download |
| Bathymetric Gradient | The angle of maximum change between cells in bathymetry grid (degrees) | Download |
| Species Locations | Coordinates for species presence locations used in the models | Download (csv) |
| Background Data | Coordinates and data for the environmental background used in the models | Download (csv) |
Species Distribution Models
| Name | Image (jpg) | Grid (ascii) | |
|---|---|---|---|
| Light-mantled Sooty Albatross |  |
Download | |
| Killer Whale (Type C) |  |
Download | |
| Antarctic Petrel |  |
Download | |
| Adelie Penguin |  |
Download | |
| Emperor Penguin |  |
Download | |
| Snow Petrel |  |
Download | |
| Crabeater Seal |  |
Download | |
| Weddell Seal |  |
Download | |
| Minke Whale |  |
Download | |
| Species Richness Index | Download | ||
| Zonation Conservation Ranking | Download |
Acknowledgments
Funding provided by the Lenfest Ocean Program and the National Science Foundation (grants OPP-0440643 and ANT-0944411).
Environmental variables where manipulated and image layers created within ArcGIS 9.3 (ESRI 2009) which was made available through a grant from the ESRI Conservation Program.
Citations
Information on modeling methodology can be found in:
Ballard, G., Jongsomjit, D., Veloz, S.D., and Ainley, D.G.. 2012. Coexistence of mesopredators in an intact polar ocean ecosystem: The basis for defining a Ross Sea marine protected area. Biological Conservation 156:72-82. http://dx.doi.org/10.1016/j.biocon.2011.11.017. PDF available here.
More information on model data sources can be found in:
Other relevant citations:
Moilanen, A., A.M.A. Franco, R. Early, R. Fox, B. Wintle & C.D. Thomas. 2005. Prioritising multiple-use landscapes for conservation: methods for large multi-species planning problems. Proc. Royal Soc. London, Series B, Biol. Sci. 272: 1885-1891.
Phillips, S.J., R.P. Anderson & R.E. Schapire. 2006. Maximum entropy modeling of species geographic distributions. Ecol. Appl. 190: 231–259.
Phillips, S.J. & M. Dudík. 2008. Modeling of species distributions with Maxent: new extensions and a comprehensive evaluation. Ecography 31 (April): 161-175.
Phillips, S.J., M. Dudík, J. Elith, C.H. Graham, A. Lehmann, J. Leathwick, & S. Ferrier. 2009. Sample selection bias and presence-only distribution models: implications for background and pseudo-absence data. Ecol. Appl. 19: 181-197.