A framework for integrating inferred movement behavior into disease risk models

Table 2 Set of potential predictor variable layers used in creating the anthrax risk map

Environmental variable (units)	Predictor name	Data source	Final Model
Soil pH x 10 in \(H_2O\)	pH	SoilGrids\(^{*}\)	X
Soil Organic Carbon Content (g/kg)	OC	SoilGrids\(^{*}\)	X
Soil Cation Exchange Capacity (cmolc/kg)	CEC	SoilGrids\(^{*}\)	X
Mean annual temperature (C\(^{\circ }\))	bio1	WorldClim\(^{\dagger }\)	X
Annual temperature range (C\(^{\circ }\))	bio7	WorldClim\(^{\dagger }\)	X
Annual precipitation (mm)	bio12	WorldClim\(^{\dagger }\)
Precipitation of the wettest month (mm)	bio13	WorldClim\(^{\dagger }\)	X
Precipitation of the driest month (mm)	bio14	WorldClim\(^{\dagger }\)
Mean NDVI	NDVI	Landsat 7\(^{\ddagger }\)
Maximum NDVI	max_ndvi	Landsat 7\(^{\ddagger }\)	X
Minimum NDVI	min_ndvi	Landsat 7\(^{\ddagger }\)	X
Range NDVI	range_ndvi	Landsat 7\(^{\ddagger }\)	X

These predictors were compiled based on their use in similar ecological niche modeling efforts of Bacillus anthracis (see [40] and [23] for more details). Several of these variables were eliminated, however, due to collinearity with other, more important, variables in the set. An ‘X’ in the ‘Final Model’ column indicates the inclusion of that variable in the final MaxEnt model. Data sources: \(^{*}\) [46]; \(^{\dagger }\) [47]; \(^{\ddagger }\) courtesy of the U.S. Geological Survey (https://espa.cr.usgs.gov/)

ISSN: 2051-3933