A county and state level vector file was downloaded from the US Census Bureau, and clipped to contain the 14 states of interest. Statistics surrounding the number of lyme disease cases per county and county populations were obtained from the CDC and the US Census Bureau. The data was imported into ArcMap, and joined to the county polygon file using a county-and-state indicator column. Points were generated by the random point generator tool, using county boundaries as the delimiting source. This produced points equal to the number of lyme disease cases recorded in 2016 by county. This point file was given XY coordinates, and exported as a csv. file. We now had our sample file to be used in MaxEnt.
The sample file represents the ‘presence localities’: is it at these points where MaxEnt will be performing tests, building training sites, and testing predictions based on environmental variables, which provide the predictive baseline.
The environmental variable data used for our present analyses came from a variety of online sources, so some additional work was needed: MaxEnt is a precise program, and all environmental raster variables have to be exactly alike in terms of extent, projection, and cell size. A Snap Raster was set under Processing Extent, and cell size and raster mask were set to match an initial raster variable in order to ensure proper raster structure.
Data which came in vector form (IPCC Climate Data, PRISM Climate Group, CGIAR Consortium) was projected to GCS NAD 1983, converted to raster form (point/polygon to raster), and extracted by masking the polygon layer of the 14 states. Raster data (World-Clim, University of East Anglia, Web Soil Survey, EarthEnv) was projected, resampled to the chosen cell size, and extracted to the 14 states’ boundaries.
To ensure that all data would come out exactly alike in terms of extent and cell size, all rasters were then referenced to a 200m resolution reference point grid, and converted back to raster form (point to raster). All rasters were then converted to ASCII form, the file type used in MaxEnt analysis.
Data chosen for our analyses was grouped into three models. Model 1 analyzed strictly environmental variables, such as temperature, precipitation, and elevation. Model 2 used environmental variables as well as vegetation factors. Model 3 incorporated deer density estimates per county, to account for the presence of hosts. These variables were chosen because they represent significant determinants of an environment’s suitability to host black-legged tick populations.
| Present Climate Variables | |||
| Name | Data Type | Coverage | Source |
| Annual Average Winter and Summer Temperatures (2020) | Point Vector; spaced 1.8km | Monthly Mean: 2020 | IPCC Climate Data |
| Average Winter Minimum Temp (2017-2018) | Point Vector; spaced 1.8km | Monthly Minimum Mean: 2017-2018 | PRISM Climate Group |
| Average Summer Maximum Temp (2017-2018) | Point Vector; spaced 1.8km | Monthly Maximum Mean: 2017-2018 | PRISM Climate Group |
| Average Summer Precip (2020) | Point Vector; spaced 1.8km | Monthly Mean: 2020 | Climate Change Global Warming Shapefiles – IPCC Climate Data |
| Average Actual Summer Vapour Pressure (1970-2000) | 30s Raster | Monthly Mean | World-Clim – Global Climate Data |
| Average Actual Winter Vapour Pressure (1970-2000) | 30s Raster | Monthly Mean | World-Clim – Global Climate Data |
|
Annual Average PET (1950-2000 average) Potential Evapotranspiration* |
Point Vector; spaced 1.8km | 50 Year Average | Global Potential Evapotranspiration – CGIAR Consortium for Spatial Information |
| Annual Aridity Index (1950-2000 average)** | Point Vector; spaced 1.8km | 50 Year Average | CGIAR Consortium for Spatial Information |
| Global Land Drought Index (1901-2016)*** | NC Raster: 5⁰ by 5⁰ grid | 110 Year Average | University of East Anglia Climatic Research Unit |
| Global High-Resolution Soil-Water Balance (1950-2000 average) Average Winter Soil Water Stress**** | Monthly Mean Raster: 30-arc-seconds | 50 Year Average | CGIAR Consortium for Spatial Information |
| Global High-Resolution Soil-Water Balance (1950-2000 average) Average Summer Soil Water Stress | Monthly Mean Raster: 30-arc-seconds | 50 Year Average | CGIAR Consortium for Spatial Information |
| DEM | Global Multi-resolution Terrain Elevation Data 2010: 7.5-arc-second | 2010 | U.S. Geological Survey – USGS Earth Explorer |
| Population (by county) | Tabular data | County Population Totals and Components of Change: 2010-2016 |
County Population Totals and Components of Change: 2010-2016 – United States Census Bureau |
| County Boundaries | Polygon Vector | 2016 Cartographic Boundary Shapefiles | Cartographic Boundary Shapefiles – United States Census Bureau |
| Soils | Raster Dataset: 30-arc-seconds | Organized by state | Web Soil Survey, National Cooperative Soil Survey |
| Herbaceous Vegetation |
Raster Dataset 30-arc-seconds |
Global 1-km Consensus Land Cover | EarthEnv : Global, remote-sensing supported environmental layers for assessing status and trends in biodiversity, ecosystems, and climate |
| Shrubs | Raster Dataset | Global 1-km Consensus Land Cover | EarthEnv : Global, remote-sensing supported environmental layers for assessing status and trends in biodiversity, ecosystems, and climate |
| Deciduous Broadleaf Trees | Raster Dataset | Global 1-km Consensus Land Cover | EarthEnv : Global, remote-sensing supported environmental layers for assessing status and trends in biodiversity, ecosystems, and climate |
| Terrestrial Ecological Regions | Polygon | World Coverage | TNC Maps from the Nature Conservancy |
| White-tailed deer density estimates across the eastern United States, 2008 | Polygon | County Density | Libraries Digital Conservancy |
| Annual Mean Temperature (1970-2000 Average) | 30-arc-second raster | Annual | World-Clim – Global Climate Data |
| Annual Precipitation (1970-2000 Average) | 30-arc-second raster | Annual | World-Clim – Global Climate Data |
| Annual Temperature Range (Mean of monthly (max temp – min temp)) (1970-2000 Average) | 30-arc-second raster | Annual | World-Clim – Global Climate Data |
| Isothermality (Mean Diurnal Range/Temperature Annual Range) (* 100) (1970-2000 Average) | 30-arc-second raster | Annual | World-Clim – Global Climate Data |
| Max Temperature of Warmest Month (1970-2000 Average) | 30-arc-second raster | Month | World-Clim – Global Climate Data |
| Mean Temperature of Coldest Quarter (1970-2000 Average) | 30-arc-second raster | Quarter | World-Clim – Global Climate Data |
| Mean Temperature of Driest Quarter (1970-2000 Average) | 30-arc-second raster | Quarter | World-Clim – Global Climate Data |
| Mean Temperature of Warmest Quarter (1970-2000 Average) | 30-arc-second raster | Quarter | World-Clim – Global Climate Data |
| Mean Temperature of Wettest Quarte (1970-2000 Average)r | 30-arc-second raster | Quarter | World-Clim – Global Climate Data |
| Mean Temperature Range (1970-2000 Average) | 30-arc-second raster | Annual | World-Clim – Global Climate Data |
| Minimum Temperature of Coldest Month (1970-2000 Average) | 30-arc-second raster | Month | World-Clim – Global Climate Data |
| Precipitation of Coldest Quarter (1970-2000 Average) | 30-arc-second raster | Quarter | World-Clim – Global Climate Data |
| Precipitation of Driest Month (1970-2000 Average) | 30-arc-second raster | Month | World-Clim – Global Climate Data |
| Precipitation of Driest Quarter (1970-2000 Average) | 30-arc-second raster | Quarter | World-Clim – Global Climate Data |
| Precipitation of Warmest Quarter (1970-2000 Average) | 30-arc-second raster | Quarter | World-Clim – Global Climate Data |
| Precipitation of Wettest Month (1970-2000 Average) | 30-arc-second raster | Month | World-Clim – Global Climate Data |
| Precipitation of Wettest Quarter (1970-2000 Average) | 30-arc-second raster | Quarter | World-Clim – Global Climate Data |
| Precipitation Seasonality (Coefficient of Variation) (1970-2000 Average) | 30-arc-second raster | Annual | World-Clim – Global Climate Data |
| Temperature Seasonality (1970-2000 Average) | 30-arc-second raster | Annual | World-Clim – Global Climate Data |
RA = radiation received at top of atmosphere
*PET = 0.0023 * RA * (Tmean + 17.8) * TD0.5 (mm / day)
TD = temperature range
The PET measurements used in the validation are calculated using the more complex Penman-Monteith model applied on direct observations of the various climatic parameters, and were obtained from the FAOCLIM 2 climate station dataset (Allen et al., 1998)
**Aridity Index (AI) = MAP / MAE
MAP = Mean Annual Precipitation\
MAE = Mean Annual Potential Evapo-Transpiration (Higher = wetter)
***Based on Palmer Drought Severity Index (negative numbers mean dry, more positive means wetter)
****Based on soil water content; derived from precipitation minus actual evapotranspiration and runoff (higher number = more moisture)
The reason for separating our study of present variables into environmental, vegetation and host categories was to gauge the model’s reaction to the inclusion of each type of information. Model 1 used data from a variety of sources, and represents our first attempt at tick suitability prediction. Model 2 & 3 used a different set of environmental data (an integrated bioclimatic dataset from World-Clim) from Model 1, and were bolstered with vegetation and host data.
Model 4 is a projection of current environmental conditions to conditions calculated to exist in 2050. Climatic variables were used exclusively in this model. Once the program was fitted correctly, MaxEnt produced AUC results, response curves, jackknife analyses, and raster images in ASCII format. Model 4 allowed us to test the general hypothesis of the northern migration of ticks and the subsequent increased risk of Lyme disease for regions further north.
| Climate Variables for Analysis of 2050 (RCP 4.5) | |||
| Name | Data Type | Coverage | Source |
| Annual Mean Temperature | 30-arc-second raster | Annual | World-Clim – Global Climate Data |
| Annual Precipitation | 30-arc-second raster | Annual | World-Clim – Global Climate Data |
| Annual Temperature Range (Mean of monthly (max temp – min temp)) | 30-arc-second raster | Annual | World-Clim – Global Climate Data |
| Isothermality (BIO2/BIO7) (* 100) | 30-arc-second raster | Annual | World-Clim – Global Climate Data |
| Max Temperature of Warmest Month | 30-arc-second raster | Month | World-Clim – Global Climate Data |
| Mean Temperature of Coldest Quarter | 30-arc-second raster | Quarter | World-Clim – Global Climate Data |
| Mean Temperature of Driest Quarter | 30-arc-second raster | Quarter | World-Clim – Global Climate Data |
| Mean Temperature of Warmest Quarter | 30-arc-second raster | Quarter | World-Clim – Global Climate Data |
| Mean Temperature of Wettest Quarter | 30-arc-second raster | Quarter | World-Clim – Global Climate Data |
| Mean Temperature Range | 30-arc-second raster | Annual | World-Clim – Global Climate Data |
| Minimum Temperature of Coldest Month | 30-arc-second raster | Month | World-Clim – Global Climate Data |
| Precipitation of Coldest Quarter | 30-arc-second raster | Quarter | World-Clim – Global Climate Data |
| Precipitation of Driest Month | 30-arc-second raster | Month | World-Clim – Global Climate Data |
| Precipitation of Driest Quarter | 30-arc-second raster | Quarter | World-Clim – Global Climate Data |
| Precipitation of Warmest Quarter | 30-arc-second raster | Quarter | World-Clim – Global Climate Data |
| Precipitation of Wettest Month | 30-arc-second raster | Month | World-Clim – Global Climate Data |
| Precipitation of Wettest Quarter | 30-arc-second raster | Quarter | World-Clim – Global Climate Data |
| Precipitation Seasonality (Coefficient of Variation) | 30-arc-second raster | Annual | World-Clim – Global Climate Data |
| Temperature Seasonality | 30-arc-second raster | Annual | World-Clim – Global Climate Data |
Next section: Results