Introduction

Purpose: Habitat suitability indices (HSIs) are being created for several game species for the first iteration of the Saskatchewan Habitat Management Plan. An HSI is a value from 0 to 1 which describes the relative capacity of a given area of land to support a selected species (1=optimal, 0= not suitable).

Game species HSI maps will be combined with other biodiversity distribution maps to perform several spatial-prioritization exercises using the Zonation software package.

Scope: This version of the Mule Deer HSI is applied to the prairie and parkland ecoregions. The HSI is calculated for each cell of a 1-km grid with a standardized extent and projection which is common to all input data for the prioritization exercise.

The HSI is meant to:

Inputs: We used 30m raster land cover data for Saskatchewan from the Agriculture and Agrifoods Canada (AAFC) Annual Crop Inventory (ACI) 2017 data as our primary input.

The 2017 ACI data has 4 natural forest cover categories. In the prairie ecozone, of the total area classified as forest (7,924 km2) 6% is classified as coniferous, 93% broadleaf, 1% mixedwood, and 0% undifferentiated.

A shrubland classification covers 6,139 km2, and is undifferentiated by species. Visual investigation finds that sparsely forested areas are classified as shrubland (e.g., sandhill areas which are covered in mature but sparse aspen/poplar).

The wetland classification (undiffertiated among wetland tyes) appears to be quite accurate and the highest resolution product available at a province-wide scale.

Example of ACI wetlands delineation near Allan hills

Example of ACI wetlands delineation near Allan hills

For a croland layer, I combined the 54 cropland type categories present in the ACI (note, many of these do not occur in SK) into a single “cropland” category.

Shuttle radar digital evlevation maps (74m scale) are included to identify areas of rugged terrain (such as creek, coullee, and draw habitats) which provide thermal cover and foraging opportunities whose benefits to mule deer may not be fully expressed by their land cover classifications alone. The elevation data were summarized by the “topographic ruggedness index” (TRI) calculated with the ‘raster’ package’s terrain() function in R.

Finally, Habib et al. 2012, showed the mule deer avoid oil & gas well heads within their range, therefore I used SK Ministry of Environment’s GIS warehouse point data on vertical well location to quantify well densities.

Implementation:

All raster manipulation was performed in R 3.5.2 with current versions of relevant pacakges (raster package version 2.8-19).

Defining Suitability Index (SI) Curves

1) Winter cover suitability indices

We used 3 variables to charecterize winter cover suitability.

  • Percent forest & shrubland cover at patch-scale (1-km2): Although mule deer can thrive in open country during summer months and shoulder seasons, forest and shrublands provide thermal cover necessary during winter and particularly in periods of deep snow accumulation for forage. The forest and shrubland layers are combined since the ACI shrubland classification tends toward heavier shrub cover in continuous stands which would provide functional similarity to forest for winter cover.

  • Percent forest cover at landscape-scale (9 km2) modifier: Although mule deer benefit from winter cover and forage provided by forest and shrub habitat types, mule deer do not tend to move deeply into heavily forest areas, therefore we adjusted the value of a habitat patch downward if it was embedded in a heavily forested landscape. For this variable we only used the forest classification of the ACI which is more representative of heavily forested areas.

  • Topographic roughness index at patch-scale (1-km2): Area with large amounts of topographic variation may provide thermal cover and foraging opportunities whose benefits to mule deer may not be fully expressed by their land cover classifications alone.

  • Well Density at patch-scale (1-km2) modifier: Because there is evidence that mule deer avoid oil & gas infrastructure, I used an index of well density (simply the count of wells in a 1-km grid cell) to adjust the value of a patch as winter cover downward. At present this variable includes all well-status categories and should be refined with guidance from someone familiar with the dataset.

These suitability indices were combined as such:

#patch scale winter cover modified by landscape context
winter.cover.SI<- (winter.SI) * forest.9km.SI
#the max SI value of winter cover or TRI modified by well density
winter.HSI<- max(winter.cover.SI, winter.tri.SI) * winter.wells.SI

SI values at any value of of the habitat variables are retrieved using linear interpolation between the breakpoints identified above (via R’s approx() function).

Interactive map of the overall winter habitat suitability index

## Loading required package: viridis
## Loading required package: viridisLite

2) Summer forage & cover suitability indices

Summer lanscape context:

  • Landscape scale vegetation communities:

  • Landscape scale topographic roughness index:

I took the max value of these two indices for the summer landscape context SI.

Patch-scale summer cover:

  • Patch-scale topographic roughness index:

  • Patch-scale summer cover suitability index:

I took the max value of each cover SI (meaning cover types are fully interchangeable.)

summer.cover.patch.SI<- max(grassland.patch.si.r, shrubland.patch.si.r,
                            wetland.patch.si.r, forest.patch.si.r, TRI.patch.si.r)

Patch-scale summer forage:

  • Patch-scale summer forage suitability index:

I took the max value of the summer forage SI:

summer.forage.patch.SI<- max(shrubland.forage.si.r, grassland.forage.si.r, cropland.forage.si.r)

Summer cover and forage suitabilities were combined as the mean value:

Summer habitat suitability interactive map:

Combining Suitability Indices for the overall HSI

Finally, I calculated overall habitat suitability indices (HSI) for each 1-km pixel by taking the mean of the winter and summer habitat suitability SIs.

Literature Cited