Modeling Habitat Suitability

African Lions, Photo Credit: African Lion and Environmental Research Trust

Background: African lions (Panthero leo) are critically endangered and are under threat from climate change. As environmental conditions change, their habitat distributions will also shift. We modeled and mapped present and future distributions of the African lion using the Wallace package in R and in ArcGIS to understand future habitat suitability.

Problem: Habitat suitability modeling under different climate scenarios will allow wildlife managers to better understand where to create protected reserves to maximize habitat suitability for lions in the future. This analysis provides priority areas for suitable lion habitat under the SSP370 climate scenario.

Approach: In Wallace, we determined African lion habitat suitability change based on lion occurrence data and environmental data that consisted of bioclimatic variables including:

• BIO1 = Annual Mean Temperature

• BIO2 = Mean Diurnal Range (Mean of monthly (max temp - min temp))

• BIO5 = Max Temperature of Warmest Month

• BIO6 = Min Temperature of Coldest Month

• BIO12 = Annual Precipitation

• BIO13 = Precipitation of Wettest Month

• BIO17 = Precipitation of Driest Quarter

We sampled 1000 background points using the environmental data to capture any environmental conditions that may have been missed in the occurrence sample data. Next, we validated the model with a training and testing data set to test the model using spatial partitioning, choosing the “block” method.

We used the Maxent module to ensure both lion occurrence and background data were used and selected the maxnet algorithm. We used a linear (LQ) feature class of our bioclim variables (temperature and precipitation). Additionally, we checked LQHP features and ran the model again. When comparing the response curves between the LQ and LQHP models, the LQHP model was not biologically plausible for the lion. The LQHP response curves showed no clear relationship between the modeled suitability and each predictor variable, indicating that this model was overfit. Therefore, we proceeded with the LQ feature class.

Next, we explored the impacts of climate change on our lion occurrence model. In the Maxent module, we used Model Transfer to create a modeled lion distribution map under the SSP370 climate scenario. This scenario is in the upper-middle part of the full range of scenarios, with an additional radiative forcing of 7 W/m² by the year 2100. We used the latest available data in Wallace and chose the climate model MIROC6 and climate scenario SSP370. Additionally, we explored climate scenario SSP126, with an additional radiative forcing of 2.6 W/m² by the year 2100, and which simulates development that is compatible with a 2°C target. Finding little difference between the two climate scenarios, we proceeded with SSP370. Next, we exported the files from Wallace and pulled the output into ArcGIS Pro. Then, we updated the symbology to be consistent with a specific style layer to visualize the changes in range distribution between present and future scenarios. Lastly, we used Raster math to calculate the change in habitat suitability between the SSP370 model and our original lion distribution model.

Results: Our results showed a change in habitat suitability for the African Lion under climate model SSP370. As shown in Figure 1, the largest area of positive change in habitat suitability is shifting poleward, as noted by the bright red in the map. Yellow indicates no change in suitability, however, blue indicates a negative change in suitability as shown in the middle and upper regions of the area.

Conclusions: Conservation actions for African lions should prioritize potential range shifts from climate change. As shown in our results, suitable habitat shifts poleward and wildlife managers should ensure that there are sufficient protected areas for lions under these future climate scenarios. This range shift could cause increased human wildlife conflict between lions and private landowners who inhabit these areas.

Figure 1. Habitat Suitability Change in African Lions. Change in habitat suitability between SSP370 and future distribution using raster analysis. The map shows a poleward shift in range: yellow indicates no change in suitability, blue denotes a negative change in suitable habitat, and red represents a positive change in habitat suitability.