9 - Habitat Suitability Mapping of Artemisia aucheri Boiss Based on the GLM Model in R

Abstract

Habitat suitability models can be a powerful tool in the prediction of spatial distribution of plant species. In the current study, the generalized linear model (GLM) was applied to model spatial distribution and suitable habitats of Artemisia aucheri Boiss, an endemic important species in mountainous parts of the northern Fars Province, Iran. The spatial locations of A. aucheri were recorded and mapped using geographic information system (GIS) tools. All identified locations of species presence were randomly split into 70% for model training and 30% for model validation. The conditioning variables, including mean annual rainfall, mean annual temperature, elevation, slope degree, aspect, geology, sand percentage, soil acidity (pH), and soil organic carbon (OC), were used to model the spatial distribution of A. aucheri using the GLM model. After preparing thematic maps, the multicollinearity test was performed to select the favorite variables. In the next step, habitat suitability maps (HSMs) were produced and categorized into four classes of low, moderate, high, and very high. The model performance was determined using the area under the curve (AUC) of the receiver operating characteristic (ROC). The results revealed that the elevation, mean annual temperature, mean annual rainfall, geology, pH, and sand percentage had significant impact on the distribution of A. aucheri in the study area. The validation of HSM produced using the GLM model showed an AUC value of 0.855 (85.5%). Based on HSM, 54.43% of the total study area had very high suitability habitats. Meanwhile, the area with low potential was 14.38%. In general, the GLM model shows a reasonably accurate approach for habitat modeling of A. aucheri in the study area. The HSMs could be a useful tool for local rangeland managers, conservational agencies, and environmental planners.