Habitat potential modelling and mapping of Teucrium polium using machine learning techniques

Assessing wildfire impact on Trigonella elliptica habitat using random forest modeling

Wildfires have a significant impact on ecosystems worldwide, especially on the degradation of arid and semi-arid rangelands. This research focuses on assessing the effects of wildfires on the habitat of Trigonella elliptica, a valuable herb species found in the central rangelands of Iran. To achieve this, the Random Forest (RF) algorithm has been deployed to predict T. elliptica habitat and fire hazard using socio-environmental variables in Yazd province, Iran. 225 fire points and 103 habitat locations were used for model training and testing. The IncNodePurity index and Probability Curves (PC) have been utilized to determine the influence of socio-environmental variables. The combination of the prediction maps of the habitat and wildfires pointed out the possible damage due to fire. The high performance of the RF model is confirmed by the area under the curve (AUC) and the true skill statistic (TSS) values (0.90 and 0.81 for the habitat; 0.92 and 0.82 for the wildfire). The importance assessment of variables revealed that elevation, slope, and precipitation are the most influential variables in the distribution of T. elliptica, while distance to roads, population density, and wind speed are the key factors affecting wildfire occurrence. In the final map, a comparison of different regions of T. elliptica habitat under fire hazard with fire-free habitats using Kruskal-Wallis and Dunn tests indicated that the fire hazard in the T. elliptica habitat is a serious concern. Since the areas with the highest fire hazard and the highest presence of T. elliptica cover approximately 2311.38 km2, neglecting these regions could lead to the gradual reduction of T. elliptica, and create conditions for secondary succession dominated by less valuable annual species. The findings of this study underscore the importance of implementing fire management strategies, protection projects, and continuous monitoring to ensure the safety and conservation of the T. elliptica habitat.

Minimal climate change impacts on the geographic distribution of Nepeta glomerulosa, medicinal species endemic to southwestern and central Asia

Medicinal plants are valuable species, but their geographic distributions may be limited or exposed to extinction by climate change. Therefore, research on medicinal plants in the face of climate change is fundamental for developing conservation strategies. Distributional patterns for a semi-endemic medicinal plant species, Nepeta glomerulosa, distributed in southwestern and central Asia was determined based on a maximum-entropy algorithm. We evaluated potential geographic shifts in suitability patterns for this species under two Shared Socioeconomic Pathways scenarios (SSP2-4.5 and SSP5-8.5) of climate change for 2060. Our models based on climatic features indicate that the species occupies montane areas under current conditions; transfer of the model to future climate scenarios indicated that suitable areas for the species will increase in general, and the species will likely track its favored set of climate conditions. But the types and degrees of these changes differ among areas. Our findings can be used to inform conservation management programs for medicinal, endemic, and endangered species that probably respond similarly to climate change in southwestern and central Asia.