The impact of climate change on suitable habitats of Dracocephalum kotschyi in central Iran

Introduction: Conservation and management programs play a vital role in managing degraded ecosystems. However, success in conservation efforts, particularly under climate change, largely depends on identifying suitable habitats for the growth and survival of target species. Dracocephalum kotschyi, a perennial medicinal plant endemic to Iran and belonging to the Lamiaceae family, is found in the highlands of central, northern, and western regions of the country. This species exhibits anti-cancer, anti-viral, anti-tumor, and bactericidal properties. This study predicts the distribution of D. kotschyi under current and future conditions (2050 and 2080) based on two general circulation models, GFDL-ESM4 and MRI-ESM2-0, across three greenhouse gas emission scenarios (SSP126, SSP370, and SSP585) in Isfahan province for the conservation and management of this valuable plant.
Material and Methods: To predict suitable habitats for D. kotschyi, data from 31 occurrence points and 19 bioclimatic variables (current and future conditions) from the CHELSA database and physiographic variables (elevation, slope, aspect from a DEM model) were utilized. A Pearson correlation test was applied to select one physiographic variable and nine bioclimatic variables for modeling. For the modeling process, the algorithms that were suitable for modeling were used in the implementation of the model. Among ten algorithms, eight were employed for predicting the distribution of this medicinal plant, including artificial neural networks, generalized additive models, generalized linear models, flexible discriminant analysis, random forests, regression and classification trees, multivariate adaptive regression splines, and envelope models. The ensemble model includes the integration of all the outputs of the models and was used to predict the favorable habitats of D. kotschyi at present and in the future with a high degree of confidence. 80% of the presence points of this species were used to generate models and 20% of this points were used to measure the performance of the models randomly with 5 repetitions.
Results and Discussion: The predictive models for D. kotschyi showed good to excellent performance, with average AUC (0.922 ± 0.061), TSS (0.821 ± 0.111), and KAPPA (0.821 ± 0.111). The species is primarily located in the northwest, west, and south of the study area. Overall, 9.20% of the area was identified as "excellent habitat suitability," 48.1% as "moderate habitat suitability," and 32.89% as "low suitability" or "unsuitable." The maximum temperature of the warmest month, elevation, and annual precipitation were the most influential variables, explaining 60.33% of the predicted distribution changes. The highest probability of occurrence (P < 0.8) is expected at maximum temperatures between 27 to 32 °C, elevations between 2000 to 3500 meters, and annual precipitation ranging from 200 to 1400 mm. Comparing current and future distributions, the area of excellent and good habitat suitability is projected to decrease significantly (from -13.3% to -99.98%) compared to current distribution, with most suitable habitats expected to be lost. Climate change may lead to an upward shift of D. kotschyi to higher elevations.
Conclusion: In general, mean annual temperature for different regions of the country would increase between 3.5-4.5°C by the year 2050 and annual precipitation would decrease in the range of 7 to 15%. Effective management of these excellent and good habitats requires accelerated decision-making and climate-adaptive operational strategies for the cultivation and restoration of this medicinal plant. Such models, if well-implemented, could play a crucial role in achieving effective conservation and management goals for these valuable medicinal species.