Astragalus verus Olivier is one of the most important native and medicinal species of Iran, which is also very important in terms of soil protection and the production of tragacanth and honey. In recent years, this valuable species faced to decline due to various reasons. Identifying the spatial pattern and the effects of climate change on plant distribution can be used to develop hypotheses about the reasons for the dryness of this species. This study aimed to determine the spatial pattern and mutual effects of live and dried Astragalus verus, and predict the potential effects of climate change on its distribution and deterioration using the CCSM4 general circulation model under RCP2.6 and RCP8.5 scenarios for the current, 2050, and 2070 periods. In this study, the 12 sites with the dominance of the A. verus and dieback of more than 30 percent were selected in Isfahan and Chaharmahal-va-Bakhtiari provinces. We established one plots with a size of 10×10 m randomly in each site for studying vegetation characteristics. In each plot, the X and Y coordinates of all species were recorded in relation to the origin of the coordinates, and one soil sample was collected exactly from the center of the plot for gradient analysis. Six other soil samples were also collected from the 0-25 and 25-50 cm depth of the soil surface in each plot randomly to eva‎luate the correlation between the soil properties and drying of the species. The weight moisture percentage, soil texture components, pH, and electrical conductivity of all soil samples were measured. In the studied sites, traces of insect larvae activities were observed and two species of wood-eating beetles were identified. The results indicated that there is no significant correlation between soil properties and the drying of the species. The relationship between soil properties and the drying percentage of A. verus was conducted by Principal Component Analysis. The percentage of soil moisture, sand, and clay content were identified as the most influential factors in the distribution of A. verus. To analyze the spatial point pattern of all species, the univariate function was used, and to analyze the mutual effects of live and dried A. verus, Ripley's K bivariate function was applied. The results showed that the spatial pattern of all the species inside the plots changed from clump to uniform and the univariate and bivariate spatial pattern of live and dry A. verus was random. To model species distribution, we selected 83 sites with the dominance of the A. verus (alive and dried species) at different elevation strata where different age classes of the species occurred and the exploitation of katira was not carried out. Nineteen bio-climatic variables on an annual scale for the period from 1979 to 2013 were obtained from the Chelsa climate database and dismo package in R software was used to produce meteorological data for the period from 2014 to 2019. The maps of three physiographic variables (elevation, slope, and slope direction) were produced and an NDVI map was also prepared with a spatial resolution of one kilometer. Spatial distribution maps of BIO1 and BIO12 were produced using the regression method, Kriging methods, and IDW in ArcGIS software. After eliminating the high-correlation layers and the unnecessary layers based on ecological consideration, the annual mean temperature (Bio1), the max temperature of the warmest month (Bio5), the precipitation of the coldest quarter (Bio19), and DEM layers were selected as the model inputs. Eight modeling methods, were used to model the occurrence and dieback of the species in a regional scale using the Biomod2 package in R software. The RF model outperformed the other models as it attained the highest values of AUC and TSS (>0.98). The RF model was employed to develop habitat suitability maps for the occurrence and decline of the species, and to develop species response curves against the most important environmental variables.

حسين بشري , مصطفي تركش اصفهاني

سعيد سلطاني كوپائي

محمد رضا وهابي , رضا مدرس , محمدعلي زارع چاهوكي