Predicting potential distribution of plant species by modeling techniques in southern rangelands of Golestan, Iran

Abstract

This paper investigates the use of ‘Ecological niche factor analysis’ (ENFA) method for modeling plant species, geographic distributions with presence-only data and ‘Logistic Regression’ (LR) methods for the investigation of plant species distribution with presence-absence data in south of Golestan. Plant density and cover, soil texture, pH, electrical conductivity (EC), gypsum contents, organic matter, lime and topography (elevation, slope and aspect) are the sample variables using the randomized systematic method. W ithin each vegetation type, the samples were collected using 10 quadrates placed at an interval of 5 m along three 50m transects. To map soil characteristics, the geostatistical method was used. According to the results in the ENFA method, variable amounts of clay and slope impact on Artemisia aucheri type, Elevation and sand content had the greatest impact on the Festuca ovina- Astragalus gossypinus type, because elevation gradients create varied climates, along with the resultant soil differentiation; promoting the diversification of plant species, and the effect of environmental factors were important amount of sand and aspect on the Bromus tomentellus type and amount of sand, elevation, and rate of organic carbon had the greatest impact on Bromus tomentellus-Festuca ovina type. Butin LR method, slope factor had the most impact on the distribution of Artemisia aucheri type .The most important factors affecting the Artemisia aucheri habitat are the degree of slope ,Sand and elevation effect of the Festuca ovina- Astragalus gossypinus type. This aspect affected the Bromus tomentellus type and Elevation affected the Bromus tomentellus- Festuca ovina type. The accuracy of the logistic regression results regarding the Kappa Index was more than ENFA method.