مدلسازي رويشگاه بالقوه گونه ي گياهي گون زرد ﴿Astragalus verus Olivier﴾ با استفاده از دو روش تحليل عاملي آشيان اكولوژيك ﴿ENFA﴾ و رگرسيون لجستيك ﴿LR﴾ در منطقه فريدون شهر استان اصفهان

چكيده انگليسي :

96 Modeling the Potential Habitat of Astragalus verus Olivier Using Ecological Niche Factor Analysis and Logistic Regression in Fereydounshahr Region Isfahan Province Mojdeh Safaei Email address mojdeh safaei@na iut ac ir Sep 18 2012 Department of natural Resources Isfahan University of Technology Isfahan 84156 83 11 Iran Degree M Sc Language Persian Supervisors Mahdi BassiriIsfahani Bassirim@cc iut ac ir and Mostafa TarkeshIsfahani Tarkesh@cc iut ac ir Abstract Restoring high value medicinal and industrial plant species is quite important in terms ofeconomical and social point of view Nowadays progresses in statistical methods and geographicinformation techniques facilitate habitat potential mapping using more advanced modelingapproaches Species distribution model can be defined as a statistical analytical algorithm whichpredicts species distribution from field observations and available environmental maps This studyaimed to generate potential distribution map of Astragalus verus Olivier a valuable medicinal plantsin Fereydunshahr rangelands in Isfahan province using Logistic regression LR and Ecological nichefactor analysis ENFA methods Soil maps were obtained from 70 soil profiles and climatic mapswere produced from 10 climatic stations data Geo statistical methods were used to convert theobtained data to digital maps with 90 by 90 meter resolutions Potential distribution map of thespecies in the study area was produced using presence absence data of 100 sites using a stratified random sampling approach For each sampling site presence and absence of the species as well as theenvironmental variables were recorded The relationships of the plant species distribution andenvironmental variables were studied using LR and ENFA methods In the next step habitat potentialmaps of the species were produced using these two methods ENFA as a profile model is one of themost frequently used techniques to generate species distribution map which uses species presence only data This method integrates geographic information system and factor analysis to producehabitat distribution map from the species occurrence and environmental data According to the results the environmental factors influencing determination of the ecological niche of the species in order ofimportance were soil electrical conductivity Caco3 content pH slope degree and monthly meantemperature The produced species distribution model had high accuracy indicated by calculatedKappa coefficient and ROC area under curve plots being 0 72 and 0 85 respectively According toENFA results 10000 hectare or 10 of the study area was recognized as suitable area for the speciesestablishment LR from discrimination group models is another modeling approach that enables toproduce distribution maps using presence and absence data According to LR results annual meanprecipitation soil clay content mean temperature of warmest season and slope degrees were the mostimportant environmental factors influencing the distribution of the species According to this modelanalysis 21000 hectare or 21 percent of the whole study area had excellent habitat condition for thespecies The produced species distribution model had acceptable accuracy as calculated Kappacoefficient and ROC area under curve plots were 0 75 and 0 5 respectively Analysis of samplingsizes on the accuracy of the studied modeling approaches revealed that ROC area under curve plotshad less variability to the Kappa coefficient and is more appropriate method to analyze the modelsoutputs accuracy According to produced potential maps of the species and response curve results appropriate Astragalus verus habitats characterized by moderate to relatively heavy soil texture 44 to48 percent clay content moderate slope 10 to 20 percent low electrical conductivity less than 0 3dS m saturated water content of 45 to 47 and pH equivalent to 7 5 The studied modelingapproaches can be used broadly to allocate appropriate areas for rehabilitation and protection

صفائي، مژده

مدلسازي پراكش جغرافيايي گونه هاي گياهي , زمين آمار , سيستم اطلاعات جغرافيايي