شماره راهنما :
1080 دكتري
پديد آورنده :
زراعت پيشه، مجتبي
عنوان :
نقشه برداري رقومي و مقياس سازي كلاس ها و برخي ويژگي هاي خاك در بخشي از اراضي استان چهارمحال و بختياري
محل تحصيل :
اصفهان: دانشگاه صنعتي اصفهان، دانشكده كشاورزي
صفحه شمار :
سيزده، [۱۶۴]ص.: مصور، جدول، نمودار
استاد راهنما :
شمس الله ايوبي
استاد مشاور :
حسين خادمي، اعظم جعفري
توصيفگر ها :
نقشه برداري رقومي , پيش بيني كلاس هاي خاك , تراكم زدائي , ژئومورفولوژي , تنوع خاك
استاد داور :
احمد جلاليان، حسين شريعتمداري، رضا جعفري
تاريخ ورود اطلاعات :
1396/08/07
كد ايرانداك :
ID1080 دكتري
چكيده انگليسي :
159Digital soil mapping and scaling of soil classes and some of soil properties in a selected area in Chaharmahal Va Bakhtiari Province Mojtaba Zeraatpisheh E mail Address zeraatpishem@yahoo com May 21 2017 Department of Soil Science Isfahan University of Technology Isfahan 84156 83111Supervisor Shamsollah Ayoubi ayoubi@cc iut ac irAdvisors Hossein Khademi hkhademi@cc iut ac ir Azam Jafari a jafari@uk ac irCollege Graduate Coordinator Mehran Shirvani Assoc Prof Department of Soil Science Collaege of Agriculture Isfahan University of Technology Department of Soil Science Collaege of Agriculture Shahid Bahonar University of KermanAbstractThe purpose of this study consists of four sections 1 the relationship between soil andlandscape 2 spatial prediction of soil classes 3 spatial prediction of some soil properties and 4 disaggregation of soil classes at scale of 1 1 000 000 in an area of 86000 ha inChaharmahal Va Bakhtiari Province In order to achieve the objectives of this study thegeomorphology map was prepared using standard methods The location of 100 profileswere found by applying conditioned Latin Haypercube Sampling cLHS and terraincovariates and also 25 profiles of legacy soil data were used In addition to profilesampling 209 surface samples were collected in order to predict the spatial distribution ofsome soil properties The results of soil landscape evolution showed that in the study areathe soil evolution influenced and controled by topography parent material and undergroundwater level in some area so that in areas with high topographic the minimum soil evolutionand in the stable or flat areas maximum soil evolution have been observed In term ofaccurate maps results showed that there is no great difference in the prediction soil mapswhich produced by digital soil mapping in the scenario with geomorphology map andwithout geomorphology map The geology map as indicator of parent material alreadyexplained the most of the variability and distribution of soil types at different soiltaxonomic levels this reason could be the low impact of geomorphology map to increasethe map purity Inclusion of the geomorphology map did slightly improve map purity andKappa Based on map purity and Kappa Multinomial logistic regression had betterperformance at higher taxonomic levels order and suborder levels however randomforest showed better performance at lower taxonomic levels great group and subgrouplevels Multinomial logistic regression was less sensitive than random forest to a decreasein the number of training observations Based on 10 fold cross validation for SOC non linear models Random Forest and Cubist had lower sdatndard deviation in R2 To predict
استاد راهنما :
شمس الله ايوبي
استاد مشاور :
حسين خادمي، اعظم جعفري
استاد داور :
احمد جلاليان، حسين شريعتمداري، رضا جعفري
