3645

3444

كارشناسي ارشد

مرتعداري

اصفهان :دانشگاه صنعتي اصفهان ،دانشكده منابع طبيعي

1386

سيزده ،104، [II]ص.: مصور ، جدول ، نمودار

ص. ع. به فارسي و انگليسي

جمال الدين خواجه الدين

عليرضا سفيانيان

13/6/86

محمد رضا وهابي

1396/02/26

كتابنامه

منابع طبيعي

مهندسي منابع طبيعي

ID3444

به فارسي و انگليسي :قابل رويت به صورت نسخه ديجيتال

Abstract Pasture utilization management with domestic animal grazing should be based onunderstanding of the special and temporal hay producing To estimate the pasture production inarid and semiarid regions needs reliable and repeatable techniques the techniques which are usefultools for managers Traditional methods of produced hay estimation have spatial and temporallimitations and can not be applied on vast areas Vegetation cover changes during a year and needsrepeated monitoring for management Under these circumstances during a year only one timeevaluation and measuring of the hay through traditional method is not enough to reach a long termmanagement program Application of remotely sensed data is a suitable method to evaluate theplant production and sequential changes detection The aims of the present research was the studyof the MODIS data capabilities to estimate the vegetation cover production on Semirom and Brojenregions as well as selecting proper vegetation indices and studying the plant production changesthrough the multi temporal data Various preprocessing including image rectification was appliedwith geo referencing of the image to a registered image with a RMSe 0 5 pixel The atmosphericand topographic corrections were applied using subtraction of dark objects method and the Lambertmethod accordingly Image processing including FCC PCA vegetation indices and supervisedclassification were employed to produce the land use and pasture production maps Field datacollection were started on June 2005 on 800 000 ha and prolonged about 4 months Variousvegetation types were sampled using stratified random sampling method 20 random samplingpoints were selected and the plant production was estimated with double checking method FourMODIS data sets were used for this study Digital data and the indices maps were used asindependent data and the field data as dependent variables The resulted models were processed onand the resulted images were categorized in 7 classes Finally the produced maps were controlledfor their accuracies Also post classification method was used to determine the 9 land use changedetections The main Kappa were estimated for the produced maps 79 61 and 72 6 The resultsconfirmed that the NDVI and SAVI maps have higher relations with the field data in addition theindices which were produced with SWIR bands have a higher relations with field data where thecover and yield are high On these sites the R square was exceed to 85 Most of the producedmaps had higher accuracies with NDVI and SAVI indices and their Kappa were very high Theland use changes are very high during the growing season which about 90 of the pasturelands During the growing seasons the most pasture products changes belongs to class 5 and 2 in theNDVI and SAVI indices map The out coming results of this study prove that the MODIS dataestimates the plant production very well Through this tools one can monitor the hay productions that is very useful for resources management as well as decision making for a logic pasture utilization

جمال الدين خواجه الدين

عليرضا سفيانيان

محمد رضا وهابي