Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Doctorat → Autriche → The use of remote sensing to evaluate and detect desert regions

Universität Wien (2011)

The use of remote sensing to evaluate and detect desert regions

Komaki, Chooghi Bairam

Titre : The use of remote sensing to evaluate and detect desert regions

Auteur : Komaki, Chooghi Bairam

Université de soutenance : Universität Wien

Grade : Doktor der Naturwissenschaften (Dr. rer.nat.) 2011

Résumé
Remote sensing plays a significant role in providing up-to-date data for the estimating of empirical indices in studying the environment, especially in drylands. The spectral and thermal bands in satellite images are also applied to calculate the indices to detect, identify, and evaluate the natural phenomena in drylands such as land degradation and desertification. In this project, for the identification of desertification in the Kashan-Qom region in Central Iran, five main indicators of desertification are used as follows : vegetation, land surface temperature, erosion, drought, and flooding ; therefore, these indices are selected as Vegetation Condition Index (VCI), Temperature Condition Index (TCI), Revised Universal Soil Loss Equation (RUSLE), and Standardized Precipitation Index (SPI), and runoff (Q), respectively. The multi-spectral satellite images of MODIS are used for the calculation of remotely sensed indices such as Vegetation Condition Index (VCI) and Temperature Condition Index (TCI). Furthermore, the ancillary data-based indices, Revised Universal Soil Loss Equation (RUSLE), and Standardized Precipitation Index (SPI), and runoff (Q), are also estimated. Then several desertification maps are produced in two models : conventional method and fuzzy model. The result of each model is also evaluated, that is, the results are assessed by the supplying of field sampling as ground truth references and the defining of error matrix. In the fuzzy modelling, a rule-based system is built by expert knowledge and data-induction method. According to the obtained results, even though the accuracy of the fuzzy model is lower than the conventional method, the fuzzy model represents the uncertainty in the classes of resulted desertification by providing a map for each class.

Mots clés : desertification / remote sensing / erosion / GIS / fuzzy modeling / MODIS / Iran / Desertifikation / Fernerkundung / Erosion / GIS / Fuzzy Modellierung / MODIS / Iran

Présentation

Version intégrale (4 Mb)

Page publiée le 28 janvier 2017, mise à jour le 14 mars 2019