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

Titre : Improving satellite remote sensing methodologies for analyzing landscape dynamics in arid environments with focus on Egypt

Auteur : Delgado Blasco, José Manuel

Université de soutenance : KU Leuven

Grade : Doctor of Science : Geography from KU Leuven 2023

Résumé partiel
Monitoring rapidly changing landscapes, especially those which cannot be easily accessed due to climate conditions, access or insecurity has been always a very chal lenging task. It is particularly timely to be able to accurately monitor such changing landscapes given the current socio-economical changes and population growth of many developing countries, as land is a limited resource and it is crucial that it is used in a planned and sustainable manner. Egypt is the perfect example of a country where the population increase linked to urban growth goes at the expense of agricultural land. Landscape dynamics in Egypt are changing very rapidly. In particular the continuous growth of urban areas, sand storms regularly covering roads and buildings, and dunes approaching villages and crop fields in the western edge of the Nile floodplain call for a continuous monitoring strategy. Satellite remote sensing is optimally suited to provide the required spatio-temporal coverage. Yet, currently used satellite remote sensing technologies have shortcomings that need to be addressed. Optical remote sensing techniques fail to detect the construction of buildings when the same materials (mud blocks) are used as their natural surroundings. Dunes need also to be regularly monitored, specially the ones that are considered hazards as they migrate towards inhabited areas or crop fields, but this has been tradi tionally done or at very small scale and using time consuming and expensive procedures. In this research, we investigated the combination of optical and microwave (SAR) satel lite data to overcome the aforementioned limitations and we propose the usage of SAR data to analyse dune migration phenomena. In this study we develop appropriate methods and provide accurate maps to facilitate landscape dynamics analysis for Egypt, focusing on urban sprawl and dune dynamics. We improved the detection of urban areas and developed an automatic method for the monitoring the dynamics of individual dunes. Our method for urban change detection overcomes the limitations of previous urban maps derived from optical sensors and improved state-of-the-art global urban layers. Additionally, our method for individual dune dynamics improves the traditional field survey methods, overcoming limitations of the optical sensor methods and is, to our knowledge, the first automatic method using SAR backscatter to monitor individual dune dynamics at dune-field scale. We employed satellite data fusion methods, artificial neural networks, and big data analysis to develop land cover classification methods applicable not only in Egypt but also in similar environments. Additionally, the developed method for dune dynamics analysis with SAR is also exportable to other dune-fields and dune types, enabling a more frequent individual dune dynamics monitoring at dune-field scale, which was not possible until now

Présentation Version intégrale (18,1 Mb)