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

Accueil du site → Doctorat → Royaume-Uni → 1991 → Development of a catchment scale erosion model for semi-arid environments and its implementation through remote sensing

University of Bristol (1991)

Development of a catchment scale erosion model for semi-arid environments and its implementation through remote sensing

Garg, Pradeep Kumar

Titre : Development of a catchment scale erosion model for semi-arid environments and its implementation through remote sensing.

Auteur : Garg, Pradeep Kumar.

Université de soutenance : University of Bristol

Grade : Doctor of Philosophy (Ph.D.) : 1991

Résumé
Water-induced soil erosion is one of the major hazards in semi-arid regions. These regions are vulnerable to erosion due to the erratic rainfall pattern, varied lithology, and low and discontinuous vegetation cover. The present study aims to develop a distributed model for the estimation of erosion and sediment yield for a catchment. The model, with its simple structure, uses a grid cell approach, incorporating the rill-interrillerosion processes and topographic characteristics of the catchment. It incorporates criteria to define rill and interrill cells within a catchment, based on threshold values of 3o slope steepness and 30&37 vegetation cover, and the concept of Hortonian overland flow. The model generates the flow network on an individual storm basis. It not only computes the net erosion in each grid cell, but also accounts for deposition of the eroded material as it moves downstream from the point of origin. The model is designed primarily for conditions which prevail in semi-arid environments. The additional feature of the model is that the input data requirements are kept to a minimum, unlike other hydrologic and erosion distributed models. The model is applied for the Albudeite catchment, about 15km north-west of the city of Murcia, south-east Spain, covering an area of approximately 19km2. The input parameters are derived by analysing data from various sources ; air-photos, SPOT and multispectral digital data, digital elevation data, and field data ( e.g. rainfall, infiltration and soil properties). An unsupervised clustering method produced encouraging results for the vegetation classification, compared with the supervised maximum likelihood method, especially when identifying bare and unvegetated natural surfaces of the catchment. The best vegetation classification for the month of April was obtained using a combination of the SPOT XS 1, XS 3 and NDVI data which produced results with 81 8&#37 accuracy at 95&#37 confidence level.

Annonce : EThOS (UK)

Proquest Dissertations & Theses

Page publiée le 25 mai 2009, mise à jour le 28 mars 2017