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

Accueil du site → Doctorat → Royaume-Uni → 1993 → Monitoring and modelling rangeland vegetation in Tunisia using satellite and meteorological data

University of Reading (1993)

Monitoring and modelling rangeland vegetation in Tunisia using satellite and meteorological data

Wellens, Jane

Titre : Monitoring and modelling rangeland vegetation in Tunisia using satellite and meteorological data.

Auteur : Wellens, Jane.

Université de soutenance : University of Reading

Grade : Doctor of Philosophy (Ph.D.) : 1993

Résumé
Normalised Difference Vegetation Index (NDVI) images generated from Advanced Very High Resolution Radiometer (AVHRR) data were used to investigate the inter- and intra-annual variations in rangeland vegetation in Tunisia. The images provide information about the vegetation status for nine agricultural years from September 1983 to August 1992. Vegetation and ground radiometric measurements were collected for ten rangeland sites during field work in 1990 and 1992 and meteorological data were also obtained for recording stations close to these sites. Strong correlations were found between estimates of percentage vegetation cover and ground radiometric measurements made at the sites, and also between AVHRR-NDVI and vegetation cover. The relationship between NDVI and estimates of above-ground green biomass was not found to be as strong. Rainfall and NDVI data for each fieldsite were compared, and significant relationships were found between the two. These indicated that there was a delay in the vegetation response to the rainfall which varied both between and within fieldsites according to the timing and distribution of rainfall within a year. The NDVI data also showed that the vegetation at several of the fieldsites remains active throughout the summer months although there is generally no rainfall at this time. A soil water-balance model to estimate the amount of moisture available to the vegetation was developed from the work of researchers in the rangelands of the USA. This estimates both the actual transpiration through the agricultural year and the ratio of actual to potential transpiration, which is closely related to yield. The outputs of the model for each fieldsite were compared with the corresponding NDVI data. A significant relationship was found between estimates of daily actual transpiration derived from the model and the NDVI. For seven of the ten fieldsites this relationship was found to be at least as strong as that between rainfall and the NDVI

Mots clés : Arid ecosystems Geography Pattern recognition systems Pattern perception Image processing Ecology Geography Pattern recognition systems Ecology

Annonce : EThOS (UK)

Proquest Dissertations & Theses

Page publiée le 26 mai 2009, mise à jour le 26 mars 2017