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

Accueil du site → Master → Canada → Integrating field sampling and remotely sensed data for the quantitative prediction of soil characteristics across a salt-affected grassland

University of Toronto (1995)

Integrating field sampling and remotely sensed data for the quantitative prediction of soil characteristics across a salt-affected grassland

Davidson, Andrew

Titre : Integrating field sampling and remotely sensed data for the quantitative prediction of soil characteristics across a salt-affected grassland

Auteur : Davidson, Andrew

Université de soutenance : University of Toronto

Grade : Master of Science (MS) 1995

Résumé
A simple local predictor model is described which allows the quantitative mapping of salinity-related soil parameters across an area of semi-arid salt-affected grasslands (Hortobagy National Park, Hungary). Predictions are made using field sampled and remotely sensed data in an uncorrelated weighted averaging scheme. The relative predicted trends in salinity across botanical classes were consistent with those drawn from field sampling, however, the predicted parameter values displayed significantly less variation than the corresponding field sampled measurements. This was explained by (i) spatial averaging ; (ii) the lack of spectral separability between botanical classes ; (iii) limitations of sensor spatial resolutions ; (iv) unrepresentative field sampling ; and (v) changing environmental conditions between the recording of satellite and field data. Despite its limitations, this approach produces a method of analyzing spatio-temporal changes in degrees of soil salinization across the grassland, and input for the driving of further grassland models.

Mots clés : Hungary, Applied sciences, Remote sensing, Geography, Biological sciences, Earth sciences Range management

Présentation (Amicus)

Page publiée le 8 mars 2015, mise à jour le 8 février 2018