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

Accueil du site → Master → Afrique du Sud → The use of remote sensing for soil moisture estimation using downscaling and soil water balance modelling in Malmesbury and the Riebeek Valley

University of the Western Cape (2014)

The use of remote sensing for soil moisture estimation using downscaling and soil water balance modelling in Malmesbury and the Riebeek Valley

Möller, Jason John

Titre : The use of remote sensing for soil moisture estimation using downscaling and soil water balance modelling in Malmesbury and the Riebeek Valley

Auteur : Möller, Jason John

Université de soutenance : University of the Western Cape

Grade : Magister Scientiae – MSc 2014

Résumé
Soil moisture forms an integral part of the hydrological cycle and exerts considerable influence on hydrological processes at or near the earth’s surface. Knowledge of soil moisture is important for planning and decision-making in the agricultural sector, land and water conservation and flood warning. Point measurements of soil moisture, although highly accurate, are time consuming, costly and do not provide an accurate indication of the soil moisture variation over time and space as soil moisture has a high degree of spatial and temporal variability. The spatial variability of soil moisture is due to the heterogeneity of soil water holding properties, the influence of plants, and land uses. The downscaling of satellite microwave soil moisture estimates and soil water balance modelling was investigated at six transects in the semi-arid, Western Cape Province of South Africa, as alternatives to in situ soil measurements. It was found that microwave soil moisture estimates compared well to in situ measurements at the six transects (study sites), with coefficient of determination (r2) values greater than 0.7 and root mean square error (RMSE) values less than 1.5%. Downscaling using the universal triangle method, performed well at 4 of the 6 transects, with r2 values great than 0.65 and low to moderate RMSE values (0.5-12%). Soil water balance modelling similarly performed well in comparison with in situ measurements at 4 of the transects with regards to r2 values (>0.6) but had moderate to high RMSE (4.5-19%). Poor downscaling results were attributed to fine scale (within 1 km) surface heterogeneity while poor model performance was attributed to soil hydrological and rainfall heterogeneity within the study areas.

Présentation

Version intégrale (19,97 Mb)

Page publiée le 11 février 2016, mise à jour le 11 juin 2018