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

Accueil du site → Master → Pays Bas → 2019 → Remote sensing assessment of land restoration interventions in South Africa

University of Twente (2019)

Remote sensing assessment of land restoration interventions in South Africa

Muchando, A. (Archford)

Titre : Remote sensing assessment of land restoration interventions in South Africa

Auteur : Muchando, A. (Archford)

Etablissement de soutenance : University of Twente

Grade : Master of Science in Geo-Information Science and Earth Observation 2019

Résumé
Restoration interventions have been implemented in the Baviaanskloof catchment following the launch of the Subtropical Thicket Restoration Programme (STRP) in 2004. This was necessitated by the need to restore landscapes that were heavily degraded due to long periods of overgrazing by livestock. Since the year 2005, over 1 000 ha of degraded land in the Baviaanskloof were put under active restoration whereby native Spekboom (Portulacaria afra) was planted to revegetate the land. Over two decades ago, some of the landowners in the Baviaanskloof have removed livestock from their farms to allow for natural vegetation regeneration. Currently, four restoration intervention types are present in the area a) Revegetation only with Spekboom without protection from livestock grazing, b) Protecting (fencing) the Spekboom revegetated areas from livestock and wild animals grazing also known as ‘thicket wide plots’, c) Planting Spekboom in livestock exclusion areas and d) Excluding livestock from degraded areas. The aim of this research was to assess the impact of these restoration interventions on green vegetation cover observed through satellite imagery. Firstly, Normalised Difference Vegetation Index (NDVI) classes that describe greenness over time were generated from Landsat satellite images from 2000 to 2018. These NDVI classes were computed using an unsupervised classification method. Six and three NDVI classes that widely occur in restoration and nonrestoration areas respectively were selected for further analysis. Secondly, the inter-annual and intra-annual (seasonal variations) trends for the nine selected NDVI classes were estimated based on average NDVI per class. The results from the trend analyses show a relatively similar trend regardless of the restoration intervention which was also compared to the non-restored areas. Lastly, the Before-After Control-Impact (BACI) comparative model was also used to assess the variability in NDVI trends between restored (impact) areas and multiple non-restored (control) sites over time. The findings show that the impact on the green vegetation cover varies across the restoration interventions. Eliminating grazing from Spekboom revegetated areas can provide the required solution in restoring the degraded subtropical thickets. This is because the Spekboom revegetated areas that are protected from either wild animals or livestock grazing have recorded a significant increase in green vegetation cover in comparison with non-restored areas. While a substantial decrease in green vegetation cover was detected in revegetated areas that are still open to livestock grazing. These findings can be used by the restoration authorities for management and planning purposes thus they can provide feedback to the restoration authorities on restoration interventions. The comparative method used in this study can be applied to any landscape in South Africa and the world at large for screening of the restoration interventions.

Version intégrale (ITC)

Page publiée le 25 septembre 2019