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

Accueil du site → Master → Afrique du Sud → 2019 → An assessment of the hydrologic response of the Keiskamma catchment to land use/cover changes

Nelson Mandela University (2019)

An assessment of the hydrologic response of the Keiskamma catchment to land use/cover changes

Sive, Mlamla

Titre : An assessment of the hydrologic response of the Keiskamma catchment to land use/cover changes

Auteur : Sive, Mlamla

Université de soutenance : Nelson Mandela University

Grade : Master of Science (Geography) 2019

Résumé partiel
The Keiskamma catchment has undergone significant land use/cover changes (LUCC) underpinned by land use policy reforms and climate change. However, the hydrological responses of the catchment to LUCC are not fully understood. This study sought to assess the hydrological response of the Keiskamma catchment to LUCC at catchment and hillslope scale using remote sensing, GIS, hydrological modelling and field experiments. Catchment scale assessments first involved LUCC mapping in IDRISI TerrSet software, using supervised image classification for two sets of multispectral imagery ; namely Landsat Thematic Mapper (TM) of 1994 and Landsat 08 Operational Land Imager (OLI) of 2016. The LUCC maps provided an indication of LUCC over time and were prerequisite land use inputs for modelling the hydrologic response of the catchment. The Soil and Water Assessment Tool (SWAT) hydrologic model was used to model the hydrologic response of the catchment to LUCC. The Sequential Uncertainty Fitting (SUFI-2) in SWAT-CUP was used to assess model performance and uncertainty analysis. The influence of rainfall on the hydrologic response of the catchment was also assessed using linear regression. One of the prominent forms of LUCC in the Keiskamma catchment, particularly central Keiskamma is P. incana shrub encroachment. Field experiments were set up to investigate the hydrologic impacts of P. incana shrub invasion at hillslope scale, as well as to validate the results obtained by the SWAT hydrologic model. Field experiments included an assessment of the Landscape Organisation Index (LOI) of the invasion, as well as assessing of surface conditions, surface runoff (L), volumetric soil water content (cm³/cm³) and sediment loss (grams) under P. incana, grass and bare-eroded areas. High image classification accuracy assessment values of 87.2 % and 87.4 % for 1994 and 2016 respectively were obtained, with a Kappa coefficient of 0.84 for both sets of imagery. Results of the study revealed a significant increase in woody vegetation encroachment, specifically shrub invasion, forest expansion in the upper parts of the catchment, as well as an increase in exotic and invasive vegetation species within the riparian zone. The SWAT model showed a good (NSE=0.69, R²=0.69 and RSR =0.56) and unsatisfactory (NSE=0.4, R²=0.4 and RSR 0.79) model performance for calibration and validation respectively.

Présentation (SEALS)

Version intégrale

Page publiée le 18 janvier 2021