Spatial-temporal mapping of Parthenium (P. HysterophoruL) in the Mtubatuba municipality, KwaZulu-Natal, South Africa.
MetadataShow full item record
Detecting and mapping the occurrence, spread, and abundance of Alien Invasive Plants (AIPs) have recently gained substantial attention, globally. Therefore, the present study aims to assess remote sensing application for mapping the spatial and temporal spread of Parthenium (P. HysterophoruL) in the Mtubatuba municipality of KwaZulu-Natal, South Africa. Parthenium is an aggressive herbaceous plant from the South and Central America that has colonized many regions of the world including Asia, Australia, and Africa. The adverse social, economic and ecological impacts of the plant have emphasized the need for a robust control programme to combat its spread. However, data for the management of the weed has been gathered by means of manual methods such as field surveys which are time and labour intensive. Alternatively, remote sensing techniques provides cost effective approach to large-scale mapping of AIPs. The first objective of the study provides an overview of advancements in satellite remote sensing for mapping AIPs spread and the associated challenges and opportunities. Satellite remote sensing techniques have been successful in detecting and mapping of AIPs, exploring their spatial and temporal distribution in rangeland ecosystems. Although they provide fine spatial information, the excessive image acquisition costs associated with the use of high spatial and hyperspectral datasets are a limitation to continuous and large-scale mapping of AIPs. The signing of the license agreement between the South African Space Agency (SANSA) and Airbus Defense and Space (ADS) has ensured a continued provision of SPOT data with improved spatial properties for South Africa. Similarly, the signing of the single licence government multi-user agreement between the South African government and SANSA has ensured free provision of SPOT data for public use in South Africa to support land change monitoring. The second objective was to determine the spatial and temporal distribution of Parthenium from 2006 to 2016 using SPOT series data in concert with Random Forest and Land Change Modeler (LCM). Findings have shown a steady decrease in Parthenium distribution over the 10-year period of the study because of the low annual rainfall experienced in the area over the recent past. Furthermore, disturbances in the soil opens vacant spaces which are susceptible to Parthenium invasion. This study has demonstrated the value of readily available multispectral SPOT series data in concert with robust and advanced non-parametric Random Forest algorithm in detecting trends and patterns in the spatial and temporal spread of AIPs.