- Title
- Remote sensing as a monitoring solution for water hyacinth (Pontederia crassipes) in the context of the biological control programme at Hartbeespoort Dam
- Creator
- Kinsler, David Louis
- Subject
- Uncatalogued
- Date Issued
- 2023-10-13
- Date
- 2023-10-13
- Type
- Academic theses
- Type
- Master's theses
- Type
- text
- Identifier
- http://hdl.handle.net/10962/424599
- Identifier
- vital:72167
- Description
- Water hyacinth (Pontederia crassipes (C.Mart.) Solms (Pontederiaceae)) is a significant aquatic weed both globally and in South Africa. Despite notable success with biological control of other invasive macrophytes, the plant remains as a problematic weed in many aquatic systems in South Africa, particularly due to the eutrophic status of many of its water systems, as well as the plant’s tolerance to cooler climatic conditions than most of its existing biological control agents. Hartbeespoort Dam, located about 30 kilometres west of Pretoria, South Africa, has been infamously infested with water hyacinth for decades, which impacts the important socioeconomic utility of the dam and functioning of natural ecological processes in the system. The dam has a long history of efforts to control water hyacinth, which include widespread herbicidal spray, mechanical removal and classical biological control programmes since the early 1990s - mostly with limited or short-lived success. However, after the introduction of a new, cold-tolerant biological control agent, Megamelus scutellaris Berg (Hemiptera: Delphacidae) in 2018 with an inundative release strategy, the water hyacinth dropped significantly from a maximum cover of about 45 percent (819 hectares) down to less than two percent (40 hectares) over a three-month period (November 2019 – January 2020). This was significant, as it marked the first successful biological control of water hyacinth in a eutrophic, temperate system in South Africa. However, due to the scale of Hartbeespoort Dam (1820 hectares) and the high spatiotemporal variation of the floating mats across time and space, quantifying and monitoring these rapid changes has proved difficult. In response to this problem, this thesis proposed a remote sensing solution to address the need for accurate, timely and readily accessible monitoring data of the water hyacinth population on the dam. Leveraging the temporally frequent (< 5 days revisit time) Sentinel-2 multispectral satellite data, as well as the powerful cloud-computing resources of Google Earth Engine, this thesis developed and deployed a relatively simple and robust index-based decision tree classification method to demonstrate the value of these technologies as an effective monitoring and analysis tool for monitoring large macrophyte infestations. To this end, several challenges had to be overcome in order to produce easily accessible data that was accurate and reliable. For example, due to the size of the Sentinel-2 Level-1C image dataset from August 2015 to March 2021 (n = 654), an automated process of filtering out clouded images was required. Additionally, the co-presence of algal and cyanobacterial blooms necessitated the development of a novel index, coined the Algae Resistant Macrophyte Index (ARMI), to deal with the challenges of accurate macrophyte detection. The high spatiotemporal variability of the floating mats meant that a typical, location-based confusion matrix as a means of assessing the accuracy of the decision tree classifier required a different approach which compared the total classified areas with higher resolution images. This thesis aims to demonstrate the utility of remote sensing tools to provide effective monitoring information to managers, researchers and other stakeholders. There is scope to expand to more areas in South Africa and beyond and may prove an invaluable tool to augment and support on-going and future macrophyte monitoring programmes.
- Description
- Thesis (MSc) -- Faculty of Science, Geography, 2023
- Format
- computer
- Format
- online resource
- Format
- application/pdf
- Format
- 1 online resource (74 pages)
- Format
- Publisher
- Rhodes University
- Publisher
- Faculty of Science, Geography
- Language
- English
- Rights
- Kinsler, David Louis
- Rights
- Use of this resource is governed by the terms and conditions of the Creative Commons "Attribution-NonCommercial-ShareAlike" License (http://creativecommons.org/licenses/by-nc-sa/2.0/)
- Hits: 1
- Visitors: 3
- Downloads: 1
Collections
RU Department of Geography
| Thumbnail | File | Description | Size | Format | |||
|---|---|---|---|---|---|---|---|
| View Details Download | SOURCE1 | KINSLER-MSC-TR23-151.pdf | 2 MB | Adobe Acrobat PDF | View Details Download |