The suitability of remote sensing for prioritising management of invasive plants in the Garden Route South Africa

Invasive tree species pose a huge problem in the Garden Route and are particularly damaging to aquatic ecosystems, including wetlands, riparian zones, lakes and estuaries. Therefore, this study aimed to determine priority areas for invasive tree species management, with a focus on aquatic ecosystems. This was achieved by using existing literature to identify priority species, based on their impact on aquatic ecosystems and their associated ecosystem services, and then testing the suitability of SPOT-6 and WorldView-3 multispectral data at detecting these focal species. The priority species identified were: Acacia cyclops (rooikrans), Acacia longifolia (long-leaved wattle), Acacia mearnsii (black wattle), Acacia melanoxylon (blackwood), Acacia saligna (Port Jackson willow), Eucalyptus camaldulensis (red river gum), Pinus pinaster (cluster pine) and Pinus radiata (radiata pine). The Random Forest classifier on SPOT-6 data achieved an overall accuracy of 62.5% and this method was consequently deemed ineffective at separating invasive tree species from other tree species in the Garden Route. The overall accuracy of the WorldView-3 classifier was higher (78.9%) but the cost of the data limited the use of more images for the detection of the focal species throughout the Garden Route. Therefore, to identify priority areas for invasive tree management, criteria derived from existing literature were input into spatial conservation planning software. The analysis identified the: Saasveld section of the Garden Route National Park, the Wilderness Lakes, Knysna Forest, Knysna Estuary, Tsitsikamma Forest around Stormsriver and a disturbed area of fynbos southeast of Kareedouw as management priorities. Currently spatial conservation planning software proved to be cost-affordable and useful tool and is recommended for invasive tree management in the Garden Route.

Identiferoai:union.ndltd.org:netd.ac.za/oai:union.ndltd.org:nmmu/vital:27538
Date January 2017
CreatorsGerolemou, Rosie Victoria, Roux, Dirk
PublisherNelson Mandela Metropolitan University, Faculty of Science
Source SetsSouth African National ETD Portal
LanguageEnglish
Detected LanguageEnglish
TypeThesis, Masters, MSc
Formatxvi, 112 leaves, pdf
RightsNelson Mandela Metropolitan University

Page generated in 0.0487 seconds