School of Earth & Environmental Sciences
Barlow, P B., A comparative study of raster and vector based approaches in vegetation mapping on Five Islands off the coast of Port Kembla., BEnviSci Hons, School of Earth & Environmental Sciences, University of Wollongong, 2018.
Remote sensing has facilitated extraordinary advances in the modelling, mapping and understanding of vegetation in remote Island ecosystems. With the unforgiving responses of Island ecosystems to anthropogenic influences, it is paramount that managerial strategies are put into place and vegetation conditions are quantified. The Five Islands group has an extensive history of anthropogenic alteration, resulting in widespread change in vegetation dynamics. In recent years introduced vegetation has overpopulated Big Island – The largest of the Island Nature Reserve – and made it inhabitable for protected burrowing and nesting seabird populations. As a result NSW National Parks and Wildlife Service (NPWS) are employing a 5-year rehabilitation plan to restore the Islands to their vegetation pre-human interference, in hope of the return of native seabird populations. This project in particular will utilize raster and vector based spatial mapping methodologies to resolve the following management questions: 1) How many Lomandra longifolia seedlings have survived, what is the health of the remaining plants and is there a correlation to distance from invasive Carpobrotus sp. 2) What are ideal Lomandra sp. densitys for areas that still need planting; 3) What vegetation is present on all Five Islands; 4) How effective was the weed treatment And 5) What is the spread of the invasive Carpobrotus sp between 2017 and 2018.
These management queries were answered through the acquisition of aerial imagery with a 4-band UAV drone camera followed by spatial analysis through ArcGIS technologies. Vector based spatial analysis included point digitization of Lomandra sp. as well as polygon digitization of exotic Carpobrotus sp. It was found that Lomandra sp. survival was low (19.74%). The digitization of exotic Carpobrotus sp. in 2017 and 2018 resulted in a calculated growth of 1.89m2/day and a correlation (r=0.47) was found between Lomandra sp. health and distance from Carpobrotus sp. The raster based supervised and unsupervised classifications of vegetation classes on the five Islands led to the distinguishing of vegetation into 5 vegetation classes determined by the dominating species. The weed treatment on Big Island was deemed highly effective with a loss of up to 88.89% of exotic species. Both raster and vector based methodologies were found to offer unique methods to spatially analyse vegetation ground coverage and can ultimately be used to solve broader managerial issues associated with delicate Island ecosystems.
Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily represent the views of the University of Wollongong.