Date of Award

2003

Degree Type

Thesis

Degree Name

Bachelor of Science Honours

School

School of Natural Sciences

Faculty

Faculty of Computing, Health and Science

First Advisor

Dr Eddie van Etten

Second Advisor

Dr Andre Hinwood

Abstract

A reliable technique to assess impacts to coastal zones is required as coastal areas across the world are under pressure from an increasing human presence. The result is a greater level of disturbance to the coast, both marine and terrestrial. This study addresses a worldwide problem but is focused to a local scale. Wedge, on the central coast of Western Australia, supports a squatter shack settlement and has done so for approximately 50 years. Over time, the number of shacks, campers, tourists, 4WD's and motorbikes has increased. In recent years this growth has been significant. Management of this area and all coastal zones require adequate methods to assess disturbance, impacts and potential problems. Species cover and environmental characteristics of a coastal heath community were studied using data from 96 quadrats collected on the dunes of the Wedge promontory. Methods for multivariate data analysis are many and varied, but few studies have used gradient analysis to assess variation from disturbance to the dune communities. This study is therefore innovative as it used both direct and indirect gradient analyses to separate natural (control from disturbed (impact) sites. Gradient analysis determines relationships between vegetation and environmental data and is therefore a more rigorous approach than traditional descriptive studies. Gradient analysis techniques of non-metric multidimensional scaling (NMDS) and Canonical Correspondence Analysis (CCA) of vegetation data separated control and impact sites according to their percentage cover and nineteen environmental variables. Although control and impacts sites were not statistically different due to natural variation, initial analysis of all sites identified the successional nature of coastal dune vegetation and resulted in a division into two major landform groups, foredune-primary dune, and swales-flats-secondary/tertiary dunes. These two groups were then analysed individually. Gradient analysis provided a clear definition of sites in disturbed areas at the local scale for this coastal site. Indirect gradient analysis gave good results for the data-set, whilst direct gradient analysis supported this and indicated a range of possible relationships to impact areas. With the simple variables used, gradient analysis provided good quality inferential results and highlighted specific vegetation species that may be used as indicators of disturbance. This technique could be applied for both assessment of variation on a local coastal scale due to disturbance (ie impact assessment) and also by using indicator species for conservation management. Gradient analysis identified and separated natural variation from human induced variation in vegetation patterns. This analysis showed that disturbed areas have different species composition and diversity, that two native species Acacia cyclops and Myoporum insulare are strongly favoured by disturbance, that the level of native species decreased and weed species numbers increased, compared to undisturbed areas. Gradient analysis proved to be an effective method to differentiate between control and impact sites and therefore would be extremely useful instrument for assessment of impact for coastal zones that are under pressure.

Share

 
COinS