Vegetation communities modelling using GIS-Integrated statistical, ecological and data models

Accad, Arnon (2004). Vegetation communities modelling using GIS-Integrated statistical, ecological and data models PhD Thesis, School of Geography, Planning and Architecture, The University of Queensland.

       
Attached Files (Some files may be inaccessible until you login with your UQ eSpace credentials)
Name Description MIMEType Size Downloads
THE17703.pdf Full text application/pdf 38.45MB 4
Author Accad, Arnon
Thesis Title Vegetation communities modelling using GIS-Integrated statistical, ecological and data models
School, Centre or Institute School of Geography, Planning and Architecture
Institution The University of Queensland
Publication date 2004
Thesis type PhD Thesis
Supervisor Dr David Neil
Dr Steve Turton
Total pages 450
Language eng
Subjects 0406 Physical Geography and Environmental Geoscience
Formatted abstract

This research was conducted in the Wet Tropics bioregion of northeastern Queensland, Australia, an area of 19,928 km2 of which 8,944 km2 were listed as a World Heritage Area (WHA) in 1988, fulfilling all four criteria required for listing. Large parts of the bioregion are inaccessible and the closed canopy in most of the area obscures the understorey vegetation, the geology, the soil, and the ground surface topography. These conditions occur in most large areas of tropical forests and prevent accurate and high precision mapping and classification of vegetation, geology, soils, elevation and other parameters using traditional methods. The aim of the research is to produce a less expensive and more accurate and realistic representation of the vegetation distribution than can be readily achieved by traditional mapping methods.

 

Traditional mapping methods (as applied in the study area by, for example, Tracey and Webb (1975) and followed by Stanton and Stanton (2001)) use high cost, labor-intensive aerial photography interpretation which can be subjective and is limited by:

 

•      the extent of remnant vegetation on the earliest available aerial photography, unless old survey records are used (Neldner et al., 1999);

 

•      differing scale and quality of aerial photography;

 

•      cloud cover and shadows; and

 

•      similar canopy features with significant floristic differences may not be delineated.


An alternative approach is proposed which integrates a data model, a statistical model and an ecological model using sophisticated Geographic Information Systems (GIS) techniques and rule-based systems to support large-scale vegetation communities' modelling. This approach is based on more realistic representation of vegetation patterns with transitional gradients from one vegetation community to another. If necessary, arbitrary, though often unrealistic, sharp boundaries can be imposed on the model by an application of statistical methods.

 

This alternative approach relies on high quality Digital Elevation Models (DEMs), and climate and soil nutrient status surfaces. The development of a fine scale DEM (80 m) for the entire Wet Tropics Region (Accad, 1996) followed by the refinement of  the climate surfaces for the Wet Tropics (Turton, Hutchinson, Accad, Hancock and Webb, 1999) provided an adequate data source to model vegetation communities of the Wet Tropics at both the regional and local scales. With the increased accuracy of DEMs over most parts of the globe and the wider development of climate surfaces using ANUCLIM (McMahon et ah, 1995) there is increasing potential for application of this vegetation modelling approach in other geographical areas.

 

Such models were developed for the Wet Tropics region and successfully applied to modelling:

 

•               the vegetation communities of the Bartle Frere 1: 50 000 scale map sheet using Principle Component Analysis (PCA) and clustering methods, which illustrated the high correlation between the massif rainforest communities and the environmental parameters;

•               the pre-clearing vegetation distribution of the Innisfail Lowland subregion using the Classification and Regression Trees (CART) modelling method (The pre-clearing distribution of all except one of the present remnant vegetation community types (Stanton and Stanton, 2002) were successfully modeled); and

 

•      the potential effect of climate change on the rainforest vegetation communities of the Innisfail Lowlands subregion, using Generalised Linear Modelling (GLM) and Generalised Additive Modelling (GAM) methods. The modelling indicates that, under the likely climate change scenarios of Walsh et al. (2001) the distributions of rainforest vegetation communities of the Innisfail Lowlands will be significantly affected.

 

This research demonstrates the capacity of vegetation modelling to extend vegetation distributions into cleared areas for which there is little data, and to further assess the distribution of rainforest vegetation communities under past, present and future climate regimes.

 

This modelling approach has good potential for wider application, including provision of:

 

•           vital information for vegetation and conservation management;

 

•           a scientific basis for rehabilitation and reafforestation of disturbed and cleared areas;

 

•            advice for agro-forestry in selection of commercial species; and
 

•           a viable method for the production of adequate vegetation maps for conservation and forestry planning of poorly-studied areas, e.g. Irian Jaya and Papua New Guinea.

       

Vegetation mapping is a fundamental prerequisite for ecosystem conservation, planning and management in most terrestrial ecosystems. Statistical modelling is rapidly becoming a key component in vegetation mapping and modelling, offering an explicit, consistent, repeatable and affordable method for extending vegetation mapping into cleared areas (Franklin, 1995; Austin et al., 2000). This research successfully develops and applies an integrated suite of such technologies in a geomorphically and ecologically complex environment.
Keyword Geographic information systems
Plant ecology -- Mathematical models

 
Citation counts: Google Scholar Search Google Scholar
Access Statistics: 82 Abstract Views, 4 File Downloads  -  Detailed Statistics
Created: Tue, 06 Dec 2011, 16:28:43 EST by Ms Christine Heslehurst on behalf of Scholarly Publishing and Digitisation Service