School of Agriculture, Food and Ecosystem Sciences - Theses

Permanent URI for this collection

http://hdl.handle.net/11343/427

Filters

2000 2
Reset filters

Settings
Statistics
Citations
Type: PhD thesis × Author: Varma, Vivek Krishna ×

Search Results

Now showing 1 - 2 of 2
  • Item
    No Preview Available
    A decision support system for sustainable forest management
    Varma, Vivek Krishna (University of Melbourne, 2000)
    The main objective of this study is to develop a decision support system for sustainable forest management that employs criteria and indicators; preferably one that enables the process to be adaptive over successive cycles of monitoring indicators, thereby capitalizing to the fullest on the criteria and indicators approach. To meet this objective, this thesis develops (1) a methodology for monitoring indicators over time, and (2) a procedure appropriate for achieving a sustainable forest resource allocation, based on a review of literature. The resulting decision support system blends the decision maker's knowledge with the information processing capabilities of the computing tools employed and is applicable on the regional scale. The application of potentially useful criteria and indicators framework faces considerable difficulties due to: the issues of integration of spatial and non-spatial data coming from a variety of sources; the choice of methodologies to analyse the large quantities of data; the incorporation of joint production characteristics of forest values; the lack of reliable and cost-effective data, and; the uncertainty associated with data, sustainability thresholds, and decision-rules. The proposed monitoring methodology addresses these issues. It comprises uncertainty-based multi-criteria evaluation of the indicators at a forest management unit level using a geographic information system. It also focuses on determining two alternative indices of sustainability. These indices are calculated by using Zadeh's fuzzy set theory and Dempster-Shafer theory of evidence to model the uncertainty in sustainability thresholds and decision-rule respectively. They proved to be similar in nature. This monitoring methodology was applied to Landsat 5 TM data for the East Gippsland Region, Australia, with the objective of examining the design of the land-use indicators over successive monitoring cycles. While confirming the usefulness of the satellite imagery as a source of data on large spatial scales, this study highlighted the importance of classification errors in reliably monitoring indicators over short periods. A case study on Aravalli Project, India, was developed to demonstrate utility of the monitoring methodology by applying a range of indicators, including some specially framed area-specific indicators, to participatory forest management. The results of both studies emphasised the need of reliable `ground-truth' data for implementing criteria and indicator framework. The approach developed for sustainable forest resource allocation and land-use planning potentially provides a basis for ensuring a `non-declining total utility' accruing from the diverse forest values to the current and future generations. It satisfactorily addresses, in a utilitarian framework, several important issues, such as incorporating social preferences while satisfying intra- and intergenerational equity concerns; efficiency; complexity; uncertainty, and ecological irreversibility. Stochastic and fuzzy/ possibilistic approaches were used to deal with uncertainty on the short and long planning horizons respectively in a linear programming model that employed the concept of aspiration levels. A heuristic procedure was developed to distribute land units spatially to various forest values. Implementation of this approach required the integration of a geographic information system, mathematical modelling software, simulation models, graphic user interfaces, specially written computer programmes, and database management system. This decision support system, based on an ethically-oriented approach, could facilitate a more informed, and probably more socially and politically acceptable, public decision-making for achieving sustainability of forest resources.
  • Item
    No Preview Available
    A decision support system for sustainable forest management
    Varma, Vivek Krishna (University of Melbourne, 2000)
    The main objective of this study is to develop a decision support system for sustainable forest management that employs criteria and indicators; preferably one that enables the process to be adaptive over successive cycles of monitoring indicators, thereby capitalizing to the fullest on the criteria and indicators approach. To meet this objective, this thesis develops (1) a methodology for monitoring indicators over time, and (2) a procedure appropriate for achieving a sustainable forest resource allocation, based on a review of literature. The resulting decision support system blends the decision maker's knowledge with the information processing capabilities of the computing tools employed and is applicable on the regional scale. The application of potentially useful criteria and indicators framework faces considerable difficulties due to: the issues of integration of spatial and non-spatial data coming from a variety of sources; the choice of methodologies to analyse the large quantities of data; the incorporation of joint production characteristics of forest values; the lack of reliable and cost-effective data, and; the uncertainty associated with data, sustainability thresholds, and decision-rules. The proposed monitoring methodology addresses these issues. It comprises uncertainty-based multi-criteria evaluation of the indicators at a forest management unit level using a geographic information system. It also focuses on determining two alternative indices of sustainability. These indices are calculated by using Zadeh's fuzzy set theory and Dempster-Shafer theory of evidence to model the uncertainty in sustainability thresholds and decision-rule respectively. They proved to be similar in nature. This monitoring methodology was applied to Landsat 5 TM data for the East Gippsland Region, Australia, with the objective of examining the design of the land-use indicators over successive monitoring cycles. While confirming the usefulness of the satellite imagery as a source of data on large spatial scales, this study highlighted the importance of classification errors in reliably monitoring indicators over short periods. A case study on Aravalli Project, India, was developed to demonstrate utility of the monitoring methodology by applying a range of indicators, including some specially framed area-specific indicators, to participatory forest management. The results of both studies emphasised the need of reliable `ground-truth' data for implementing criteria and indicator framework. The approach developed for sustainable forest resource allocation and land-use planning potentially provides a basis for ensuring a `non-declining total utility' accruing from the diverse forest values to the current and future generations. It satisfactorily addresses, in a utilitarian framework, several important issues, such as incorporating social preferences while satisfying intra- and intergenerational equity concerns; efficiency; complexity; uncertainty, and ecological irreversibility. Stochastic and fuzzy/ possibilistic approaches were used to deal with uncertainty on the short and long planning horizons respectively in a linear programming model that employed the concept of aspiration levels. A heuristic procedure was developed to distribute land units spatially to various forest values. Implementation of this approach required the integration of a geographic information system, mathematical modelling software, simulation models, graphic user interfaces, specially written computer programmes, and database management system. This decision support system, based on an ethically-oriented approach, could facilitate a more informed, and probably more socially and politically acceptable, public decision-making for achieving sustainability of forest resources.