Natural Resources, School of

 

Date of this Version

3-2013

Citation

Schapaugh, A.W. 2013. New Tools for Quantitative Decision Analysis in Applied Ecology and Conservation. Ph.D. Dissertation. The University of Nebraska.

Comments

A Dissertation Presented to the Faculty of the Graduate College at the University of Nebraska In Partial Fulfillment of Requirements For the Degree of Doctor of Philosophy, Major: Natural Resource Sciences (Applied Ecology), Under the Supervision of Professor Richard A.J. Tyre. Lincoln, Nebraska: March, 2013

Copyright (c) 2013 Adam W. Schapaugh

Abstract

Scientists have generated a massive body of theory aimed at predicting and managing the impacts of anthropogenic activities on populations, species, and ecosystems. Transforming this research into knowledge that informs complex decision-making problems remains a major challenge in environmental management and conservation. My dissertation research aims to address this issue through the development and application of mathematical and statistical models. I integrate tools, concepts, and techniques from ecology, applied mathematics, computer science, and statistics to build structured decision-making frameworks for spatial prioritization, resource allocation, and optimal scheduling. I also tackle several of the technical challenges limiting the utility of such tools in practice, and seek to make them accessible to other scientists and decision-makers. Much of my research is motivated by the interest in land acquisition as an in situ conservation strategy. In Chapters 1 and 2, I develop an integrated reserve selection framework for spatial priority-setting and optimal investing. The framework combines Bayesian methods and Markov decision theory in the context of making land acquisition decisions. A second focus of my research focuses on overcoming several of the technical and computational challenges of utilizing Markov decision processes (MDPs) in the context of real-world planning. In Chapter 3, I introduce and test and class of approximation algorithms developed in the artificial intelligence community to simply and solve MDPs with large state spaces. In Chapter 4, I develop a novel method that uses information-gap (radius of stability-type) models to represent uncertainty in the state transition function of an MDP. Rather than requiring information about the extent of parametric uncertainty at the outset, this method addresses the question of how much uncertainty is permissible before the optimal policy would change. Finally, in Chapter 5, I develop a pair of sensitivity metrics for info-gap decision analysis. Both sensitivity metrics are an essential addition to the robust optimization toolkit, providing a systematic approach for identifying weaknesses in an info-gap decision analysis. They are also needed quantities in the effort to make sound, defensible decisions.

Advisor: Richard A. Tyre