I address three aspects of conservation planning and environmental management, introducing new information and quantitative methods for better conservation decision-making. In chapter 1, I model the ecosystem service of recreation in Puget Sound, WA, to understand the spatial heterogeneity in recreational usage, and determine how water quality variation affects visitation. Visitation model results illustrate how visitation relates to various amenities and costs. In a fixed effects analysis, visitation was negatively related to water quality while controlling for seasonal trends and time-variant factors. This indicates people are responding to changes in water quality, and an improvement would yield an increase in the value of recreation. The results are useful for the prioritization of water quality improvement and other natural resource management projects.

In chapter 2 I compare optimal to heuristic selection algorithms in a utility maximization problem. I also quantify the efficiency of the algorithm when site availability is a stochastic process reflective of land use change. I show the optimal IP formulation outperforms the greedy heuristic by 8%. When site availability is a stochastic process with fewer available parcels, the IP solution is again higher than the greedy, and the accumulated utility is 70% - 88% of the full IP solution. This study is unique in its use of integer programming in a multicriteria utility maximization approach, and has many potential applications in conservation planning and natural resource management.

In chapter 3, I model conservation easement costs to determine how easement value affects conservation priorities. Cost information improves conservation efficiency and effectiveness, yet is not often considered. Similarly, conservation easements are frequently used by conservation organizations, but rarely in large scale planning. I model easements using spatial hedonic models with observed transactions and agricultural assessment data. I then prioritize conservation actions with each data set, and find differences in priorities. Since development rights and agricultural value can be quite different at a given site, easements are not a simple fraction of total value, justifying the extra modeling steps. If the full cost is used in planning instead of easement value, priorities may be incorrect or inefficient at meeting conservation targets.