- Advancing the understanding of natural resource management is an important step in mitigating the effects of human activity on the environment, and ensuring efficient outcomes for many sectors of the economy. As humanity’s role in the natural world becomes better understood, the importance of interdisciplinary modeling has grown in leaps and bounds. This is evidenced by the rise of fields such as bioeconomics, the economics of climate change, and the increasing influence of “societal dimensions” departments in universities around the country. It is becoming evident that a holistic understanding of feedbacks between the natural and economic realms is crucial for developing the research agenda of tomorrow. In addition, advances in computing resources have made research questions previously restricted by their computational complexity viable for analysis. Both of these developments bode well for interdisciplinary modeling; however, much of these developments remain unrealized in the literature. For instance, the continued utilization of large scale earth system models such as the Community Earth System Model (CESM) for impact studies (e.g. Law et al., 2018) has highlighted the importance of representing the social systems accurately within the model. Despite this, the use of natural resource models that are consistent with economic theory are nowhere to be found amongst the many modules of CESM, or other similar models. Instead, economic models are used to inform the input datasets of these models, which is rigorous but unsatisfying once one realizes that this approach completely fails to capture the feedback between the natural and social systems that intuition tells us is there. The lack of such modeling also precludes running sophisticated policy experiments within CESM and her sister models. These policy experiments, with their robust representations of physical processes, can be better positioned to examine the effect of these policies on a variety of outcomes, both environmental and economic than what currently exists. This is in addition to the fact that there are still many aspects of policy design that are unexplored in natural resource management. The details about the design of environmental policies, especially those targeting the private provision of ecosystem benefits, must be fine tuned to achieve an optimal outcome. One particular aspect of policy design that is understudied in the literature is that of the duration of contracts for ecosystem service programs. Many policies currently in practice base the duration of the contract on environmental goals of the policy. However, economic incentives could change the impacts of the policy should the duration be changed. The efficient design of policies depends on the feedbacks between social and natural systems. Though models such as CESM can address uncertainties about future effects of climate change and disturbance, it is a deterministic model of natural resources. In reality, natural resources effectively behave in a stochastic manner. This results in management strategies that require substantial investments in monitoring and learning, as good information is crucial for optimal management. This has led to many studies examining adaptive management of natural resources, and learning in systems such as fisheries (Kling et al. 2017), livestock management (MacLachlan et al., 2017), and regulatory enforcement (White, 2005). There is a substantial gap in what the literature addresses. Previous studies ignore the role of price stochasticity, as well as stochasticity in other observable variables, in determining the optimal learning strategy of natural resource owners. This is a more generalized description of natural resource management that has implications far outside of private natural resource management. This dissertation advances the the design and application of modeling techniques in natural resource management, as well as theory behind these models. In what follows, we analyze the feedback between natural and social systems in forestry. We show that the forest sector adapts to disturbance events such as wildfire or pine beetle outbreaks through shifting harvests to different areas. This model has the potential to improve the representation of social systems within large scale earth system models, and to allow for economic policy experiments on a larger scale than what has been previously observed in the literature. We explore the economics of contract duration within a forest-based carbon offset program, which is the first time such a question has been addressed through modeling. It also contributes to current discussions of implementing forest-based carbon offsets in Oregon’s carbon abatement plan. This dissertation achieves an advancement of the economics of information in partially observable resource systems by solving a model of forest management where the volume of timber is observed imperfectly, and observations are costly and noisy. In Chapter 1, I introduce the common themes of the dissertation, and provide an overview of what is to follow. The natural resource system this work addresses is primarily forestry. In particular, it focuses on the issues surrounding ecosystem service provision and management within private forestry. In Chapter 2, I construct a partial equilibrium (PE) model of the forest sector in the western United States. The model is spatially explicit, and overcomes issues involving its solve time by utilizing a novel algorithm that simulates an auction between agents in the model. Furthermore, the model can be coupled to CESM in order to obtain a more realistic representation of biological processes and climate change relative to what is available to forest sector models currently. The realism of the model is aided by the incorporation of numerous datasets such as land ownership and transportation costs. The model is unique in its scale, and is solvable over a larger range and with a higher resolution than other forest sector models. It also has a realistic depiction of the ecology of forestry through its ability to couple to CESM. This model is particularly useful for modeling the feedback between the natural system of the forest and economic system of the forest sector. Specifically, it’s beneficial for understanding the impact of forest disturbances on the economy, and how that shapes future disturbance patterns. The results suggest that in the short run, the spatial distribution of harvests changes substantially, with the difference in overall harvests growing over time due to the effects the disturbances have on mill capacity and profitability. We also utilize our model for understanding the impacts of policies specifically addressing disturbance vulnerability, as well as the impacts of state-level policies and how those may affect the surrounding region. In Chapter 3, I utilize a regional forest sector model of western Oregon in order to analyze the effects of changing the duration of forest-based carbon offset contracts. The model is a spatially explicit model that tracks both sawtimber and pulp production, as well as price levels and mill capacities. It keeps track of the amount of timber being exported as well, and average management decisions such as rotation lengths. The model is applied to scenarios that vary in the duration of the contract as well as the price of the carbon, which is fixed during the model run. Whereas previous studies have examined the effects of these contracts on the Oregon forest sector (Latta et al., 2011), no study has yet addressed the role of contract duration on enrollment and program performance. We find that market forces stabilize the amount of carbon being removed from the landscape every time step. This analysis is useful in serving as a critique of current approaching to contracting for forest-based carbon offset programs such as the one in California by showing that alternative contract lengths are capable of higher levels of sequestration over given time periods. In Chapter 4, I construct a model of forest management under state uncertainty that optimizes both the timing of harvest as well as measurement of the forest resource, known as “inventory”. Forest resources, along with practically every other natural resource, exhibit state uncertainty – uncertainty about the present state of the resource. Oftentimes natural resources are only observed when investments are made in measurement of the resource. Furthermore, a perfect measurement of the resource is oftentimes infeasible, either for reasons having to do with the biology of the resource or because it is cost prohibitive. In this chapter I solve the forest manager’s problem under state uncertainty as a continuous-state Mixed Observability Markov Decision Process (MOMDP). I find that the optimal timing of learning is influenced not just by price level, but surprisingly by price stochasticity as well. Chapter 4’s innovation is that it presents the first continuous state model of natural resource management under state uncertainty that includes price stochasticity. For a majority of natural resource management problems, price stochasticity plays an important role, and the results from this project allow us to understand how it influences not just harvest timing, but the optimal investments in measurement and learning. We find that learning is valuable. Using an empirical model of forest growth that captures its natural stochasticity, we are able to calculate the costs associated with state uncertainty when inventory is not an option. We find that conducting costly yet accurate inventories in an optimal way greatly reduces the burden of state uncertainty, and increases the value of the stand through improved management. This chapter also presents the first model of forest inventory that is grounded in microeconomic theory. The expansion of interdisciplinary research as well as the availability of new computational techniques in the field of economics have resulted in opportunities for researchers looking to address difficult problems in natural resource economics. My dissertation is a combination of methodological advances, as well as inquiries into potential policy applications. I hope that what follows from here will aid both future researchers interested in similar topics, as well as policymakers with questions about the design of schemes targeting private forest landowners. The extensions and limitations of all of these studies will be discussed as they are presented. Because of the methodological nature of much of this dissertation’s content, the possibility exists to greatly expand on what has been done here in future studies.