Estimation and adaptive management of dynamic disease and non-native species risk

The complexities of disease and non-native species introductions have led economic modelers to employ simplifying assumptions that smooth over very real difficulties experienced by managers of these ecosystems. Management under uncertainty is almost always the norm in these systems due to the partially observability of most pests and pathogens. In this thesis, we explore extensions to existing methodologies that allow new questions regarding this uncertainty to be addressed. Each chapter explores a different context and approach that suits the available data and biophysical realities.;We begin by examining the management of an endemic infectious disease when prevalence is unknown and may only be partially observed through testing. The intricacies associated with accounting for uncertainty in a dynamical system are addressed using an extension of the Partially Observable Markov Decision Process (POMDP) framework to a continuous state-space (prevalence). The results highlight the value of a manager's responsiveness to test results when selecting future expenditures on broad-based controls (disease testing), and suggest that ignoring the available information leads to excessive expenditures when suppressing prevalence. We also find the surprising result that an adaptive management approach that simply ignores transmission (which is a danger when transmission is slow) chooses to cease testing if they believe prevalence is sufficiently low, which precludes future adjustments. As a result, prevalence takes off and losses abound.;The introduction of non-native species and pathogens via trade of numerous agricultural products represents a significant externality. We relate the regionally disaggregated import values of plant materials with the discoveries of a particular order of non-native insect species (Hemiptera) using a time series of introduction spanning from the 998-2012 and a new time-series of import values spanning from 1854-2012. This relationship is separately estimated using a world-aggregated and a regional model of introductions, both integrated with a model of delayed discoveries. The results suggest significant regional differences in the current risk of imports, and provide estimates regarding the magnitude of these differences. The results also demonstrate the importance of considering changes in the composition of region of origin when evaluating the effect of import regulation.;Individuals' behavioral responses to infectious disease outbreaks fundamentally alter the nature of transmission. We explore social avoidance (or distancing) behavior using information on observed infections and television viewership within an epidemiological framework. We find that avoidance (as measured by viewership) plays a significant role in disease transmission, but only constitutes one margin of adjustment. Furthermore, the ability of viewership to capture avoidance qualitatively changes during the course of an outbreak, becoming less reliable later in the outbreak. The effect of alternative avoidance behaviors, individual hygiene choices and public interventions are identified in aggregate using public announcements of school closures, and have significantly more explanatory power than viewership alone. This research, however, represents a significant advancement toward incorporating behavioral data into what had historically been a theoretical framework.