Pseudorandom walks in ecological analysis: Capturing uncertainty for better estimation and decision making

As scientists, we have to familiarize ourselves with ways of measuring both what we don\u27t know and our confidence in what we do know. In a broad sense, we refer to these measures as uncertainty. The presumption is the more knowledge we have the less uncertain we are and the more confident we become in what we know. In this document I have tried to show a few ways in which uncertainty can manifest itself in the ecology of a species and how that impacts our ability to make decisions about wildlife management. There are two main types of uncertainty that plague ecological analysis: epistemic uncertainty and linguistic uncertainty. Epistemic uncertainty can generally be thought of as uncertainty in what we know. Linguistic uncertainty, on other hand, arises from the imprecision of language. I dealt with epistemic uncertainty in the first three chapters. First, I confronted uncertainty that arises in wildlife surveys due to imperfect detection and from spatial variation in where animals are found. To deal with this, I developed a statistical model that is capable of estimating detection rates and capturing spatial variation in count data. In the first chapter I applied this model to estimating the abundance of endangered bird species in western Nebraska. Next, I addressed the idea of reducing parametric uncertainty about endangered species abundances and distributions by designing better surveys. I analyzed the tradeoff between investing localized effort in a few locations and spreading effort out among more locations. I also considered whether going to the same sites year after year provided better information than choosing new sites where we expect to find individuals. I also dealt with a kind of uncertainty that is not very common to ecologists: Knightian uncertainty. This kind of uncertainty arises in situations where uncertainty is not measurable because we know too little about a system. I used an information gap approach to analyze a set of conservation decisions that must be made despite great uncertainty. Finally, I addressed the notion of linguistic uncertainty in conservation criteria and how this potentially impacts the allocation of money in conservation programs. My hope is that these tools and analyses provide some use to ecologists and wildlife managers