Introduction

Monitoring abundance means estimating trends in abundance, usually through time though occasionally across space or with respect to some other variable. Estimating temporal trends in abundance of animal populations is a common objective in both applied and theoretical research. The most common design involves surveys in the same locations run once or more per year for several years. The data are often collected using ‘index methods’ in which the counts are not restricted to well-defined plots or if they are the animals present are not all detected and the fraction detected is not known. Results are usually summarized by calculating the mean number of animals detected per plot, route, or some other measure of effort, during each period. These means are then plotted against time (Fig. 8.1). When the counts come from complete surveys of well-defined plots, then the Y-axis is in density units. In the more common case of index data, the Y-axis shows the number recorded per survey route or some other measure of effort. We assume (or at least hope) that a 5% change in survey results indicates a 5% change in population size but we have no direct measure of absolute density.

The analysis of trend data raises several difficult issues. First, ‘the trend’ can be defined in different ways, and the choice among them may be difficult and subjective. Second, statistical difficulties arise in estimating precision because the same routes are surveyed each year so the annual means are not independent. Third, use of index methods, rather than complete counts on well-defined plots, means that change in survey efficiency may cause spurious trends in the data.