Plot Mapping: Estimating Densities of Breeding Bird Territories by Combining Spot Mapping and Transect Techniques

Abstract

-Originally proposed by J. T. Emlen in 1977, plot mapping combines the accuracy of territory or spot mapping with the efficiency of transect or plot count techniques. This method can be used to estimate densities of animal species that advertise their presence on mutually-exclusive territories. Plot mapping can be used with transects or sample plots, with fixed boundaries or speciesspecific variable boundaries. It is especially well-suited for censusing breeding songbird communities. For each species, a probability of detection is estimated and subsequently used to adjust song counts to territories present. The variance of the density estimate provides a measure of its precision. Estimating densities of territories of breeding birds based on counts of singing individuals is confounded by sources of error that may be grouped into two categories: (1) unknown rates of sound attenuation with increasing distance from the observer, and (2) unknown and varying rates of singing by the bird. Methods that have been proposed for dealing with the first type of error (Emlen 1971, Burnham et al. 1980, etc.) involve mathematical relationships between sound attenuation and distance for variable-width transects and variable-radius circular plots. The problem can be circumvented, however, by choosing a strip transect width (or circular plot radius) small enough that the probability of detecting a singing bird is always equal to one. Such narrow fixed-width strips and fixed-radius circular plots require only that the observer determine whether or not the detected bird is within the sample area. Emlen (1977) has been the only one to address the second source of error: birds do not sing all the time. The probability of detecting a territorial bird even nearby is not equal to one if the bird is silent and inactive; in some habitats, in fact, that probability may be zero. Any methodology that fails to account for undetected territories will consistently underestimate densities. Emlen (1977) proposed a simple intuitive method for estimating territory densities from song detections, which entailed calculating an instantaneous sound detection frequency and then using it to adjust the density estimate for territories that were not detected. The method is applicable to both variableand fixed-distance survey techniques. Surprisingly, Emlen's method has received little attention from field researchers and apparently even less from theoretical biometricians. In a recent symposium devoted entirely to counting birds (Ralph and Scott 1981), few participants acknowledged Emlen's (1977) technique for dealing with nondetected birds at short distances, and only one (Tilghman and Rusch 1981:202) reported ever having used the method. My purpose here is to describe an expansion ofEmlen's technique, report on its application, and encourage its use and evaluation in the field. I have tried to quantify density of territories rather than density of individual birds because (1) density of territories probably depends upon habitat quality (which is often what one wishes to study), (2) density of territories is probably constant throughout a breeding season in spite of mortality, as the frequently observed, almost immediate reoccupation of abandoned territories seems to suggest, and (3) some territorial pairs may have helpers, confusing the numerical relationship between singing males and individuals.