The route-regression method has been used in recent years to analyze data from roadside surveys. With this method, a population trend is estimated for each route in a region, then regional trends are estimated as a weighted mean of the individual route trends. This method can accurately incorporate data that is unbalanced by changes in years surveyed and observer differences. We suggest that route-regression methodology is most efficient in the estimation of long-term (>5 year) trends, and tends to provide conservative results for low-density species.
Most of the surveys presently used to estimate population trends on a large geographic scale depend upon repeated visits to a number of randomly selected routes or monitoring points. As these surveys cannot be analyzed by modeling annual mean densities among routes within a region, no natural annual index of population density exists for the region. We discuss two possible methodologies for estimating annual indices of abundance. In the context of the route-regression methodology, in which trends are estimated for each route and regional population trends are estimated as weighted averages of route trends, it is possible to find average residual distances between the predicted trends on each route and the actual...