Details: Trend Maps 1966-1996
In this analysis, the trend at any point was estimated as a weighted average of trend information from nearby survey routes containing information from the species. Trend on these routes was estimated as a yearly change, using the Link and Sauer (1994) estimating equation procedure. The weights in the average are the standard route-regression weights (Geissler and Sauer 1990, Link and Sauer 1994), and an inverse distance from the point, the allow for decreased influence as distance from the point of interest increased.
First, all routes within 250 Km of the point of interest were evaluated. If no routes on which the species was observed were encountered in that area, the point was omitted. If fewer than 5 routes were encountered, the search radius was extended to 350 Km and the larger sample of routes was used in the analysis, but no more than 15 routes on which the species was observed were used in the average.
The resulting map was then trimmed to match the boundaries of the species range, as estimated from the relative abundance map.