Analysis of counts from monitoring guillemots in Britain and Ireland
(1) This analysis of monitoring data is based on twenty-seven British and Irish colonies where individual guillemots (Uria aalge) were counted repeatedly in clearly defined areas of cliff (plots) 1975-82. It offers advice on the planning of future schemes. (2) In general, numbers at colonies in north Scotland and north-east Britain increased from 1975 to 1982 but some Welsh and Irish colonies decreased. (3) Plots within a colony tended to show similar long-term trends, but there were some exceptions, particularly in colonies where overall numbers were decreasing. (4) At thirteen of the twenty-two colonies with significant long-term trends there was significant year-to-year variation about the trend. Much of this variation seemed haphazard but there was some evidence of a general slowing down in the rate of increase towards the end of the period. (5) Within-year variation of counts was measured using the coefficient of variation (standard deviation expressed as a percentage of the mean). For plots containing 200 or more birds the value was usually 8-9%. (6) Counts at about half the colonies showed marked within-season trends but, in most cases, their numerical contribution to the coefficient of variation was relatively small and the patterns of trends at any colony were not consistent between years. (7) Within each colony the constituent plots displayed similar within-season trends. (8) A simple statistical model for the various components of spatial and temporal variation is given in Appendix 1, which also develops a method for testing for differences in annual changes among constituent plots at a colony. (9) In all colonies combined, there were 107 instances where year-to-year changes were measured on two or more constituent plots. Of these, fifty showed significant differences between plots in annual changes. This suggests that plot differences are to be expected in any monitoring scheme. These can be allowed for, and their effects quantified, by appropriate replication of study plots.