# Survey Design & Analyses

Design - selecting the survey plots

The CBS is based on a random stratified approach. The Republic was divided into eight regions, and 10 km squares (based on the Irish National Grid) were randomly selected within each, and allocated in sequence. For each 10 km square selected, the 1 km square at the extreme southwest corner is surveyed. Those with less than 50% land, e.g. coastal areas or lake shores, have been excluded, leaving some 700 possible survey squares. The survey aims to achieve coverage of the same 1 km squares each year, ideally by the same observer, although there is likely to be some changeover of survey participants.

The ideal survey route within each 1 km square comprises two parallel lines, each 1 km in length about 500 m apart and about 250 m from the edge of the square. For practical reasons there is often deviation from the ideal route. Each 1-km transect is divided into five 200 m sections, at which level all information is collected.

## Analysing the data

The total numbers of adult birds of each species detected in each 1 km square were calculated for each year. The maximum of the two counts (from early and late visits) was used as the annual measure of relative abundance for each species. Annual population indices were calculated using *TRIM *(Trends & Indices for Monitoring Data), a program used for the analysis of time series of counts with missing observations. Counts are modelled as a function of square (site) and year effects, with interpolated estimates for site-year combinations with missing data. The stratified sampling design results in unequal representation of regions across Ireland, so annual counts were weighted by the inverse of the proportion of the area of each region that was surveyed that year. Further details about TRIM can be found here.

Population trends for species occurring on a mean of 30 or more squares over the duration of the survey were estimated by examining the overall rate of annual change, as caution is urged because of the low precision associated with sample sizes smaller than 30.

Population change is usually displayed in the form of indices, where the results from one season are set to some arbitrary figure, usually 1 or 100, and index values are calculated for all other seasons according to how each relates to the base season. A constant rate of decline is exponential when illustrated. For example, if a population is declining by 50% each year, then if the initial index is 1, the index at timepoint 2 is 0.5, at timepoint 3 is 0.25. If the population doubles each year, the index values for the respective timepoints are 2, 4 and 8. Index values are thus measures of relative abundance for a species, and usually the relationship between this and the absolute abundance is unknown.

The mean annual change was estimated by fitting a regression line through the data. Trends were calculated across all habitats. Trends were also produced for a number of bird groups (defined by species of similar habits and habitats) by calculating the geometric means of the annual indices of the respective species.