A Comparison of Trends and Geographical Variation in Mammal Abundance in the Breeding Bird Survey and the National Gamebag Census.

Abstract

In the UK, monitoring schemes and surveys for mammals vary markedly in methodologies and geographical and temporal extent among species, mainly due to the specialised protocols required to detect and identify most species. In this report, we compare results from two very different current annual monitoring schemes - the BTO/JNCC/RSPB Breeding Bird Survey (BBS) and the GWCT National Gamebag Census (NGC). Both schemes cover the entire UK, have been running for at least 15 years, and provide quantitative measures of abundance for multiple species, with considerable but not complete overlap in species coverage. This makes it possible to compare both temporal and geographic patterns of abundance and to explore possibilities for combining results across the two schemes.
Of nine species (brown hare, mountain hare, rabbit, grey squirrel, fox, red deer, fallow deer, roe deer and Reeves muntjac) for which sufficiently reliable population trends could be generated for both schemes between 1995 and 2009, there were no significant differences in the population trend between schemes. Both schemes indicated a significant increase in grey squirrel. The BBS also revealed significant increases in three deer species (red, roe and muntjac) and a significant decline in rabbits over this period. With the exception of red deer, the changes in numbers from NGC over this period were similar but not significantly so. This provided justification for generating a joint trend for each of these nine species using data from both schemes that took into account the statistical variation within each scheme which varied among species. BBS and NGC differ in methodologies as well as in geographical coverage (although both cover the UK) and this provides a process for generating an agreed trend for consistent reporting, particularly in the context of statutory reporting requirements.
Spatial maps of relative abundance and change were produced for Great Britain only due to limited data coverage in Northern Ireland. They provide a useful visualisation of geographical patterns, and reveal concordance at very broad but not finer-scale resolutions. This is due to important differences in the design of the two schemes, the extent and distribution of the sampling sites, differences in coverage across the season, and differences in detectability despite using the same spatial modelling procedures. It is therefore not recommended to combine data from BBS and NGC in spatial maps using these spatial modelling approaches. Although clearly it would be possible to combine data without modelling - such as collating evidence of presence over particular time periods - this approach would not be able to make use of the predictive capacity of models needed for data based on a sampling design.