AbstractAimThe aim was to identify the primary drivers of compositional change in breeding bird assemblages over a 40‐year period.LocationBritain.Time periodFrom 1970 to 2010.Major taxa studiedBirds.MethodsUsing morphological trait measurements and a dataset of presence–absence data for British breeding birds surveyed in 10 km × 10 km hectads during two time periods, we calculated temporal taxonomic and functional beta diversity for each hectad alongside the change in species richness, mean nearest taxon distance (MNTD) and mean pairwise distance (MPD). We also estimated potential drivers of beta diversity, including climatic and land‐use and land‐cover (LULC) change variables, elevation and assemblage species richness in 1970 (1970rich). We used random forest regressions to test which variables best explained compositional change in the assemblages. We also assessed spatial taxonomic and functional change by analysing multiple‐site beta diversity and pairwise dissimilarities between time periods.ResultsInitial (1970) species richness was the most important predictor (highest importance score) across all models, with areas characterized by higher initial richness experiencing less assemblage change overall. The coordinates included to capture spatial autocorrelation in the data were also important predictors of change. Most climate and LULC variables had relatively low explanatory power; elevation and average temperature were the most influential. All metrics increased slightly with increasing elevation, except for species richness change and MPD, which decreased.Main conclusionsThe composition of British breeding bird assemblages changed substantially between 1970 and 2010. Spatial heterogeneity increased, both taxonomically and functionally. We show evidence that hectads with larger assemblages have been buffered from temporal diversity change and that those at higher elevations changed more in composition than those at lower elevations. Overall, coarse‐resolution climate and LULC explained only small to moderate amounts of variation, suggesting that stochastic assembly change or finer‐scale drivers might be drivers of temporal changes in assemblage composition.