BackgroundThe Belly River Group of southern Alberta is one of the best-sampled Late Cretaceous terrestrial faunal assemblages in the world. This system provides a high-resolution biostratigraphic record of terrestrial vertebrate diversity and faunal turnover, and it has considerable potential to be a model system for testing hypotheses of dinosaur palaeoecological dynamics, including important aspects of palaeoecommunity structure, trophic interactions, and responses to environmental change. Vertebrate fossil microsites (assemblages of small bones and teeth concentrated together over a relatively short time and thought to be representative of community composition) offer an unparalleled dataset to better test these hypotheses by ameliorating problems of sample size, geography, and chronostratigraphic control that hamper other palaeoecological analyses. Here, we assembled a comprehensive relative abundance dataset of microsites sampled from the entire Belly River Group and performed a series of analyses to test the influence of environmental factors on site and taxon clustering, and assess the stability of faunal assemblages both temporally and spatially. We also test the long-held idea that populations of large dinosaur taxa were particularly sensitive to small-scale environmental gradients, such as the paralic (coastal) to alluvial (inland) regimes present within the time-equivalent depositional basin of the upper Oldman and lower Dinosaur Park Formations.ResultsPalaeoenvironment (i.e. reconstructed environmental conditions, related to relative amount of alluvial, fluvial, and coastal influence in associated sedimentary strata) was found to be strongly associated with clustering of sites by relative-abundance faunal assemblages, particularly in relation to changes in faunal assemblage composition and marine-terrestrial environmental transitions. Palaeogeography/palaeolandscape were moderately associated to site relative abundance assemblage clustering, with depositional setting and time (i.e. vertical position within stratigraphic unit) more weakly associated. Interestingly, while vertebrate relative abundance assemblages as a whole were strongly correlated with these marine-terrestrial transitions, the dinosaur fauna does not appear to be particularly sensitive to them.ConclusionsThis analysis confirms that depositional setting (i.e. the sediment type/sorting and associated characteristics) has little effect on faunal assemblage composition, in contrast to the effect of changes in the broader palaeoenvironment (e.g. upper vs. lower coastal plain, etc.), with marine-terrestrial transitions driving temporal faunal dynamics within the Belly River Group. The similarity of the dinosaur faunal assemblages between the time-equivalent portions of the Dinosaur Park Formation and Oldman Formation suggests that either these palaeoenvironments are more similar than characterized in the literature, or that the dinosaurs are less sensitive to variation in palaeoenvironment than has often been suggested. A lack of sensitivity...