TAPHONOMIC CONSTRAINTS ON FOLIAR PHYSIOGNOMIC INTERPRETATIONS OF LATE CRETACEOUS AND TERTIARY PALEOCLIMATES

Foliar physiognomic interpretations of Late Cretaceous and Tertiary palaeoclimates. such as those derived from Leaf Margin Analysis, usually rely on taxon-based observations from the forest canopy. In these analyses, Southern Hemisphere rainforests are sometimes considered analogues of North American Late Cretaceous vegetation. Past analyses have not considered taphonomic biases. Taphonomic influences on specimen-based and taxon-based approaches to foliar physiognomic analyses of palaeoclimate are assessed here using forest-floor and stream-bed leaf litter from modern Australian rainforests. Forest floor leaf litter is markedly local in origin and is taxonomically biased towards leaves from canopy trees. Litter has smaller leaves than expected from the canopy, using either specimen- or taxon-based observations. The Leaf Size Index (LSI) is moderately robust to taphonomic bias, with a correction factor of 3:2 for canopy LSI:forest-floor LSI. Leaf margin type is highly variable using specimens or when total species numbers are low. The accuracy of palaeotemperature interpretations based on leaf margin type alone (Leaf Margin Analysis) are therefore strongly affected by taphonomic biases and local floristics. Direct comparisons between canopy and fossil leaf assemblages may therefore misrepresent the palaeoclimate of the original standing vegetation. Fluvially transported assemblages may exaggerate this bias. However, a "foliar physiognomic signature" is defined for these modern leaf beds using leaf-size, shape and margin type from specimen-, rather than taxon-based observations. Multivariate analysis (Discriminant Analysis) of the foliar physiognomic signature of these leaf beds can differentiate between leaf beds from each of the forest-types examined. Examination of modern leaf assemblages allow an a priori analysis of taphonomic influences on foliar physiognomic character, in contrast to ad hoc hypotheses used in comparisons based on canopy observations, and potentially offer more reliable reconstructions of Late Cretaceous and Tertiary palaeoclimates from leaf megafloras than have previously been shown.