Paleoclimate from fossil plants and application to the early Cenozoic Rocky Mountains

Wladimir Köppen called vegetation "crystallized, visible climate," and his metaphor encouraged paleobotanists to climb the chain of inference from fossil plants to paleovegetation to paleoclimate. Inferring paleovegetation from fossils has turned out to be very difficult, however, and today most paleobotanical methods for inferring paleoclimate do not try to reconstruct paleovegetation as a first step. Three major approaches are widely use to infer paleoclimate from plant fossils: 1) phylogenetic inferences rely on the climatic distributions of extant relatives of fossils, 2) morphological inferences use present-day correlations of climate with plant morphology (e.g, leaf shape, wood anatomy), and 3) chemical inferences rely on correlations between climate and the stable isotopic composition of plants or organic compounds. Each approach makes assumptions that are hard to verify. Phylogenetic inference depends on accurate identification of fossils, and also assumes that evolution and/or extinction has not shifted the climatic distributions of plant lineages through time. On average this assumption is less valid for older time periods, but probably it is not radically wrong for the early Cenozoic. Morphological approaches don't require taxonomic identification of plant fossils, but do assume that correlations between plant form and climate have been constant over time. This assumption is bolstered if the ecophysiological cause of the morphology-climate correlation is well understood, but often it isn't. Stable isotopic approaches assume that present-day correlations between isotopic composition and climate apply to the past. Commonly the chemical and physiological mechanisms responsible for the correlation are moderately well known, but often the variation among different taxonomic and functional groups of plants is poorly characterized. In spite of limitations and uncertainties on all methods for inferring paleoclimate from fossil plants, broad patterns emerge from analysis of early Cenozoic floras from the Rocky Mountain region. Paleocene climates across the region were warm with warm winters. Mean annual temperature estimates vary from 10-18 °C depending on the time and place, and ground-freezing climates occurred only north of 40-45 °N. Plants and sedimentary environments suggest low altitude deposition, though floras are not as homogeneous as once thought, suggesting barriers existed. Eocene climates were warmer, with mean annual temperature estimates of 14-25 °C, and ground-freezing climates occurring only north of the Canadian border. Paleobotanical evidence for substantial paleoelevations in basinal areas is weak, but volcanic terrains to the west preserve floras that suggest higher paleoelevations, even in the early and middle Eocene. The terms "frost-free" and "tropical" have sometimes been used to describe Eocene climate and vegetation of the northern U.S. Rocky Mountains, but are probably not justified, with the possible exception of the the warmest early Eocene hyperthermal events at low paleoelevation.