Abstract: Past climates experienced by fossil plant species have often been inferred based on the environmental requirements of their evolutionarily nearest living relatives (NLR). Here we have combined paleoclimatic estimation using NLRs with ecolological niche modeling (ENM) and have demonstrated the combined approach by inferring the Eocene environment of Aesculus L. (Hippocastanoideae, Sapindaceae), a genus of woody eudicots with extant species generally preferring temperate climates. Specifically, we applied ENM-NLR to four Eocene floras in which Aesculus fossils are known to occur: McAbee and Princeton (British Columbia), Republic (Washington State), and Spitsbergen (European Arctic Circle). Additionally, we used ENM-NLR to estimate whether paleoenvironmental conditions were suitable forAesculus at a fifth flora, Copper Basin (Nevada), where the fossil record of the genus is less clear. We generated models for all NLRs in Genetic Algorithm for Rule Set Production using georeferenced data from the Global Biodiversity Information Facility and Bioclim environmental parameters. For each fossil flora, the best models of individual NLRs were summarized into composite community models, which were taken to represent climatically analogous modern communities and used to infer the paleoclimates of the fossil localities. Our results are generally consistent with previous studies that used other methods to estimate paleoclimates and suggest that McAbee, Princeton, Republic, and Spitsbergen had temperate environments. For the Copper Basin flora, our results show ranges of environmental variables that may be too broad for predicting whether Aesculus was present. Our study appears to be the first to combine the NLR approach and ENM to infer paleoclimates.

Key words: Aesculus, arcto-Tertiary, Copper Basin, ecological niche modeling, Eocene, fossils, GARP, Okanagan Highlands, Spitsbergen.