Evaluating the sensitivity of paleoclimate results using different approaches of leaf morphotype resolution: A case study using the Eocene Florissant paleoflora, central Colorado

We tested the sensitivity of paleoclimate estimates to different resolutions of leaf morphotype distinction by grouping one paleofloral collection three ways and comparing the results. We examined the different morphotype resolutions using a recently collected Eocene (34 Ma) fossil leaf assemblage from Florissant Fossil Beds National Monument Collection Site 9 (CS-9, Florissant, Colorado), an exposure that yields one of the most diverse sets of dicotyledonous angiosperm (‘dicot’) morphotypes yet analyzed from a single site within this well-studied formation. We grouped this collection using what are herein referred to as ‘Splitter’ (number of morphotypes n = 129), ‘Intermediate’ (n = 96), and ‘Lumper’ (n = 53) morphotyping approaches, which differed in their degree of delineating foliar morphotypes. Leaf Margin Analysis (LMA) and Leaf Area Analysis (LAA) were used to obtain paleotemperature and paleoprecipitation estimates from the morphotype datasets and Climate Leaf Analysis Multivariate Program (CLAMP) was employed to examine further parameters. We perform an additional paleoclimatic analysis of CS-9 in which we divide the site into three stratigraphic sections, and finally we perform an analysis on sample collections above and below a volcanic event horizon within CS-9. Whole collection results across the three morphotype resolutions indicate a mean annual temperature between 10.8 and 12.5 °C, and annual precipitation between 61.8 and 130.7 cm at the time of deposition. The ‘Splitter,’ ‘Intermediate,’ and ‘Lumper’ approaches provide very similar estimates by all methods of analysis and suggest that morphotype resolution specificity has very little effect on climatic results using these leaf physiognomic techniques.