Warm (Not Hot) Tropics During the Late Paleocene: First Continental Evidence

Over the last two decades paleoclimatologists have debated if Paleogene climates were globally warmer than present or if warming was largely at middle and high latitudes. The scarcity of data from equatorial latitudes, and from tropical continental areas in particular, has made it difficult to reconstruct latitudinal gradients confidently. We collected fossil leaves from the late Paleocene Cerrejon Formation, approximately 60 million years old, located in the coal mines of Cerrejon, Guajira Peninsula, Colombia (11.13 N, 72.57 W). The Cerrejon paleoflora is the oldest-known record of Neotropical rainforest. Eight hundred and fifty-nine fossil leaves (823 of dicots) were sorted into 41 morphotypes. We then used leaf margin analysis to estimate mean annual temperature (MAT) and Leaf Area Analysis to estimate mean annual precipitation (MAP). 74% of the dicot leaf morphotypes are entire-margined, giving an estimated MAT of 23.8°C ± 2.1°C. The mean leaf size category for the dicots is mesophyll, and the estimated MAP is 324 cm/year +140 cm -98 cm. The estimated MAT is slightly cooler than for nearby modern climate stations. Preservational biases in favor of stream-margin and wetland plants with toothed leaves tend to result in underestimates of MAT using leaf margin analysis, however the effect is seldom more than 5°C. We conclude that low-latitude South America was unlikely to have been warmer than the modern tropics during the late Paleocene, thus suggesting a very low latitudinal temperature gradient. Three mechanisms might have left the tropics relatively cool while warming the middle and high latitudes: 1) increased poleward ocean heat transport, 2) polar stratospheric clouds, or 3) an unknown forest-microclimate temperature feedback. High levels of atmospheric CO2 alone are not likely to have produced the low latitudinal temperature gradient we infer for the late Paleocene.