Colder springs and warmer, wetter summers during the late glacial climate of central North America: inferences from stable isotopes δD and δ13C and wood anatomy of sub-fossil oak wood

Background/Question/Methods

Pollen data from sediment cores indicate that vegetation assemblies across much of North America during the late-glacial interval (LGI, 17-10 cal ka BP) were dominated by plant species in mixes that are largely not existent today. These no-analog conditions have made it difficult to specify climatic conditions during the LGI. We report here on stable isotope ratios δD and δ¹³C of cellulose from tree-rings of modern bur oaks (Quercus macrocarpa Michx.) sampled at three locations across its range and compare those to the same isotope signals from sub-fossil oaks. Cellulose δD primarily records precipitation origin and condensation temperature. Other physiological influences on cellulose δD can be accurately estimated through parameterization of a mechanistic model. Cellulose δ¹³C is indicative of the ratio of leaf-intercellular [CO₂] (c_i) to atmospheric [CO₂] (c_a), which in turn is controlled by photosynthetic rates and plant water status. We use these modern stable isotope proxies from 79 sub-fossil oak log cross-sections to infer environmental conditions during the LGI. These logs were recovered from stream sediments in northern Missouri and are ¹⁴C-dated to 9.9-13.63 cal ka BP. We then compare isotopic signals to a wood-anatomy based proxy for spring temperature from the same sub-fossil oaks.

Results/Conclusions

Sub-fossil oak proxy records show broad synchrony with Midwestern pollen data and the Greenland ice core record. With the exception of Younger-Dryas period (11.6-12.9 cal ka BP) sub-fossil cellulose δD was often greater during the LGI than cellulose δD from modern Missouri oaks. Given δD inputs of modern Missouri precipitation, a mechanistic model of cellulose δD cannot replicate these higher values for parameterizations within the modern climatic envelope for bur oak. Therefore, precipitation had higher dD, indicating summers over central North America were characterized by more warm and wet air masses from the Gulf of Mexico. δ¹³C values indicate oaks growing during the LGI had lower c_i/c_a values compared to modern Missouri oaks, attributable to greater growing season irradiance. Peaks in c_i/c_a preceding the start and end of the Younger-Dryas suggest these periods may have been particularly wet. Despite apparently warmer and wetter summer conditions, our wood-anatomy based proxy suggests spring temperatures were cooler during the LGI. This juxtaposition of opposing spring and summer climatic signals support predictions by global climate models of greater seasonality during the LGI. Hence, these paleo-records confirm that a novel climate accompanied the non-analog vegetation types of central North America during the LGI.