Expanded response-surfaces: a new method to reconstruct paleoclimates from fossil pollen assemblages that lack modern analogues

Abstract

Pollen-based paleoclimatic interpretations of late-glacial to early Holocene climates (17–9 ka) in Midwestern North America are hampered by samples that lack modern analogues. Unresolved questions include the magnitude and direction of temperature seasonality (i.e. were these climates more or less seasonal than present) and the temporal changes in precipitation. Central to the no-analogue problem is the truncation of modern pollen-climate relationships for abundant late-glacial taxa such as Fraxinus. Here we present a new method called the expanded response-surface (ERS) method, developed to reconstruct climates from no-analogue pollen assemblages and applied to a high-resolution late-glacial pollen record from Crystal Lake, Illinois to test hypotheses about late-glacial climates. The key assumptions central to the ERS method are: (1) plant species and pollen abundances follow symmetrical unimodal distributions along climatic gradients, (2) taxa with truncated distributions in modern climate space occupy a subset of their fundamental niche, and (3) expansion of truncated distributions by mirroring around the distributional mode recovers the portion of the fundamental niche not realized in the modern climate space. With the ERS method, we expanded modern pollen-climate relationships by mirroring pollen abundances for each taxon around a mode defined with respect to four climate axes (mean winter temperature, mean summer temperature, mean winter precipitation, and mean summer precipitation). The ERS method reconstructed past temperatures and precipitation during the height of no-analogue conditions (14 160–12 370 cal yr BP) for 37% of the Crystal Lake samples where techniques that employed only modern observational data found matches for only 13% of the fossil samples. The total climate space of the expanded taxa set allowed analogue matches under more seasonal-than-present climates with higher-than-present precipitation. The ERS climate reconstructions for the height of no-analogue conditions indicated cooler-than-present summer and winter temperatures, similar-to-present seasonal range in temperatures, higher-than-present winter precipitation, and similar-to-present summer precipitation. These results thus suggest that high moisture availability helped drive the formation of the Midwestern no-analogue communities with high Fraxinus nigra abundances, but do not show higher-than-present temperature seasonality notwithstanding the higher-than-present insolation seasonality at this time. During the no-analogue late-glacial interval, Picea mariana, F. nigra, and Larix stands probably grew on low-lying, poorly drained soils in the Crystal Lake region; whereas Abies, Picea glauca, Quercus, and Ostrya/ Carpinus grew on upland positions with better soil drainage.