The Modern Analog Technique (MAT) has become a standard tool in pollen based climate reconstructions. The method assumes that the primary determinant of the distribution of vegetation is climate. This assumption is valid at regional to subcontinental scales. At more local scales, however, other ecological factors such as soil quality impose spatial patterns on vegetation. We tested the robustness of pollen-based climate histories at 10 sites in northwestern Wisconsin on substrates ranging from glacial till to coarse sand. Potential modern analogs included 150 sites from the western Great Lakes region and 450 from northeastern North America from the NEOTOMA database. All had at least 19% of pine grains differentiated between jack-red pine type and white pine type so that undifferentiated pines could be divided proportionately into ecologically important categories. Sites south of 40o latitude, west of 100o, and east of 97o latitude were eliminated to avoid congenerics not found in this region. Climate reconstructions were based on squared chord distance (SCD) analogs between fossil and modern pollen samples. Modern pollen samples with SCD <0.07 were considered good analogs. Climate reconstructions were based on a weighted average of the modern climate of analog sites.
Results/Conclusions
Sites on mesic soils derived from till tended to have cooler temperature (1-2 oC) and higher precipitation (~5%) reconstructions than nearby sites on coarse sand. This offset was often, but not always, of similar degree throughout the core, suggesting that soil texture creates a systematic bias in reconstructed climate due to the types of vegetation it supports. The sensitivity of vegetation to climate change also varied among sites. For instance, two of the three sites on fine sand (intermediate site quality) responded sharply (~5% drop in reconstructed precipitation) to a severe drought ~1500 years ago (documented from a 14C-dated sand lens). All three fine sand sites also had a reconstructed drop in winter temperature at this time. Initial results from sites on coarse sand and mesic till do not show a strong response to this drought. Our results demonstrate the importance of factors such as soil texture in the sensitivity of vegetation to climate change, and therefore the reconstruction of climate from pollen. Further testing is needed to determine whether sites on intermediate quality soils are more susceptible to climatic changes in general, but soil texture should be considered more explicitly in pollen based climate reconstructions.