PS 26-137 - Environmental and historical controls on Fagus grandifolia settlement-era distributions in the upper Midwest

Tuesday, August 8, 2017
Exhibit Hall, Oregon Convention Center
Megan Seeley1, Simon Goring2 and Jack W. Williams2, (1)University of Wisconsin, Madison, (2)Geography, University of Wisconsin-Madison, Madison, WI
Background/Question/Methods

Fagus grandifolia occurs throughout the eastern US and reaches its western limit in eastern Wisconsin. While currently no theory fully answers why F. grandifolia does not expand across Wisconsin, two leading theories propose environmental filtering and dispersal limitation as the causal factors. Prior modeling work suggests that lake-effect snow and soil characteristics are key controls on F. grandifolia distributions in northern Michigan, while prior paleoecological research suggests that F. grandifolia did not arrive in eastern Wisconsin until 6,000 years ago. This study uses new settlement-era datasets of F. grandifolia constructed from the Public Land Survey (PLS) to model environmental and dispersal controls on F. grandifolia distribution in Wisconsin and Michigan, a bordering state with comparatively widespread F. grandifolia populations. With this new PLS dataset, Natural Resource and Conservation Service soil maps, and PRISM historical climate data (for 1895-1924), we created a species distribution models. Using the PRISM data, we reconstructed the 1895-1924 upper-Midwest climate. We ran several experiments with an SDM R package to test the transferability of models developed separately for Wisconsin and Michigan F. grandifoliapopulations. We use a spatial process to represent the hypothesis that the Wisconsin population distribution is in part controlled by dispersal limitation.

Results/Conclusions

Historical climate and soil physical properties differ between Wisconsin and Michigan, with Michigan soils being sandier than Wisconsin’s on average. Michigan had, on average, 62 mm more snow than Wisconsin. High snowfall was observed throughout Michigan with the highest proportion falling in the upper peninsula, north-central lower Michigan, and southwest corner. Wisconsin experienced some lake-effect snow along its western border, but the effect was restricted to within roughly 60 km of Lake Michigan. Despite Michigan’s substantial snowfall, it experienced about 20 mm less precipitation annually than Wisconsin. On average, winter temperatures were lower in Wisconsin by 3.5 degrees C. Other environmental variables such as potential evapotranspiration, actual evapotranspiration, and growing degree-days displayed a north-south gradient and were similar between the states. Initial simulations with the SDM package in R indicate a strong environmental control on F. grandifolia distributions and partial transferability between the Michigan and Wisconsin models. Work is underway with the dispersal limitation simulations.