Historical vegetation reconstruction using spatial modeling on pre-Euro-American data to inform ecosystem restoration in the Driftless Area of the U.S. Midwest

Grubh, Kumudan
Major Professor
Lisa A. Schulte
Committee Member
Journal Title
Journal ISSN
Volume Title
Research Projects
Organizational Units
Journal Issue
Natural Resource Ecology and Management

The potential future outcomes of ecological restoration and management activities can be partially understood by investigating reference conditions and historical processes that contributed to the variability of past ecosystems. Reconstructing pre-Euro-American vegetation from Public Land Survey (PLS) records provides spatially precise and extensive reference information from which such variability estimates can be developed. We thus mapped pre-Euro-American vegetation cover within the Driftless Area of the U.S. Midwest using PLS records. Our mapping employed a four step process. First, trees ambiguously designated by the Public Land surveyors were identified to species using logistic regression models. Next, conditionally specified regression models that used species-level data along with relevant environmental predictor variables were used to predict relative dominance of respective tree species at a finer resolution (0.64 km2) than that provided by the original dataset (2.56 km2). Cluster analysis was then used to classify

regression output into vegetation types. Lastly, vegetation types were and mapped using a geographic information system. Outputs from this process revealed that the Driftless Area was historically covered by prairies and oak savannas--dominated by bur (Quercus macrocarpa), white (Q. alba), and black (Q. velutina) oak--interspersed with woodlands and forests, suggesting the absence of oak-hickory ecosystems as typical of the central hardwood region. Prairie made up

42% of the area; savannas 35%; closed forests 14%; and open woodlands 7% of the land cover. Mesic forests in the north and south-east were dominated by sugar maple (Acer saccharum), elm

(Ulmus spp.), American basswood (Tillia americana), white oak, and ironwood (Carpinus caroliniana and Ostrya virginiana). Pine (Pinus spp.) was dominant in the north-east and partly in the east. Aspen (Populus spp.) dominated mostly in the northern part, to the west of the mesic and

pine forests. The Mississippi and Wisconsin River corridors consisted of mostly prairie and savannas composed of a mixture of tree species including silver maple (A. saccharinum), northern pin oak (Q. ellipsoidalis), willow (Salix spp.), bur oak, birch (Betula spp.), white oak, black oak, and white ash (Fraxinus americana). On the whole, the dominance by prairie or savanna ecosystems, and dominance of oak species, indicates that fire played a key role in mediating historical

landscape patterns and ecosystem processes in the Midwest Driftless Area. In addition to aiding in the establishment of restoration and management goals, our maps can be used as baseline data

in investigating interactions between vegetation and disturbance regimes, and in quantifying vegetation change over time. Furthermore, our vegetation reconstruction techniques can be replicated in other geographical areas with similar vegetation data.