COS 124-8
Two centuries of forest change observed using a topographically-based presettlement map

Thursday, August 13, 2015: 4:00 PM
325, Baltimore Convention Center
James M. Dyer, Department of Geography, Ohio University, Athens, OH
Todd F. Hutchinson, Northern Research Station, USDA Forest Service, Delaware, OH

With the expansion of Euro-American settlement, followed by abandonment of agricultural lands and altered disturbance regimes, deciduous forests of eastern North America have undergone dramatic changes in the last several hundred years. Witness trees recorded in presettlement surveys provide a record of forest conditions in Appalachian Ohio c.1800, and can serve as a baseline for change assessment.  The goal of this study is to develop a map of forest associations for a five-county area, and assess changes in the contemporary forest.  A challenge to developing a presettlement map, especially in this topographically-complex study area, is the low spatial resolution at which witness trees are recorded.   Our assumption is that presettlement vegetation pattern will be correlated with topography, since topographic position influences moisture demand and moisture stress at a site. Therefore, we first define spatial sampling units, then aggregate witness tree data within those units. Using a fine-scale DEM, we create classes based on slope, aspect, topographic position, available water capacity and soil pH. In order to determine if different topographic settings support distinctive vegetation associations, a cluster analysis joins “topographic samples” into hierarchical groups based on similarities in the relative densities of constituent species.  To assess change from our baseline, we obtained Forest Inventory and Analysis (FIA) plot data, organized to most closely match our presettlement record.  We used the Sørenson index to assess compositional similarity between forest groups.


6233 witness trees were transcribed from two surveys and plotted within GIS; this number was reduced to 4006 by removing duplicates found in the records, trees in riparian settings, severely disturbed sites (e.g., strip mines),  and trees <25.4 cm diameter. Cluster analysis suggests two presettlement forest associations: ridges, vs. slopes + valleys.  Both were dominated by white oak and similar compositionally, but ridges had a greater abundance of black and chestnut oak, while slopes + valleys had greater abundance of white oak, blackgum, and sugar maple.  Comparison with FIA data reveals a dramatic decline in oaks and a notable increase in red maple.  The Sørenson index confirms significant differences in presettlement and contemporary forests.  Furthermore, the distinction between ridges and slopes + valleys has become more pronounced in the contemporary forest.  It would appear that tree associations of lower moisture-stress sites have exhibited a greater compositional shift in the past 200 years, which appears consistent with expected changes caused by the cooler and wetter climate of the 20th century.