Skip common site navigation and headers
United States Environmental Protection Agency
Great Lakes
Begin Hierarchical Links EPA Home > Great Lakes > EcosystemsUpland Ecosystems > Oak Savanna Conferences > 1993 Midwest Oak Conferences > Larry G. Brewer
Aquatic Ecosystems
EPA Region 5 Critical Ecosystems
Ecosystem Funding
Great Lakes Basin Ecosystem
Great Lakes Biological Diversity
Green Landscaping
Rivers and Streams
Upland Ecosystem



1993 Proceedings of the Midwest Oak Savanna Conferences

1993 Proceedings of the Midwest Oak Savanna Conferences


Figure 1. The Oak Openings of Northwest Ohio
Figure 2. Fifteen Remnant Sites Surveyed in 1989
Figure 3. Baseline Stem Density and Age Class in Mary's Savanna for Quercus alba (QUAL), Quercus velutina (QUVE), Sassafras albidum (SAAL), and Prunus serotina (PRSE)
Figure 4. Stem Density and Age Class in Mary's Savanna after Three Prescribed Burns
Figure 5. Average Depths of O1 and O2 Horizons in Mary's Savanna Pre-and Post-Burn

Larry G. Brewer 
3-D Environmental 
781 Neeb Road, #5 
Cincinnati, OH 45233 

Michelle T. Grigore 
Metroparks of Toledo 
5100 West Central Ave. 
Toledo, OH 43615  

A cooperative management and monitoring program was established in 1988 between The Nature Conservancy and Metroparks of the Toledo Area to restore degraded oak savanna remnants in the Oak Openings geologic area in northwest Ohio. The project provided detailed community descriptions and baseline data for monitoring the progress of management in oak savanna units. Historic surveys were utilized to produce a plant community map for pre-European settlement conditions in Oak Openings Preserve Metropark. Macroplots, microplots, vegetation structure, site characteristics, age classes of trees and photomonitoring provided a complete picture of the present-day community. Preliminary results of three years of management with prescribed burning has shown a reduction in Sassafras albidum, Prunus serotina, Quercus alba, and Quercus velutina in the sapling class and a gradual thinning of the organic soil horizon. Carex pennsylvanica and Peteridium aquilinum exhibited dramatic increases in the herbaceous layer. Results will continue to be monitored at five-year intervals. 



The former "Oak Openings" of northwestern Ohio (Moseley, 1928) were located southwest of Toledo on 130 square miles of glacial lake sand deposits (Forsyth, 1959) (Fig. 1). Prior to Euro-American settlement, they consisted of 43% oak savanna (4-43 trees/hectare), 27% wet prairie (<1 tree/hectare), 23% oak woodland (44+ trees/hectare) and 6.7% oak barrens (1-3 trees per hectare) (Brewer and Vankat, 1995). In the Oak Openings the oak savannas, barrens and oak woodlands were located on dune ridges, while wet prairies were found in interdunal areas that were subject to flooding during the spring (Mayfield, 1969). 

Although oak savanna was formerly the dominant plant community in the “Oak Openings”, cessation of fire, fragmentation, and incompatible land uses by man has resulted in only a handful of oak savanna remnants. Many of the remnants exist in a 3,000-acre natural area known as Oak Openings Preserve Metropark. Despite its decline, the Oak Openings area of northwest Ohio still supports over 85 state rare, threatened and endangered plant species; more than any other area in the state (Easterly, 1979, Mc Cane and Burns, 1984). 

In 1988, The Nature Conservancy (TNC) and the Metropolitan Park District of the Toledo Area (Metroparks) jointly funded a study to obtain baseline data on oak communities in the Oak Openings Metropark by establishing permanent macroplots. This information is being used to determine the present condition of oak savanna remnants in the park, as well as changes following management with prescribed burns and selective cutting. TNC also plans to use the data in their efforts to characterize oak barren/savanna communities. 


United States General Land Survey notes taken in 1821 were utilized to produce a pre-European settlement plant community map for Oak Openings Preserve Metropark and determine the extent of oak savanna within the park boundaries. Fifteen historic sites were then surveyed for possible extant savannas by the senior author (Fig. 2). These sites were evaluated on their canopy structure and abundance of prairie understory species. Six remnant stands were chosen for macroplot studies based on their restoration potential. 

Permanent multiple releve macroplots, 20 x 25 meters, were placed within each stand to characterize the plant community. Each macroplot included six 2x2 meter microplots, to further refine data on the herbaceous layer and organic horizon. Vegetation composition, structure, and age class (canopy cover, stratum percent cover, DBH of all stems) was recorded for each macroplot as well as soil characteristics (organic horizon depth, presence of charcoal in soil), topography, and evidence of disturbance. Photomonitoring points were established at the southwest corner of each macroplot to provide a visual estimate of the progress of savanna restoration. Twelve macroplots were initially established in 1988. An additional 38 were established between 1989 and 1990. (Funding for the 1989 and 1990 study was provided by Division of Natural Areas and Preserves of the Ohio Department of Natural Resources.) 

These macroplots will be resampled at five year intervals. Metroparks will use a subset of the data to monitor progress towards the compositional and structural goal of 50% or less oak canopy cover and a herbaceous layer dominated by prairie grasses, sedges and forbs typical of oak savanna. 

Data from the macroplots was analyzed by the Midwest Regional Office of The Nature Conservancy (White et al., 1989) using multivariate techniques. DCA Ordinations and Twinspan analysis were used to classify the vegetation and relate stands to one another and other Oak Barren/Savanna communities in the Midwest. 


Analysis by Brewer (1989) and Brewer and Vankat (1993) of the 1821 pre-European settlement surveys of Oak Openings Preserve showed that 1447 hectares of oak savanna/woodland were present within the park boundaries, with an average tree density of 15 trees per hectare in oak woodland and 3 trees per hectare in oak savanna. The oak woodland was found primarily on the east side of Swan Creek where it was protected from fire. The wet prairies had fewer than one tree per hectare and comprised 51 hectares of parkland in 1821. Although these wet prairies were small, many of them were connected to larger prairies to the northeast and southwest. The oak savanna/woodland consisted of 50% black oak Quercus velutina (black oak), 44.8% Quercus alba (white oak), 3.4% Fraxinus pennsylvanica (ash-mostly green ash), and 1.72% Salix spp. (willow). 

By 1900, nearly all savanna trees present during the pre-European settlement surveys in the “Oak Openings” had been cleared for timber, cropland and grazing (Mayfield 1969). One exception was a 15 hectare area on the north side of the preserve. This community has succeeded to forest, but still has numerous large white oaks-- some nearly 400 years old. By the turn of the century, most savannas on the poor, sandy soils of the “Oak Openings” had been abandoned. Many reverted to oak barrens/savanna; however, with the absence of fire, most of these areas are now woodland and forest. Today oak savanna scattered throughout the original “Oak Openings” covers less than 250 hectares. 

By 1989, the Oak Openings Preserve Metropark was composed of 178 hectares of planted conifers, 95 hectares of red maple swamp (former wet prairie), 532 hectares of oak forest and woodlands, and 692 hectares of open fields, roads, streams and lakes. Two hundred hectares of these open fields are beginning to succeed to oak savanna (Brewer 1989). 

Analysis of macroplot data divided the six remnant stands in Oak Openings Preserve into two plant communities-- Oak Barrens and well drained, rich Woodland/Savanna (White et al., 1989). The communities were characterized as follows: In the Oak Barrens stands, Quercus velutina dominates the canopy and subcanopy, accounting for 95% of the basal area. Quercus alba is present in small numbers in the canopy, and Prunus serotina in the subcanopy. Dominant tall shrubs and saplings include Quercus velutina, Prunus serotina, and Sassafras albidum. The low shrub layer is dominated by Vaccinium angustifolium (early low blueberry), V. pallidum (late low blueberry), Gaylussacia baccata (huckleberry), and lesser amounts of Rosa carolina (pasture rose), Sassafras albidum, Quercus velutina, and Prunus serotina. Carex pennsylvanica (early sedge) and Pteridium aquilinum (bracken fern) cover 25 to 50% of the herbaceous layer within the stands. Lower cover classes of prairie forbs are also present including Lupinus perennis (lupine), Tephrosia virginiana (goat’s rue), Lithospermum caroliniense (puccoon), and Baptisia tinctoria (wild indigo). 

The Oak Woodland/Savanna sites are dominated by Quercus alba and Quercus velutina in the canopy. Prunus serotina, Acer rubrum (red maple) and Sassafras albidum dominate the subcanopy, while Hamamelis virginiana (witch hazel), Sassafras albidum, and Acer rubrum form the major components of the tall shrub/sapling layer. The low shrub layer is composed mainly of Hamamelis virginiana, Acer rubrum, Sassafras albidum, and Celtis occidentalis, with lesser amounts Vaccinium angustifolium, V. pallidum, Prunus serotina, and Rosa carolina. Detailed site descriptions for each of the six stands are given in Brewer (1989). 

In the Oak Barrens there was concern that Quercus velutina saplings were being eliminated by shaded and competition from high densities of Sassafras albidum and Prunus serotina. Shade and litter on the surface also appeared to be contributing to the lower diversity in the understory. The Quercus velutina and Quercus alba canopy trees in the Oak Woodland/Savanna showed even less promise of replacing themselves, with Prunus serotina and Acer rubrum being the dominant trees in the subcanopy. Prescribed burning was chosen to reduce canopy cover and stem densities of the smaller size classes to mimic the natural disturbance that maintained the savanna community at settlement. In one stand the larger Prunus serotina and Sassafras albidum trees were selectively cut to reduce canopy cover. 

Because of the brief period available for burning in northwest Ohio, not all of the stands had been subjected to fire by 1992. However, one high quality oak barren/savanna site, known as Mary's savanna, was burned during the Fall of 1988, and and during the Spring 1989, 1990, and 1992. Preliminary analysis of data from the first three years of burning in this stand and the adjacent control plots indicate some progress. 

Macroplot data from 1988 and 1991 in the burn unit show a significant decrease in stem density of all species in the 1 to 5 cm. and 5 to 10 cm DBH range following management (Fig. 3 and Fig. 4). Sassafras albidum and Pinus strobus (white pine) were eliminated from both these size classes and Quercus alba and Quercus velutinafrom the 1 to 5 cm class. Prunus serotina stem density in this smallest size class was 975/ha in 1988 and only 15/ha in 1991. Photomonitoring reflected this as a gross change in canopy cover and visibility in the tall shrub layer. Many of Quercus velutina formed grubs which grew vigorously when the area was burned. Sassafras albidum sprouts also increased, and more annual burns will be necessary to deplete the stored food in the roots of this species. 

In this short time interval it is much more difficult to assess the effect on the herbaceous layer. Many of the prairie forbs are present in low densities and, although they appear to be increasing their dominance, few have reached the next higher cover class. New species have appeared within the macroplot but not yet in the herbaceous microplots, where a slight increase is most likely to be noted. 

A one way ANOVA, comparing average percent cover for species in microplots in the burned and unburned units indicated significant differences in Carex pennsylvanica and Pteridium aquilinum from pre- to post-burn in the burn unit (P> .05) (Table 1). There was no significant difference between years in the cover of these two species in the unburned unit. An increase in Vaccinium sp. and sassafras albidum sprouts was also seen in the microplots with fire as compared with the control unit, but this difference was not significant. 

The effect of fire on the organic horizon was also studied by measuring the depth of the O1 and O2 layers in 1988, 1989, and 1990 in microplots in the burned and unburned (control unit) portion of St. Mary’s savanna and the control unit. There was a a gradual decrease in fresh and decomposed leaf litter in the burn unit, especially in the O1 horizon which decreased from 5.5 cm. to 1 cm. between 1989 and 1990 (Fig. 5). 


The releve macroplots established prior to management in remnant oak savannas provided baseline data and a means to monitor the success of oak savanna restoration efforts in reducing canopy cover and increasing the dominance of prairie species. Sampling at five year intervals was planned to provide enough time to note differences in the plant communities. However, after three years of prescribed burning in one stand, some conclusions have been produced. 

Stem densities of all woody species decreased in the 1 to 5 and 5 to 10 cm DBH range. The Quercus velutina saplings have formed grubs which grow vigorously when fire is removed from the site for a year. When food resources of less desirable species such as Sassafras albidum and Prunus serotina are depleted and canopy cover decreased, burns will be made less frequently to allow the oak grubs to grow as canopy replacements. 

The herbaceous layer is responding slowly to fire in the oak savanna management unit. We suspect that reduced canopy cover and the burning of the O1 soil horizon increased the nutrients available to rhizomatous plants such as Carex pennsylvanica and Pteridium aquilinum. These two species doubled their percent cover in the burn unit. The O2 horizon, which overlays the mineral soil, is also being slowly reduced by fire. At present, thickness of this layer may prevent the establishment of savanna herb seedlings, many of which require mineral soils on which to germinate. We hypothesize that continued reduction in the O2 horizon will produce an increase in the desired herbaceous savanna species, especially if they still exist in the seed bank or as root stock, and do not have to be dispersed in from long distances. 

Analysis of macroplot data from Oak Openings Preserve Metropark in northwest Ohio and savanna remnants in Indiana, Ontario, Wisconsin and Minnesota indicated a floristic gradient from Minnesota to Ohio (White et al., 1989). The Metropark stands were most closely related to stands in northern Indiana and Ontario. They hypothesized that management for disturbance in the Metropark oak barrens would increase understory diversity and similarity to northern Indiana oak barrens stands located in the Indiana Dunes National Lakeshore. Continued monitoring of the permanent macroplots established in Oka Opening Metropark remnants will be undertaken at five year intervals and analyzed for shifts towards a species composition toward greater similarity to that of the Indiana remnants. Since there are differences in the two sites, it is unclear how close in composition to the Indiana Dunes oak barrens the Metropark oak barrens will become. The Indiana Dunes oak barrens are pristine in that they have continued to be burned at an average interval of 11 years (Henderson and Long 1984). while the oak barrens at the Oak Openings Metropark have not been burned for over half a century. In addition, the top soil (A horizon) in the Indiana Dunes have not been disturbed, while it has been eliminated in many parts of the Metropark as a result of unsuccessfull farming attempst. Furthermore, the hydrology has changed significantly in the Oak Openings Metropark since the turn of the century due to draining (Mayfield 1969). 

The macroplot data will also help to evaluate the success of management in restoring a canopy cover to lessthan 50% as well as greater dominance of prairie forbs and graminoids. As we move towards this goal, continued monitoring of this recovering oak savanna may also give us insight into community dynamics; understanding that could benefit other midwestern oak savanna restoration efforts. 


Brewer, L. G. 1989. Vegetation Changes in the Oak Savannas and Woodlands of Northwestern Ohio. Final Report, submitted to Ohio Dept. of Natura Resources Division of Natural Areas and Preserves. 65pp. 

Brewer, L. G. and J. L. Vankat. 1993. The natural vegetation of the Oak Openings of Northwest Ohio. Map. 

Easterly, N. W. 1979. Rare and infrequent plant species in the Oak Openings of northwest Ohio. Ohio J. Sci. 79:51-58. 

Forsyth, J. L. 1959. The beach ridges of northern Ohio. Ohio Geol. Surv. Infor. Circular 25. 10 pp. 

Mayfield, H. 1976. Changes in the natural history of the Toledo region since the coming of the white man. Metropolitan Park District of the Toledo Area. 17 pp. 

Mc Cance, R. M., Jr. and J. R. Burns (Ed.) 1984. Ohio endangered and threatened vascular plants: abstracts of state-listed taxa. Division of Natural Areas and Preserves, Dept. of Natural Resources, Columbus, Ohio. 635 pp. 

Moseley, E. L. 1928. Flora of the Oak Openings. Ohio Acad. Sci. Special Paper 2: 79-134. 

White, M., K. Chapman and M. Huffman. 1989. Vegetation Analysis of the Toledo Metroparks Oak Openings for Plant Community Characterization and Monitoring. The Nature Conservancy, Midwest Field Office. Minnesota. 42 pp. 


Carex pennsylvanica  
Pteridium aquilinum  
Gaylussacia baccata 
Vaccinium vacillans  
Vaccinium myrtloides  
Vaccinium angustifolium  
Quercus velutina  
Quercus alba  
Rosa carolina  
Prunus serotina
Acer rubrum  
Sassafras albidum  
Panicum lanuginosum  
Tephrosia virginia 
Polygonatum biflorum
Potentilla simplex  
Lithospermum croceum 
Commandra umbellata  
Euphorbia corollata
Lysimachia quadrifolia  
Smilax glauca
Liatris aspera
Solidago nemoralis  
Baptisia tinctoria
Arenaria stricta  
Helianthus divaricatus
Lupinus perennis
Pyrus coronaria  
Hamamelis virginiana  
Elymus canadensis  
Rumex acetosella 
Rubus flagellaris
Carex depauperatum 
Poa pratense 
Polygonatum conaliculatum
Andropogon scoparius


Table 1. Average Percentage Cover of Plant Species in Microplots



Figure 1. The Oak Openings of Northwest Ohio

Figure 1. The Oak Openings of Northwest Ohio

Figure 2. Fifteen Remnant Sites Surveyed in 1989

Figure 2. Fifteen Remnant Sites Surveyed in 1989

Figure 3. Baseline Stem Density and Age Class in Mary's Savanna for Quercus alba (QUAL), Quercus velutina (QUVE), Sassafras albidum (SAAL), and Prunus serotina (PRSE)
Figure 3. Baseline Stem Density and Age Class in Mary's Savanna for Quercus alba (QUAL), Quercus velutina (QUVE), Sassafras albidum (SAAL), and Prunus serotina (PRSE)  

Figure 4. Stem Density and Age Class in Mary's Savanna after Three Prescribed Burns

Figure 4. Stem Density and Age Class in Mary's Savanna after Three Prescribed Burns

Figure 5. Average Depths of O1 and O2 Horizons in Mary's Savanna Pre- and Post-burn.

Figure 5. Average Depths of O1 and O2 Horizons in Mary's Savanna Pre- and Post-burn.



Begin Site Footer

EPA Home | Privacy and Security Notice | Contact Us