Skip common site navigation and headers
United States Environmental Protection Agency
Great Lakes Ecosystems
Begin Hierarchical Links EPA Home > Great Lakes EcosystemsUpland Ecosystems > 1993 Oak Savanna Conferences > Kim A. Chapman
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 Conferences1993 Proceedings of the Midwest Oak Savanna Conferences



  • Kim A. Chapman, Mark A. White
    Mary R. Huffman, Don Faber-Langendoen 
    Midwest Heritage Task Force
    The Nature Conservancy
    1313 Fifth Street SE, Room 314
    Minneapolis, MN 55414
    Tel: (612) 331-0750

    Oak barrens landscapes offer great opportunity for the restoration of large areas of savanna in the upper Midwest. Located on poor soils of outwash, lake plains and dunes, and high or irregular moraines, oak barrens consists of a complex mosaic of natural communities and vegetation structures. These include savanna and more closed-canopy woodlands and scrub, dry and moist oak sites, moist prairies, wetlands, and disjunct populations of plants of the Atlantic coastal plain. Fire and drought strongly influence vegetation structure, while local site factors (topography, soils, water bodies, wetlands) control both fire frequency and intensity, as well as species composition. Frequent low-intensity fires, probably more common historically on level ground near prairies, result in little tree death, producing a savanna or woodland structure with little woody understory. Infrequent high-intensity fires, probably historically more common in areas of rough topography, water bodies or swamps, result in much tree death, producing a dense canopy during fire-free periods, and scrub following a fire. Despite tree death, oaks persist as "grubs" due to their ability to resprout, and readily replace canopy trees during fire-free intervals. Several rare species occur in oak barrens, the most notable being the Karner blue butterfly. Degradation of oak barrens takes place when fires are prevented, hydrology is altered, land is cleared for agriculture or mined for sand, and residential and urban development takes place. Oak barrens are apparently not heavily damaged by grazing, which seems to maintain open conditions in the absence of fire. Restoration should address these factors, as well as consider the mosaic of several plant communities and vegetation structures that characterize oak barrens. Areas of 1,000-2,000 acres are perhaps minimal to encompass this variety, and provide for the habitat needs of some animal species. Recovery of disturbed lands around high quality barrens should be undertaken, given the ability of many oak barrens plant species to persist in the soil or recolonize bare ground. While fire is the most important management tool, manual cutting techniques are frequently used to accelerate removal of trees. Monitoring of canopy and understory cover, as well as groundlayer diversity, will help direct management towards goals set for an oak barrens landscape.



    Oak barrens have a canopy layer that ranges from a sparse, scattered canopy (generally from 5 to 30 percent cover - sometimes termed a savanna physiognomy), where the ground layer may vary from open, herbaceous dominated to scrub oak dominated, to a more closed canopy (30 - 80% - sometimes termed a woodland physiognomy). The canopy layer is dominated by one or two of the following species: Quercus velutina (black oak), Quercus ellipsoidalis (northern pin oak), Quercus macrocarpa (bur oak), or rarely Quercus marilandica (black jack oak). Quercus alba (white oak) often occurs as a canopy tree in the eastern range, but is seldom dominant, except in woodland conditions. Tree height varies from 5 to 15 meters. A subcanopy layer composed of the preceding species, or in some cases Prunus spp, Cornus spp., Sassafras albidum, Carya spp., and Corylus americana may be present. The low shrub/scrub layer may contain ericaceous shrubs, Rosa spp., and any of the preceding woody plants. The herb layer is dominated by graminoids (e.g, Carex pensylvanica, Andropogon gerardii, Schizachyrium scoparium); forbs tend to increase as woody cover increases. Composition and structure vary in this type depending on geography, site characteristics, and disturbance history. The open oak barrens can be divided into two main types, Black Oak Barrens and Bur Oak - Northern Pin Oak Barrens. Northern Pin Oak Barrens of northern Wisconsin and Michigan form a variant of the Black Oak Barrens. Additional woodland types, not discussed in detail here, include White Oak - Black Oak Woodland. 


    The Oak Barrens occur in the prairie-forest transition region of the Midwest. Black Oak Barrens range from southern Michigan, the glaciated portions of Ohio, Indiana, Illinois, south of the tension zone through Wisconsin and southeastern Minnesota, and westward to the glaciated regions of Iowa, Missouri, Kansas and Nebraska (Fig. 1). Bur Oak - Northern Pin Oak Barrens occur in central and western Minnesota, and perhaps west-central Wisconsin (Fig. 1). Topography varies from level to gently rolling, and landforms from outwash plains to dune systems on the lake plain. 

    The barrens communities generally occur on well-drained coarse-textured sandy, infertile soils derived from glacial outwash, high gravelly/sandy moraines, or lake plain dune systems (Curtis 1959). Soils vary from almost pure sand to sandy loam to loamy sand. In the driftless area of Wisconsin, Minnesota, Iowa and Illinois, the communities occur on unglaciated, residual soil over sandstone bedrock. 

    Landscape Position, Origin and Type
    An understanding of the oak barrens landscape mosaic is essential for designing long-lived reserves. The next section outlines basic principles of oak barrens landscape structure and process. The exemplary oak barrens landscapes from which much of this report was derived are: Toledo Oak Openings in northwestern Ohio, Indiana Dunes National Lakeshore along Lake Michigan in northern Indiana, and Cedar Creek Natural History Area and Allison Savanna in east-central Minnesota (these two sites encompass the same oak barrens system). 

    Oak Barrens are dominant on portions of a) glacial lakeplains where aeolian dunes formed at old shorelines (Gordon 1969, Hanson 1976)(Fig. 2), and b) glacial outwash plains of sands and gravels, also subject to dune formation (Archambault 1987, Homoya et al. 1985) (Fig. 2). Dune ridges at southern Lake Michigan and southeast Lake Huron are closer and more regularly concentric than along the west shores of Lake Erie and Lake Huron. 

    Oak Barrens were present, but less common, in c) ice-contact (kettle-kame) and gravelly/sandy end moraines, where the ice sheet melted in place, piled unsorted materials in knobs and ridges, and created kettleholes where chunks of ice were buried, then melted (Archambault 1987, Chapman 1984). While the hydrological interactions in Oak Barrens are not well understood, seasonally high water plays an important role in lower areas (Cole and Pavlovic 1987). Oak Barrens on upper slopes may provide important locations for infiltration and creation of subsurface flow of water to lower, moister portions of the mosaic (S. Apfelbaum pers. comm.). 

    Location of plant communities in the oak barrens landscape mosaic

      • a)  Driest places in the landscape--dune and ridge tops, south and west-facing slopes, steep slopes, upper slopes--support oak barrens. Deep sands without layers that impede drainage support oak barrens on flat ground. Black oak dominance shifts to bur oak-northern pin oak from south-central Wisconsin northwestward, with concomitant understory changes. Oak Barrens also occupies lower, level landscape positions westward from south-central Wisconsin. At the University of Wisconsin Arboretum, near the transition zone, the barrens oaks are mostly hybrids of black oak and northern pin oak (Kline 1989, pers. comm.).

        b)  Lower dune ridges, lower slopes, dry flats between ridges, and north slopes support white oak or white-black oak woodland and savanna with bracken fern and Baptisia spp. in the east, changing to northern pin oak woodland and bur oak savanna west of south-central Wisconsin (Archambault 1987, Cole & Pavlovic 1987, Chapman & White pers. obs., Bakowsky 1988). Red oak may occur, particularly on north slopes. Clay lenses or mineral cementations often occur within the B soil horizon, improving soil moisture (Grigal et al. 1974, Archambault 1987). 

        c)  Moist dune valleys and sand flats just above maximum flood level support moist sand prairie (Cole & Pavlovic 1987, White & Madany 1978) with characteristic composition (e.g., Scleria triglomerata, Hypericum kalmianum, Liatris spicata, Polygala sanguinea, Aletris farinosa, Baptisia spp., Spiraea spp.). Depth to water table on lakeplains January to April is 0.15-0.45m (Stone et al. 1980), in June 1.1-1.5m (Cole and Pavlovic 1987). This community diminishes inland from the Great Lakes and westward from central Wisconsin. Populus tremuloides readily occupies sites of moist sand prairie because of a water table near the surface (Cole and Pavlovic 1987, Grimm 1984). In southeastern Michigan outwash plains, black oak woodland exists on moist sand prairie (Archambault 1987). 

        d)  Wet prairie, wet meadow, and marsh occupy channels and depressions in lakeplain dune formations and in outwash (Moran 1978, Cole & Pavlovic 1987, Chapman & White pers. obs.). Sites are flooded in spring (Grigal et al. 1974, Cole & Pavlovic 1987). Marshes occupy the deepest portions of channels anddepressions. In Minnesota outwash, water in marshes was 0.5-1.5m deep throughout the year (Grigal et al. 1974). Government land surveyors recorded depths of most marshes on the lakeplain in the Chicago region at 0.3-0.9m. June water table in lakeplain wet prairie was 0.5-1m below the surface (Cole & Pavlovic 1987). Quercus palustris (Pin oak) and Populus tremuloides occupy the highest sites in wet prairie (Cole & Pavlovic 1987, Chapman & White pers. obs.). Pin oak flats are found in southern Michigan and southwestern Ontario, but disappear northwest of Illinois. 

    Other communities: Water levels in wetlands fluctuate in these outwash plains and sandy glacial lake beds. Consequently, some gently sloping shorelines harbor communities of plants, many of them annuals, which respond favorably to cyclic flooding and draw-down, including species associated with the Atlantic Coastal Plain, carried here by migration along glacial lakeshores and waterfowl transport (Keddy & Reznicek 1982, McLaughlin 1932, Peattie 1922). 

    In southern Michigan, southern Wisconsin, and east-central Minnesota, peatlands supporting bog, bog forest, and fen can occupy depressions in outwash plains, to the exclusion of wet prairie and marsh. 

    In the prairie-forest transition zone the climate can support prairie, forest, or barrens vegetation. Vegetation pattern is most closely correlated with variation in physical factors of the landscape: topography, edaphic factors and their influence on disturbance regime (Anderson 1983). The variation in composition and structure of Oak Barrens is related to topographic and edaphic variability which influence nutrients, soil moisture, humidity, insolation, and to the effects of adjacent fire resistant communities on the movement of fire. These factors along with macroclimate affect fire frequency and intensity (Grimm 1984). 

    Soil Characteristics and Topographic Factors
    Oak Barrens soils have low fertility, organic matter content, and moisture retention capacity (Grigal et al. 1974, White 1983, Cole and Pavlovic 1987). Bray (1960) determined that oak barrens soils have low nutrient content compared with forest and prairie soils, noting that black oak and white oak foliage have low calcium content which may not provide sufficient buffering to prevent leaching of nutrients from the soil. Cole and Pavlovic (1987) reported that soil moisture and fertility increase from high elevation dry oak woods to mesic oak woods to mesic prairie. Tilman (1983, 1984) determined that nitrogen was the major limiting nutrient in the sandy soils of Cedar Creek Natural History Area (CCNHA) in east-central Minnesota, which contains oak barrens. Experimental work at CCNHA by Tilman (1983, 1984), Inouye et al. (1987), and Wedin and Tilman (1992) indicate that C-4 grasses characteristic of prairies and savannas are good competitors at low nitrogen and litter levels. White (1986) found significant differences in species distributions related to soil series at Cedar Creek. Quercus ellipsoidalis peaked in the Zimmermann series while Quercus macrocarpa peaked in the Sartell series. Species characteristic of dry open sites (e.g., Stipa spartea, Dichanthelium oligosanthes) peaked in the Sartell series and were negatively correlated with Q. ellipsoidalis overstory. The Sartell series is a Typic Entisol (dry B horizon, no clay layer in B2 horizon) with lower water content at field capacity than the Zimmerman series, and Alfic Entisol (clay layer in B2 horizon) (Grigal et al. 1974). 

    Cole and Pavlovic (1987) demonstrated that elevation and slope aspect class differentiated dry oak woods from mesic oak woods. The highest elevation sites were all dry oak woods. High westerly slopes had the lowest stem density, while northern exposures had the highest basal area composed of Acer rubrum, Q. alba, and Sassafras albidum. Shrub cover changed with differences in slope aspect and insolation: exposed slopes were dominated by Vaccinium pallidum, V. angustifolium, and Gaylussacia baccata, while northern exposures had high cover of Prunus serotina, Cornus foemina, and Toxicodendron radicans

    Climate and Drought
    Westerlies prevail in the temperate region, carrying fire across landscapes from west to east. Local moderation of that regional pattern may take place at a mesoclimatic level near the Great Lakes (Albert et al. 1986). Bacone et al. (cited in Cole & Pavlovic 1987) believed that a shift from black oak to white oak dominance in barrens at the south end of Lake Michigan was due to moist, cool winds from the lake moderating fire behavior at the east end of their study area. Other lakeplain sites may experience similar mesoclimatic effects. 

    Climate, particularly the periodic occurrence of drought, has played a role in the distribution and persistence of oak barrens. Kline and Cottam (1979) suggested that climate shifts in the last 500 years may have affected the persistence of some oak systems. Whitford and Whitford (1971) indicated that droughty conditions on sandy soils may be a more significant factor maintaining oak barrens in central Wisconsin than fires. Faber-Langendoen and Tester (1993) found that drought, perhaps in combination with rapid moisture changes, probably caused high mortality rates for mature oak trees in parts of the barrens at CCNHA that had sparse canopy cover. 

    Landscape Patterns and their Effects on Disturbance Regimes
    Plant communities in a landscape influence regional fire patterns. Forested peatlands, and rich moist forests are generally not flammable, while prairies and savannas catch and carry fire readily. Marshes are intermediate in flammability between these forests and prairies or barrens, depending on the amount of precipitation and time of year. The variable physiognomy of oak barrens can be partly explained by the influence of landscape pattern and adjacent plant communities (Bowles et al. 1994). Barrens near open water and fire-resistant plant communities experienced fewer, more intense fires than barrens away from these communities. Such sites cycled from scrub to woodland or forest, back to scrub. Next to fire sources, i.e. level prairie damped by few wetlands, barrens burned frequently but with low intensity, creating and maintaining savanna and woodland physiognomies with little underbrush. Intermediate situations created a savanna physiognomy in the canopy, and brushy grassland beneath. 

    In Minnesota, almost exclusively Indian-set fires swept entire landscapes in spring or fall, except in fire-safe areas behind rivers, streams, lakes, and rough topography (Grimm 1984). Presettlement vegetation on the Anoka Sand Plain in east-central Minnesota illustrates the influence of landscape patterns and adjacent plant communities on the distribution of oak barrens. i) Oak barrens sites next to fire-resistant portions of the landscape (e.g. open water, tamarack-cedar swamps) were described by government surveyors in 1854-56 as "Oak Brush", "Grub Oak", "Brushland," and "Brushy Prairie." ii) Away from these fire sinks, but next to marshes and wet meadows (which burned easily only in the fall), oak barrens sites were described as "Black Oak with Grub Oak understory", "Scattering Black Oak with Scrub Oak" and "Timber scattering Black and Bur Oak and Oak Brush." iii) Next to level prairie with few marshes or wet meadows (a fire source that always burned), oak barrens sites were described as "all of line in Prairie w/scattered Bur and Black Oak", "Bur Oak Openings" and "Bur Oak and a Few Black Oak, no Underbrush." 

    Conversely, small wetland sites in a matrix of oak barrens were mapped by surveyors as marsh and wet meadow. But marsh and wet meadow sites adjacent to fire-resistant plant communities sometimes supported tamarack swamp. 

    Disturbances: Fire Effects
    The effect of adjacent vegetation described above produces, in schematic fashion, the following conditions: i) infrequent, high intensity fires, ii) fires of intermediate frequency and intensity, and iii) frequent, low intensity fires. Nuzzo (1986), citing work by Grimm (1981), Henderson (1982) and Thor & Nichols (1973), concluded that low intensity fire creates woodland with little canopy death and an open understory, while infrequent, high intensity fire kills canopy trees and produces scrub. Subsequent frequent burning does not kill the oak scrub, and may increase stem density (Grimm 1981, Curtis 1959). Many historical and modern accounts state that no burning at all produces a closed canopy in 20-40 years, possibly taking longer on sandy ridgetops and south-facing dune slopes (Cole & Pavlovic 1987). Lack of fire in fire susceptible areas leads to increased canopy and subcanopy cover, invasion by fire intolerant species (P. serotina, S. albidum), increased litter layer and fuel loads, and decreased herb layer diversity. Increased tree cover in some cases will lead to more mesic site conditions: higher humidity, lower soil temperature, lower wind speeds, greater soil moisture retention. 

    The role of fire in oak barrens has been documented and investigated in numerous studies (e.g., Curtis 1959, White 1983, Anderson and Brown 1983, Henderson and Long 1984, Grimm 1984, Tester 1989, Faber-Langendoen and Davis 1995). In general fire maintains open canopy, and depending on frequency and intensity either open or scrubby understory. In the absence of fire Oak Barrens may succeed to closed dry forest (Curtis 1959, White 1986). 

    White (1983, 1986) determined in a central MN oak barrens (CCNHA) that low intensity prescribed fireat 1 to 3 years intervals removed woody cover in the sapling and shrub layers but had little effect on large canopy trees (>30 cm dbh). Graminoid and forb species increased in importance with this fire regime. Quercus ellipsoidalis peaked on unburned sites, while Quercus macrocarpa peaked on burned sites. Tester (1989) examined the effects of fire frequency on vegetation at the same site. Tree density and basal area decreased with greater burning frequency, and most of the reduction occurred in the 5 to 25 cm dbh size classes. Species richness increased with fire frequency. Results suggested that 2 year burn intervals produced the greatest ground layer diversity. Percent cover of true prairie species (graminoids, forbs, and shrubs) tended to increase. Introduced grasses and forbs as a whole showed no significant change in cover with respect to the number of burns. However, initial cover values were less than 1% at most sites. 

    Henderson and Long (1984) examined the recent fire history of two black oak woodlands in northern Indiana and determined that fire frequency and intensity strongly influenced composition and structure. Infrequent, high intensity fires created an open overstory and a scrubby sapling layer of black and white oak and high cover of herbaceous species. More frequent low intensity fires were associated with higher canopy cover, higher density, lower basal area and lower herb and shrub cover. Curtis (1959) hypothesized that scrub oak barrens were subject to infrequent catastrophic fires which removed the canopy and created a dense scrub oak structure. 

    Anderson and Brown (1983) studied the effects of a prescribed fire in an open oak barrens and adjacent closed forest in Northern Illinois. They found no mortality or damage to large, isolated trees in the barrens while in the closed forest tree mortality was high. Tree mortality in the forest was attributed to high fuel accumulations. They hypothesized that isolated trees in a savanna are in part protected by 1) shade which inhibits graminoid and forb productivity, resulting in low fuel loads, 2) oak leaves are swept away by wind leaving small amounts of leaf fuel. Anderson and Brown (1986) proposed that the maintenance of closed forests in fire susceptible areas was precluded by fuel buildups of quantities sufficient to allow fires to cause high tree mortality and convert closed forest to savanna. However, isolated trees may be more susceptible to drought effects (Faber-Langendoen and Tester 1993). 

    Disturbance: Grazing Effects
    Livestock grazing had several impacts on oak barrens vegetation. Grazing probably reduced fuel loads and changed fire behavior, possibly leading to less frequent, lower intensity fire and higher canopy cover and tree density. Over-grazing can weaken native grasses and allow invasion of exotic and non-native weedy species. Curtis (1959) and others stated that post-settlement grazing in the absence of fire prevented increases in tree density in midwestern oak barrens and savannas. However, Whitford and Whitford (1971) found in a central Wisconsin stand that early post-settlement increase in tree density corresponded with grazing. They cite Dyksterhuis's (1957) hypothesis that grazing promotes tree reproduction by weakening the herb layer species. Neverthelesss, grazed savannas often remain open under heavy grazing, and succeed to closed woodlands once grazing is stopped (Henderson 1989, pers. comm.). Grazing may both increase the rate of tree establishment and, under heavy grazing, increase seedling mortality. 

    Overstory-Understory Relationships
    Curtis (1959) and Bray (1960) viewed barrens and other open treed communities as transitional between prairie and forest and not as a distinct association. Bray (1960) in an ordination of understory vegetation determined that the order of stands was correlated with a soil moisture gradient from dry to moist and with increasing canopy cover. He stated that xerophytic species were the least shade tolerant, while species characteristic of more moist habitats were more shade tolerant. Bray (1958) noted that, at a given soil moisture level, increasing shade increased the number of mesic species that were found. Bray (1960) suggested that xerophytic species' adaptations to prevent moisture loss (e.g. wax coatings, thickened cuticle and epidermis) may limit light absorption by photosynthetic organs, leaving these species less competitive in shaded habitats. Bray (1958) described a rapid decrease in light intensity with any degree of shading by leaf or branch canopy, leading to an increase in soil moisture and a decrease in soil temperature beneath the savanna canopy. 

    White and Chapman (1989) found a strong correlation between species distribution and canopy cover in oak barrens and oak woodland in northern Ohio. Species composition changed from dry site, disturbance-adapted species to those characteristic of shaded, mesic sites along gradients of open to closed canopy and dry todry-mesic soil moisture. 

    In general, species characteristic of dry open habitats and of more closed mesic sites are distributed in oak barrens stands relative to canopy cover, site moisture status and disturbance history. 

    Presettlement Vegetation Descriptions
    Early workers described barrens as an open parkland with widely scattered oaks and a prairie understory (Stout 1944, Bray 1960). This savanna physiognomy characterized some deep soil oak openings and barrens communities, but probably was not representative of the variation in these communities during the presettlement period. 

    Recently a new model has emerged which describes a more variable physiognomy for oak barrens. Grimm (1981) determined that the prevailing physiognomy for oak barrens in the Big Woods region of Minnesota was scrub oak. He described the vegetation as dense thickets of brush and scrub oak, where scrub oak patches occurred in a mosaic with grassland. "True savanna" (widely spaced trees with a grassland understory) was present but was not widespread. 

    Grimm (1981, 1984) examined studies of presettlement vegetation from other areas in the prairie-forest border region and determined that oak barrens formed a heterogeneous vegetation group. In this model, most trees were small, as few reached sufficient size to resist stem killing by frequent fires. Trees weakened by fire were susceptible to pathogens. Sprouts grew from oak grubs in the understory. Frequent fires kept oak sprouts low in stature, but increased numbers of sprouts. A few years without fire might lead to formation of dense thickets with scrub oak overtopping herbaceous vegetation. 

    An examination of land survey records of extant oak barrens and other areas supports the view of a heterogeneous physiognomy. Survey line descriptions from CCNHA in central Minnesota cited earlier [Landscape patterns and their effects on disturbance regimes] provide a good example of this: "scattering black and burr oak, oak brush", "black oaks with scrub oak understory", "burr oak and few black oak, no underbrush", "brushy prairie". Bacone et. al. (1979) provided these descriptions of presettlement Oak Barrens in northern Indiana: "shrubby yellow and jack oak barrens" 'white and yellow oak barrens" "rolling with oak". In southern Michigan descriptions ranged from open park like savanna with large scattered trees to small scrub oak barrens (Chapman 1984). Descriptions from the Kitty Todd preserve in northern Ohio refer to "scrub oak, oak brush, soil 3rd rate" (Huffman 1989, pers. comm.). 

    Savannas on Walpole Island in Lake Erie have been burned annually for the past 100 years. These stands have both the classic scattered canopy structure with large widely spaced oaks and an understory of prairie species, as well as woodland structure. Oak grubs are present but remain small as a result of the annual fires (Bakowski 1989, pers. comm.). 

    In summary, oak barrens growing on dry, infertile soils were heterogeneous, ranging from dense scrub oak to open, scattered canopies, depending on the fire regime. The fire regime was influenced by fire breaks, topo-edaphic factors, climate, and native American activities (Grimm 1981, 1984, Anderson and Brown 1986). 

    Invertebrate Fauna
    The fragmented and degraded nature of midwestern oak barrens has resulted in the decline of many insect species which are characteristic of dry, open habitats or are obligate to barrens host plants. 

    Panzer (1988) suggests that prairie-restricted insect species are extinction prone because of fluctuating population, poor dispersal, and patchy distribution. Competition, changes in habitat that decrease its quality, predation and excessive burning can contribute to their extirpation. Timing and frequency of burning may be important for maintaining invertebrate fauna. 

    Oak barrens represent important habitat for a number of rare Lepidoptera. The following rare Lepidoptera are characteristic of oak barrens habitat and deserve special management concern: Lycaeides melissa samuelis (Karner blue), Incisalia iris (Frosted Elfin), Erynis persius (Persius Dusky Wing), Viola pedata (Regal Fritillary). A recent publication by Andow et al. (1994) brings together information on the biology, status, and management of Karner blue. For information on species composition of other invertebrates see Panzer (1987, 1988), Panzer and Gnaedinger (1986), and Panzer and Stillwaugh (1987). 

    Vertebrate Fauna
    More information is needed on the role of vertebrate fauna in oak barrens. Davis et al. (1995) present an integrated model showing how fire affects pocket gopher populations that in turn, through their soil mounds, affect the persistence of certain plant species. Vertebrates represent special problems in preserve design and management of oak barrens because of the large areas required to maintain minimum viable populations of larger mammals. Other articles in this Proceedings provide lists of vertebrate species. 


    Fire suppression
    The primary threat to the oak barrens community is fire suppression. The most serious consequence of fire exclusion is increased woody cover (Curtis 1959, Anderson and Schwegman 1982, Aldrich 1888, and Gilbert 1873). Numerous forms of degradation stem from fire suppression, which results in increased woody cover, shading and litter effects, as follows: 

    • increased density of woody species characteristic of oak barrens and of moister, more shaded conditions (Curtis 1959, Cottam 1949, Aldrich 1888, Gilbert 1873);
    • loss of characteristic savanna or woodland structure;
    • long term fire suppression leading to closed canopy forest and inadequate oak regeneration (Brewer 1989, Abrams 1992);
    • reduced species diversity in the understory (White 1986, Tester 1989);
    • dominance shifts in understory species toward lower cover by graminoids, more by forbs and woody plants (White 1986,  Pavlovic et al. 1984);
    • disappearance of some Lepidoptera, such as the Karner blue, as host plants decline in flowering or seed production;
    •  reduced vigor of understory species;
    •  increased likelihood of canopy oak mortality when fire re-enters system because of high fuel accumulations (Albert pers.  comm., Huffman & Chapman, pers. obs., Henderson and Long 1984, Anderson and Brown 1983, 1986)

    Early accounts of changes in barrens describe dramatic increases in woody cover as a result of cessation of Indian burning (Cottam 1949, Aldrich 1888, Gilbert 1873, Vogl 1961, Chapman 1984), though in oak barrens systems these changes may have been slower (Whitford and Whitford 1971). Aerial photographs taken at TNC's Kitty Todd Preserve and Toledo Oak Openings Metropark and of TNC's Allison Savanna Preserve and the Cedar Creek Natural History Area (CCNHA) between 1938-1940 indicate that the barrens remained sparsely covered by tree canopy well into the settlement period (Huffman, Chapman, pers. obs.; Faber-Langendoen and Davis 1995) Evidence from field observations, interviews with past landowners and neighbors, as well as the photos themselves, confirm that these areas were pastured, grazed and/or hayed, which may have kept the barrens open (Gary Reese pers. comm., Joe Croy pers. comm., Rick Johnson pers. comm.). 

    The herb layer of oak barrens is generally resistant to grazing. Curtis (1959) noted that prairie was most resistant to grazing at wet and dry extremes and most susceptible to replacement at the mesic center. Tilman (1987) suggests that cool season exotics have higher nitrogen requirements than dry prairie grasses and thus are limited by low nitrogen levels on dry, infertile sites. Overgrazing may weaken native grasses and allow increased abundance of cool season exotics such as Poa pratensis and P. compressa

    Hydrologic Changes
    Before settlement of North America, beaver were abundant on first through fourth order streams where they could impound flowing water and create a supply of young aspen, willow, alder and other woody plants for winter food (Naiman et al. 1988). In northern Minnesota beaver build 2-3.9 dams/km of stream, and on the north shore of the Gulf of St. Lawrence they build 8.6-16 dams/km of stream. Beaver populations in the Midwest were decimated in the mid-1800s by the the fur trade; hence, their effect in oak barrens landscapes hasnot been felt for over 100 years. 

    Beaver may not have occupied certain lakeplain barrens that lacked flowing streams (Cole & Pavlovic 1987). But Hubbard (1838) was told by Indians that beaver, abundant on the southeast Michigan lakeplain but gone by 1807, created wet prairies by damming streams. On lakeplains and outwash plains with streams, beaver would dramatically influence the landscape by i) expanding wetland area, and ii) creating barriers to fire. Because flooding creates and maintains wet prairie and moist sand prairie, beaver benefited these communities. Since active impoundments contain some open water year-round and even abandoned dams retain spring meltwater, fires could not cross beaver-influenced wetlands easily, and perhaps not at all in spring. 

    Alteration of hydrology in barrens landscapes by removal of beaver, ditching and tiling may not seriously affect Oak Barrens per se because they are located higher on the landscape (Cole and Pavlovic 1987, Archambault 1987). However, oak barrens landscapes are studded with moist sand prairies, wet prairies, wet meadows, and marshes. These communities co-occur with specific soils, depths to water table, and perhaps amount of organic matter (Cole and Pavlovic 1987, Huffman pers. obs.). Early accounts indicate dramatic increases in woody cover on prairie portions of the landscape mosaic following ditching and drainage (Davis 1908). 

    Agriculture and Sand Mining
    The most open portions of Oak Barrens attracted early farmers because clearing was unnecessary. In many cases, open areas were farmed for brief periods unsuccessfully and abandoned. Because of poor fertility, only rarely are barrens cultivated today. 

    Studies at CCNHA on old fields which were formerly oak barrens indicate low nitrogen levels may be responsible for slow invasion of abandoned fields by oaks (Tilman 1987). Documentation of the rapidity of invasion of oaks on other abandoned farmlands would be of interest. 

    Borrow pits and sometimes other sand-mining operations are found in oak barrens landscapes. This is more likely to occur in lakeplain systems where dunes are larger. Oak barrens on southeast Lake Michigan sand dunes have been destroyed by sand mining (Chapman pers. obs.). Portions of Kitty Todd Preserve were mined (Huffman pers. obs.). 

    Residential and Industrial Development
    Because of rolling topography, fairly high promontories, and a relatively large number of wetlands, oak barrens landscapes make attractive home sites. Because of low agricultural value, barrens landscapes are also common locations for industrial development. Roads, septic fields, and water supply systems are easily built. Lakeplain systems are under tremendous pressure because all are located near large urban centers along the Great Lakes. Outwash systems are less threatened. 


    Preserve Size
    Evidence presented here suggests that plant species of oak barrens can persist through cycles of canopy closure and canopy removal. High species numbers can be attained with small areas. Consequently a preserve for the oak barrens plant community need only be large enough to burn safely and absorb legitimate visitors as well as less benign use by preserve neighbors. Most high quality oak barrens remnants are smaller than 100 acres, but with proper management can reasonably be expected to protect the plant community for a hundred years or more. 

    Animal populations may not persist, however. As discussed above, characteristic butterflies of oak barrens have declined over most of their ranges. Schweitzer and Rawinski (1987) note that small community occurrences almost always have degraded Lepidoptera fauna in northeastern pine-oak barrens. They also note that even sites larger that 1000 acres have lost species. The decline of Lepidoptera in oak barrens is related to: 1) loss of larval food plants because of increased shading resulting from fire suppression (cf. Shuey et al. 1987) 2) barriers to migration (e.g., highways, fences, parking lots) (Panzer 1988). Persistence of butterflies over the last 100 years is linked to larger areas where mixture of open and closed canopy is found. Panzer (1988) makes3 general recommendations for maintaining populations of restricted insect species 1) eliminate unnatural landscape features, 2) maximize habitat diversity, noting that because of dichotomous life histories insects often require habitat mosaics, 3) burn with restraint. 

    An assessment of area needed for preserves can be done using minimum viable population estimates. Assuming a minimum viable population 200 interbreeding adults, and with or without an additional 600 other individuals, the density per acre of animals can be used to calculate a minimum acreage needed for population persistence. Although these are gross estimates they give some indication of preserve sizes needed to maintain some types of faunal diversity: 

    Eastern Chipmunk  35  130 
    13-Lined Ground Squirrel  40  160 
    Ornate Box Turtle  100  400 
    Red Backed Vole  200  800 
    Badger  245,700  828,000 

    The essence of an oak barrens is its variable physiognomy in time and space. As has been shown, before settlement this dynamic resulted from the complex interplay of fire frequency and intensity, fire source areas and fire barriers, woven into the landscape mosaic and played out over vast expanses. That pattern of diversity no longer exists, and oak barrens preserves are less secure as a result. Mimicking the landscape diversity pattern is possible. It may require areas of 1000 to 2000 acres for management units large enough to achieve a physiognomy and composition suitable for some targeted animals. 

    Increasing Preserve Size using Degraded Lands
    Larger preserves increase diversity by accommodating animal species and increasing the number and size of different patches of natural communities. In a fire-prone community, larger preserve size buffers against the negative effects of catastrophic fire. Movement of species from deteriorating to high quality habitat can take place within a single preserve. Finally, selection pressures on species originate within the preserve, rather than in response to forces emanating from the human-dominated landscape around it. 

    Restoration of barrens landscapes has greater potential for rapid success than for most other Midwest ecosystems. The recovery potential of degraded barrens is good. Moreover, large blocks of degraded barrens often abut high quality oak barrens, or lie next to abandoned fields and other disturbed areas where oak barrens groundlayer species persist. 

    Building on a high quality core, preserve design should strive for integration of large blocks of closed oak barrens with refugia of barrens groundlayer species. It should also strive for restoration around a high quality core of large blocks of degraded barrens showing good recovery potential (diagnostic or constant species of barrens are present, albeit rare and struggling). 

    Many barrens landscapes are artificially drained. Beaver activity has been eliminated for well over 100 years. Since the high seasonal water table of lakeplains and outwash plains was a significant factor on the presettlement landscape, preserve design should consider the feasibility of blocking or filling ditches andbreaking tile fields. Local engineering authorities may oppose changes in drainage if they affect homesites, wells, pastures or commercial installations. Mosquito control agencies target ponded water. Increased demand for mosquito control could harm invertebrate members of the barrens community. Despite these obstacles, restoration of former water levels is a good preparation for global climate change, and enough benefits derive from improved groundwater condition that a good case can be made for hydrological restoration. 

    Fire Considerations
    Preserve design should consider the feasibility of burning large portions of a preserve at one time. Design must protect areas sensitive to smoke, such as busy roads, airports, and villages. Consult local weather stations to assess prevailing winds during burn seasons. Smoke buffers of one-fourth to one-half mile are considered sufficient by some managers. 


    Results of management treatments on degraded oak barrens in the Midwest indicate good potential exists to restore natural community composition and structure. Little is known about the recovery potential of faunal populations. 

    Recovery with Fire
    Twenty-five years of low intensity prescribed burning at CCNHA in central Minnesota removed woody plants from the understory and increased groundlayer species richness that included grasses and forbs characteristic of prairie and savanna communities (White 1983, 1986, Tester 1989). However this burning regime had little effect on large trees (>30 cm dbh). At nearby Allison Savanna, fire management using a variety of fire return intervals demonstrated that low frequency burns in spring every 1 to 3 years were successful in reducing canopy cover of the oaks, albeit slowly (Faber-Langendoen and Davis 1995). If the fire return interval was 4 or more years then canopy continued to increase. Ultimately, drought, or a combination of drought and fire, may be more important in effecting the longevity of the large isolated trees (Faber-Langendoen and Tester 1993). 

    Results from Kitty Todd Preserve, northern Ohio (Huffman unpublished data), central Wisconsin (Blewett 1978, Holtz and Howell 1982), and northern Wisconsin (Vogl 1961) indicate rapid change in herb layer composition toward barrens species with fire and/or cutting treatments. 

    A catastrophic fire in 1971 turned oak woods into dense oak-aspen scrub at Kitty Todd Preserve. In 1987, 96% of the area was covered by oaks, aspen, and 15-foot cherries. Greenbriar dominated the understory. After a fall 1987 burn, 45 barrens understory species were recorded, and canopy cover was reduced to 41% (Huffman unpubl. data). 

    In degraded barrens these barrens species may originate from 1) perennial plants present as rhizomes in the soil or as understory plants of low vigor, 2) viable seeds dormant in the seedbank until canopy and litter cover are removed, and 3) a seed source located nearby in stands of high quality barrens vegetation. 

    Recovery After Soil Disturbance
    Today some abandoned farm fields are refuges for heliophytic barrens understory plants which otherwise have disappeared from unplowed, but overgrown barrens in the area. Significant populations of barrens species can be found in abandoned farm fields in time periods ranging from less than 5 to 20 years (Huffman and Chapman pers. obs.). However, both native and non-native weedy species, such as Polytrichum commune, Daucus carota, Melilotus alba, M. officinalis, Oenothera biennis, Solanum carolinense, Rumex acetosella, Poa pratensis, Poa compressa, and Agropyron repens are likely to invade and/or dominate a site. At TNC's Kitty Todd Preserve, abandoned farm fields are home for over 50 Ohio rare plant species, many of them oak barrens groundlayer species, and three state-endangered Lepidoptera species. 

    Disturbance of the soil which removes litter and underbrush can restore groundlayer species in oak barrens. An aspen thicket, bulldozed in 1985, was later purchased by TNC and found to harbor the highestconcentration of rare species in the Kitty Todd Preserve (Huffman 1988). A fire-break bulldozed through another aspen thicket next to an abandoned field at Kitty Todd produced populations of Lupinus perennis, Tephrosia virginiana, Comandra umbellata, Asclepias tuberosa and Helianthemum sp. Populations of the rare butterflies Erynnis persius and Incisalia iris colonized lupine in this fire-break the next year (Shuey et al. 1987). 

    A strip that removed two to four inches of soil was bulldozed through thick underbrush in an oak barrens on the Saginaw lakeplain northeast of Bay City, MI in 1978. In 1982, Asclepias tuberosa, Comandra umbellata, Salix humilis, Helianthus spp., and Aster spp. were observed growing in the bulldozed strip beneath the closed oak canopy, but not elsewhere (Chapman pers. obs.). Whether a comparable effect could be induced by fire on these sites is not known, nor is it clear to what degree only certain barrens species may benefit from bulldozing. 

    Sand pits at Kitty Todd and Oak Openings Metropark provide habitat for species associated with shallow ponds with fluctuating water levels, a minor but significant community of oak barrens landscapes (Huffman pers. obs.). Plant species include those disjunct from the Atlantic Coastal Plain. Sand pits expose the water table and provide suitable habitat for species which have disappeared from their former habitat because regional water tables have been lowered. 

    Powerline right-of-ways, roadways, firebreaks, and even bulldozed areas are refuges for shade-intolerant barrens plants (Huffman and Chapman pers. obs.). 


    Biological monitoring is needed to assess effectiveness of management treatments and progress toward stated goals. Goals are developed by comparing present vegetation composition and structure with a "model" of the natural community. Divergence from the model directs management. For example, a natural community model may indicate that a site should have an open canopy structure with low woody cover and density, but the site currently has 80% canopy cover and 40% subcanopy cover. The intent of management is to reduce canopy cover less than 30% and subcanopy to 10% cover. Monitoring would focus on measuring movement toward these goals. 

    Management Objectives
    Goals suggested here are only approximate, as cover and composition vary with local site conditions. 

    Objective 1: Create and maintain very open (savanna) and open understory (woodland) conditions on dry or fire prone sites. For monitoring management accomplishments in oak barrens use the following: 


    a) canopy cover  5-60% 
    b) woody understory cover  0-20% 
    c) groundlayer cover  high 
    d) native groundlayer diversity  high 
    • Method 1, parameters a and b: Accurate estimates of woody cover can be made using aerial photographs. Do baseline measures from pre-treatment aerial photos. Post-treatment monitoring should be done in the growing season following the treatment year. Photos from the Agricultural Stabilization and Conservation Services (ASCS) can be used for baseline cover estimates, however since these are taken at infrequent intervals their use may be limited. Transparencies of the Soil Conservation Service (SCS) yearly photos may be suitable. A variety of measurement methods are possible such as the random dot planimeter, and digitizing. For information on using aerial imagery consult Brewer and McCann (1982).

    • Method 2, parameters a and b: Establish at minimum a 20x25 m macroplot to measure diameter at breast height (dbh) for stems > 5 cm dbh. Smaller stems are counted and placed in sapling class. Macroplots could be 20x50 m or 80x50 m, depending on overstory stem density, in which case the 20x25m plots could become subplots. Smaller stems could be counted in randomly or systematically selected subplots. Size class distributions can be examined for changes in density and basal area.

    • Method 3, parameters a and b: Stand basal area can rapidly estimated using forest mensuration techniques such as the Bitterlich angle gage. See Mueller-Dombois and Ellenburg (1974) for a discussion of these methods. 

    • Method 4, parameters a, b, c, and d: Establish permanent 20x25 m macroplots (or larger - see Method 2) on representative sites. Estimate woody cover by stratum class. Woody cover could also be monitored using permanent photo points. 

    For the herb layer estimate cover of constant and diagnostic species in a 5x20 m subplot in the center of the 20x25m plot, or use 15-25 0.5x1.0 m (or 0.25x0.5 m) microplots systematically located within either a single 20x25 macroplot or, if a larger plot size was used, spread over the entire plot. 

    • Method 5, parameter b: Percent cover of shrubs and saplings can be estimated using two line intercepts crossing the macroplot from east to west and north to south. 

    • Method 6, parameters c and d: Line transects with systematic or randomly located microplots are often used to monitor frequency and cover of herb layer species. A stratified random plot placement can be used in which plots are randomized within mapped vegetation patches. Estimate cover of constant and diagnostic species in the herb layer.

    Objective 2: Create and or maintain scrub oak barrens on sites prone to low frequency, high intensity fires, with low canopy cover (5 to 30%), high cover of oak species in tall shrub/sapling and low shrub layers (30%), and high graminoid cover (75%). 

    a) canopy cover  5-30% 
    b) woody understory cover  3-50% 
    c) graminoid cover  >75% 
    • Method 1, parameters a and b: see method 1 above.
    • Method 2, parameters a and b: see method 2 above.
    • Method 3, parameters a and b: see method 3 above.
    • Method 4, parameters a, b and c (see also method 4 above): establish permanent 20x25 m macroplots on representative sites. Estimate woody cover in each stratum class. Cover of undesirable woody species should also be estimated (eg. Prunus serotina, Acer rubrum, Populus tremuloides). Permanent photo points within the macroplot could also be used to monitor changes in woody cover. 
      • For graminoid cover: use 5x20 m subplot, or use 15-30 0.5x1.0 m (or 0.25x0.5 m) quadrats systemically located, estimate total graminoid and non-graminoid cover in the herb layer, estimate cover of dominant graminoids, and character and constant species that are graminoids.
    • Method 5, parameter b: see method 5 above.
    • Method 6, parameter c: line transects with systematic or randomly located microplots are often used to monitor frequency and cover of herb layer species. We recommend stratified random plot placement in which plots are randomized within mapped vegetation patches. Estimate cover of grasses and non-grasses, and of constant and character graminoid species.

    General recommendations: 

    • 1) plots should be resampled during the growing season following management treatments. Full species composition should be resampled at five year intervals.
    • 2) baseline data on cover or frequency for all plant species should be taken prior to management treatments. 

    Biological Monitoring Programs
    Kitty Todd Preserve, TNC Land Steward OH: Rare plant and animal species monitoring, Oak Openings Metropark, Toledo OH: Macroplots used to monitor effects of prescribed fire on canopy cover and ground layer composition. Indiana Dunes National Lakeshore, Porter IN: Effects of prescribed fire on canopy cover, tree and shrub density, ground layer composition. Individual trees are tagged and monitored. Helen Allison Savanna and CCNHA, MN: Macroplots used to monitor canopy cover and herb layer response to prescribed fire. Tagged or mapped trees in large macroplots are being used to monitor recruitment, growth and mortality, and systematically placed quadrats are also being used to monitor the ground and canopy layers. University of WI, Steven's Point (Alan Haney): Macroplots are being used to monitor effects of prescribed burning in a variety of sites across Wisconsin and Illinois. Allegan State Game Preserve, Michigan: Burn management is being monitored. 


    1) What are fire effects on rare faunal populations? A number of rare lepidoptera have host plants occurring on barrens. We need more information on the biology, ecology and the effects of standard management practices on these species. 

    2) What are the effects of different burn intervals, intensities and seasons on vegetation and rare fauna? Fire history of oak barrens is essential. 

    3. a) What is the nature of seasonal water changes? What is the role of seasonal high water tables on the vegetation mosaic? How does the water table fluctuate seasonally? Models of hydrologic systems are needed for individual preserves. 

    b) What is the nature of annual water changes? What is the role of periodic droughts? 

    4) What roles do seedbanks, vegetative reproduction, and external seed sources play in restoration of degraded oak barrens? 

    5) How rapidly do barrens species recolonize abandoned agricultural land? How does the kind and degree of disturbance influence recolonization? 

    6) Information is needed on minimum viable population size, and area/habitat requirements for characteristic barrens animal species. 


    Oak barrens should be managed using natural processes that shaped them--fire and water. In managing barrens diversity, the location, number and size of each kind of barrens physiognomy and composition should be considered. Wetland patches in the barrens landscape should also be brought into the equation as integral "filters" of the regional fire pattern. Barrens management should restore good numbers of large patches which represent different barrens physiognomies. This approach provides habitat for shade-tolerant and light-demanding plant species, and large blocks of different kinds of uniform habitat for animals. 

    Application of different intensity fires, in fall and spring, under cool and hot fire conditions, creates the necessary diverse physiognomies. Infrequent, canopy-replacing fires in dense woody cover may produce scrub; annual running ground fires in woodland with heavier graminoid groundlayer may produce open, scattered tree canopy. Intermediate conditions arise from intermediate prescriptions. 

    Spring burns carry irregularly through areas of the barrens landscape influenced by high spring water table. Patchy fire leaves refugia for invertebrates, other species, and may permit oak seedling establishment. Late spring and summer burns probably do greater harm to woody plants than fall or early spring burns. Many wetlands, moist sand prairie, pin oak flats, and other low areas may be too wet for spring burning. Schwegmanand McClain (1985) observed that fall fires burn wet prairies and marshes which are often too moist to burn in the spring. Analysis of weather conditions at the Kitty Todd Preserve indicate that fire weather occurs most frequently in the fall (Huffman pers. comm.). Fall may have been the typical burn season in northwest Ohio (Aldrich 1888). An historical account reported an Indian-set fire in a snowless December and another author indicated Indians set fires in fall in southern Michigan (Chapman 1984). Schwegman and McClain (1985) state that presettlement fires in Illinois occurred frequently in the fall. However, in southern Wisconsin, where cool/wet falls are frequent, this may not be the case (R. Henderson 1989, pers. comm.). 

    Because data on the response of barrens to different fire treatments is scanty, all new information is useful and should be recorded and reported. Managers should take detailed records on each burn, such as weather conditions before and after the burning, fuel characteristics, and fire behavior parameters. Fire observations must be coupled with information on vegetation response. If possible, a monitoring station should be established within the treatment area before the first treatment and a second station could be established in an areas not to be treated. 

    To produce barrens structure without a brushy understory, successive low-intensity burns are necessary. Use annual burns during the restoration phase of management until the height and density of woody sprouts meets a stated goal. For more rapid results on badly degraded sites, cutting may be required to augment annual burning. Sprouts allowed to grow a full season are able to develop larger and larger root systems each year, aggravating a woody plant problem. Once a structural goal is met, fire return interval can be relaxed to once every 1-3 years. 

    To produce scrub barrens with a graminoid understory, high-intensity fires at long intervals are necessary. A single fire under extreme conditions set at Jasper-Pulaski State Fish and Wildlife Area (IN) removed virtually all oaks from an almost closed canopy, producing a mosaic of dense oak scrub and graminoid patches (White & Chapman pers. obs.). Intervals between fires may be as great as 15 years. 

    Fire breaks should be clear of dead fuels. (Leaf blowers have been used successfully for leaves and mowed fine fuels.) In barrens, snags within falling distance of fire breaks should be knocked or sawed down to ensure burn crew safety. Barrens fires can require long hours of mop-up; dead limbs high in old trees may ignite unexpectedly and throw sparks into non-burn units. Old snags can smolder throughout the growing season (A. Steuter pers. comm.) If fire management is successful, live woody cover is converted to downed heavy fuels, which can accumulate and require intensive fire break preparation and mop-up. 

    Burn crews should be properly equipped with hardhats, Nomex suits, and radios in burn units large enough to lose sight of crew members. Burns must be carefully planned and executed according to all applicable guidelines established by the managing agency. 

    Mowing can be used in areas where fire is undesirable, e.g. small areas set aside as butterfly refugia. Mowing dramatically changes stand structure, however, by eliminating shrubs and setting back oak regeneration. It is possible to steer a tractor around oak saplings and some shrubs, but the result is artificial. Nevertheless, grassland patches can be a desirable addition to a barrens landscape, especially if they sustain butterfly populations. Mowing should be done in late fall or winter to minimize impacts on herbaceous species. Properly timed mowing controls some weedy understory species, (e.g. Melilotus officinalis). 

    Girdling can control clonal species that resprout vigorously, such as aspen (see TNC 1987 for an extensive review of treatment methods for Populus spp.). Girdling also can be effective on cherry, maple, and possibly sassafras. Girdling is done with a variety of tools, including a hatchet and modified leaf spring, or "leaf spring spud" (R. Johnson pers. comm.). Girdling is easiest and most effective in spring when sap is running and cambial cells balloon with fluids. Before or after spring, it is more difficult. Some managers combine girdling with various herbicide applications. One technique is to girdle the tree the first year and herbicides the second (Ballard pers.comm.). 

    Selective Cutting
    Although selective cutting is labor-intensive, it is the commonest method used to restore oak barrens. Selective cutting runs the gamut from occasional hand-clipping in butterfly refugia, to removal of large maples and aspen using chainsaws and chippers. Disturbance to the ground and disposing of cut debris are drawbacks of this method. Brush has been scattered in grassy areas where fire or decay eventually consumes it; piled and later removed; chipped and hauled at greater expense; and piled and burned, which sterilizes soil and provides an ashy substrate for weed colonization (Packard, Croy, Huffman, Ladd pers. comm.). 

    Grazing was a significant force keeping barrens open from the time of settlement to the mid-1900's, as evidenced by aerial photographs and land use history. Cattle relish aspen sprouts, and pasturing could control aspen where it is a problem (Heitlinger pers. comm.). Cole and Pavlovic (1987) postulate that browsing by elk and bison controlled woody growth in a Lake Michigan barrens landscape. Elk and bison grazing reduced woody growth in the national parks and forests of the western United States; however, fencing and shepherding costs are high, and grazing effects on barrens groundlayer species are not well documented. Evidence suggests that barrens are relatively resistant to exotic species invasion associated with grazing. In any case, this technique may prove useful in the future. Agricultural science and range management journals and experts should be consulted and a grazing plan put together before implementing this technique. 


    Figure 1. Distribution map for the two major types of open oak barrens in the Upper Midwest

    Figure 1. Distribution map for the two major types of open oak barrens in the Upper Midwest: Bur Oak -Northern Pin Oak Barrens (western type - primarily MN) and Black Oak Barrens (eastern type). Within the Black Oak Barrens type, two areas on glacial lake plain are outlined separately. Location of high-quality sites that are nominated (dots) and dedicated (circles) National Natural Landmarks are shown: 1) Agassiz Dunes, 2) Cedar Creek-Helen Allison, 3) Chippewa River sites, 4) Ft. McCoy, 5) Illinois Beach, 6) Kankakee Sand Plains site, 7) Indiana Dunes, 8) Allegan, 9) Shakey Lakes. Other site (triangle): 10) Toledo Metroparks-Kitty Todd.

    Figure 2. Position of Plant Communities in Barrens Landscapes in Outwash Plains (top figure) and Glacial Lake Plains (bottom figure).

    Figure 2. Position of plant communities in barrens landscapes in outwash plains (top figure) and glacial lake plains (bottom figure). a) dune and ridge tops and slopes, b) lower dune ridges, lower slopes and dry flats, c) moist dune valleys and sand flats just above maximum flood level, d) channels and depressions where sites are flooded in spring.


    Abrams, M.D. 1992. Fire and the development of oak forests. Bioscience 42: 346-353. 

    Albert, D. A., et al. 1986. Regional landscape ecosystems of Michigan. Sch. Nat. Res., Univ. Mich., Ann Arbor. 32pp + map. 

    Aldrich, L.C. 1888. The History of Henry and Fulton Counties, Ohio. D. Mason and Company, Syracuse, NY, 713 pp. 

    Anderson, R.C. 1983. The eastern prairie forest transition an overview. Pp. 8692, in R. Brewer (ed). Proc. Eighth N. Amer. Prairie Conf. Western MI University, Kalamazoo, MI. 

    Anderson, R.C. and J. Schwegman. 1982. The response of southern Illinois barren vegetation to prescribed burning. Trans. Ill. Acad. Sci. 64: 287-291. 

    Anderson R.C., and L.E. Brown. 1983. Comparative effects of fire on trees in a Midwestern Savannah and adjacent forest. Bull. Torrey Bot. Club 1:8790. 

    Anderson, R.C., and L.E. Brown. 1986. Stability and instability in plant communities following fire. Amer. J. Bot. 73: 364368. 

    Andow, D.A., R.J. Baker, and C.P. Lane. 1994. Karner blue butterfly: a symbol of a vanishing landscape. Miscellaneous Publication Series 84-1994. Minnesota Agricultural Experiment Station, St. Paul, MN. 222 pp. 

    Archambault, L. 1987. Classification of oak ecosystems in southeastern Michigan. Ph. D. Dissertation, University of Michigan, Ann Arbor. 159+ pp. 

    Bacone, J.A., et al. 1979. Pre-settlement vegetation of the Indiana Dunes National Lakeshore. Unpublished Report submitted to the Science Division, Indiana Dunes National Lakeshore. 

    Bakowsky, W. 1988. The phytosociology of Midwestern Oak Savanna in the Caroline an Region of Southern Ontario. M.Sc. Thesis, Dept. of Botany. Univeristy of Toronto. 

    Blewett, T. 1978. Prairie and savanna restoration in the Necedah National Wildlife Refuge. Pp. 154-157 in D.C. Glenn-Lewin and R.Q. Landers, Jr. (eds.). Fifth Midwest Prairie Conference Proceedings, Aug. 22-24, 1976. Iowa State University, Ames, Iowa. 

    Bowles, M.L., M.D. Hutchison, and J.L. McBride. 1994. Landscape pattern and structure of oak savanna, woodland, and barrens in northeastern Illinois at the time of European settlement. Pg. 65-74 in Fralish, J. et al. (eds.). Proceedings of the North American Conference on Barrens and Savannas. Oct. 15-16, 1994. Illinois State University, Normal, IL. 

    Bray, J.R. 1958. The distribution of savanna species in relation to light intensity. Can J. Bot. (36): 671681. 

    Bray, J.R. 1960. The composition of savanna vegetation in Wisconsin. Ecology 41:721732. 

    Brewer, L. 1989. A survey of the oak barrens and savannas of the Oak Openings Preserve Metropark. Report submitted to The Nature Conservancy, Ohio Chapter, Columbus, Ohio. 

    Brewer, R. and M.T. McCann. 1982. Laboratory and field manual of ecology. Western Michigan University. Saunders College Publishing, New York, NY. 

    Chapman K.A. 1984. An ecological investigation of native grassland in southern Michigan. Unpublished Masters Thesis, submitted to the graduate college of Western Michigan University, Kalamazoo, Michigan. 

    Cole, K.L. and N.B. Pavlovic 1987. Howes Prairie, A remnant of the Indiana Dune Prairie preserved by periodic flooding. Unpublished manuscript, Science Division, Indiana Dunes National Lakeshore. 

    Cottam, G. 1949. The phytosociology of an oak woods in southwestern Wisconsin. Ecology 30:271287. 

    Curtis, J.T. 1959. The vegetation of Wisconsin: an ordination of plant communities. University of Wisconsin Press, Madison, WI. 658 pp. 

    Davis, C.A. 1908. The native vegetation of Tuscola county. Tenth Ann. Rpt. Geol. Survey Mich., p.290336. 

    Davis, M.A. et al. 1995. Multiple disturbances and the dynamics of pattern and persistence in a sand plain oak woodland (manuscript submitted) 

    Dyksterhuis, E.J. 1957. The savannah concept and its use. Ecology 38: 435-442. 

    Faber-Langendoen, D. and J.R. Tester. 1993. Oak mortality in sand savannas following drought in east-central Minnesota. Bull. Torrey Bot. Club 120: 248-256. 

    Faber-Langendoen, D. and M. Davis. 1995. Effects of fire frequency on tree canopy cover at Allison Savanna, east-central Minnesota. Natural Areas Journal (in press) 

    Gilbert, G.K. 1873. Geology of Fulton County. Ohio Geol. Surv. Report. 1:567572. 

    Gordon, R. B. 1969. The natural vegetation of Ohio in pioneer days. Bull. of the Ohio Biological Survey, news series. vol. III, No.2. 

    Grigal, D.F. et al. 1974. Soils of the Cedar Creek Natural History Area. Miscellaneous Report 123-1974. Agr. Exp. Stn. University of Minnesota, St. Paul, MN. 

    Grimm, E.C. 1981. an ecological and paleoecological study of the Big Woods region of Minnesota. Phd thesis, University of Minnesota, Minneapolis, Minnesota. 

    Grimm, E. C. 1984. Fire and other factors controlling the big woods vegetation of Minnesota in the midnineteenth century. Ecol. Monogr. 54(3): 291311. 

    Hanson, J.D. 1976. A phytosociological study of the oak savanna in southeastern North Dakota. MS Thesis. NDSU, Fargo. 135pp. 

    Henderson, R. 1982. Vegetation - fire ecology of tallgrass prairie. Natural Areas Journal 2: 17-26. 

    Henderson, N. R. and J. N. Long. 1984. A comparison of stand structure and fire history in two black oak woodlands in northwestern Indiana. Bot. Gaz. 145(2): 222228. 

    Holtz, S. L., and E. A. Howell. 1982. Restoration of grassland in a degraded woods using the management techniques of cutting and burning. In: R. Brewer, (ed.). Proc. of the Eigth North American Prairie Conf.: 1982, Aug. 14: Kalamazoo, MI. 124129. 

    Homoya, M. A. et al. 1985. The natural regions of Indiana. Proc. Indiana Acad. Sci. 94:245268. 

    Hubbard, B. 1838. Report of the state geologist. In. G.N. Fuller (ed.). Mich. State Geol. Ann. Rpts. 1-7, 1837-1844, (1928). House Doc. No. 24, Michigan Legislature, Lansing, MI. 

    Huffman, M. R. 1988. Management plan for the Kitty Todd Preserve Lucas County, Ohio. Unpublished, The Nature Conservancy. 1504 W. 1st Ave. Columbus, Ohio, 43212. 

    Inouye, R. S. et. al. 1987. Old field succession on a Minnesota sand plain. Ecology 68(1): 1226. 

    Keddy, P. and A.A. Reznicek. 1982. The role of seed banks in the persistence of Ontario's coastal plain flora. American Journal of Botany 69: 13-22. 

    Kline, V.M. and G. Cottam. 1979. Vegetation response to climate and fire in the driftless area of Wisconsin. Ecology 60: 861-868. 

    McLaughlin, W.T. 1932. Atlantic coastal plain plants in the sand barrens of northwestern Wisconsin. Ecological Monographs 32: 335-383. 

    Moran, R.C. 1978. Presettlement vegetation of Lake County, Illinois. Pp 12-18, in D.C. Glenn-Lewin and R.Q. Landers, Jr. (eds.). Fifth Midwest Prairie Conference Proceedings, Aug. 22-24, 1976. Iowa State University, Ames, Iowa. 

    Mueller-Dombois, D. and H. Ellenberg. 1974. Aims and methods of vegetation ecology. John Wiley and Sons, New York. 

    Naiman, R.J. et al. 1988. Alteration of North American streams by beaver. Bioscience 38: 753-762. 

    Nuzzo, V. A. 1986. Extent and status of midwest oak savanna: pre-settlement and 1985. Natural Areas Journal 6(2): 636 

    Panzer, R. 1987. The grasshoppers, katydids, and walking sticks (Orthoptera: Acrididae, Tettigoniidae, Phasmatidae) of the prairies and savannas of Illinois. Trans. Ill. Acad. Sci. Supplement 80:63. 

    Panzer, R. 1988. Managing prairie remnants for insect conservation. Natural Areas J. 8(2):8390. 

    Panzer, R., and R. Gnaedinger. 1986. The butterflies and skippers of Hoosier Priarie Nature Preserve, Lake County, Indiana. Report submitted to the Indiana Division of Nature Preserves, Indianapolis, Indiana. 

    Panzer, R. and D. Stillwaugh. 1987. A survey of the insects of Tefft Savanna Nature Preserve, with special emphasis on butterflies, grasshoppers, and katydids. Indiana Division of Nature Preserves, Indianapolis, IN. 35p. 

    Pavlovic, N.B., et al. 1984. Status report concerning the study of the effects of fire and flooding on the Howes Prairie ecosystem. Indiana Dunes National Lakeshore. 

    Peattie, D. C. 1922. The Atlantic coastal plain element in the flora of the Great Lakes. Rhodora. 24:5770. 

    Schwegman, J.E. and W.E. McClain. 1985. Vegetative effects and management implications of a fall prescribed burn on an Illinois hill prairie. Natural Areas Journal 5: 4-8. 

    Schweitzer, D. and T. Rawinski. 1987. Northeastern Pitch Pine/Scrub Oak Barrens Element Stewardship Abstract. The Nature Conservancy, Eastern Regional Office, Boston, MA. 

    Shuey, J.A., et al. 1987. Butterflies that are endangered threatened, and of special concern in Ohio. Ohio J. Sci. 87(4): 98106. 

    Stone, K.L. et al. 1980. Soil Survey of Lucas County, Ohio. USDA-SCS. Ohio DNR-Division of Lands and Soils and Ohio Agric. Res. Devel. Ct. U.S. Gov. Printing Office. 139 pp. + maps. 

    Stout, A.B. 1944. The Bur oak openings in southern Wisconsin. Trans. Wisconsin Acad. Sci. 36:141161. 

    Tester, J.R. 1989. Effects of fire frequency on oak savanna in eastcentral Minnesota. Bull. Torrey Bot. Club 116: 134-144. 

    The Nature Conservancy. 1987. Populus spp. Element Stewardship Abstract. Midwest Regional Office, Minneapolis, MN. 

    Thor, E. and G.M. Nichols. 1973. Some effects of fire on litter, soil and hardwood regeneration. Proceedings of the Annual Tall Timbers Fire Ecology Conference, No. 13. March 22-23, 1973, Tallahasse, Florida. 

    Tilman, D. 1983. Plant succession and gopher disturbance along an experimental gradient. Oecologia 60:285292. 

    Tilman, D. 1984. Plant dominance along an experimental nutrient gradient. Ecology 65:14451453. 

    Tilman, D. 1987. Secondary succession and the pattern of plant dominance along an experimental nitrogen gradient. Ecological Monographs, 57:198214. 

    Vogl, R. J. 1961. The effects of fire on some upland vegetation types. Ph.D. Thesis. University of Wisconsin. 103 pp. 

    Wedin, D. and Tilman. 1992 Nitrogen cycling, plant competition, and the stability of tallgrass prairie. Pp. 5-8 in Daryl D. Smith and C.A. Jacobs (eds). Proceedings of the Twelfth North American Prairie Conference, August 5-9, 1990. University of Northern Iowa, Cedar Falls, IA. 

    White, A.S. 1983. The effects of thirteen years of annual prescribed burning on a quercus ellipsoidalis community in Minnesota. Ecology 64:10811085. 

    White, A.S. 1986. Prescribed burning for oak savanna restoration in central Minnesota. USDAForest Service Res. Pap. NC266; North Central Forest Experiment Station, St. Paul, MN. 12 p. 

    White, M.A.and K.A. Chapman. 1989. Vegetation analysis of the Toledo Metroparks oak openings for community characterization and monitoring. Unpublished report, The Nature Conservancy, Midwest Regional Office. 

    White, J. and M. Madany. 1978. Classification of natural communities in Illinois. In Natural Areas Inventory technical report: Vol. I, survey methods and results, p.311405. Ill. Nat. Areas Invent., Urbana, IL. 

    Whitford, P.B. and K. Whitford. 1971. Savanna in central Wisconsin, U.S.A. Vegetation 23:7779.


    Begin Site Footer

    EPA Home | Privacy and Security Notice | Contact Us