Jump to main content.



Editors: Mark K. Leach, University of Wisconsin, Madison Laurel Ross, The Nature Conservancy

Contributors:  Don Faber-Langendoen, The Nature Conservancy Richard Henderson, Wisconsin DNR Ken McCarty, Missouri DNR, Division of State Parks Steve Packard, The Nature Conservancy Brian Pruka, University of Wisconsin, Madison

Table of Contents

Introduction: A Call to Action

Figure 1. The presettlement distribution of Midwestern oak savannas and woodlands (based on Nuzzo 1994). The delineation of boundaries is somewhat arbitrary and within the zone, boundaries between oak ecosystems and other ecosystems may have been either gradual or abrupt depending upon local environmental conditions related to topography, soils, ignition sources, and other factors (Anderson 1983). Many "outliers" are too small to indicate on a map of this coarse resolution. 

This Midwest Oak Ecosystems Recovery Plan calls for establishing a system of reserves that support a full array of species, communities, and processes by restoring and managing working examples of now rare, but once wide-spread (Figure 1), ecosystems. Successful recovery depends both on good science and common sense, including the recognition that people are integral parts of Midwestern ecosystems. Although ecosystem recovery efforts are inherently local, they will often involve several levels of government, the private sector, non-profit organizations, and private citizens. These efforts benefit from cooperation, decentralized planning, and citizen action. 

Oak ecosystem recovery and management are not new. There are excellent examples to learn from and to build upon. Local and regional recovery plans are in various stages of development throughout the Midwest. These range in size and scope from plans for small areas in parks and arboreta, to state wide or multi-state plans. Some plans have a narrow purpose, such as compliance with the Endangered Species Act (e.g., Bleser & Leach 1994), others aim for research, conservation, or aesthetics. The Midwest Oak Ecosystems Recovery Plan hopes to serve these other plans as a conceptual umbrella by providing a regional perspective. 

Challenges and Opportunities

Less than two centuries ago oak ecosystems were an integral part of the Midwestern landscape. Since that time the complex biological webs of animals and plants have been fragmented or replaced by agricultural fields, highways, and cities. In a 1985 survey, Nuzzo (1986, 1994) located relatively high-quality oak savanna that totaled approximately 0.02 percent of the presettlement extent of the community. Perhaps millions of acres of degraded but potentially recoverable oak ecosystems remain, although these have lost both biological diversity and functions. Regionally, once common species have become extirpated or extinct. In short, the once healthy ecosystems that covered this land no longer support their original diversity of life. 

People are increasingly aware of the appalling rate of global species extinctions and the loss of natural vegetation. However, it is important for us to understand however that the loss of biological diversity (Wilson 1992) does not occur only in exotic places such as tropical rain forests. In the American Midwest as well, we witness extinctions, the loss of species and vegetational diversity, and the breakdown of life-bearing ecosystems that such losses indicate. 

Such were the concerns in 1993 that brought to Chicago over 100 scientists, natural resource managers, restorationists, and others to freely discuss ideas about how to best address the challenge of restoring Midwestern oak ecosystems (Botts et al.1994). There were certainly disagreements among the participants on the details, but a general consensus emerged that it is a priority to recover sizeable areas of oak savanna and woodland ecosystems. 

We face an urgent need for action to restore and preserve functioning ecosystems but our understanding of how these systems function is incomplete. One approach to this dilemma is to actively work to restore and manage remnant communities now, with flexible, continually-reevaluated plans that incorporate the uncertainties and react to breakthroughs in the knowledge base on oak ecosystems. (Shively & Temple 1994). As Reed Noss (1994) urges, it is time for conservation biologists to stick their necks out and state openly what they do know, offering guidance to society for conserving biodiversity in the face of uncertainty. (Noss 1994). 

Pioneering efforts in ecological restoration throughout the region are invaluable guides. Every state in our region has undertaken oak savanna recovery projects of various kinds and sizes. Through this work researchers and practitioners are developing more a sophisticated--and more realistic--understanding of the complexities of these communities. Very important advances have been made in the study of vegetation, mammals, birds, herptiles, invertebrates, and the role of disturbance and stress. Knowledge is accumulating so rapidly that researchers and practitioners hold annual conferences to discuss and disseminate new findings. Major teaching institutions now offer training in the techniques of ecological restoration including oak ecosystems. As new projects emerge thousands of people, professionals and volunteers alike, are gaining first-hand experience at the nuts-and-bolts of recovery efforts. 

Today in the Midwest, millions of acres of degraded oak-associated vegetation remnants persist. Government agencies, non-profit organizations, research institutions, private enterprises, and individuals are developing the policies and techniques necessary to restore and maintain these rare biological systems. For these efforts to succeed, we must establish attainable goals and approaches as well as assemble substantial financial and human resources. 

This document covers many kinds of oak ecosystems over a large geographic area, therefore it does not provide fine-scale, site-specific information. The discussion focuses primarily on oak communities, but much of what is written here can also be applied to closely related ecosystems such as prairies and pine-oak barrens. It does offer a broader perspective that might be helpful to local efforts by 1) offering regional recovery goals and possible actions toward reaching those goals; 2) identifying some key players in implementing this plan with suggestions on fostering cooperation; 3) summarizing information on the ecology of oak savannas and woodlands; and 4) providing examples of ongoing recovery efforts. 

Recovery Goals

Oak ecosystem recovery throughout its range is the overall goal--the restoration of all natural ecosystem functions and processes, including a natural range of biotic community types for a full array of native species, and conditions in which natural selection can operate (adapted from Shively & Temple 1994). Management should be carried out using a landscape scale ecosystem approach that includes a long-term perspective, and an emphasis on preserving ecological structure and function. (Henderson, 1993) 

Goal 1: Achieve broad consensus for a regional oak ecosystems recovery effort.

  1. With extensive and inclusive participation from all relevant entities, produce and widely distribute a unifying document which will offer a range-wide perspective to local recovery efforts. 
  2. Agree upon a time frame for implementing steps toward recovery goals. 
  3. Gain support for this work in every state and province in the region, and among all levels of government, particularly those entities which hold or manage land.

Goal 2: Establish a networked system of reserves that captures the full array of oak ecosystem species, communities, and processes, and that conserves viable populations of all plants and animals known to inhabit them. 

  1. Recover the full range of variation in structure and composition (i.e. not just on unproductive soils). Recovery efforts should be located in all physiognomic provinces. Recovery of at least two to three percent of the land within each province may be possible and/or desirable in some cases. 
  2. Set specific recovery goals within local jurisdictions. Examples to date include: Missouri State Parks are working towards recovery of 62,000 acres of savanna and woodlands (McCarty IN PRESS); natural resource managers in Wisconsin have drafted plans proposing to recover 2 to 3 percent of the original acreage of oak ecosystems (Henderson 1995). 
  3. Plans should be designed to maintain viable populations of all extant species characteristic of Midwest oak ecosystems before they reach threatened or endangered status and to achieve full recovery of all federally listed oak ecosystem species on protected land. 
  4. Identify and include buffer lands as needed in this network through incentives and education.

Goal 3: Complete inventory and mapping of oak ecosystems. 

  1. Locate all the best (largest, highest quality, most intact) remnants of all target community types, as best we know them, and target them for both protection and management/recovery. 
  2. Specifically include small sites in planning when they are the most intact or best available or are located in strategically important places to function as ecological "stepping stones" to connect larger parcels. These may add diversity to regional restoration efforts. 
  3. Identify sites that could be connected or consolidated into larger preserve systems. 
  4. Coordinate the conservation of oak ecosystems biodiversity with the conservation of other midwestern ecosystems. 

Goal 4: Establish a network of pilot restoration projects with intensive monitoring.

  1. Use this network to evaluate the efficacy of specific restoration methods and to accumulate an experiential basis for predicting how systems will respond to various treatments. 
  2. Develop communication systems to make this information widely available to practitioners and planners. 
  3. Work toward standardization of monitoring protocols to foster better communication among practitioners and the scientific community. 
  4. Develop criteria for rapid bioassessment protocols.

Goal 5: Foster the development of influential local stewardship groups and education networks.

  1. Develop community outreach programs designed to significantly increase levels of public participation in recovery projects. 
  2. Develop education and training programs for natural resource managers. These should involve training in ecological concepts, transdisciplinary thinking, and continuous learning (Henderson, 1993). 
  3. Develop, produce and distribute a management handbook based on the best available technology. 
  4. Create regional teams comprised of natural resource managers, citizen stewardship volunteers, and communications professionals which would develop educational materials and distribute information about that region's projects and progress to a broad audience. Develop communications pathways among these teams. 
  5. To the extent possible, establish recovery projects in areas that involve citizens from the full range of the midwest's human cultures and populations specifically including both metropolitan and rural areas.

Goal 6: Develop incentives and information in order to involve and engage the private sector (e.g. nurseries, timber harvesters, farmers, land developers, recreation land owners, etc.).

  1. Encourage recovery projects on private lands through changes in property tax laws (Shively & Temple 1994). 
  2. Identify sites to investigate sustainable but economically viable practices on privately owned land. Incorporate representative sites into the pilot restoration network (see goal 4 above). 

Goal 7: Develop a regional cooperative burning policy.

  1. Develop a public education strategy to increase understanding and awareness of the critical need for prescribed fire in natural areas. 
  2. Examine and evaluate air quality standards as they apply to controlled burning. 
  3. Expand, upgrade and intensify prescribed burn training programs for resource managers at all levels. 
  4. Explore the potential for extending federal land planning support to all landowners including state and private.

Goal 8: Develop a research agenda.

  1. Identify gaps in knowledge about plant community associations, classification, and status of declining and rare species and remnants. Prioritize these questions when possible. 
  2. Determine necessary size, characteristics, and management needs of buffer areas. 
  3. Assess the effectiveness of restoration technology on an ongoing basis and provide scientifically sound input to assist in management decisions. 
  4. Employ research to develop new management techniques and to refine existing techniques, directed at expanding our understanding of ecosystems and at identifying options for future management. 
  5. Expand our knowledge base about the impact of management decisions with respect to genetic diversity. 
  6. Develop a research agenda for exotic/aggressive species to identify future problems, design control strategies for the most dangerous weeds (list to be compiled), evaluate ecological impacts as well as effectiveness of herbicides, and map occurrences. 

Goal 9: Establish coalitions and partnerships dedicated to conservation of oak ecosystems both locally and regionally.

  1. Identify, support, and promote existing and potential partnerships and accomplishments. 
  2. Develop communications systems to coordinate research among all regional agencies and states.

Goal 10: Investigate the role of native and domestic herbivores in managing and restoring functioning savanna ecosystems.

  1. Identify and determine the extent of existing savanna reserves being managed to include elk, bison and perhaps cattle. 
  2. Select pilot sites in distinct savanna landscapes as models to introduce elk or bison for research and demonstration sites. 


Figure 2.  Savanna and Woodland Classification Systems (adapted from Attachment 1 in Botts et al. 1994, originally based on a sketch by Don Faber-Langendoen).

Curtis (1959) did not use percent cover to distinguish prairie from savanna; savanna was defined as having at least one and prairie as having less than one, mature tree per acre. Citations to other systems are as follows: UNESCO (1973); INIA, Illinois Natural Areas Inventory (White & Madany 1978); NBPP, North Branch Prairie Project, Illinois (no citation); TNC (Faber-Langendoen, Appendix A this document). Henderson (in Botts et al 1994) distinguished forest from woodland as follows. Forests: multilayered with well-developed woody understories; prevalence of spring ephemerals and fire sensitive herbs; absence or reduced importance of summer and fall blooming groundlayer plants; infrequent to rare fire. Woodlands: two-layered with little or no woody shrub or understory layers; summer and fall bloomimg groundlayer plants prevalent and productive; frequent, but low intensity, fire.  



For clarity, we define several of the terms used in this document. Readers should be aware that the nomenclature of vegetation is troublesome due to ambiguity, differences in local usage, and historic shifts in meaning. Our oak ecosystems can be thought of as forming a continuum from prairie with no large trees through savanna and woodland to a closed canopy forest (Figure 2). Various authors have placed the words prairie, savanna, woodland, and forests in different places along the continuum. Words such as "prairie," "savanna," "woodland," "forest," "oak opening" and "oak barrens" can only be defined arbitrarily because these blend one into the other. As Gleason observed long ago, a sharp line cannot be drawn between most ecosystems. 

Except where noted, these words, prairie, savanna, woodland, and forest, are used broadly, recognizing that they overlap. 

Those interested in the historical use of some of these terms to describe Midwestern vegetation may want to read papers by Bowles & McBride (1994), White (1994, 1995), and Wood (1994). 


In the past ecologists and conservationists have used several systems to classify Midwestern vegetation, but today there is no single widely-used and accepted classification system. At the 1993 workshops there was considerable discussion among many participants about developing such a system. In response, Faber-Langendoen (1994) proposed a classification system for savannas and woodlands that was included in the 1993 pre-conference paper (M. Anderson et al. 1993) and reprinted in the 1994 Draft Plan (Botts et al. 1994). That system has been expanded and updated with the participation of state Natural Heritage offices. It is still being reviewed and refined, but is presented here as Appendix A to aid communications and to generate discussion. 

Attempts at classifying ecosystems are confronted by fundamental and practical dilemmas (Klijn 1994). The system presented here is designed to aid region-wide conservation planning. Other classification systems may better suit particular needs. 

Visions of Ecosystem Recovery: Values and Processes

The overall goal presented in this plan is to restore and maintain a network of native ecosystems large enough and diverse enough to insure the perpetuation of native biological diversity characteristic to each ecosystem and to maintain other crucial ecosystem functions. 

The emerging vision of conservation strategies in the Midwest and elsewhere has been described as landscape-oriented (Neave et al. 1994). This approach is based on the principles of ecology; it is regional in scale, process-oriented, ecosystem-oriented, and includes all elements of the landscape, including protected and unprotected lands, as well as humans and our activities; it is rooted in the stewardship ethic which maintains that human have responsibilities in nature (Leopold 1949, Loesser-Small 1992). Humanity has lived within and in turn has influenced the oak-ecosystem landscape for several millennia. The work of burning, gathering and planting seeds, and controlling weeds can be the basis for new ceremonies and festivals celebrating the human community's renewed partnership with nature (Turner 1992, Jordan 1992, Jordan 1995). 


The Ecology of Midwestern Oak Ecosystems


The study of ecosystems is a relatively young science; it has been only 60 years since the term was coined by the English botanist A.G. Tansley (1935). A system may be defined as an integrated entity whose overall properties are different from the properties of its elements; ecosystems, therefore, are plant, animal, and human communities, their abiotic habitats, and their interactions and processes. 

Scientists since Tansley have been stimulated to increase our understanding of the nature of ecosystems, but have not developed a set of unifying concepts (Schulze & Zwolfer 1987). Instead there are pluralistic approaches (Schulze & Zwolfer 1987). Here, in our effort to describe oak ecosystems we take an applied approach, focusing on vegetation since that is the component of oak ecosystems best understood and often the first concern of restoration efforts. 

Less than 10 years ago, in her excellent status report on Midwest oak savannas, Nuzzo (1986) wrote "Little scientific information exists about oak savanna." In the last few years, there has been remarkable growth in the study of oak ecosystems as well as conferences held in Chicago in 1993, Normal, Illinois in 1994, and Springfield Missouri in 1995. Many new studies have begun in response to information needs for conservation and restoration (e.g., see Fralish et al. 1994). 

Oak Ecosystem Ecology and Biogeography

Origin and Distribution

Oak savanna and woodland ecosystems have probably existed in North America for about 20-25 million years (Barry & Spicer 1987). Their composition and geography changed continually during that time with variations in climate. As the Wisconsin Ice Sheet retreated (around 15,000 to 10,000 years ago), the advance of species varied with dispersal mechanisms and the suitability of habitat. During the past 13,000 years the climate varied greatly, with a warming period until about 8,000-7,000 BP (before present) and a cooling trend beginning around 3,500 BP (Stearns 1995). It is widely believed that during the warmer periods fires increased and prairies and savannas expanded; during the cooler periods fires were less common and forests expanded (Clark 1988, Stearns 1993, 1995). An apparent exception, perhaps due to changing fire management practices, occurred on the northern Indiana dunes, where pine forests gave way to prairie/oak savanna during a minor cooling period A.D. 264 to 1007 (Cole & Taylor 1995). A more-or-less stable mosaic of prairie, oak savanna, woodland, and wetlands may have occupied a continuous band running north to south and covering millions of acres for the past several thousand years (Henderson 1995). During earlier, cooler glacial advances, savannas and woodlands probably existed further to the south and west. 

Midwestern oak ecosystems lie geographically between the forests of the east and the prairies of the Great Plains (see Figure 1). At a basic level, one could say that the pre-settlement distribution of grasslands, savannas, and forests, are related to the general climate patterns, with, from east to west, precipitation decreasing and the frequency of drought increasing (Transeau 1935, Dyksterhuis 1957, Curtis 1959, Lauenroth 1979, Risser et al. 1981). This east-west climate pattern also describes a fire regime gradient. In the east infrequent but catastrophic fires simultaneously terminated and initiated long-lived species (the fire-initiated systems of Vogl 1977). Moving west there were increasingly frequent fires which prevented excessive fuel buildup and controlled invasion of fire intolerant species (Vogl 1977) that otherwise would have resulted in severe fires which destroy vegetation. Along this east-to-west gradient both the diversity and density of trees decreases. From north to south, there is also a change in species, having to do with glacial history, dispersal mechanisms, responses to seasonal temperatures, day-length, etc. 

At finer spatial scales the vegetation of the Midwest could be described as a mosaic of communities including savannas, grasslands, wetlands, and forests. A considerable body of literature describes how the landscape pattern of vegetation relates to local climate, soils, topography, distance and direction to fire breaks, proximity to Indian villages, etc. [e.g., Weaver 1954, Weaver and Albertson 1956, Anderson & Anderson 1975, Hole 1976, Kline and Cottam 1979, Leitner et al. 1991, Bowles and McBride 1994, Will-Wolf and Montague 1994, Nyberg IN PRESS). 

Midwest oak ecosystems varied greatly in composition and structure. In a very simplified view, Midwest oak vegetation grades to the west into prairies; to the northwest into aspen and jack pine parkland; to the northeast into pine barrens and mixed forest: and to the east and southeast into oak-hickory and mixed pine-oak forest. Oak savannas occur as far south as Mexico (Gonzalez 1986). 

Physical Environment 

Climate, soils, and topography are important physical factors influencing the distribution and dynamics of oak ecosystems. Since we are discussing ecosystems over a large geographic area it is worth noting differences and similarities found. 

Climate. Midwestern climate is very seasonal, with hot summers and cold winters and unpredictable precipitation patterns. From east to west across the Midwest there is a moisture gradient, with decreased total precipitation and more droughts toward the west. Day length and length of growing season varies considerably north to south. 

Substrate and Hydrology. Oak savannas and woodlands may have occurred on many different terrestrial soil types of the Midwest, including "wetland," "prairie," and "forest" soils. Curtis (1959) noted that savannas often occupied the border between prairie and woodland/forest. Dyksterhuis (1957) noted that savannas occurred on prairie or forest sites and were sometimes of anthropogenic origin. Different kinds of plant communities (e.g., prairie and woodland) could occur on similar soils where fire regimes differed, such as on different sides of fire breaks (e.g., Leitner et al. 1991, Will-Wolf & Montague 1994). 

Oak savannas and woodlands grew across a broad range of hydrologic and moisture conditions, from dry (e.g., post oak central dry barrens) to moist (e.g., swamp-white oak bottom lands). Most high quality remnants of oak savanna and woodlands are on dry, often sandy, nutrient-poor soils. Many former savannas on richer and more mesic soils were converted to agriculture or, in the absence of fire, were invaded and altered by fire-intolerant species (Curtis 1959, McCune & Cottam 1985). Oak savannas occurring on wet to wet-mesic soils have been little studied (Curtis 1959, Hujik 1993, Solecki 1993, Leach 1994). Very little is known about the relationship between the conservation of these hydric communities and the alterations of hydrologic conditions by draining and water-level regulation. 

Disturbance and stress

Midwestern prairies, oak savannas, and oak woodlands are disturbance-maintained systems (Vogl 1977). As defined in this plan, disturbance is an action that quickly removes some living biomass (see Huston 1994). A disturbance seldom affects all components of a community equally. Some may be merely stressed, some unaffected, and others enhanced. In this report, fire is always described as a disturbance. 

Historically, three major causes of disturbance within oak ecosystems were fire, drought, and herbivory. Other, perhaps less important disturbances included disease (e.g. oak wilt), floods, tornadoes and other high winds, and ice storms. 

Species diversity is often correlated with disturbance both positively and negatively (Huston 1994). Several studies have found that the highest levels of species diversity occur when disturbances are not too severe and their frequencies are "intermediate" (Louckes 1970, Auclair and Goff 1971, Connell 1978, Huston 1979, 1994). The same disturbance regime may be highly destructive to one vegetation guild (e.g. trees) but "intermediate to another (e.g. grasses), and thus have complex effects on the diversity of vegetation overall. 

Prairies, oak savannas, and oak woodlands are fire-maintained communities, dependent upon frequent (annual to about once every decade) fires to prevent fuel buildup and control woody plant invasion. In these communities fine fuel (plant debris) often accumulates faster than it can decompose adding annually to the litter layer (Vogl 1977). Litter buildup decreases primary productivity and can alter species composition (Old 1969, Peet et al. 1975, Rice & Parenti 1978, Hulbert 1988). Fire consumes litter and stimulates vigorous growth of fire-adapted species. Fire can also control non-adapted species. 

In the absence of fire, the Midwest climate and soils could support non-oak forests. Instead, throughout much of the Midwest, regular fires maintained prairie and oak ecosystems (Gleason 1913, Sauer 1950, Curtis 1959, Daubenmire 1968, Vogl 1974, Bragg and Hulbert 1976, Grimm 1984, Anderson 1983). Without fire there is widespread potential for oaks to be replaced by more shade-tolerant species (e.g., Christensen 1977, Lorimer 1984, McCune & Cottam 1985, Nowacki et al. 1990). However, this successional pattern varies among regions and soil conditions and is more common on mesic rather than on xeric sites (Whitford & Whitford 1971, Abrams 1986 1992, Host et al. 1987, McCune & Cottam 1985, Nowacki & Abrams 1991). 

Midwest oaks are fire-maintained species (see Cottam 1949, Curtis 1959, Anderson 1983, Lorimer 1985, Reich et al. 1990, Will-Wolf 1991, Abrams 1992, Cole & Taylor 1995). Paleobotanical studies consistently reveal oak pollen to be associated with layers of charcoal (Watts 1980, Davis 1985, Winkler 1986, Abrams 1992). Lorimer (1985) offers these reasons why oaks are favored by fire over other trees: thick bark, sprouting ability, resistance to rotting after scarring, and the suitability of fire created seedbeds for seed germination. Reich et al. (1990) found that burning oak saplings stimulated higher concentrations of leaf nitrogen, which may increase rates of photosynthesis. 

Besides species-specific plant traits, several other factors affect the response of woody plants to fire: plant size, season of the year; fire intensity and duration, and other factors (Vogl 1974, Anderson 1982, Anderson 1983). Tree mortality is also affected by fuel conditions near the trunk (Anderson and Brown 1983, 1986). Few mature trees of most species are killed by "cool" ground fires (White 1983, Abrams 1988, Reich et al 1990), although some species, particular paper birch are very fire sensitive. The greatest inter-species difference in mortality from "cool" fires appears among saplings and seedlings. For example, following a spring burn in central Wisconsin, Reich et al. (1990) recorded a sapling mortality of 67 to 100 percent of black cherry, paper birch, and big-toothed aspen, but no mortality in Hill's oak saplings. In Iowa's abandoned fields, fire can aid the recruitment of oak (Crow et al. 1994). Fire-free periods can be important for oaks to reach large, fire-resistant size (Haney & Apfelbaum 1990). Extremely "hot" fires may damage all trees regardless of species. 

There is limited evidence for determining what was the "natural" fire regime in Midwestern ecosystems. In most areas there are too few trees of sufficient age to use dendrochronologies as a basis for estimating ancient fire frequencies (return intervals). Three fire scar studies in Missouri date fires as far back as 1740. Prior to settlement by Anglo-Americans in 1850, post oaks growing on a ridge in the Mark Twain National Forest reveal a minimum mean return interval of 2.8 years (Cutter & Guyette 1994). Similar studies at Caney Mountain Wildlife Refuge revealed a minimum mean return interval of 4.3 years (Guyette & Cutter 1991) and 3.2 years for cedar glades on the Ava Ranger District (Guyette & McGinnes 1982). 

Both people and lightning played important roles in establishing and maintaining Midwestern oak ecosystems. In the Midwest people regularly ignited fires for at least the last 5,000 years. The reasons they set fires and the extent of the burns have been discussed by Stewart (1951, 1956), Curtis (1959), Pyne (1983, 1984), and others. 

Abrams (1990) lists these adaptations in oaks to drought: deep roots, xeromorphic leaves, low water-potential threshold for stomatal closure, and the ability to adjust osmotically. Oak species differ in their drought tolerance. Oaks can also maintain relatively high rates of photosynthesis during droughts when they have the ability to out-survive other tree species on nutrient poor sites (Reich & Hinckley 1980). One possible explanation for oaks' water efficiency is that they may substitute isoprene for water as a leaf coolant (Sharky & Singsaas 1995). 

Water relations may be extremely important in the distribution of ground layer species. Bray (1958) reported some plant species growing in wet and wet-mesic prairies also grew under trees in drier savannas. Such mesic-to-xeric shifts probably result from lower heat loads and lower rates of evapotranspiration found in partial shade (Bray 1958, Pruka 1994a). 

The great mass of small animals, including invertebrates, that consume plant parts undoubtedly had a great influence on pre-settlement Midwest vegetation. Savannas and grasslands have high ratios of plant productivity to plant biomass, and "nearly all the productivity is palatable and accessible to a wide size range of animals, without need for specialization for climbing trees (p. 458, Huston 1994). 

Grazing animals can interact strongly with fire, since both fire and grazers consume the same resource (Huston 1994), and because fire stimulates new, nitrogen rich, highly palatable growth. The predators of grazing animals would also affect plant-herbivore interactions, as has been studied in African grasslands by McNaughton (1985). 

Since settlement, pasturing in some areas has maintained savanna-like groves of open-grown oak trees, but the native ground layers have been mostly replaced by Eurasian species. The oaks themselves often fail to regenerate. Deer and other grazers/browsers are well-known to decrease the regeneration of some woody species (Alverson et al. 1994). These animals can also stimulate woody growth by damaging herb-layers and soil (Dyksterhuis 1957). In some cases, light pasturing has preserved certain native herbs, and has kept areas free of brush and tree saplings. In places, former pasture lands have potential for ecosystem recovery (e.g., Bronny 1989, Pruka 1994a). Light grazing by livestock for short durations may have a small role in recovery efforts (Henderson 1995), as may mechanical mowing (Swengel & Swengel 1995). 

Much more research is needed into the possible benefits of grazing, by both native and domestic ungulates, as a recovery tool. Livestock managers increasingly use rotational grazing techniques, where some part of a pasture is grazed while the remainder "rests" (Undersander et al. No date). Rotational grazing methods might be designed to benefit both biodiversity and farmers (Agriculture Ecosystems Research Project, No date) particularly in "buffer" areas surrounding preserves. 

Not until large contiguous areas of native vegetation are restored and re-populated with native grazers and their predators, can we observe directly the complex interactions between landscapes, disturbance, grazing, and predation. 

The now extinct passenger pigeon periodically had a significant effect on oak ecosystems (Blockstein & Tordoff 1985). Historic reports describe millions of nests covering nearly every tree across several counties. Trees sometimes broke from the weight. Below the nests several inches of excrement could accumulate. The birds could consume most of the acorn crop, depriving other animals of this food. On the other hand, passenger pigeons were abundant prey for predators. 

Biological Diversity

Biological diversity comprises all levels of natural variation from molecular and genetic to landscape patterns, including species (Huston 1994). These levels are not independent, but interact in complex ways to produce the great natural variation found on earth. 

In our efforts to conserve biological diversity, it is important to recognize that not all components of diversity are equally threatened. There is great variation among species in their level of dependence upon remnant or restored communities. Some native species are well-adapted for survival in our highly-altered landscape. These include common weeds like tall goldenrod and horseweed. Other species are dependent upon specific habitats not found outside of remnants or restorations i.e., the "conservative species" of Swink & Wilhelm (1994) and the "remnant dependent" species of Panzer et al. (1995). Some species require very large areas of habitat. 

The effectiveness of ecosystem recovery efforts in the preservation of biodiversity will depend upon how well the needs of remnant-dependent species are met. Remnant-dependent species have been identified for some taxa in some areas (e.g., Chicago region vascular plants, Swink & Wilhelm 1994; certain Chicago region insects, Panzer et al. 1995). Comprehensive lists would greatly benefit recovery efforts, but require much more research before completion. 

Landscape Diversity

The pattern of vegetation found on the landscape results from complex ecological interactions operating at different spatial patterns and temporal scales. (Turner 1987). Several studies using data from original land surveys have shown complex relationships between the distribution and density of various tree species (often interpreted as representing different community types) and various other factors such as topographic relief, soil texture, fire breaks, and ignition sources (e.g., Anderson & Anderson 1975, Anderson 1991, Will-Wolf & Montague 1994, Nyberg IN PRESS, Bowles et al. 1994). For example, tree densities were typically higher on steeper north and east slopes, and closer to fire breaks. 

Oak savannas and oak woodlands were often part of a complex landscape mosaic. For example, Palos Township, Cook County, Illinois at the time of settlement was described by Nyberg (IN PRESS) as containing five major community types, each covering between 8 and 36 percent of the land area. Pierce (1954) described the natural or presettlement vegetation at Cedar Creek Natural History Area, Anoka and Isanti Counties, Minnesota as a patchwork constrained by physical features but also greatly affected by human fires. 

Vascular Plant Diversity

Fire adapted communities which are frequently burned often have high species diversity, particularly of shrubs, grasses, sedges, and forbs (Huston 1994). The great bulk of the region's native, non-aquatic plant species grow, at least sometimes, in association with oaks. Within a single stand of open-grown oaks the light regime at the ground layer is highly variable with certain points a few meters apart receiving over 90 percent and less than 10 percent of integrated potential sunlight (Leach & Givnish IN PREP). Scattered trees also cause other micro-habitat conditions to be heterogeneous at small scales including soil temperature, soil moisture, and humidity (Ko 1990, Ko & Reich 1993). 

Another question raised in recent years is the degree of specialization of the flora of oak ecosystems. In Wisconsin, Curtis (1959) reported very few species reached their peak abundance in oak openings. However observations in restored savannas suggest that the original vegetation may have been more diverse and specialized than indicated by Curtis (Packard 1988a, 1988b; Bronny 1989; Clewell 1989; Leach 1994; Pruka 1994a, 1994b; Henderson 1995). 

Animal Diversity

Animal diversity is high in productive systems. Because primary productivity is high in prairies and savannas and much of the product is available, prairies and savannas should support a large mass of prey species, which in turn, should support a diversity of fauna at higher trophic levels. 

The diversity of certain taxa (e.g., birds, herptiles, some invertebrates) is correlated with structural heterogeneity which is relatively high in savannas. For example, many cold-blooded species benefit from a mosaic of sun and shade which is important for thermal regulation. Fry (1983) noted that bird diversity of tropical savannas was very high as the result of structural heterogeneity and the high availability of plant seeds and insect prey, condition also expected to be high in Midwestern savannas. 

Without doubt, the ecological roles of animals, both large and small, were very important in creating and maintaining savanna/woodland ecosystems. These relationships among trophic levels needs continued study (see Risser 1995). 

Almost surely, invertebrates were the largest component of biodiversity in oak ecosystems (Thomas Givnish, pers com), yet little is known about their characteristic habitats, their natural assemblages, or their current status. (Henderson 1995) Many invertebrates are very selective in their food choices; as host plants have suffered declines so too have the invertebrates. It is likely that many invertebrate species of oak ecosystems were lost or are now very rare (Henderson 1995). Givnish (pers com) suggests that we are unlikely to gain many further insights into the assemblages of invertebrates until large areas of savanna are reconstructed. 

Midwestern oak ecosystems are likely to be important refugia for invertebrates. In a thirteen year study in the Chicago area Panzer et al. (1995) concluded many, but less than one quarter of the prairie- and savanna-inhabiting insects depend on remnant habitat for survival. Habitat specificity was highest among certain groups such as flower moths (Shinia spp.) and root boring moths (Papaipema spp.) Where surveys have been conducted for Lepidoptera, for example in the Chicago region (Panzer et al. 1995), Fort McCoy Military Reservation, Wisconsin (Maxwell & Ferge 1994), and the Oak Openings in Ohio (Shuey 1994), numerous rare species have been found. Ohio is the only Midwestern state to complete a state-wide survey of all Lepidoptera species. That state's oak savanna communities support their largest assemblage of imperiled moths and butterflies (Shuey 1994). At least some savanna insects, such as the Karner blue butterfly, federally listed as endangered, appear to have a metapopulation structure that depends upon a patchwork of fire over an extensive area (Givnish et al. 1988, Shuey 1994). 

Conservation of invertebrates offers special management challenges. For example, the Karner blue butterfly is dependent on a single fire-adapted host plant, wild lupine. The butterfly populations however seem vulnerable to fire in all life stages. Thus fire appears at once to improve and maintain habitat while damaging populations. The solution may be to manage a patchwork of sites, burning only a portion of each them. Population sizes within a patch may fluctuate greatly, perhaps crashing after a fire but recolonization of the improved habitat quickly rebuilds the population. For more information on management of insect species with dynamic metapopulations see Arnold (1983), Givnish et al (1988), Shuey (1994). The U.S. Fish and Wildlife Service recovery plan for the Karner blue butterfly is expected to be published in late 1995 or early 1996. 

Savanna fires can benefit herptiles in at least two ways (Hay 1995) 

Sample and Mossman (1994) estimated that in the prairie-savanna-woodland landscapes about 100 breeding species regularly occurred in presettlement Wisconsin. For the entire prairie-forest transition Herkert (1994) lists 83 breeding species, ten of which are listed as threatened or endangered in at least one midwestern state: Bachman's sparrow, Bell's vireo, Bewick's Wren, golden winged warbler, Lark sparrow, loggerhead shrike, long eared owl, merlin, prairie warbler, and Swainson's hawk. 

Many savanna birds are now common birds of cities, suburbs, and farms (e.g., blue jay, black-capped chickadee, American robin) while other savanna bird species are rare or decreasing (e.g. barn owl, Cooper's hawk) 

Only one savanna bird species has gone extinct, the passenger pigeon most likely from dove-hunting (Blockstein & Tordoff 1985). Other birds have been extirpated from large portions of their previous ranges within the Midwest. In several areas the American turkey was extirpated, but successfully reintroduced. In Missouri, bird diversity is highest in more open communities, lowest in what we might call over-mature (i.e. fire suppressed) oak-hickory forest (Evans and Kirkman 1981). The red-cockaded woodpecker and Bachman's sparrow are now rare or extirpated in Missouri (McCarty IN PRESS) The decline of other species (e.g. red-headed woodpecker, northern flicker, towhee, field sparrow, and bobolink) may be due to habitat losses in the midwest as well as problems with wintering grounds (Sample & Mossman 1994). In Wisconsin, species that may require special habitat management include the prairie chicken, sharp-tailed grouse, red-headed woodpecker, loggerhead shrike, Bell's vireo, field sparrow, and bobolink (Sample & Mossman 1994). 

Many of the mammal species that once were part of Midwestern oak ecosystems are still doing well today. The fox squirrel, cottontail rabbit, woodchuck, and white-tailed deer are abundant in or around our cities, suburbs, and farms. In places, some of these species are overabundant and can cause problems in natural areas. Others, such as the timber wolf, bison, elk, bobcat and black bear, have been extirpated from large areas of their former range or have had their numbers greatly reduced. 

Timber wolves, mountain lions, black bears and other top predators are re-invading their former ranges. Mountain lions are now, or recently were in southern Wisconsin and northern Illinois. Timber wolves from northern Minnesota have ventured into the savanna region as far as Portage, Wisconsin, although Wisconsin DNR experts think packs are unlikely to form except in areas with few roads and people. 

Other life forms. 
Very little is known about the diversity, distribution and abundance within oak ecosystems of fungi, slime molds, lichens, algae, bacteria, and other small organisms. Benjamin et al. (1989) found a decrease in vesicular-arbuscular-mycorrhizal fungi across a community gradient from sand prairie to closed black and blackjack oak forest in Mason County, Illinois. Studies at the University of Illinois, Normal continue on the ecology of mycorhizal fungi (Roger Anderson, pers com). At the Morton Arboretum Gerould Wilhelm and Doug Ladd are studying Midwestern lichens. Greg Mueller and Methvan are studying the non-lichenized fungi at the Field Museum in Chicago. 

Genetic Diversity

The preservation of a diversity of remnant populations may be very important in maintaining genetic diversity. Even small remnants may be valuable sources of genetic variation. Restoration planners should when possible use local sources when reintroducing plants and animals (Howell 1988, this document, Appendix C

Ecosystem Processes, Functions, and Values

The term processes refers to the interactions between ecosystem components, both biotic and physical, and the flow of energy and matter through the system. There have been few attempts to relate abiotic and biotic processes in Midwestern savannas (Risser 1988). A number of studies have related Midwestern vegetation to climatic conditions (e.g., Dyksterhuis 1957, Risser et al. 1981) and to fires and to soils (e.g., Gleason 1913, Curtis 1959, Hole 1976, Abrams 1992). Bray (1960b) and Ovington et al. (1963) measured aspects of plant productivity and biomass in Minnesota corn fields, prairies, savannas, and woodlands/forests. Oak woodlands were estimated to have a greater biomass and greater estimated annual net productivity than were prairies or savannas. 

Given the global rise in atmospheric carbon, it seems sensible to increase our understanding of how natural systems cycle and sequester carbon. The carbon content below ground of oak savannas and woodlands must once have been considerable and their destruction and subsequent degradation of soil would have released carbon to the atmosphere. Large scale oak ecosystem recovery might be a prudent component of a global strategy to control the rise in atmospheric carbon. 

Risser (1988) states, "[The] paucity of information [on the relation of abiotic and biotic process in savannas] is particularly unfortunate since understanding the interplay between trees and the herbaceous layer would be valuable, and because the climatic transitional zone should lead to powerful tests of ecosystems processes operating at the margins of both grasslands and forests." Again Risser (1988) concludes, "Finally, the North American savannas offer exciting opportunities for explaining relationships between abiotic variables and ecosystem properties. Intricate relationships have been demonstrated among climate, vegetation structure, forage quality, and herbivory in African savannas (McNaughton 1985), and similar processes are undoubtedly to be found in North America." 

Threats to Oak Ecosystems

Loss of Biological Diversity

Region wide, oak savanna and woodland ecosystems are highly fragmented with less than 0.02 percent of the original area remaining (Nuzzo 1985). Existing remnants are often highly degraded, especially on relatively rich soils. Preservation alone of remnants is not sufficient to conserve their biological diversity of remnants over the long term. Handse. 

The loss of existing high quality examples of oak ecosystems and the loss of areas with high recovery potential continues. Global threats to ecosystem function and ecosystem biodiversity were assessed by Soule (1991). He suggested that in wealthier countries the significant impacts are due to exotic species, pollution, climate change and habitat fragmentation. While Midwestern landscapes may be threatened by other factors, the list provides a useful framework for consideration of research priorities (Boyle 1994). 

Timber harvesting could be a sustainable enterprise consistent with ecosystem management, but much more research into the best methods for such harvesting is needed. Much too often, landowners do not receive the financial and technical support and encouragement needed to manage lands for long-term values rather than short-term economic gain. For example when farm prices drop, farmers are apt to "mine" their oaks (Smith 1992) with no plan for regeneration. One challenge of recovery efforts is to find ways to assure sustainable timber harvests (and economic incentives for landowners) while maintaining ecosystem processes and native biodiversity. Cutting of oaks for timber can be highly compatible with biodiversity conservation since those communities are so "disturbance" dependent. Overstocking is probably more of a problem than overharvesting in many cases. The challenge is to make the harvest work for oak reproduction and biodiversity in general. In addition, changes are needed in property tax laws to reflect the ecological value of land and not solely its value as a commodity (Shively & Temple 1994). 

Loss of Ecosystem Processes

The loss of our native plant and animal communities has caused alterations in carbon cycling, water cycling, energy flows, soil building, and other ecosystem functions. How these changes affect the sustainability of the biosphere is not always known. However, prudent approaches to environmental problems may often include an ecosystem restoration component. 

For example, over the last 100 years flood damage along the Mississippi River has been steadily increasing. The floods of 1993 caused $16 billion in damage (Hey & Philippi 1995). The restoration of native prairie-savanna-woodland vegetation could play an important role in diminishing flood problems by: a) intercepting and holding precipitation; and b) increasing the water holding capacity of soils by replenishing their organic content (Hey & Philippi 1995). 

The Recovery Process


The recovery of oak savannas and oak woodlands requires the protection and management of high quality remnants and the restoration of degraded areas. Successful recovery efforts are more and more likely as partnerships continue to develop among government agencies, public institutions, the private sector, and citizens. 


The terms "conservation," "preservation," "restoration," "management," and "recovery" are generally understood, but are not used consistently in the literature. Before continuing the discussion, a framework is provided here, based largely on the definitions given by Howell (1988) and Howell & Jordan (1991). 

"Conservation" refers to a broad set of activities undertaken to ensure the continuance of ecological resources. Conservation can be thought of as a continuum with "preservation" at one end and "restoration" at the other with much overlap. "Preservation" activities take place in situations where the most desired elements are present, and can include such strategies as acquiring land, securing easements, and negotiating management agreements. "Restoration," in contrast, implies a return to a previous condition, state, or function including both natural processes and species composition. Restoration activities may include, for example, reintroduction of extirpated species, creation of prairie on abandoned farm fields, the return of fire to fire-maintained communities, or the removal of alien buckthorn from a remnant. Note that the species and processes that are restored are those that were believed to have been formerly present. Both preservation and restoration imply long-term commitments to active management. 

"Recovery" describes conservation activities that are designed to make viable once again populations, species, communities, and ecosystems, which are endangered or imperiled. Recovery of oak ecosystems involves preserving intact communities and restoring degraded and lost communities to insure the continuance of natural selection and other ecosystem processes. 

Recovery Planning

In this section some of the planning issues of particular interest to oak ecosystem recovery are introduced. Many of the citations given here provide more detailed discussions. The Board of Directors of the Society for Ecological Restoration has approved a set of policies regarding restorations included as Appendix C. The Society of Ecological Restoration also publishes two journals "Restoration and Management Notes" and "Restoration Ecology." Many of the citations found throughout this plan may also be useful. Island Press will soon (spring 1996) publish a book on restoration methods for tallgrass prairies and savannas. 

Ecological Models

The science and art of restoration is still young with a number of different approaches being tested and developed as we learn. Some restorationists strive to recreate as nearly as known the communities of plants that occurred at the site prior to its degradation. Others may focus on the preservation of populations and process under present conditions. 

In actual practice, however, restorationists are concerned with very similar ecological models, processes, and techniques and thus success can be measured similarly: how closely does the restored community match the measurable attributes of other pristine or remnant communities? For example, in a savanna restoration the measurable goal may be modeled from compositional observations in remnants which might have, for example, 80 to 100 native vascular plant species per acre with at least 75 "conservative" or "remnant dependent" species present. In addition to plant composition, vegetation structure, animal composition, and ecosystem processes may be modeled. 

Unfortunately, for some Midwestern oak ecosystems (e.g., lowland savannas, tall-grass mesic savannas) specific models have been lacking because no or few pristine or remnant examples were available to study (Curtis 1959, Packard 1988a, 1988b, Solecki 1993). Restorationists don't always know how close their mental pictures of ecological models come to describing canopy structure, species composition, and other elements of the community, although a great deal is being learned from restored areas. 

Models for restorations need not be static. Historically some sites may have cycled between nearly closed oak woodland to very open savanna and back again, depending on stochastic fluctuations in fire frequency and intensity and other factors. Recovery efforts should not "freeze the vegetation in time," but very large preserves are required to allow natural fire frequencies and climatic fluctuations to determine vegetation patterns (Roger Anderson, pers. com.). 

In the absence of clearly sound structural and compositional models, restoration planners often chose to restore missing ecosystem processes (most importantly fire), remove or control exotics, and restore, where possible, hydrologic processes. The results should be monitored for changes in structure, composition (especially of "conservative" species), and other ecosystem functions. Adding or removing individuals or species and alternating fire prescriptions may be done according to the best judgment of the restorationists. This approach would be particularly enlightening on larger preserves where fire-landscape interactions can be allowed to develop. 

Inventories and Assessment

Region wide, strategic recovery efforts would benefit from knowledge of the best opportunities for the recovery of different kinds of oak ecosystems. What opportunities exist for large-scale projects? how large should they be? how connected should they be? where can oak communities be restored on nutrient rich soils? where can viable populations of elk and bison be restored? Region-wide inventories of potential recovery sites have not been attempted yet. The Midwest Regional Office (MRO) of the Nature Conservancy in Minneapolis is in the process of gathering, classifying, and mapping information on the occurrence of oak ecosystem elements (both uncommon species and remnant communities). Those efforts are limited by incomplete field data; more field investigations are needed to identify potential recovery sites. An example of MRO's efforts is given in Appendix B for three upper Midwestern ecosystems: Northern Bur Oak Openings, Northern Black Oak Barrens, and Northern White Oak-Bur Oak Woodland. These are given to further discussion on the best ways to continue regional recovery planning. 

Success of recovery efforts at any one site depends a great deal on initial conditions: the more degraded, the more restoration will cost, the more time it will take, and the less complete it will be (Howell 1988, McCarty IN PRESS). Many field biologists and restorationists are experienced and skilled at recognizing easily restorable remnants. Their expertise can be used when evaluating sites for restoration. As the ecosystem degrades and habitat changes, persisting species are useful indicators. Recently, Pruka (1995; reprinted here as Appendix D) published an indicator species method for identifying restorable oak savanna/woodland remnants in Wisconsin. Another useful indicator of recoverable oak ecosystems is the presence of native oaks in any life stage. Open grown trees with widely-spreading lower branches (even if those branches are dead) indicate that the canopy once had an open structure (Cottam 1949). 

Prioritizing: Working with Limited Resources

Restoration and Management

Restoration activities need to be customized for the unique qualities and purposes of each site. A few commonly considered items are highlighted. 

Purposes and goals. What are the purposes of the restoration? What ecological models will be used as a guide? Is there a commitment for long term monitoring and management? 

Site conditions. What were past land uses? What are the present conditions? What is the structure and composition of current vegetation? Has succession from oak savanna or woodland to, for example, sugar maple forest progressed to the point where restoration may be difficult and expensive? Will there need to be remedial planting? What kinds of animals use the site? What are the abiotic features? 

Models. What kinds of plant and animal communities previously occupied the site? What was the physiognomy of that vegetation as deduced from soils, topography, the surrounding landscape, and historical records? What was the fire regime of those communities? What are the present (and future) biotic and abiotic conditions that affect the choice of future communities? Are there nearby and affordable sources for plant and animal reintroduction? Will the chosen model emphasize composition, structure, or processes (e.g., fire regime)? 

Size. Size may be the primary factor when considering how many species and ecosystem properties the site can support. Adjoining land use also affects those properties, especially as a source of invading species, nutrient inputs, and alterations of hydrology. Most restorations would benefit from being nested within buffer areas which could be semi-restored areas that limit undesired inputs and increase the effective size of the restoration. 

Fire. What actions need to be taken to assure that prescribed fire regimes can be realized legally, safely, efficiently, and without causing problems with neighbors? 

Unwanted species. Which undesirable species (plants and animals) are present? How will each of these be dealt with? How will new invaders be identified and dealt with? 

Defensibility. How will the restored ecosystem be protected from human abuse? 

Use. How will people use the site? How will decisions be made about what uses are and are not appropriate? What kinds of research will be encouraged or permitted? 

Viability. What is the outlook for long-term viability of the restoration? Who will provide stewardship in 20, 50, or 100 years?

Restoring and Managing

Common methods in oak ecosystem restoration include: returning the tree and shrub canopy to a prescribed structure, implementing a remedial fire regime, and reintroducing missing species. Elements removed are generally exotic species and over-abundant native trees and shrubs, but could also be drain tiles, ditches, buildings, and pavement, etc. In many cases total removal may be an unrealistic goal and control becomes the goal. These methods return the groundlayer environment to the condition suitable for recovery of the groundlayer community. Reintroduced plant species are usually herbaceous perennials, but can include shrubs (e.g., Bowles et al. 1994) and trees. 

Prescribed burning is perhaps the single most important management tool in the restoration and management of oak ecosystems (Ladd 1991, McCarty IN PRESS). Using fire in this way is not new. For example, at the Cedar Creek Natural History Area, Anoka County, Minnesota, fire has been used to restore and maintain oak savannas since 1964 (Irving 1981). 

The mechanisms by which fire affects vegetation are complex (see Ecology section and citations therein). The benefits of fire include: allowing more light to the ground layer by top-killing some trees, saplings, and shrubs; removing accumulated debris from the ground surface; and providing regeneration opportunities for some species. Possible negative impacts of fire may include the killing of desired species, particularly invertebrates (Swengel 1994). Ill-timed fires may also kill desirable plant at vulnerable life stages. 

Burning of degraded savanna remnants generally results in the increase of small-scale richness and diversity of plant species (Bacone 1981, Wilhelm & Master 1994, McCarty IN PRESS). Sometimes fires appear to allow for natural, unassisted recovery of missing species (e.g., Apfelbaum & Haney 1991, Packard 1988a, Hubner & Leach IN PRESS). 

In Missouri, as elsewhere, many restorable remnants have a thick growth of young trees and shrubs. The Missouri strategy for restoring degraded savannas and woodlands focuses on two important steps (McCarty IN PRESS): 1) reopening the midstory to allow light to the ground layer, and 2) reinstating fire regimes. In most cases, removing most of the 30 to 50 year old understory trees has solved the structure problem. Fires can create and maintain a more open midstory, control woody resprouts, stimulate the expansion and reproduction of herbaceous plants, remove surface litter, and provide germination sites. Some undesired plants are more effectively controlled with herbicides. After the initial clearing of the midstory and during the early stages or recovery, prescribed fires are frequent, often annual. As the groundlayer recovers and new woody growth is reduced, fire frequency can be reduced. During the early restoration period, fires tend to be light and spotty. Later, fire frequency decisions are made to balance fuel loading (and hence potential fire intensity) against advances in new woody growth. 

In remnants free of dense woody understories, such as those that have been grazed, restoration efforts can usually proceed simply by reintroducing fire. 

McCarty (1995) described the typical phases of recovery observed in Missouri: 1) After canopy thinning and initial fire(s) the groundlayer is dominated by sprouts of oaks and hickory, with shrubs and woody vines. 2) Soon the more easily spreading herbaceous species become common. 3) Later, a strong grass and forb groundlayer (with eight to ten or more species per quarter-meter square and very little bare ground visible) still includes robust tree sprouts and shrubs. 4) The more "conservative" herbs expand and spread. The number of years required to pass through these phases depends a great deal on initial conditions, including the presence of various herbaceous species. Good sites may reach phase three in several years while poor ones may take a decade or more to reach phase two. 

Typically, several fires are needed to sufficiently control woody growth. Once a sapling is top-killed, its multi-stemmed resprouts are even more likely to be top killed again, apparently because the stems trap leaves and other flammable debris held loosely around the stems (Henderson 1983). Sometimes certain woody plants increase following an initial burn, but decrease after the second (Bacone 1981). Fire mortality varies among species and life stages (see ECOLOGY SECTION). Many restoration projects use herbicides to control undesired plants. 

Much more experience is needed before we know the "best" fire regimes for maintaining all the desired components of fragmented savannas and woodlands. It is important, however, that fires not stop completely. For example, at a site in Minnesota after four years without fire, American hazel had largely grown back to its pre-burn distribution, density, height, and weight (Axelrod & Irving 1982). 

If fire alone does not move a remnant into phase 2 or 3, the site may have been damaged beyond the power of fire alone to restore. In this case, species introductions may be made (see e.g., Packard & Balaban 1994). With experience, restorationists will often know before the first fire which areas will need remedial plantings and which will not. In many situations it may be best to scatter seeds immediately following initial fires or brush removal, thus preempting a possible weed problem. Rather than recovering an entire site at once, restorations may proceed piece-meal, restoring sections at a time and using the more mature sections as seed sources. 

McCarty (IN PRESS) provides further information on the methods and costs or canopy restoration in Missouri. To reduce dense midstories, an expensive but effective approach is to hire crews to thin, treat stumps and burn the slash. This may cost up to $1200 per acre for the worst sites, $50 to $175 per acre for sites invaded by cedar. In Lake County, Illinois, large machine-mounted choppers clear large areas at relatively low costs (James Anderson pers com). Elsewhere, the power of volunteers is invoked for this purpose. A more effective but more hazardous method is a hot initial fire timed when the trees have budded in the spring and when burn conditions approach the limits of a typical management fire prescription. Such burns work best on steep terrain where convection columns magnify the effects. Autumn fires may achieve similar results (Gerould Wilhelm, pers. com.). 

A limiting factor in many restorations is the availability of seeds and planting stock. Using local ecotypes may help preserve genetic-diversity and assure that plants are suited to local conditions (Howell 1988). Producing sufficient quantities of seeds and planting stock often requires nursery propagation. Often restoration sites themselves can serve as seed sources for new plantings. 

Restorationists often lack information on what the floristic composition of savanna understories was historically, especially on loamy soils (Givnish 1993, Leach 1994). The widely cited accounts of Bray (1955, 1958, 1960) and Curtis (1959) on oak savanna composition may be misleading (Packard 1988a, 1988b, Givnish 1993, Henderson 1995). Recent and ongoing research into remnant and restored savannas (e.g., Pruka 1994a, 1994b, Leach 1994, Packard & Balaban 1994) is providing information on which species may be appropriately planted in different locations along soil gradients (between sites) and light gradients (within sites). It may take many years for planted restorations to attain the horizontal vegetation pattern of natural remnants (Leach 1994), therefore planted communities should be used with caution as models only. 


Monitoring is an integral component of restoration and management (Appendix C). Management activities should incorporate information on changing conditions caused by those activities. In addition, funding sources often require that objective standards of success are met. Research objectives often indicate quantitative observations. Prerestoration data should be taken when possible, but restoration of degraded or critically threatened sites should not be held hostage to obtaining baseline data (Botts et al 1994). 

Monitoring methods should be repeatable and efficient. Haney & Apfelbaum (1994) recommend using fixed transects with nested plots to sample tree, shrub, and herbaceous layers. Typical monitoring methods may periodically measure plant community attributes of canopy structure, community composition, status of alien species, and fuel loads. Commonly, animal use of the site, particularly of birds, is documented. 

Monitoring plans for each restoration should be agreed to in advance and customized to the purposes and goals of the restoration. Monitoring approaches should be flexible to adjust to changing needs and conditions. Ideally, information should be gathered from different sites by the same methods for useful comparisons. 

Species are not all equal as indicators of restoration progress. Indicator guilds or functional groups (e.g., nitrogen fixers), exotic species, rare species, "conservative" or "remnant-dependant" species, plant biomass or productivity, and amphibian biomass have been suggested as pragmatic indicators of ecosystem "health." In the past, lists of indicators have been developed haphazardly, often reflecting the interests of specific user groups and value systems, rather than according to more broad ecological criteria. 

One proven method for evaluating and assessing natural quality is the Floristic Quality Assessment describe in detail by Swink and Wilhelm (1994). This method depends upon the a priori assignment of conservatism coefficients to each native vascular plant species. Areas of the Midwest for which such coefficients have been determined include the Chicago region, Michigan, and Missouri. 

Research Priorities

The Research Working Group of the Midwest Environmental Roundtables' Interagency Cooperation on Ecosystem Management (ICEM) has been considering the research needs of recovering oak savanna and woodlands. The Working Group prioritized a list of 43 research questions published in Botts et al (1994; p. 113). That list is reprinted in this document as Appendix E in order to further discussion. Members of the Working Group continue to discuss how recovery efforts can be best served by inter-agency cooperation on research. The Working Group can be reached through its coordinator, Richard Whitman, National Biological Survey, 1100 N Mineral Springs Road, Porter, Indiana, 46304; 219-926-8336. 


Oak savannas and oak woodlands once dominated much of the Midwestern landscape, but now examples are highly fragmented and often highly degraded. Many successful conservation efforts have demonstrated that degraded remnants of these ecosystems are restorable. Several large areas (thousands of acres each) and many small areas (some a few acres or less) have been identified as likely candidates for recovery. 

This Midwest Oak Ecosystems Recovery Plan promotes current and future efforts to restore these ecosystems to insure their preservation and the continued natural selection of their component species. This plan is the product of workshops, discussions, and communications among hundreds of scientists, natural resource managers, and restoration practitioners. It attempts to aid local recovery efforts by providing a region-wide perspective that includes ecological background on these ecosystems and their recovery and gives examples of recovery efforts. The plan also suggests certain goals, strategies, and possible actions that will move recovery efforts forward. 

Several preserves of tens-of-thousands of acres or larger will help preserve ecosystem processes and large numbers of species, including those requiring large habitats. Many smaller restorations will protect both rare species and genetic variation. Recovery on both public and private lands is important to these efforts. 

Literature Cited


Local Navigation

Jump to main content.