1994 Proceedings
North American Conference on Savannas and Barrens

The results reported here reflect ongoing work to conserve and restore the biodiversity of a degraded tallgrass savanna, one of the rarest major natural communities in North America. The work was performed by the North Branch Prairie Project, a volunteer group, on land owned by the Forest Preserve District of Cook County. The savanna is an important natural community in the midwest (Curtis 1959; Nuzzo 1986). In the 1970s, the Illinois Natural Areas Inventory (INAI) (White 1978) searched for classic savannas (called "Oak Openings" by Curtis) and found essentially no undisturbed remnants. According to the definition used by Curtis, canopy cover in savanna could reach a maximum of 50%. For oak communities of greater than 50% Curtis recognized dry and dry-mesic forest. Mesic and wet-mesic oak stands of greater than 50% canopy were not treated as separate from maple forest by Curtis.

The INAI recognized savannas with canopies up to 80%. This definition helped conservation efforts focus on a component of the continuum that was not obvious in Curtis' classification. But the INAI definition seemed to imply that even with the denser canopies, the dominant herbaceous vegetation would be prairie grasses. Our efforts to restore such land (the 50% to 80% canopy savanna, referred to here as "closed savanna") became inadvertently experimental, and we began collecting data systematically only after the management process had begun. Because there were no known healthy models and only fragmentary historic information on which to base our restoration efforts, we planted closed savanna areas with seed of species drawn from a variety of nearby habitats including prairie, brushland, woodland edges, and interior oak woodlands. We hypothesized that a diverse turf of native herbs would eventually coalesce in a relatively stable community and would compete successfully with the species of weeds and brush associated today with degraded woodlands. We further hypothesized that some of those species would be relatively conservative ones. The herb community on the site after seven years of restoration was still changing rapidly, but the vegetation showed trends that may have some implications for future savanna research and conservation.

STUDY AREA

Vestal Grove, the study area, is within the 36 ha (90 A.) Somme Prairie Grove Forest Preserve in Northbrook, Illinois. This preserve spans the west-facing slope of the Deerfield Lobe of the Lake Border Moraine. According to the Public Land Survey of 1838 the indigenous landscape contained scattered oaks. Soils are Markham silt loam (2 to 5 percent slopes) and Beecher Silt Loam (U.S.D.A. 1979).

Vestal Grove itself measures approximately 200 by 75 meters and is surrounded by dense brush and old-field pasture vegetation. The trees of the grove are mostly bur oak (Quercus macrocarpa) with lesser numbers of Hill's oak (Quercus elipsoidalis), shagbark hickory (Carya ovata) and black cherry (Prunus serotina). Prior to management, the understory was overwhelmingly European buckthorn (Rhamnus cathartica) with little herbaceous vegetation of any kind in the deep shade beneath it.

Initial restoration efforts (1983 and 1984) consisted of spring burning. The oak leaf fuel supported fires with flame-lengths of 0.1 to 0.25 m. These fires were eventually sufficient to topkill most of the buckthorn in the grove. The fires did not carry through the non-oak brush borders to the pasture grasslands. The soil surface remained mostly bare during these two years. Few native species typical of relatively stable habitats appeared. However, large numbers of individuals of alien and native weedy species were increasingly evident.

METHODS

Beginning in fall 1985, transect sampling of the herb layer was begun. At this same time we began an annual seed broadcast in an attempt to restore a native herbaceous matrix and to determine which species were best adapted for the early stages of the restoration of such lands. About two thirds of the grove was burned in each year. Weeds thought to be a threat to the restoration were reduced by pulling and scything; these included garlic mustard (Alliaria officinalis), burdock (Arctium minus), briers (Rubus sp.), and tall goldenrod (Solidago altissima).

Seed was gathered from within a radius of 25 miles. Species included all those in the North Branch Prairie Project mesic and wet mesic prairie seed mixes and new mixes created for mesic and wet-mesic areas of both "open" and "closed" savanna. We defined open savanna as that in the 10% to50% canopy range and closed savanna as that in the 50% to 80% canopy range. Because the open savanna areas initially had insufficient fuel to burn in most years, this portion of the continuum on this site was much slower to respond to our restoration efforts. This paper, accordingly, focuses on the 50% - 80% part of the continuum. Seed from the mixes representing all three continuum components (i.e., prairie, open savanna, closed savanna) was broadcast throughout the burned area in the weeks following the spring or fall burn throughout the period (1985-1991) reported on in this paper.

In 1986 four random transects were established in Somme Prairie Grove for the sampling of a wide variety of biota (trees, shrubs, herbs, cryptogams, invertebrates, birds, and small mammals). One of these traversed the center of Vestal Grove on its long axis. Circular herbaceous plots on this random line were 1 m2 and at 10 m intervals. All herbaceous and those woody plants with foliage below one meter were recorded by species and an estimate of cover. In order to represent both early season and later season flora, this transect was sampled in late spring and late summer. To make a single sample, the species lists from the two samplings were combined and, if two different cover values were recorded, the larger was used. Fortunately, it possible for us to assemble approximate pre-seeding vegetation data for this permanent transect. A transect arbitrarily placed and sampled in late summer 1985 was by chance parallel to and just a few feet from the random permanent transect. Thus, the 85/86 sample combines the very similar data from the two nearby transects.

RESULTS

Two sets of data are presented here. In each case, the data are summarized from the 19 quadrats that fell within the 50% - 80% canopy portion of the study area. Table 1 compares importance values for the twenty species with the highest such values (cumulatively in the three samples) at the start of seeding and after three and five growing seasons.

The data in Table 2 compares four measures of the biodiversity of the quadrats at five times between 1985 and 1991. In each case the average value for the nineteen plots is given.

In the case of the first measure, all species, the number given includes all vascular plant species found in a quadrat, alien and native, herbaceous and woody so long as those woody species have foliage in the plot within one meter of the ground.

The second measure, native species, includes only those species recognized as native to the Chicago region by Swink and Wilhelm (1979).

The third measure, NARI, is the Natural Area Rating Index as described by Swink and Wilhelm, with the modification that the value for each quadrat is computed as if that quadrat were an entire natural area. The NARI is an attempt to characterize the quality of vegetation through a formula that yields higher numbers in proportion to the number of native species and the degree to which those species are characteristic of high quality natural communities. For use in the formula, each species in the flora has been subjectively assigned a number proportional to the conservativeness of that species as judged by Swink and Wilhelm (1979).

The fourth measure, weighted NARI, computes a number similar to the above except that the contribution of a given species to the quadrat average is proportional to the percent cover recorded for that species in the plot.

DISCUSSION

The data in Table 1 indicate that the most weedy (as reflected by their low "coefficients of conservatism") are in rapid decline. Examples include European buckthorn and nightshade (Solanum dulcamara).

Most species of relatively low conservatism also are in decline. But some including tall goldenrod (Solidago altissima) and blackberry (Rubus allegheniensis) are rapidly increasing. Such highly aggressive native species are thought by some restorationists to represent a serious threat in the early stages of attempts to restore highly degraded savannas.

The most conservative of these twenty species (cc > 4), with some exceptions, are increasing rapidly. Conservative increasers include sweet black-eyed susan (Rudbeckia subtomentosa), purple joe pye weed (Eupatorium purpureum), and wood reed (Cinna arundinacea). All the rapidly increasing and relatively conservative species in Table 1 were absent from the grove according to the plant lists made before burning and prior to seeding. No characteristic prairie species (with the possible exception of the Rudbeckia) appears in the list of the twenty most important herbs in this closed savanna, despite the broadcast of prairie seed throughout the grove. Rudbeckia does not occur in high quality prairies in northeastern Cook County. Seed for this species was found and gathered only at disturbed remnant savannas.

Table 2 gives the average species richness or natural quality of the nineteen 1 m2 quadrats that fell within the closed savanna. Numbers of species and native species initially increased and then fell, with the number of native species in 1991 nearly twice that in 1985.

Values for the NARI and the weighted NARI have continued to climb. The 1991 values are respectively 4 times and 6 times the 1985 values, a striking and persistent trend which suggests that these figures may reflect phenomena not revealed in the figures for numbers of species and may also suggest possible confirmation of the hypothesized trend toward a relatively stable turf with good numbers of conservative species.

ACKNOWLEDGMENTS

This work would have been impossible without the fine work of the burn teams, seed gatherers and other restoration practitioners of the North Branch Prairie Project. Transects were designed and laid out with the help of Steven Apfelbaum and Alan Haney. Data was gathered by the authors, Steven Apfelbaum, Jane Balaban, and Laurel Ross. Statistics were calculated and analyzed by John and Jane Balaban.

LITERATURE CITED

Curtis, J. 1959. The Vegetation of Wisconsin: An Ordination of Plant Communities. University of Wisconsin Press, Madison, WI.

Swink, F. and G. Wilhelm. 1979. Plants of the Chicago Region. The Morton Arboretum. Lisle, IL.

Nuzzo, V. A. 1986. Extent and Status of Midwest Oak Savanna: Presettlement and 1985. Natural Areas Journal. 6:6-36.

U.S.D.A. 1979. Soil Survey of DuPage and part of Cook Counties, Illinois. Illinois Agricultural Experiment Station Report No. 108.

White, J. 1978. Illinois Natural Areas Inventory Technical Report. Illinois Natural Areas Inventory. Urbana, IL.

Table 1.  Importance values for the twenty most important species in 19 1 m2 quadrats in Vestal Grove (cc = coefficient of conservatism from Swink and Wilhelm 1979).