1993 Proceedings of the Midwest Oak Savanna Conferences
A PRE- AND POST-BURN VEGETATION DATA AND MANAGEMENT PLAN FOR RECONSTRUCTING A SAVANNA
During the Illinois Natural Areas Inventory, no intact, deep-soil, mesic savanna was found in the state, and only one high quality dry-mesic savanna was found. It is important that we start to study the remnants of these areas while some still exist, so that reconstruction efforts will have some logical basis. The present study was undertaken to develop a management plan, and start the development of a mesic savanna community at Forest Glen Preserve, Vermilion County, Illinois. The study involved a pre-burn inventory of the vegetation, and the changes that occurred after burning. It also involved post-burn survival of woody species. These studies indicated that oaks are well adapted to surviving burns by resprouting from the base or trunk. Also, most shrub species, although top-killed after fire, do resprout, as do many weedy tree species and woody exotics. A set of guidelines have been developed for savanna reconstruction.
Defining the word savanna, as used by natural areas botanists, is a very difficult task. A broad description would be an area with open-growing trees, growing as scattered individuals or in small groves, with an herbaceous, primarily grassy understory (Bray 1955). During the Natural Areas Inventory of Illinois the definition used was based on canopy cover and soil moisture (White 1978). Canopy cover ranged from 10-80%, soil moisture was described as mesic, dry mesic, and sand. Much controversy over the definition has taken place since then. Reviewing the General Land Office survey records adds to the confusion for the surveyors did not use a standardized terminology. Different surveyors used different terminology for the same community and the same terminology to describe different vegetation types. Some of these terms were oak openings, oak barrens, oak scrub, brush prairie, scattered timber, open woodland and prairie with scattered trees (Nuzzo 1986). Anderson and Brown (1983) and Anderson (1991a), and many others, proposed that savannas are not a separate ecosystem but an ecotone between two ecosystems, prairie and forest; to elaborate, a transition zone between the western prairies and the eastern deciduous forests. When considering the transient, fluxuating nature of the savanna, this seems most appropriate, for this community is totally dependent on outside factors for it's existence. To condense this into a definition, savannas are an ecotone dominated by oaks, having between 10-80% canopy cover, with or without a shrub layer, and an herbaceous, predominantly grassy ground layer composed of species associated with both prairie and forest communities. It is fire dependent, and in the absence of fire or extreme droughty conditions will convert into a forest community.
Some of the factors that shaped or created savannas were fire, slope, climate, and soil (Anderson 1991b). The fire frequency determined whether trees would be able to grow large enough to withstand being top killed, thus growing beyond the seedling stage. In areas where fire occurs annually, few trees got past seedling stage. If fires occurred every 3-5 years some trees would be able to reach a size at which they were fire resistant. Early settlers noted that within twenty to forty years after the frequent fires were stopped, the savannas appeared as closed oak forests.
Slope, topographic relief, controlled the spread of the fire. Down hill slopes do not carry fire well. When vegetation burns there is a rising convection air current from the heat that is being produced. Fire will move up hill quickly, lifted by its own current, conversely when moving down hill the rising air current works against the fire, sometimes causing it to burn out. Also, land with greater slope tends to have more drainage (i.e. streams) than flat land. These act as fire breaks, reducing the fire frequency in the area.
Climate also works in conjunction with fire. During times of drought vegetation will be dryer especially in the fall, and will burn hotter and more completely than in a wet year. During wet years water in depressions and low lying areas causes the fire to burn out or skip areas, resulting in less complete burns and more chance for trees to grow.
Poor soils, like that of sand savannas, do not support the large biomass growth that deep soils support. Fires through these areas would be less intense, allowing for trees to grow. Fires through deep soil areas that support large biomass growth would be of higher intensity, killing the trees.
Pre- and post-burn vegetation studies were done on a savanna reconstruction site at Forest Glen Preserve, Vermilion County, Illinois. Herbaceous frequency data and woody frequency and density data were taken during the summer and fall of 1991. The site was burned in March of 1992. The measurements were repeated in the summer and fall of 1992. Fisher's Exact Test was run on both the frequency and density data.
Sixty-six herbaceous species were recorded for the study area. Of these, seventeen showed significant frequency differences. Those significantly increased after the burn were: Agalinis tenuifolia, Ambrosia artemisiifolia, Cassia fasciculata, Carex sp., Desmodium glabellum, Dichanthelium acuminatum, Euthamia graminifolia, Lespedeza virginica, Potentilla simplex, Solidago juncea, Sorghastnim nutans, and Trifolium pratense.
After the burn, bare spots of soil were created where litter had been. The annuals and more aggressive species were able to colonize theses areas. Those that were significantly decreased after the burn were: Dactylis glomerata, Dianthus armeria, Eupatorium serotinum, Hypericum punctatum, and Oenothera biennis. The first four are not fire tolerant species, however the decrease of Oenothera biennis was unexpected.
Only two species of shrubs significantly increased in frequency after the fire (Table 1). These were Rubus allegheniensis (blackberry) and Rubus flagellaris (dewberry). Most of this increase was due to root sprouts. There were no significant decreases.
Trees that were less than 2.5cm in diameter at breast height were considered seedlings, those that were 2.5cm and larger were considered saplings. Juglans nigra (black walnut) was the only tree seedling which showed a significant difference in frequency, it decreased after the burn. It is not a fire tolerant species (Table 1). Both Juglans nigra and Prunus serofina (wild black cherry) decreased significantly in density after the burn. They are not fire resistant species. Sassafras albidum and Malus ioensis (wild crabapple) increased significantly in density after the burn, mainly because of root sprouts.
Only one saplings tree species showed a significant difference after the burn (Table 1). Prunus serotina was significantly deceased in frequency and density. Again this tree is not fire resistant.
The total stem numbers for the tree seedlings were not significantly different after the burn. However the stem numbers for the tree saplings greatly decreased after the burn. This reflected the number that were top killed.
During the summer of 1992, several species of trees on the savanna site were surveyed to determine their ability to survive and or resprout after the burn. This data was compiled into four categories: (1) not killed, lower limb damage sometimes occurred; (2) killed, not resprouting the next spring; (3) top-killed, resprouting the next spring from the base; and (4) top-killed, resprouting the next spring from the trunk (Table 2).
In general, the oaks were relatively resistant to the fire, particularly once the individuals reached a height of 2 to 4 m. Quercus macrocarpa (bur oak) was the most resistant, with all of the individuals over 1 m in height surviving and not being top killed. Quercus alba (white oak) individuals also survived the fire with relatively little damage once reaching a height of 3 m. Small individuals were commonly top killed, but did resprout from the base or the trunk. The remaining oak species (shingle, red and black) were less resistant, with some 4 m tall individuals being top killed.
Juglans nigra (black walnut) exhibited patterns similar to the shingle oak. Below 3m the majority were top killed resprouting from the base and trunk. At 3-4m and > 4m the individuals were not killed. This species shows a great ability to resprout and appears to be fire resistant above 3m.
During the Natural Areas Inventory of Illinois, there were no grade A, deep soil, mesic savannas identified in the state (White 1978). In fact Nuzzo reported in 1986, that to date, no intact grade A, deep soil, mesic savannas had been located in the eight midwestern states she surveyed. For this reason it is imperative that the existing remnants be restored and reconstruction efforts started. A list of ideas to be considered and steps to be taken has been compiled to help those interested in starting such efforts.
Anderson, R. C. 1991a. Savanna concepts revisited. BioScience 41:371.
Anderson, R. C. 1991b. Presettlement forests of Illinois. Pages 9-19 in G. V. Burger, J. E. Ebinger, and G. S. Wilhelm, eds. Proceedings of the oak woods management workshop. Eastern Illinois University, Charleston, Illinois.
Anderson, R. C. and L. E. Brown. 1983. Comparative effects of fire on trees in a midwestern savanna and an adjacent forest. Bulletin of the Torrey Botanical Club 110: 87-90.
Bray, J. R. 1955. The savanna vegetation of Wisconsin and the application of the concepts of order and complexity to the field of ecology. PhD Thesis. Univ. Wisconsin, Madison, WI.
Nuzzo, V. A. 1986. Extent and status of midwest oak savanna: Presettlement and 1985. Natural Areas Journal 6:6-36.
White, J. 1978. Illinois natural areas inventory technical report, VoL I: Survey methods and results. Illinois Natural Areas Inventory, Urbana, Illinois. xix+426p.