AN ECOLOGICAL AND PALEOECOLOGICAL STUDY OF THE VEGETATION IN THE BIG WOODS REGION OF MINNESOTA.

摘要

The Big Woods was a large area of deciduous forest along the prairie-woodland border in south-central Minnesota. Records of the original land surveys were used to reconstruct the presettlement vegetation, and fossil-pollen analysis was used to reconstruct the vegetation for the past several thousand years.;Computer plots of bearing trees were used to map the vegetation of the Big Woods region. The characteristic trees of the Big Woods were elm, basswood, ironwood, sugar maple, ash, hickory, and butternut. Although the Big Woods has been described as a sugar maple-basswood forest, the most abundant tree was elm. Basswood and ironwood were widespread. Sugar maple had a more restricted distribution, occurring in the areas of lowest fire frequency. Oak scrub occurred on rolling, well-drained topography south and east of the Big Woods; while aspen-parkland occurred on level, poorly drained topography west of the Big Woods. The vegetation is poorly correlated with the regional climatic gradients. Fire frequency strongly controlled the overall pattern of the vegetation. The primary variables influencing fire frequency were water bodies, topography, and the existing vegetation. Firebreaks sharply formed the prairie-woodland border. Fire-tolerant oak and aspen woodlands formed a buffer zone between the Big Woods and prairie, except where very effective firebreaks formed sharp boundaries between them. The oak woodland along the prairie-woodland border has frequently been described as savanna; however, the oak-dominated vegetation in the Big Woods region was dense scrub. A review of the literature reveals that the interpretation of savanna in other areas is questionable.;The pollen and sediment stratigraphy of three sites was investigated--Wolsfeld Lake and Wolsfeld Woods Marsh in the northeastern Big Woods and French Lake near the western margin of the Big Woods. Most of the Big Woods was prairie during the mid-Holocene. Oak woodland began invading prairie about 5000 years ago in the eastern Big Woods, then expanded westward, becoming established at French Lake 2400 years ago. Oak woodland persisted until 300 years ago, when ironwood, elm, basswood, and eventually sugar maple expanded and became dominant. The changes from prairie to oak woodland and from oak woodland to "bigwoods" must have required reductions in fire frequency, probably caused by increased precipitation and decreased temperature. About 100 years before the expansion of bigwoods sediment influx increased at the coring site in Wolsfeld Lake and continuous peat began to accumulate in Wolsfeld Woods Marsh, supporting the hypothesis that increased precipitation triggered the expansion of bigwoods. A conceptual model is presented, in which climatic changes trigger vegetation changes, but interactions among the existing vegetational pattern, fire, and topography cause climatic thresholds for vegetational change to vary in space and time.;The bearing-tree data in the land-survey records have been widely used for reconstructing vegetation. Contrary to the assumption of many ecologists, however, the bearing trees were not necessarily the closest trees to the corner posts but, in fact, were a biased sample of the vegetation. The purpose of bearing trees was to facilitate relocation of the corners, and the surveyors chose the trees most likely to achieve this purpose. The most common misuses of the data have been purported statistical analyses of surveyor bias and calculations of absolute tree density. These require the unlikely and untestable assumption that trees were randomly arranged throughout the entire area from which the trees are a sample. Reliable calculations of absolute tree density also require the assumption of no surveyor bias. Despite these limitations, if used with a proper perspective towards their original purpose, the land-survey data are extremely valuable for reconstructing vegetation.