Vegetation and Tree Species Patterns near the Northern Terminus of the Southern Floodplain Forest

The bottomland forest on Horseshoe Lake Island, located on the Mississippi alluvial plain in Alexander County, Illinois in comprised of 2 stands, one relatively undisturbed and the other which is recovering from disturbance in the late 1800s or early 1900s. In both stands vegetational structure, gradient relationships, diversity and size—class characteristics were studied. In the old—growth stand, 35 soil—site variables were measured or estimated for interpretation of vegetational patterns and species distributions. Both direct and indirect gradient analyses were evaluated for use in floodplain forest studies and ultimately the indirect approach with multiple regression interpretation was used to construct a gradient model of the vegetation. Basal area and density of the old—growth and floristic composition of both stands are typical of much of the original Southern Floodplain Forest. The coenocline was divided into 3 segments, mesic, transition and wet with the first and last corresponding to the ridge and hardwood bottoms described by Braun (1950). The important species on the ridge bottoms are Acer saccharum, Asimina triloba, Liquidambar styraciflua, Ulumus rubra, Quercus rubra and Tilia americana and are associated with well—drained soils and infrequent flooding. The transitional segment is dominated by Asimina triloba, Liquidambar styraciflua, Quercus michauxii, Quercus muhlenbergii and Ulmus americana with Quercus pagodaefolia and Quercus shumardii as associates on moderately heavy and poorly drained soils with intermediate flooding. The hardwood bottom sites are dominated by Acer rubrum, Fraxinus pennsylvanica, Liquidambar styraciflua and Ulmus americana and may be flooded for several months each year and have heavy textured, poorly drained soils. Multiple regression analysis with ridge regression revealed that Fraxinus americana and Acer saccharum have wider tolerances to flooding and poor aeration than previously reported. Importance Values of several species were related to duration and depth of flooding, soil mottling and soil texture indicating that distributions are affected by a site—inundation, soil drainage—aeration complex. Diversity of the tree stratum is comparable with that found in some of the mixed mesophytic forests elsewhere. Due to competition from Asimina triloba and Aesculus discolor and possibly some environmental limitations due to the northern location of the stands, the shrub—sapling stratum is depauperate. Along the coenocline from mesic to wet sites density and equitability decrease while basal area increases. Size—class analysis indicates that both the negative power and negative exponential functions fit the size—class distribution for the 2 stands but that neither is entirely appropriate. The old—growth stand is characterized by 3 groups of individuals each with different mortality rates which give a plateaued deviation from each model. The secondary stand is more balanced and does not exhibit a well developed plateau. A comparison of individual species reveals that various regeneration patterns are present which range from uneven—aged (sized) in the case of Acer saccharum to even—aged for Fagus grandifolia, Quercus rubra and Ulmus rubra. The plateaued pattern in the old—growth stand does not reflect past disturbance but apparently is a characteristic of this forest. The principal mode of regeneration in this stand is gap—phase replacement.