Printer Friendly

Changes in the floodplain forest vegetation of the lower Wisconsin river over the last fifty years.

INTRODUCTION

Temperate river-floodplain forest systems represent some of the most diverse and productive habitats on earth (Brinson, 1990; Decamps, 1996; Ward et al., 2002). Additionally, they provide many ecologically and socially valuable functions, such as flood protection, water filtration, habitat for wildlife, migration corridors and recreation (Costanza et al., 1997; Decamps, 1996; Naiman and Decamps, 1997; Petts, 1990; Pinay et al., 1991; Postel and Carpenter, 1997; Ward and Stanford, 1995). The high level of species and functional diversity result directly from periodic flood events that characterize natural systems. However, systems with human alteration of rivers and floodplains often experience reductions in the occurrence and magnitude of flood events (Dynesius and Nilsson, 1994). Further, many forested floodplains have been cleared for other uses (Tockner and Stanford, 2002). Remaining floodplain forests have experienced other anthropogenic disturbances, such as logging, introduced disease, increased herbivory and invasive species (Naiman and Decamps, 1997; Petts, 1990; Tickner et al., 2001). Understanding how these factors have altered floodplain forest ecology is essential to the proper conservation and management of these remaining forests. In this study, we examine how the vegetation of the floodplain forests of the Lower Wisconsin State Riverway has responded to multiple ecological and anthropogenic factors during the last 50 y.

In natural river-floodplain forests, periodic flooding events create a varied microtopography that includes natural levees, ridges, swales and backwaters (Brinson, 1990; Hupp and Osterkamp, 1996). Small differences in elevation from site to site translate into large differences in the frequency, duration and magnitude of flooding. As flood tolerance is the primary factor determining the vegetation at a given site in these systems, the variation in flooding levels across the floodplain allows a wide range of species with different levels of flood tolerances to establish and coexist within a relatively small area (Barnes, 1978; Brinson, 1990; Kozlowski, 2002b; Menges and Waller, 1983; Naiman and D6camps, 1997). Flooding events also result in sediment deposition that alter local elevation (Hodges, 1997). Differential deposition rates across the floodplain also influence species composition at a given site. Thus, in contrast to many upland forests where light availability and species' shade tolerance often drive forest ecology (Oliver and Larson, 1996), flooding and sediment deposition primarily determine floodplain forest composition and succession. Flooding can even alter an individual species' shade tolerance (Hall and Harcombe, 1998). Thus when the natural flooding regime is reduced, light availability and a species' shade tolerance should become much more important for species regeneration and succession in floodplain systems.

Floodplain forest dynamics are more complex than simple interactions between flood and shade tolerance, particularly where humans are active in the forests. Natural disturbances such as wildfires and wind throws have historically played a role in these forests. Floodplain forests are also subject to timber harvests, which can influence light availability, species distribution and regeneration (Deconchat and Balent, 2001; Kozlowski, 2002a; Lockaby et al., 1999; Messina et al., 1997). Introduced diseases can alter community composition: in the Northern U.S., Dutch elm disease decimated elm populations, which historically were a major species in temperate floodplain forests (Burns and Honkala, 1990; Dunn, 1986; Hughes and Cass, 1997). Increases in herbivore populations, such as the white-tailed deer (Odocoileus virginianus) can also affect community composition (Waller and Alverson, 1997). Studies have shown that browsing by white-tailed deer can hinder tree regeneration and alter forest composition (Boerner and Brinkman, 1996; Rooney, 2001; Shimoda et al., 1994), and at least one study demonstrates browsing tolerance to be more important than shade tolerance in forested areas with excessively high deer densities (Collins and Carson, 2003). Finally, the establishment of invasive species can alter forest composition and in some cases, hinder regeneration of native species (Hughes and Cass, 1997; Knutson and Klaas, 1998; Tickner et al., 2001).

In recent decades, there has been a growing recognition of the need to conserve remaining floodplain forest complexes for both their ecological and social values (Decamps, 1996; Kozlowski, 2002b; Ward et al., 2002). To do so properly, we need to understand how forests respond to the factors mentioned above. This study investigates the changes in the tree, understory and ground layers of several of the temperate floodplain forests of the Lower Wisconsin over the last 50 y. Our analysis addresses the questions of how the composition and structure of the vegetation have changed in the last 50 y, in particular, with respect to species' flood and shade tolerances, species' herbivore defenses and the presence of invasive species.

METHODS

STUDY AREA

The Lower Wisconsin flows westward from the last dam on the Wisconsin River 143 km to the Mississippi River (Fig. 1). Its climate is continental, with cold winters and warm summers. The mean annual temperature is 7.8 C, ranging from -6 C in the winter to 20 C in the summer (Lindstrom and Young, 2002); average annual precipitation is 825 mm. The average river discharge at Muscoda, located approximately halfway between the last dam and the confluence with the Mississippi, is 247 [m.sup.3]/s. Annual floods typically occur in spring, in concert with snowmelt.

[FIGURE 1 OMITTED]

The Lower Wisconsin flows freely and contains one of the largest complexes of floodplain forests left in the state (Wisconsin Department of Natural Resources, 1988). Forests occupy approximately 50% of the river's floodplain and, in contrast to other areas upstream, their extent has remained relatively constant since the 1930s (Freeman et al., 2003). Curtis and his students found that the floodplain forests in the 1950s were typically initiated by black willow (Salix nigra) and eastern cottonwood (Populus deltoides) on recently scoured or deposited substrates (Curtis, 1959; Ware, 1955). In forest gaps, river birch (Betula nigra) and swamp white oak (Quercus bicolor) also functioned as early successional species. When river dynamics allowed, these communities were succeeded by communities dominated by American elm (Ulmus americana), green ash (Fraxinus pennsylvanica var. lanceolata) and silver maple (Acer saccharinum). Woody understory was sparse in the 1950s (Ware, 1955). According to Curtis (1959), the ground layer of these wet forests was dominated by wood nettle (Laportea canadensis), poison ivy (Toxiocodendron radicans), river bank grape (Vitis riparia), white grass (Leersia virginica) and Virginia creeper (Parthenocissus quinquefolia). In contrast, the ground layer of the wet-mesic forests was dominated by Virginia creeper, wood nettle,jewelweed (Impatiens capensis), jack-in-the-pulpit (Arisaema triphyllum) and blue marsh violet (Viola cucullata).

Although the Lower Wisconsin flows unregulated, 26 dams and 21 reservoirs lie on the Wisconsin upstream. The last and largest of the reservoirs were completed in the 1950s (Krug and House, 1980). Studies by Krug and House (1980) and Pfeiffer (2000) show that river regulation has decreased flooding on average 10-15% since river regulation was completed in the 1950s, sufficient to alter the vegetation of the floodplains. At the same time, the forests of Lower Wisconsin have experienced other anthropogenic impacts. Timber harvests have been common practice in the region since European settlement. Dutch Elm disease arrived in the 1950s and devastated regional elm populations (Dunn, 1986). Invasive species are also affecting the Lower Wisconsin, including European buckthorn (Rhamnus cathartica), Asiatic honeysuckle species (Lonicera spp.) and reed canary grass (Phalaris arundinacea). Finally, Wisconsin's deer populations have grown dramatically since methodical surveys began in 1960 (McCown and Wallenfang, 1998). In the Southern Farmlands region, which includes the Lower Wisconsin River, deer populations doubled between 1980 and 2001 (Rolley, 2005). In 2001, the overwintering surveys reported deer densities of 9.3.-19.3 deer/[km.sup.2] along the Lower Wisconsin from, above the state's management goal of 7.7-11.6 deer/[km.sup.2] for these areas (Wisconsin Department of Natural Resources, 2002).

For this study, we resampled five sites along the Lower Wisconsin where Curtis and his students surveyed the vegetation of the floodplain forests in the early 1950s (Table 1). Three of these sites have experienced significant disturbances since Curtis's surveys: the forests at Ferry Bluff and Helena experienced differing levels of timber harvests; the forest at Arena was partially cleared for a parking lot, which required us to shift part of our sampling into the forests adjacent to Curtis's site.

DATA COLLECTION

Historical data.--We acquired historical data from Curtis's 1950s survey from the Department of Botany at the University of Wisconsin-Madison. Ware (1955), a student of Curtis, reported that sites were chosen if the stands were greater than 10 acres (4.05 ha), were of natural origin and had had no recent disturbance. In order to avoid edge effects, they did not sample within a 25-yard (22.9 m) buffer along the edges of the forest. For most stands, they sampled 40 points. For sites judged to be uniform, they only sampled 20 points (Lone Rock). At Helena, they sampled in two areas, for a total of 60 points. They defined the tree layer as those trees with a diameter at breast height (dbh) of greater than 9.9 cm (12 [in.sup.2]). They employed a plotless sampling technique, the random pairs method, to survey the tree layer of the floodplain forests (Cottam and Curtis, 1956). For each tree, they recorded species and dbh. For each pair, they recorded the inter-tree distance, which permits the estimation of stem density. For the understory, they recorded the species of the four woody stems nearest to the sampling point smaller than 9.9 cm basal area (Ware, 1955). The exact methodology for the understory is ambiguous, as some points lacked data on some or all of the four stems and none of the data included shrub species. Further, they did not record distance measures for the understory, rendering it impossible to calculate an absolute estimate of its density. For the herb layer, they recorded all species found in 1-[m.sup.2] plots located at the sampling points. In all, they sampled 20 herb plots in each site. Since they did not sample herb plots at Arena, we excluded this site from the analyses of the herb-layer.

Data collection for this study.--For this study, we resurveyed the five Curtis sites in the summer of 2001, with changes in sampling locations as noted in the study site section above. We defined the tree layer as trees with a diameter at breast height (dbh) greater than 10 cm and the understory as stems taller than 1.5 m with a dbh of 1-10 cm. We surveyed the tree layer using the point-quarter method, another plotless method considered comparable, but slightly superior to the random pairs method (Cottam and Curtis, 1956; Ware, 1955). It provides data on species, dbh and distance of the individual from the sampling point. We sampled all understory stems in 1 x 10 m quadrats centered on each sampling point, with the long side of the quadrat running parallel to the flow of the river. We recorded herb species present in two separate 1-[m.sup.2] plots randomly located in the quadrants surrounding each sampling point. We located sampling points along transects using a stratified random sampling design outside of a 25 in buffer at the edges of the forest. In all, we sampled 40 points per site.

Data analysis--tree and understory layers.--Since the number of sampling points differed at two sites (Helena and Lone Rock), we used STATA to create a random sample of appropriate size from the larger 2001 sample. Thus, we drew 40 points randomly from the 60 Curtis sampled at Helena and 20 points randomly from the points sampled at Lone Rock. To analyze the changes that have occurred since the 1950s, we calculated summary data statistics for each site and in each time period, including mean basal area per tree, basal area per hectare, tree stein density and understory stem density (2001 only). To test for a difference in the mean basal area per tree across time, we log-transformed the data and performed a t-test. For each time period, we also calculated several measures of diversity for both the tree layer and the understory: species richness (S, number of species per site), the Shannon-Weaver index (H'), and Pielon's J, a measure of evenness (Ludwig and Reynolds, 1988). We calculated importance values (IV) for each species for each time period. For the tree layer, the IV is the sum of relative density and relative dominance (basal area); for the understory, the IV is the sum of relative density and relative frequency. We group two sets of species for this analysis. As Ware (1955) noted that species identified early in their study as Fraxinus americana, were likely F. pennsylvanica (F. americana was much more common in Eastern Wisconsin), we lump occurrences of F. americana with F. pennsylvanica in the 1950s data. Similarly, Ware noted that Ulmus americana and U. rubra often hybridize and are difficult to identify to species in the field. Thus, we lumped both Ulmus species together and analyze them as such for both the 1950s and 2001 data.

We examined changes in importance for the three most dominant species per site and time period, as well as for three groups of species: characteristic pioneer species (as defined by Curtis, 1959) historical dominant canopy species (Curtis, 1959) and mesic species not typically associated with lowland/floodplain forests in southern Wisconsin. We also analyzed the change in the appropriateness of the community species mix for a wetland environment. To accomplish this, we calculated a community wetland coefficient using the individual species wetland coefficient for region 3, as defined by US Fish and Wildlife (National Plant Data Center, 2002). This number ranges from -5 (obligate wetland species) to 5 (obligate upland species). From these values, we calculated a community weighted average for each time period/site, based on the corresponding importance value for a given species.

Data analysis--ground layer.--We analyzed the ground layer by site in a similar fashion to the other two vegetation layers. No data existed for the Arena site from the 1950s, so we excluded that site from ground layer analysis. Since the sampling effort for this study resulted in four times as many sampling plots per site as the Curtis study, we again used STATA to choose 20 random plots from the 80 sampled at each site in 2001. For all of our analyses, we grouped certain species together in order to avoid potential differences in species identification across time periods. We lumped all species of Carex (sedges) and Helianthus (sunflowers), respectively, as Curtis only identified the abundance of specimens to genus in the 1950s data. We also lumped Viola spp. (violets); without flowers, these specimens were hard to identify to species. We calculated several measures of diversity: mean species richness (S, number of species per site), the Shannon-Weaver index (H') and Pielou's J. We examined changes in the five most dominant species (determined by plot frequency) per site. We also examined changes in the overall frequency of species that Curtis (1959) identified as modal species for southern lowland forests. Curtis defined modal species as those species that associate with a specific habitat type more than other habitat types in Wisconsin (in this case, the southern lowland forests, which included the floodplain forests of the Lower Wisconsin River and other smaller systems, but also covered the hardwood swamp subtype). We used Jull (2001) to examine changes in species based on deer browsing preference. Lastly, we calculated a community wetland coefficient in the same manner we did for the tree and understory layers; for the ground layer, we used relative frequency per site to weight the values for each species.

For all of these analyses, we related observed historical changes to the history of tree harvests recorded at the individual sites. Nomenclature follows Gleason and Cronquist (1991). For all statistical analyses, we employed an alpha level of 0.05.

RESULTS

Site characteristics.--The five sites sampled present a range of anthropogenic disturbance levels at a variety of elevations in the floodplain. The sample sites at Helena and Ferry Bluff appear to have experienced greater levels of (recorded) human disturbance through vegetation removal since the sampling by Curtis and his students, Arena less so (Table 1). The sites at Lone Rock and Wyoming Bluffs in contrast have not experienced timber harvests since before the Curtis sampling. (NB: subsequent tables and figures show sites arranged in order of decreasing levels of human disturbance). Sites also differed in mean elevation (Fig. 2): Helena, Arena, and Lone Rock all lay between 3.2 and 3.3 m above river bed, Ferry Bluff lay at mean elevation of 2.2 [+ or -] 0.08 m (SE) and Wyoming Bluffs at about 1.9 [+ or -] 0.07 m above the river bed.

Forest characteristics.--All sites surveyed have experienced a significant decrease in the mean basal area of adult trees: ranging from a 0.059 [m.sup.2] decrease at Ferry Bluff to a 0.014 [m.sup.2] decrease at Lone Rock (test statistics in Table 2). No trend was evident with respect to human disturbance (tree removal), although Wyoming Bluffs, the site with the longest history without timber harvests, also had the largest mean basal area. Changes in basal area per hectare were more variable, ranging from a decrease of 34.4 [m.sup.2]*[ha.sup.-1] at Ferry Bluff to an increase of 18.7 [m.sup.2]*[ha.sup.-1] at Helena. Again there was no apparent trend with history of tree harvests. Tree density follows a similar trend, ranging from a decrease of 119 stems per ha at Ferry Bluff to an increase of 268 stems per hectare at Helena. Although historical data were not available to assess changes in understory stem density, the current levels range from 225 stems per hectare at Wyoming Bluffs to 2750 stems per hectare at Helena.

[FIGURE 2 OMITTED]

Changes in tree layer species diversity generally corresponds to human disturbance (Table 3). The two harvested sites, Helena and Ferry Bluff, had the greatest H' and both had increased since the 1950s. Within the tree harvest categories, H' seemed to be greater at the higher elevation sites. Trends in understory species diversity were similar with respect to elevation, but not with disturbance, although the two extreme H' values occurred on the most and least disturbed sites (Table 3). Likewise, diversity at the ground layer reflected the elevational influence, but did not show any clear trends with human disturbance. Unlike the tree and understory layers, the four sampled ground layer sites all experienced increases in diversity, owing primarily to increases in species richness at each site; only one site (Ferry Bluff) also experienced a significant increase in species richness per sampling quadrat (+3.8, P < 0.001).

Species composition--tree layer.--The composition of the tree layer has undergone dramatic changes since the sampling by Curtis (Table 4 and Fig. 3a-c). In the 1950s, the sites were dominated by a mix of characteristic pioneer (river birch and swamp white oak) and species historically considered dominant (sensu Curtis-silver maple, green ash and elm). In 2001, silver maple, green ash and swamp white oak remain dominant species; silver maple had become strongly dominant at Arena and Wyoming Bluffs. All but the Wyoming Bluff site experienced a large increase in species not historically associated with lowland forests in Wisconsin: Northern red oak (Quercus rubra), bitternut hickory (Carya cordiformis) and hackberry (Celtis occidentalis). Considering species groupings, pioneer species have for the most part experienced a large decline, with only minor increases noted for cottonwood (Helena), swamp white oak (Arena) and river birch (Wyoming Bluffs). The historical canopy dominants have generally experienced declines as well, although silver maple, as previously noted, presents a more mixed picture, with increases at the less disturbed sites and decreases on the disturbed sites. The grouping of new (i.e., not dominant historically) and/or typical mesic to upland species has experienced large increases, as Figure 3c demonstrates. Basswood (Tilia americana), a more mesic species reported in the lowland forests in the past, has generally lost importance since the 1950s. Trends at Wyoming Bluffs tend to run contrary to those at the other sites, with one pioneer species (river birch) and two historical canopy dominants (ash and silver maple) increasing in importance since the 1950s.

[FIGURE 3 OMITTED]

Species composition--understory.--Substantial changes have also taken place in the understory layer (Table 4 and Fig. 4 a-c). As with the tree layer, almost all species dominating the understory in the 1950s were characteristic of the Curtis' southern lowland forests. In contrast to the earlier time period, the understory of three of the five sites (Helena, Ferry Bluff, Arena) is now dominated by two invasive species, the native prickly ash (Zanthyloxum americanum) and the exotic European buckthorn. Lone Rock is dominated by bitternut hickory, as it was in the 1950s although much more so. The Wyoming Bluffs understory is now heavily dominated by green ash. All but one pioneer species were either absent or showed declines in the understory layer at all sites. Swamp white oak experienced increases at Ferry Bluff and Arena, but experienced a substantial decline at Lone Rock. Similarly, most historical canopy dominants also experienced large declines in the understory layer since the 1950s; only green ash showed a large increase at one site (Wyoming Bluffs). In contrast, bitternut hickory hackberry and the invasives, European buckthorn and prickly ash, showed large increases across many of the sites. Again, Wyoming Bluffs was the exception.

Ulmus species.--A closer examination of changes in Ulmus species in the tree and understory layers revealed general decreases throughout the study sites (Table 5). Overall basal area per hectare decreased at all sites but Helena, where no change was evident; average basal area per individual decreased substantially at all sites (Fig. 5). The changes in density of Ulmus in the tree and understory layers varied more widely than basal area. Absolute Ulmus tree densities increased at the sites that had experienced timber harvests, but decreased at the others. Relative density of Ulmus in the understory decreased at all sites except Lone Rock.

Species composition--ground layer.--An examination of the five most frequent ground species reveals some substantial shifts at most sites. In the 1950s, most dominant species at all sites were modal species; by 2001, all sites but Wyoming Bluffs experienced a decline in the number of dominant modal species (Table 6) and their overall frequency (Fig. 6). Also noteworthy is the increase in the presence of exotic species (Table 7), which lists several species that were absent from the 1950s records. In the 1950s, only one non-native species, bouncing bet (Saponaria officinalis), was recorded and only at one site (Lone Rock). The dominant species in the ground layer also appear to have shifted to species less likely to be browsed by deer (species bolded in Table 6), as these have increased in both number and frequency.

[FIGURE 4 OMITTED]

Despite the higher sampling effort in 2001, we did not encounter several species identified by Curtis in the 1950s: dotted smartweed (Polygonum punctatum), arrow-leaved tearthumb (P. sagittatum) and several grasses: brome (Bromus spp.), bluejoint (Calamagrostis canadensis), panic grass (Panicum spp.), common woodreed (Cinna arundinacea) and big bluestem (Andropogon gerardii).

Community wetland coefficients.--Figure 7 presents the change in the overall wetland coefficient for each site from the 1950s to 2001. Two sites in the tree layer show a shift to species with higher wetland coefficients (i.e., species typical of non-wetland areas), Helena and Ferry Bluff. Arena and Lone Rock's site coefficient appear to be similar across time periods, while Wyoming Bluffs' has become wetter. The changes in the understory vegetation are more dramatic. All sites but Wyoming Bluffs have become less wet, with Helena, Ferry Bluff and Arena all possessing positive coefficients. Wyoming Bluffs again exhibits a large negative shift in its site coefficient. The ground layer results also reflect these trends: Helena and Ferry Bluff have experienced increases in their coefficients, with Helena's becoming positive. Lone Rock has remained relatively constant, and Wyoming Bluffs has become more negative.

DISCUSSION

TREE AND UNDERSTORY LAYERS

Curtis (1959) and Ware (1955) felt that succession within the lowland forests of southern Wisconsin did not proceed past the stage where American elm, green ash and silver maple had become dominant. They felt river dynamics would typically reset a site, before any further successional change might take place. Indeed our analysis of the 1950s data for the Lower Wisconsin shows that the sites possessed a varying mix of what Curtis termed pioneer and later-successional species. However, since that time several major changes have taken place: flooding on the Lower Wisconsin has decreased, Dutch elm disease has devastated the adult elm population and the density of a major herbivore, the white-tailed deer, has increased dramatically. As the results from this study suggest, these changes have likely altered the ecological dynamics on the Lower Wisconsin, which, in turn, are interacting with timber harvests and invasive species to result in a very different floodplain forest community.

At all sites other than Helena, silver maple remains a canopy dominant (exceptionally so at Arena and Wyoming Bluffs); in some cases, ash and elm are also present. The dominance of silver maple is common across Midwestern floodplain forests with altered flood regimes (Barnes, 1997; Knutson and Klaas, 1998). However, silver maple in particular, shows little to no regeneration across the sites, suggesting that the current dominance is a transitory aspect. In place of these historical dominants, we see the increase in several species not historically dominant in the Lower Wisconsin, most notably bitternut hickory and hackberry. These two species, which have increased in importance in both the tree layer and the understory, are typical of floodplain forests, and they are generally associated with later successional stages in other systems. As such, they may represent more permanent additions to the Lower Wisconsin canopy and demonstrate the progression to a stage not previously observed. Burns and Honkala (1990) listed bitternut hickory as a common associate of subclimax and climax forests, particularly in more southern forests in the US. Peet and Loucks (1977) considered it to be a shade tolerant, later-successional species. Robertson et al. (1978) found bitternut hickory a canopy dominant in second growth floodplain forests in southern Illinois, where it was also recruiting successfully. Hackberry is also considered a shade tolerant, intermediate successional species (Yin, 1998). Ware (1955) noted that hackberry was a co-dominant species at two higher elevation sites in the Wisconsin floodplain. It also often replaces its southern counterpart, sugarberry (Celtis laevigata), in the northern extent of the Sugarberry-American Elm-Green Ash forest type, an intermediate successional floodplain forest type typical of the Lower Mississippi Valley (Burns and Honkala, 1990; Hodges, 1997). This forest type is capable of self-replacement and can persist 200-300 y on the Lower Mississippi.

[FIGURE 5 OMITTED]

Hall and Harcombe (2001) noted that in natural floodplains, "growth" (a characteristic of flood tolerant pioneer species) represents a better strategy for saplings than "survivorship" (more characteristic of shade tolerant species). The general decrease in pioneer and historical dominant species (all relatively flood tolerant) and the increase in species with lower flood tolerances, such as bitternut hickory, hackberry and Northern red oak, at four of the five sites (excluding Wyoming Bluffs, which we discuss later) suggest that survivorship is becoming the more common strategy in this region. This reflects the loss of natural flooding dynamics in the floodplain and a shift towards later-successional communities and, in instances where Northern red oak is present, a system more reflective of upland communities. Other studies have found similar results on other rivers experiencing altered flooding regimes, such as the Chippewa, the Mississippi, the Allegheny and the Rhine (Barnes, 1997; Cowell and Dyer, 2002; Knutson and Klaas, 1998; Schnitzler, 1994; Yin, 1998).

[FIGURE 6 OMITTED]

The impact of timber harvests at Helena and Ferry Bluff (and potentially the indirect impacts at Arena, through increased light in the forest interior and edge disturbances) appear to have accelerated changes in community composition. Other than the loss of the elm species, the non-harvested sites show relatively little decrease in the role of the historical dominants; in fact, both silver maple and green ash experienced increases in their importance values. On the harvested sites, the community is different. The diversity index of these harvested sites is markedly greater. The removal of adult silver maple and green ash has allowed bitternut hickory and hackberry--both shade tolerant species--to become dominant more quickly (at Helena and Ferry Bluff, respectively). Since the harvests were selective (Table 1), shade-tolerant saplings in the understory were likely already present and able to take advantage of the canopy openings, preventing the large-scale establishment of many pioneer/shade-intolerant species. Further, timber harvest regulations have also helped maintain swamp white oaks along the Lower Wisconsin by prohibiting oak harvesting and protecting oak regeneration. Thus, although Curtis considered swamp white oak a pioneer species, it has remained an important part of the canopy at several sites.

Dutch elm disease functioned similarly to timber harvests by creating canopy gaps. Dutch elm disease resulted in large-scale adult elm mortality after the 1950s and affected all sites with elm populations. Since elms were a dominant species of the tree layer at all sites except Helena in the 1950s (where an earlier timber harvest likely reduced their importance), Dutch elm disease resulted in a major loss for these communities. The remaining elm population has generally been reduced and consists of much smaller, younger trees (Fig. 5), linked likely to the young age at which elms reproduce and their shade tolerance (Parker and Leopold, 1983). The few increases observed result from the proliferation of juveniles, which due to the disease will likely not persist long.

Although impossible to determine from the Curtis data, it is likely that the understory density has increased greatly at the four sites (excluding Wyoming Bluffs). The fact that 30% of the understory sampling points lacked data (and consequently stems) in the 1950s suggests understory densities were low. Ware (1955) also noted that saplings were generally sparse in the understory in southern Wisconsin. Comparisons with other understory density estimates from the literature also indicate that the current understory density is abnormally high (Hall and Harcombe, 2001; Robertson et al., 1978). Although the timber harvests may have also played a role, results from Hale (2004) indicated that the understory densities of non-harvested and harvested sites in the LWSR are similar. Another possible cause is the loss of elms. Dunn (1986) found that a loss of as few as five elms per hectare can greatly increase the level of shrubby growth in forests of southeastern Wisconsin; the loss of elms at the three non-harvested sites ranges from 17 to 101 trees per ha, but the likely harvest of elms at Helena and Ferry Bluff makes these data harder to interpret. Since the two main invasive species, prickly ash and European buckthorn are not considered particularly flood tolerant, the decreased flooding regime is also a likely contributing factor, particularly given the results at Wyoming Bluffs.

Another likely factor contributing to the proliferation of the invasives is the overabundance of white-tailed deer in the forests of southern Wisconsin. As mentioned earlier, Wisconsin has experienced a dramatic rise in white-tailed deer population and deer densities in the Lower Wisconsin area generally exceeded management goals. Studies have demonstrated how increased levels of herbivory by deer have altered forest compositions (Rooney, 2001; Waller and Alverson, 1997) Two of the dominant understory species, prickly ash and European buckthorn, possess a large numbers of sharp thorns (along the branches and at the tips of the branches, respectively) that can deter browsing herbivores, such as deer, and possibly confer a competitive advantage to such species in the current floodplain environment. Although the literature lacks studies investigating browse levels of these two species, two studies suggest that heavy deer browsing promotes the growth of other Zanthoxlyum species (Moog et al., 2002; Shimoda et al., 1994). Further, both prickly ash and buckthorn contain toxins, which should reduce their attractiveness to herbivores (Ju et al., 2001; Lichtensteiger et al., 1997).

In addition to contributing to the spread of the invasives, several studies have shown that deer herbivory can reduce tree regeneration in floodplain forests and may thus, help explain the decrease in species such as silver maple and green ash (Boerner and Brinkman, 1996; Liang and Seagle, 2002; Shimoda et al., 1994; Sweeney et al., 2002). In a study by Kost et al. (1998), deer browsing on silver maple severely limited the researchers' ability to analyze silver maple in their project. The study by Liang and Seagle (2002) found that deer browsing increased mortality and decreased annual growth in green ash. They also demonstrated that herbivory can alter successional processes and hasten the development of shade-tolerant phases. Thus, herbivory might also contribute to the success of more shade-tolerant species such as bitternut hickory.

[FIGURE 7 OMITTED]

GROUND LAYER

In general, the ground layer in these forests varies annually depending on the flood regime (Curtis, 1959), which renders the task of analyzing long-term changes in composition more complex than is true for the understory or tree layer. Nonetheless, our findings point towards shifts in composition that reflect decreased flooding, invasive species and increased levels of herbivory.

Using the species that Curtis identified as modal species for southern Wisconsin wet forests (i.e., those species whose had their highest frequency of occurrence in these forests), we analyzed how the ground layer in the 2001 compared to that measured in the 1950s. Modal species such as Canadian honewort (Cryptotaenia canadensis), American germander (Teucrium canadense) and calico aster (Aster lateriflorus) were common dominants of these sites in the 1950s, and modal species in general occurred in over a quarter of the plots sampled at each of the sites. By 2001, the role of modal species had decreased at all sites but Wyoming Bluffs. In their place, species less characteristic of solely wetland environments such as Virginia creeper (Parthenocissus quinquefolia), jumpseed (Polygonum virginianum) and cleavers (Galium aparine) have become more common (Barnes, 1978; Bell, 1974; Menges and Waller, 1983). The decreased flooding along the Lower Wisconsin is a plausible explanation for the apparent increase and domination of species more typical of less flooded areas, since the composition of the herb layer in floodplain forests is driven primarily by microtopographic differences in elevation and the resulting differences in flood regimes (Barnes, 1978; Bell, 1974; Menges, 1986; Siebel and Bouwma, 1998).

While our data provide some evidence that the composition of these forests is transitioning from a ground layer characteristic of wet forests to one characteristic of more mesic forests, it is evident that the decreased flood regime is not the only factor driving the observed changes. Invasive species have increased in the ground layer (Table 7). At Wyoming Bluffs two invasives, reed canary grass and moneywort, have become the most dominant species in the ground layer. These species are considered 'undesirable' from a conservation perspective and could potentially influence the vegetation composition of these forests. One possible consequence is the apparent loss of two species typical of lowland habitats, dotted smartweed and arrow-leaved tearthumb (Menges and Waller, 1983). These species are found in similar mictotopograpic sites as the two invasive species mentioned above. In addition, the spread of reed canary grass has possibly come at the expense of other grass species, which have become less abundant. Its spread could also be problematic for future tree regeneration (Fierke and Kauffman, 2005; Knutson and Klaas, 1998).

While we do not have evidence that the increase of these invasive species and the decrease in other native species is a causal relationship, it is, nevertheless, an interesting one. Invasive species represent a major threat to biodiversity worldwide and have the potential to alter vegetation composition (Alvarez and Cushman, 2002; Callaway and Aschehoug, 2000), hydrology and geomorphology (Tickner et al., 2001; Zavaleta, 2000), and disturbance regimes (D'Antonio and Vitousek, 1992).

Given studies in other habitats (Rooney et al., 2004; Russell et al., 2001), there is reason to believe that deer herbivory is also a factor driving the composition of these forests. The results shown in Table 6 lend weight to this argument, as the number of dominant species not typically browsed by deer (Jull, 2001) has increased at each site, which suggests that resistance to deer browsing may convey a greater competitive advantage than it did 50 y ago. Although reed canary grass was not identified by Jull (2001) as a species that deer avoid, a recent study by Kercher and Zedler (2004) indicated that reed canary grass responds well to grazing. These findings suggest that deer may play a role in the observed changes in the composition of the ground layer and the spread of some invasive species.

SUCCESSIONAL PROCESSES

The often contradictory situation at the Wyoming Bluff site provides some interesting insights into the processes occurring along the Lower Wisconsin. Our qualitative observation during field sampling was that Wyoming Bluffs was by far the wettest of our sample sites, which is corroborated by the results of the community wetland coefficients for all three layers. Moreover, the results from the wetland coefficients suggest that this site has become wetter since the 1950s, which seems to contradict the influence of the decreased flooding regime. A model developed by Hodges (1997) may provide an explanation.

Hodges (1997) discussed a model of floodplain forest succession based on floodplain systems in the southeastern United States (Lower Mississippi River and Coastal Plain systems). A key component of this model is the role of sedimentation that occurs during floods and which increased the local elevation. On sites subject to higher levels of sedimentation, elevation eventually becomes raised enough to remove the site from most of floodplain's hydrological influences, which results in succession to the hickory-oak climax. However, Hodges also notes that interruptions in the local hydrology, such as through dam building, could result in the reversal of successional processes. Although we have no data on sedimentation on any of our sites, this could well explain the results we found at the Wyoming Bluffs site, which appeared to have grown wetter and at least maintained its successional status, if not reverted somewhat.

A significant difference between the Wyoming Bluffs and other sites is the elevation (see Fig. 2). If the sedimentation hypothesis applies to Wyoming Bluffs, then the other sites either have likely been subject to lower sedimentation rates or had gained sufficient elevation that decreased flooding has possibly accelerated or induced further succession on sites where it had not previously been recorded. In a future investigation of this hypothesis, it would be important to look at the flood regime generated by the upstream dams. Although the overall flood regime has decreased, the low flows of the rivers have been augmented (Wisconsin Department of Natural Resources, 1988), which could result in more frequent flooding of the lowest lying areas of the floodplain, despite decreased flooding elsewhere in the floodplain. The difference in elevation between Wyoming Bluffs and Ferry Bluff, the next lowest site (Fig. 2), may thus represent an important boundary between these two different flooding regimes that has important management implications.

The changes associated with the altered hydrologic regime combined with the other discussed disturbances (timber harvests, introduced species, increased herbivory) have altered the composition and structure of the Lower Wisconsin floodplain forests. Understanding the role each of these factors play in community dynamics is important for the management and conservation of these forests, which have regional ecological significance and are part of an important state protected area, the Lower Wisconsin State Riverway (Wisconsin Department of Natural Resources, 1988). The Riverway was created, in part, to protect the natural resources of the Lower Wisconsin River Valley. The future composition of the forest and the biota that depend upon it are uncertain. In the face of such uncertainty, an adaptive approach to the forest's management will be necessary (sensu Lee 1993). Our results suggest that the understory may be the most useful layer as a gauge of the current trajectory of a forest for management purposes; however, all layers should be monitored. A restoration of a more normal flooding regime on the river would be the best way to reverse succession in this area. In lieu of that, managers should concentrate efforts on those factors most within their control: timber harvests, deer populations and invasive species.

Acknowledgments.--We would like to thank C. Abel, M. Carneiro, A. K. Dreyer, J. Kreyling, C. Noguiera and G. Vorhes for able research assistance; B. Larget for statistical advice; the guidance of our respective committees (Hale: D. Field, N. Langston, E. Stanley and P. Zedler; Alsum: S. Hotchkiss, E. Stanley and J. Zedler); and the staff of the LWSR for their time and assistance. Portions of this work were supported by the NSF IGERT Grant 9870703 (Human Dimensions of Social and Aquatic System Interactions); by the Wisconsin Academy of Sciences, Arts and Letters Lois Almon small grants program; the UW Department of Botany John Jefferson Davis Memorial Fund; and the US Department of Education FIPSE award PJ116J990028A. All remaining errors are our own.

SUBMITTED 29 JANUARY 2007

ACCEPTED 2 APRIL 2008

LITERATURE CITED

ALVAREZ, M. E. AND J. H. CUSHMAN. 2002. Community-level consequences of a plant invasion: effects on three habitats in coastal California. Ecol. Appl., 12:1434-1444.

BARNES, W. J. 1978. The distribution of floodplain herbs as influenced by annual flood elevation. Wisconsin Acad. Sci. Arts, and Letters, 66:254-266.

--. 1997. Vegetation dynamics on the floodplain of the lower Chippewa River in Wisconsin. J. Torrey Botan. Club, 124:189-197.

BELL, D. T. 1974. Studies on the ecology of a streamside forest: Composition and distribution of vegetation beneath the tree canopy. Bull. Torrey Botan. Club, 101:14-20.

BOERNER, R. AND J. BRINKMAN. 1996. Ten years of tree seedling establishment and mortality in an Ohio deciduous forest complex. Bull. Torrey Botan. Club, 123:309-317.

BRINSON, M. 1990. Riverine forests, p. 87-141. In: A. E. Lugo, M. Brinson and S. Brown (eds.). Ecosystems of the World 15: Forested Wetlands. Elsevier, New York.

BURNS, R. M. AND B. H. HONKALA. 1990. Silvics of North America, Volume II. U.S. Department of Agriculture, Washington, DC.

CALLAWAY, R. M. AND E. T. ASCHEHOUG. 2000. Invasive plants versus their new and old neighbors: a mechanism for exotic invasion. Science, 290:521-523.

COLLINS, R. J. AND W. CARSON. 2003. Do succession models predict the right pattern for the wrong reason: shade vs. herbivore tolerance? Ecological Society of America 88th Annual Meeting.

COSTANZA, R., R. D'ARGE, R. DE GROOT, S. FARBER, M. GRASSO, B. HANNON, K. LIMBURG, S. NAEEM, R. V. ONEILL, J. PARUELO, R. G. RASKIN, P. SUTTON AND M. VAN DEN BELT. 1997. The value of the world's ecosystem services and natural capital. Nature, 387:253-260.

COTTAM, G. AND J. T. CURTIS. 1956. The use of distance measures in phytosociological sampling. Ecology, 37:451-460.

COWELL, C. M. AND J. M. DYER. 2002. Vegetation development in a modified riparian environment: Human imprints on an Allegheny River wilderness. Ann. Associ. Am. Geog., 92:189-202.

CURTIS, J. T. 1959. The Vegetation of Wisconsin. University of Wisconsin Press, Madison, Wisconsin.

D'ANTONIO, C. M. AND P. M. VITOUSEK. 1992. Biological invasions by exotic grasses, the grass/fire cycle and global change. Ann. Rev. Ecol. and Syst., 23:63-87.

DECAMPS, H. 1996. The renewal of floodplain forests along rivers: a landscape perspective. Verhandlungen der Internationalen Vereiningung fur Theoretische und Angewandte Limnologie, 26:35-59.

DECONCHAT, M. AND G. BALENT. 2001. Effect of logging on vegetation at a fine scale. Ann. For. Sci., 58:315-328.

DUNN, C. 1986. Shrub layer response to the death of Ulmus americana in southeastern Wisconsin lowland forests. Bull. Torrey Botan. Club, 113:142-148.

DYNESIUS, M. AND C. NILSSON. 1994. Fragmentation and flow regulation in the Northern third of the world. Science, 266:753-762.

FIERKE, M. K. AND J. B. KAUFFMAN. 2005. Structural dynamics of riparian forests along a black cottonwood successional gradient. For. Ecol. and Manag., 215:149-162.

FREEMAN, R. E., E. H. STANLEY AND M. G. TURNER. 2003. Analysis and conservation implications of landscape change in the Wisconsin River floodplain, USA. Ecol. Appl., 13:416-431.

GLEASON, H. A. AND A. CRONQUIST. 1991. Manual of Vascular Plants of Northeastern United States and Adjacent Canada. The New York Botanical Garden, Bronx, New York.

HALE, B. 2004. Conservation in temperate river-floodplain forests: a comparative analysis of the Lower Wisconsin State Riverway and the Middle Elbe Biosphere Reserve. Ph.D. dissertation thesis, University of Wisconsin-Madison, Madison.

HALL, R. B. W. AND P. A. HARCOMBE. 1998. Flooding alters apparent position of floodplain saplings on a light gradient. Ecology, 79:847-855.

--AND--. 2001. Sapling dynamics in a southeastern Texas floodplain forest. J. Veget. Sd., 12:427-438.

HODGES, J. 1997. Development and ecology of bottomland hardwood sites. For. Ecol. and Manag., 90:117-125.

HUGHES, J. W. AND W. B. CASS. 1997. Pattern and process of a floodplain forest, Vermont, USA: Predicted responses of vegetation to perturbation. J. Appl. Ecol., 34:594-612.

HUPP, C. R. AND W. R. OSTERKAMP. 1996. Riparian vegetation and fluvial geomorphic processes. Geomorphology, 14:277-295.

JU, Y., C. C. STILL, J. N. SACALIS, J. G. LI AND C. T. HO. 2001. Cytotoxic coumarins and lignans from extracts of the northern prickly ash (Zanthoxylum americanum). Phytotherapy Res., 15:441-443.

JULL, L. G. 2001. Plants not favored by deer. Report A3727. University of Wisconsin-Extension, Madison, Wisconsin.

KERCHER, S. M. AND J. B. ZEDLER. 2004. Multiple disturbances accelerate invasion of reed canary grass (Phalaris arundinacea L.) in a mesocosm study. Oecologia, 138:455-464.

KNUTSON, M. AND E. KLAAS. 1998. Floodplain forest losses and changes in forest community composition and structure in the Upper Mississippi River: a wildlife habitat at risk. Nat. Areas J., 18:138-150.

KOST, D. A., J. P. VIMMERSTEDT AND J. H. BROWN. 1998. Topsoiling, ripping, and fertilizing effects on tree growth and nutrition on calcereous minesoils. For. Ecol. and Manag., 103:307-319.

KOZLOWSKI, T. T. 2002a. Physiological ecology of natural regeneration of harvested and disturbed forest stands: implications for forest management. For. Ecol. and Manag., 158:195-221.

--. 2002b. Physiological-ecological impacts of flooding on riparian forest ecosystems. Wetlands, 22:550-561.

KRUG, W. R. AND L. B. HOUSE. 1980. Streamflow model of Wisconsin River for estimating flood frequency and volume. USGS Water-Resources Investigations Open-File Report 80-1103. United States Geological Survey.

LIANG, S. Y. AND S. W. SEAGLE. 2002. Browsing and microhabitat effects on riparian forest woody seedling demography. Ecology, 83:212-227.

LICHTENSTEIGER, C. A., N. A. JOHNSTON AND V. R. BEASLEY. 1997. Rhamnus cathartica (Buckthorn) hepatocellular toxicity in mice. Toxicol. Path., 25:449-452.

LINDSTROM, S. AND J. YOUNG. 2002. Madison climate trends reflected in changed "normal" statistics, University of Wisconsin-Madison.

LOCKABY, B. G., C. C. TRETTIN AND S. H. SCHOENHOLTZ. 1999. Effects of silvicultural activities on wetland biogeochemistry. J. Environ. Qual., 28:1687-1698.

LUDWIG, J. A. AND J. F. REYNOLDS. 1988. Statistical Ecology. John Wiley and Sons, New York.

MCCOWN, W. AND K. WALLENFANG. 1998. Wisconsin's Deer Management Program. SS-931-98. Wisconsin Department of Natural Resources, Madison. 39 p.

MENGES, E. S. 1986. Environmental correlates of herb species composition in five Southern Wisconsin floodplain forests. Am. Midl. Nat., 115:106-117.

--AND D. M. WALLER. 1983. Plant strategies in relation to elevation and light in floodplain herbs. Am. Nat., 122:454-473.

MESSINA, M. G., S. H. SCHOENHOLTZ, M. W. LOWE, Z. WANG, D. K. GUNTER AND A. J. LONDO. 1997. Initial responses of woody vegetation, water quality, and soils to harvesting intensity in a Texas bottomland ecosystem. For. Ecol. and Manag., 90:201-215.

MOOG, J., H. FELDHAAR AND U. MASCHWITZ. 2002. On the caulinary domatia of the SE-Asian ant-plant Zanthoxylnm myriacanthum Wall. ex Hook. f. (Rutaceae) their influence on branch statics, and the protection against herbivory. Sociobiology, 40:547-574.

NAIMAN, R.J. AND H. DECAMPS. 1997. The ecology of interfaces: riparian zones. Ann. Rev. Ecol. and Syst., 28:621-658.

NATIONAL PLANT DATA CENTER. 2002. The PLANTS Database, Version 3.5, Natural Resources Conservation Service.

OLIVER, C. D. AND B. C. LARSON. 1996. Forest Stand Dynamics. John Wiley and Sons, New York.

PARKER, G. R. AND D. J. LEOPOLD. 1983. Replacement of Ulmus americana L. in a mature east-central Indiana Woods. Bull. Torrey Botan. Club, 110:482-488.

PEET, R. K. AND O. L. Loucks. 1977. A gradient analysis of southern Wisconsin forests. Ecology, 58:485-499.

PETTS, G. 1990. Forested river corridors: a lost resource, p. 12-34. In: D. Cosgrove and G. Petts (eds.). Water, Engineering, and Landscape. Belhaven Press, London.

PFEIFFER, S. M. 2000. Groundwater/surface water interactions in a lowland savannah on the lower Wisconsin River floodplain. M.S. Thesis thesis, University of Wisconsin-Madison.

PINAY, G., H. DECAMPS, E. CHAUVET AND E. FUSTEC. 1991. Functions of ecotones in fluvial systems, p. 141-170. In: R. J. Naiman and H. Decamps (eds.). Ecology and Management of Aquatic-Terrestrial Ecotones. UNESCO-Paris and Parthenon Publishing Group, Paris.

POSTEL, S. AND S. CARENTER. 1997. Freshwater ecosystem services, p. 195-214. In: G. C. Daily (ed.). Nature's Services: societal dependence on natural ecosystems. Island Press, Washington, DC.

ROBERTSON, P. A., G. T. WEAVER AND J. A. CAVANAUGH. 1978. Vegetation and tree species patterns near the Northern terminus of the Southern Floodplain Forest. Ecol. Mono., 48:249-267.

ROLLEY, R. E. 2005. White-tailed deer population status 2005. Wisconsin Department of Natural Resources, Madison. 8 p.

ROONEY, T. P. 2001. Deer impacts on forest ecosystems: a North American perspective. Forestry, 74:201-208.

--, S. M. WIEGMANN, D. A. ROGERS AND D. M. WALLER. 2004. Biotic Impoverishment and Homogenization in Unfragmented Forest Understory Communities. Conserv. Biol., 18:787-798.

RUSSELL, F. L., D. B. ZIPPIN AND N. L. FOWLER. 2001. Effects of white-tailed deer (Odocoileus virginianus) on plants, plant populations and communities: A review. Am. Mid. Nat., 146:1-26.

SCHNITZLER, A. 1994. European alluvial hardwood forests of large floodplains. J. Biogeog., 21:604-623.

SHIMODA, K., K. KIMURA, M. KANZAKI AND K. YODA. 1994. The Regeneration of Pioneer Tree Species under Browsing Pressure of Sika-Deer in an Evergreen Oak Forest. Ecol. Res., 9:85-92.

SIEBEL, H. N. AND I. M. BOUWMA. 1998. The occurrence of herbs and woody juveniles in a hardwood floodplain forest in relation to flooding and light. J. Veget. Sci., 9:623-630.

SWEENEY, B. W., S. J. CZAPKA, A AND T. YERKES. 2002. Riparian forest restoration: Increasing success by reducing plant competition and herbivory. Restor. Ecol., 10:392-400.

TICKNER, D. P., P. G. ANGOLD, A. M. GURNELL AND J. O. MOUNTFORD. 2001. Riparian plant invasions: hydrogeomorphological control. Prog. Phys. Geog., 25:22-52.

TOCKNER, K. AND J. A. STANFORD. 2002. Riverine flood plains: present state and future trends. Environ. Conserv., 29:308-330.

WALLER, D. AND W. ALVERSON. 1997. The white-tailed deer: a keystone herbivore. Wildl. Soci. Bull., 25:217-226.

WARD, J. V. AND J. A. STANFORD. 1995. The serial discontinuity concept. Regulated Rivers: Res. and Manag., 10:159-168.

--, K. TOCKNER, D. B. ARSCOTT AND C. CLARET. 2002. Riverine landscape diversity. Freshwater Biol., 47:517-539.

WARE, G. H. 1955. A phytosociological study of lowland hardwood forests in Southern Wisconsin. Ph.D. Dissertation thesis, University of Wisconsin, Madison.

WISCONSIN DEPARTMENT Or NATURAL. RESOURCES. 1988. Final Enviromnental Impact Statement: Proposed Lower Wisconsin State Riverway. Wisconsin Department of Natural Resources, Madison, Wisconsin.

--. 2002. The state of the Lower Wisconsin River Basin. PUBL WT-559-2002. Wisconsin Department of Natural Resources, Madison, Wisconsin. 456 p.

YIN, Y. 1998. Flooding and forest succession in a modified stretch along the Upper Mississippi River. Regul. Riv.-Res. & Manag., 14:217-225.

ZAVALETA, E. S. 2000. Valuing ecosystem services lost to Tamarix invasion in the United States. In: H. A. Mooney and R. Hobbs (eds.). Invasive Species in a Changing World. Island Press, Washington, DC.

BRACK W. HALE (1)

Gaylord Nelson Institute for Environmental Studies, University of Wisconsin-Madison, 53703

AND

ESTHER M. ALSUM AND MICHAEL S. ADAMS

Department of Botany, University of Wisconsin-Madison, 53703

(1) Corresponding author: current address: Franklin College Switzerland, Via Ponte Tresa 29, 6924 Sorengo, Switzerland; e-mail: bhale@fc.edu Telephone: +41-91-985-22-60, FAX: +41-91-994-41-17
TABLE 1.--Sites within the LWSR surveyed by Curtis in the 1950s

              Survey
Site          points   Comments (a)

Helena (HL)     60     Harvest in 1940s; Overstory
                       harvest in 1977

Ferry           40     Large timber removal in early
Bluff (FB)             1980s; 1994 harvest of silver
                       maple and green ash (dbh >36 cm)

Arena (AR)      40     Partial clearing after 1950s
                       for parking lot construction

Lone            20     Last harvest in 1932
Rock (LR)

Wyoming         40     Last harvest in 1896
Bluffs (WY)

(a) Source: Wisconsin Department of Natural Resources records.
NB: Harvest records were often incomplete, data above represent
best available information

TABLE 2--Site characteristics in 2000. Change from 1950s in
parentheses, when available (t-test results for difference in mean
basal area across time periods provided)

Characteristic                 HL               FB

Mean canopy (%)               84.6             78.1
Mean basal area          0.079 (-0.021)   0.084 (-0.059)
  per tree ([m.sup.2])     t = 3.58,        t = 7.05,
                            P<0.001          P<0.001
Basal area per
  hectare                 30.6 (+18.7)     24.4 (-34.4)
Tree density (per ha)      386 (+268)       291 (-119)
Understory density
  (per ha)                    2750             2025

Characteristic                 AR               LR

Mean canopy (%)               76.3             78.0
Mean basal area          0.084 (-0.022)   0.074 (-0.014)
  per tree ([m.sup.2])     t = 5.24,        t = 2.02,
                            P<0.001          P=0.046
Basal area per
  hectare                 25.9 (+1.2)      29.1 (+1.1)
Tree density (per ha)      310 (+76)        395 (+75)
Understory density
  (per ha)                    2325             950

Characteristic                 WY

Mean canopy (%)               73.4
Mean basal area          0.093 (-0.054)
  per tree ([m.sup.2])     t = 3.25,
                            P=0.001
Basal area per
  hectare                 19.7 (-25.2)
Tree density (per ha)      212 (-92)
Understory density
  (per ha)                    225

TABLE 3.--Diversity measures for tree layer in 2001 (Change from
1950s in parentheses). S = species richness, H' = Shannon-Weaver
index, J = Pielou's evenness index

Layer Measure               HL               FB

Tree
  S                      10 (+1)         9 (+4)
  H'                   2.04 (+0.14)   1.81 (+0.38)
  J                    0.89 (+0.02)   0.83 (-0.06)
Understory
  S                      12 (0)          7 (+2)
  S (shrubs)*             4              1
  H'                   1.99 (+0.13)   1.55 (+0.56)
  J                    0.80 (+0.05)   0.80 (+0.18)

Ground
  Mean S per quadrat   10.6 (-0.9)     7.1 (+3.8) (a)
  S per site             62 (+15)       34 (+15)
  H'                   3.73 (+0.26)   3.04 (+0.45)
  J                    0.90           0.86 (-0.02)

Layer Measure               AR               LR

Tree
  S                      10 (+1)         9 (+1)
  H'                   1.58 (-0.27)   1.57
  J                    0.69 (-0.16)   0.72 (-0.04)
Understory
  S                      10 (+1)         9 (+3)
  S (shrubs)*             2              3
  H'                   1.96 (+0.27)   1.53 (+0.22)
  J                    0.85 (+0.08)   0.70 (-0.04)

Ground
  Mean S per quadrat        NA         8.6 (+0.4)
  S per site                NA          45 (+10)
  H'                        NA        3.48 (+0.23)
  J                         NA        0.91 (0)

Layer Measure               WY

Tree
  S                       8 (+2)
  H'                   0.96 (-0.46)
  J                    0.46 (-0.33)
Understory
  S                       4 (-4)
  S (shrubs)*             1
  H'                   1.18 (-0.22)
  J                    0.85 (+0.18)

Ground
  Mean S per quadrat    6.1 (+0.1)
  S per site             34 (+11)
  H'                   3.12 (+0.40)
  J                    0.88 (0.02)

* Not recorded during in Curtis analysis of understory;
does not include Zanthyloxum amencanum

(a) Change significant, P [less than or equal to] 0.05

TABLE 4.--Dominant species of tree layer by site (IV 200).
Code represents first two letters of genus followed by
first two letters of species (e.g., Acsa = Acer saccharinum).
Species not mentioned in text: Acne: Acer negundo (box elder);
Cephoc: Cephalanthus occidentalis (buttonbush); Frni: Fraxinus
nigra (black ash)

             HL             FB             AR

Tree Layer

1950s   Acsa (52.5)    Acsa (85.6)    Acsa (70.5)
        Qubi (48.4)    Frpe (49.3)    Ulmus (33.3)
        Frpe (27.9)    Ulmus (24.8)   Beni (31.0)

2001    Caco (45.5)    Acsa (64.5)    Acsa (106.1)
        Qubi (36.6)    Ulmus (51.2)   Frpe (26.7)
        Quru (32.9)    Ceoc (21.8)    Qubi (21.6)

Understory Layer

1950s   Frpe (45.4)    Acsa (81.0)    Frpe (80.4)
        Ulmus (28.7)   Ulmus (70.4)   Beni (38.0)
        Frni (27.4)    Frpe (8.8)     Ulmus (29.2)

2001    Zaam (70.1)    Rhca (98.3)    Zaam (60.6)
        Caco (36.2)    Ceoc (23.3)    Frpe (44.0)
        Ceoc (19.5)    Zaam (23.3)    Ceoc (21.6)

             LR             WY

Tree Layer

1950s   Qubi (96.8)    Ulmus (74.4)
        Ulmus (35.9)   Acsa (74.3)
        Acsa (19.8)    Frpe (20.9)

2001    Qubi (93.3)    Acsa (137.8)
        Caco (35.5)    Frpe (37.8)
        Acsa (32.4)    Ulmus (16.0)

Understory Layer

1950s   Caco (62.2)    Acsa (61.9)
        Qubi (57.0)    Ulmus (42.3)
        Frpe (15.3)    Tiam (21.3)

2001    Caco (103.7)   Frpe (105.6)
        Zaam (20.4)    Cepoc (47.2)
        Acsa (16.2)    Acne (23.6)
        Ulmus (16.2)   Ulmus (23.6)

TABLE 5.--Changes in Ulmus spp characteristics since 1950s by site

                       HL      FB      AR      LR       WY

Basal area
  ([m.sup.2]/ha)        0.0    -2.0    -3.7    -6.9    -14.5
Tree density
  (stems per ha)       10.6     5.4   -24.5   -17.2   -100.8
Understory relative
  density (%)         -11.6   -18.9    -8.2     7.9     -2.0

TABLE 6.--Top five most frequent groundlayer species with absolute
frequency by site and time period. Bolded species indicate species
not typically browsed by white-tailed deer (Dull, 2001). Starred
species indicate wet forest modal species (Curtis, 1959)

         HL

1950s    Laportea canadensis *          0.9
         Cryptotaenia canadensis *      0.85
         Galium obtusum                 0.8
         Dioscorea villosa              0.75
         Aster lateriflorus *           0.6
         Solidago gigantea              0.6

2001     Carex sp.                      0.75
         Parthenocissus quinquefolia    0.6
         Laportea canadensis *          0.55
         Galium aparine                 0.5
         Smilax hispdda                 0.5

         FB

1950s    Laportea canadensis *          0.8
         Carex sp.                      0.4
         Boehmeria cylindrica *         0.25
         Leersia virginica *            0.25
         Toxicodendron radicans *       0.25

2001     Laportea canadensis *          0.9
         Polygonum virginianum          0.85
         Geum canadense                 0.75
         Ranunculus hispidus            0.65
         Carex sp.                      0.45

         LR

1950s    Solidago gigantea              0.7
         Teucrium canadense *           0.7
         Menispermum canadense *        0.6
         Zanthoxylum americanum (a)     0.6
         Helianthus sp.                 0.5

2001     Parthenocissus quinquefolia    0.75
         Rubus occidentalis             0.60
         Menispermum canadense *        0.5
         Boehmeria cylindrica *         0.45
         Dioscorea villosa              0.45

         WY

1950s    Cryptotaenia canaaensis *      0.95
         Laportea canadensis *          0.75
         Cuscula gronvii                0.6
         Rudbeckia laciniata *          0.55
         Carex sp.                      0.45
         Ranunculus abortivus           0.45

2001     Phalaris arundinacea           0.7
         Lysimachia nummularia          0.55
         Laportea canadensis *          0.5
         Carex sp.                      0.45
         Ranunculus hispidus            0.45

(a) Zanthoxylum americanum was still present in
the 2001 survey with a frequency of 0.4

TABLE 7.--Frequency of non-native species in the ground layer in 2001.
None of listed species was recorded in the 1950s data

Species                  HL     FB     LR     WY

Convolvulus arvensis                  0.10
Lysimachia nummularia   0.10   0.15   0.05   0.55
Phalaris arundicacea    0.05   0.05          0.70
Rhamnus cathartica             0.25
Taraxicum officinale    0.10          0.05
Vinca minor             0.20   0.05          0.15
COPYRIGHT 2008 University of Notre Dame, Department of Biological Sciences
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2008 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Hale, Brack W.; Alsum, Esther M.; Adams, Michael S.
Publication:The American Midland Naturalist
Date:Oct 1, 2008
Words:9796
Previous Article:Landscape and local factors affecting Northern white cedar (Thuja occidentalis) recruitment in the Chequamegon-Nicolet National Forest, Wisconsin...
Next Article:Field confirmation of laboratory-determined lower temperature tolerance of transgenic and wildtype zebra danios, Danio rerio.
Topics:

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters |