Vegetation and elevational gradients within a bottomland hardwood forest of southeastern Louisiana.
Bottomland hardwood forests are a characteristic plant community of the Mississippi River deltaic plain of southeastern Louisiana with species belonging to the genera Quercus, Celtis, Liquidambar, Ulmus, Fraxinus, Carya, Acer and Sabal (Penfound and Howard, 1940; Braun, 1950; Montz, 1972). Such forests typically develop on natural levee ridges and narrow streambanks built up by periodic depositional activity of river distributaries (Spearing, 1995).
In southeastern Louisiana, forested levee ridges delimit drainage basins and, in coastal regions, protect interior wetland marshes from erosion (Spearing, 1995). Levee ridges bordering inactive distributaries are subject to gradual erosion and subsidence; this results in more frequent and prolonged flooding and promotes a gradient of plant communities from bottomland forest to swamp and then marsh.
Several authors have described examples of bottomland vegetation, attempting to characterize an apparent zonation of tree species caused by slight changes in ground elevation and concomitant hydroperiod (e.g., Penfound and Hathaway, 1938; Penfound, 1952; Shelford, 1954; Hosner and Minkler, 1963; Bell, 1974; Bell and del Moral, 1977; Nixon et al., 1977; Hupp, 1982, 1983; Hupp and Osterkamp, 1985; White, 1983. 1993). However, few studies have attempted a quantitative correlation between species distribution and elevational gradients (for examples see Bell, 1974; Bell and del Moral, 1977: Smith, 1996).
The present study characterizes a bottomland (old regrowth) forest bordering an inactive distributary in southeastern Louisiana. Although our principal objective is to describe the local distribution of mature woody species over an elevational gradient, we also quantify the distribution of woody saplings and understory. We examine the proliferation of a recently introduced exotic species (Sapium sebiferum), and whether understory native species affect its sapling growth and regeneration. Invasive species have damaged and severely altered native habitats and species compositions around the world, and our study site may be no exception. We hope to determine whether or not natural conditions (e.g., shade competition from native understory species) will eventually inhibit the continued invasion and increasing density, of Sapium sebiferum at our site. We also compare the elevational distributions of tree species with expected distributions based upon regional wetland indicator data, and briefly assess the predictive power of such data at the local level.
The study site is part of the Bayou Sauvage Ridge (BSR), a natural levee bordering Bayou Sauvage and located within the Bayou Sauvage National Wildlife Refuge, approximately 32 km east of New Orleans (Business District). The Refuge is bordered on the north by Lake Pontchartrain, on the east by the Chef Menteur Pass and on the south by Lake Borgne [ILLUSTRATION FOR FIGURE 1 OMITTED]. The study site lies between U.S. Highway 11 and the Maxent Flood-Control Levee, an area of roughly 92 ha. Although the highest elevation is only about 1 m above sea level, the ridge is sufficiently well drained to support a bottomland forest. Mean annual temperature is about 20 C, with summer averages around 27 C, and winter averages around 12 C (White, 1983; U.S. Fish and Wildlife Service, 1994). Rainfall averages 162 cm per year (Penfound and Hathaway, 1938; White, 1983; U.S. Fish and Wildlife Service, 1994). The highest average monthly rainfall of 20 cm occurs in July (White, 1983).
Until about 200 years ago Bayou Sauvage was an active distributary of the Mississippi River, contributing sediment to the Saint Bernard lobe of the Mississippi River Delta (Spearing, 1995). Typical of such seasonally flooded distributaries, heavier sediment accumulated along the stream course to form parallel levee ridges, while finer sediments and organic material accumulated behind the levees in shallow lagoons and swales. Elevation of such levee ridges increases sharply along the fore-ridge or "stream-front" side, and decreases gradually on the back-ridge or levee "backlands" that border adjacent marshes or open lagoons. The depositional activity and elevational differences promote notably different soil composition at the fore-ridge and marsh edges of BSR with Sharkey silty clay loam at the former and Gentilly muck at the latter (Trahan, 1989).
Before 1886, the BSR was clear-cut for sugar cultivation (Penfound and Howard, 1940), but trees and much of the palmetto community on the back-ridge may have been spared because soils were too wet for agriculture. Today, the BSR supports a mature second-growth forest roughly 110-y-old. Areas adjacent to BSR are heavily disturbed, and much of the regrowth vegetation has been destroyed. Furthermore, a cypress (Taxodium distichum) swamp community bordering the BSR back-ridge (Blind Lagoon) was almost completely removed in the early 20th Century. Nevertheless, the hardwood forest at BSR, with live oaks draped with Spanish-moss and an understory of dwarf palmettos, is probably the oldest forest community in the New Orleans region.
Vegetation. - Woody vegetation was sampled along twelve 2 m wide belt transects (Cottam and Curtis, 1956; Cox, 1976; Causton, 1988). Transects were oriented roughly perpendicular to the long axis of BSR, extending from the margin of Bayou Sauvage to the edge of an open marsh surrounding Blind Lagoon ([ILLUSTRATION FOR FIGURE 1 OMITTED]; Table 1) and varied in length from 190510 m. All transects were sampled between May and December 1995. Transect origins were spaced approximately 185 or 285 m apart, depending upon the curvature of the ridge. Sampling began with the first woody stem (sapling or adult) at the bayou end of the transect and ceased with the last woody stem bordering the marsh. Each woody plant rooted within a transect was identified to species (Brown, 1945; Radford et al., 1968), and its position (meters from transect origin) noted. Diameter at breast height (DBH) was measured for all stems over 2 m tall and these individuals were considered adults. Woody saplings (0.3-2 m tall) were identified to species, counted and their transect positions noted. Palmettos [TABULAR DATA FOR TABLE 1 OMITTED] (Sabal minor) were counted and their cover (0-100%) was estimated using a line intercept method (Barbour et al., 1987).
Elevation. - Topographical changes along each transect were assessed by measuring elevation at 5 m intervals using a hydraulic leveling device (HLD) consisting of a water-filled three-liter, transparent plastic jug with an attached, transparent, flexible hose 34 m long (fitted with a removable air-tight clamp to prevent water loss during movement of the HLD). The jug was attached to a tree at some convenient height and the hose was moved along the transect; the distance between the water level in the HLD and the ground was measured with a meter stick. By carefully repositioning the jug, the same water level in the HLD could be maintained along the length of each transect. The height of the water level in the HLD was measured relative to the surface water level of Bayou Sauvage; the fluctuating water level of the bayou was measured relative to the top of a concrete weir structure located at the Maxent Levee [ILLUSTRATION FOR FIGURE 1 OMITTED]. Elevational measurements taken along all transects were then standardized against the average measured height of Bayou Sauvage relative to the top of the weir. Relative elevations along each transect were computed using the formula:
[E.sub.0-n] = A - [B.sub.0-n] + (D - C)
Where A = height of water level maintained in the hydraulic leveling device (HLD) relative to surface of Bayou Sauvage; [B.sub.0-n] = distance from ground to water level in HLD measured at 5 m intervals along each transect; C = distance from top of the permanent weir structure to the surface of Bayou Sauvage on days when transect elevations were recorded (i.e., a measure of Bayou Sauvage water level fluctuations); D = average of distance C for all twelve transects; and [E.sub.0-n] = ground elevations along transect relative to average height of Bayou Sauvage. For the purpose of determining the elevational amplitudes of plant species, individuals were assigned the elevational reading made closest to them along the transect.
Data analyses. - Since we did not employ plot-based sampling in the field, it was necessary to subdivide the vegetation data from the twelve belt transects into either contiguous 2 x 5 m plots (856 in 8562 [m.sup.2], BSR total area) or 2 x 25 m plots (171 plots in 8562 [m.sup.2]) for analysis. Frequency (number of plots in which a species occurs) was computed using 2 x 5 m plots, other computations (e.g., correspondence analysis) using 2 X 25 m plots. For general vegetation description, data were analyzed using standard methods to obtain density. (number of individuals/ha), relative (%) density, dominance (basal area/ha), relative (%) dominance, abundance (total number of individuals of each species found within the total area), relative (1%) abundance, frequency, relative frequency and importance values (relative density + relative dominance + relative frequency) (Cox, 1976; Barbour et al., 1987; Causton, 1988). StatMost (Datamost Corporation, Salt Lake City, UT, 1994-1995, Version 2.50) and Minitab (Minitab Incorporated, State College, PA, 1993, Release 9) statistical software were used for all computations.
In order to assess species associations using correspondence analysis (COA; see Ludwig and Reynolds, 1988 for a detailed description) and the influence of Sabal minor cover on sapling abundance the 171, 2 X 25 m plots were used, since this plot size was large enough to show variation in the data. Species of especially low abundance (Morus spp., Persea palustris, Taxodium distichum and Cephalanthus occidentalis) or of broad elevational distribution (Sapium sebiferum, Sabal minor) were not included in the COA. To interpret the COA ordination of the 171, 50 [m.sup.2] plots and 17 woody species, COA components were related with the environmental variable, elevation, using Pearson Product-Moment correlations (Ludwig and Reynolds, 1988). Linear regression and Pearson Product-Moment analyses were used to test for correlation between palmetto cover and abundance of Sapium sebiferum and Diospyros virginiana saplings in the plots.
Canopy. - Adult individuals of 24 woody species were found within the BSR sample (Table 2). Sapium sebiferum, an introduced exotic, dominates the site with 2128 adults counted in the belt transects (2485 stems/ha). The importance value of this weedy species far exceeds that of any native hardwood species in the study area. Among adult native trees, Quercus virginiana and Celtis laevigata codominate; although their relative abundances and corresponding frequencies are low in comparison with those of S. sebiferum, they have high dominance because of large trunk sizes. The principal subdominant is Salix nigra. Acer rubrum, Diospyros virginiana and Daubentonia drummondii are well represented, but their low dominance can be attributed to small trunk diameters. Taxodium distichum is represented by a few large individuals that apparently escaped earlier logging. Sabal minor has the highest density in the understory throughout the BSR study site (Table 2). The dominant hardwood shrub is Ilex decidua and thickets of multiple-trunked individuals were found on every transect.
Elevational gradients. - Ground elevations at BSR (relative to average measured height of water in Bayou Sauvage) varied from -52.2 cm (transect 1) to +111.3 cm (transect 4) (Table 1). The bayou and marsh ends of the transects are frequently inundated, while land above ca. 60 cm are flooded only after exceptionally heavy rains (pers. obs.). The higher portions of the BSR support the greatest number of woody species (Table 3). Native hardwoods characteristic of higher elevations (median elevation [greater than]57 cm; minimum elevation [greater than]28 cm) are Ulmus rubra, Quercus virginiana, Myrica cerifera, Cornus drummondii, Celtis laevigata, Liquidambar styraciflua, Ilex vomitoria, Persea palustris and Quercus nigra (Table 3). Other species such as Daubentonia drummondii, Salix nigra, Taxodium distichum and Cephalanthus occidentalis do not occur at elevations above 52 cm and are most common below 15 cm, often growing in standing water. Although S. nigra and D. drummondii are restricted to low elevations, they rarely occur together; S. nigra is largely confined to the fore-ridge margin of Bayou Sauvage, and D. drummondii to the margin of the back-ridge marshes. Species such as Acer rubrum, Fraxinus americana, Baccharis halimifolia, Diospyros virginiana, Gleditsia triacanthos, Carya aquatica and Ilex decidua are most often found at intermediate elevations (38 cm to 58 cm) on the back-ridge (Table 3). Cornus drummondii and Gleditsia triacanthos have the narrowest elevational amplitudes; relatively narrow amplitudes in other species may be due to small sample sizes. Sapium sebiferum and Sabal minor have the widest elevational ranges, occurring at all elevations except in standing water.
Woody saplings and understory. - Only one-third of all tree species at the BSR study site produce substantial numbers of saplings, with Sapium sebiferum producing more (3916 per ha) than all other species combined. Among native species saplings of Daubentonia drummondii are most abundant. The number of saplings of understory tree species (e.g., Diospyros virginiana, Ilex decidua, I. vomitoria, Baccharis halimifolia) is higher than for overstory trees, except for Acer rubrum and Fraxinus americana, In general, species with the greatest density of adults also have high numbers of saplings. Also, median elevations of saplings (Table 4) correspond well with median elevations of adult trees.
Sabal minor covered 46% of the total area sampled; among 856 plots, 227 had 100% S. minor cover, 307 had varying percentages and 322 plots had no S. minor cover. No significant correlation was found between percent S. minor cover and sapling abundances of either Sapium sebiferum or Diospyros virginiana. Saplings of other species were not sufficiently abundant to warrant similar analysis.
Plant communities. - The COA on the transformed belt transect data (171, 50 [m.sup.2] plots including 17 woody species) yielded coordinates for both the 171 plots and the 17 species; [TABULAR DATA FOR TABLE 2 OMITTED] [TABULAR DATA FOR TABLE 3 OMITTED] each set of coordinates was analyzed for community structure at BSR [ILLUSTRATION FOR FIGURE 2 OMITTED]. The 2nd and 3rd components of the COA explained approximately 70% of the variance in the data and only these were used for ordination interpretation.
The COA ordination of the 17 woody species revealed one large clump of all species, except Salix nigra and Daubentonia drummondii [ILLUSTRATION FOR FIGURE 2 OMITTED], which are usually present as solitary tree species in low-elevation plots at opposite ends of each transect. The species coordinates on the 3rd COA component significantly correlated with the measured median elevations of the species (r = -0.74, P [less than] 0.001) whereas the species coordinates on the 2nd COA component did not significantly correlate with elevation (r = 0.17, P [greater than] 0.5), but may instead reflect soil differences.
The COA ordination of the 171 plots revealed a similar pattern to the one described above, but with some additional resolution [ILLUSTRATION FOR FIGURE 2 OMITTED]. Plots 29, 31, 55, 56, 150, 151, 73, encircled as "low-ridge" [ILLUSTRATION FOR FIGURE 2 OMITTED] contained species most commonly found at intermediate elevations (e.g., Diospyros virginiana, Acer rubrum and Ilex decidua), whereas plots encircled as "high-ridge" contained more plots having high-ridge species as dominants. Similar to COA ordination of the 17 species described above, the plot coordinates on the 3rd COA component showed a significant negative correlation with median plot elevations (r = -0.63; P [less than] 0.001). The plot coordinates on the 2nd COA component did not correlate significantly with elevation (r = -0.33; P [greater than] 0.01), but may be related to edaphic features. Separation of Daubentonia drummondii along the 2nd axis of the ordination may be due to the influence of soil. The vertical separation is suggestive, however, given the disjunct distributions (occurring in plots at opposite ends of each transect) and similar elevation ranges of Salix nigra and D. drummondii (Table 3; see Discussion).
TABLE 4. - Summary of elevational data for saplings. Species are listed by median elevation: all elevations are in cm Species Min Max Mean Median Range Salix nigra -9.7 23.3 4.2 3.8 33.0 Carya aquatica 7.8 24.3 14.5 11.3 16.5 Cephalanthus occidentalis 13.3 13.3 13.3 13.3 0 Daubentonia drummondii -3.7 46.8 12.8 13.8 50.5 Sapium sebiferum -9.7 105.8 33.8 30.3 115.5 Ulmus rubra 16.9 57.8 37.4 37.4 40.9 Baccharis halimifolia 1.8 68.3 35.8 38.8 66.5 Acer rubrum 3.8 81.3 38.2 39.3 77.5 Fraxinus americana 8.3 99.8 46.1 49.3 91.5 Sambucus canadenis 36.3 72.3 52.6 49.3 36.0 Quercus virginiana 54.3 54.3 54.3 54.3 0 Diospyros virginiana 2.8 111.3 55.9 54.8 108.5 Liquidambar styraciflua 39.8 72.8 61.0 63.3 33.0 Ilex decidua 15.0 111.3 59.2 63.8 96.3 Morus alba 72.8 72.8 72.8 72.8 0 Ilex vomitoria 16.9 108.8 68.0 73.1 91.9
The COA ordinations of species and plots (Fig. 2) suggested the presence of two communities, each with one dominant woody species, and a third loosely circumscribed and more inclusive community of 15 species. The communities were delineated as follows: (1) Salix nigra community characteristic of bayou margins, (2) Daubentonia drummondii community bordering open marsh and (3) ridge-forest community, falling between the bayou and marsh, and tentatively divisible into two subcommunities, "high-ridge" and "low-ridge" [ILLUSTRATION FOR FIGURE 2 OMITTED].
The ridge-forest community, is comprised of species whose individual elevational ranges overlap considerably (Table 3), and whose median elevations differ only slightly (ca. 40-75 cm) (Table 3). Such overlap made difficult the division of the ridge-forest community despite the tentative existence of "high-ridge" and "low-ridge" as shown in the COA ordination of the 171 plots ([ILLUSTRATION FOR FIGURE 2 OMITTED]; cf., Discussion); indeed not all species within plots uniformly conformed to this separation. We ran a COA on the 50 [m.sup.2] plot data after removing Salix nigra and Daubentonia drummondii to test for additional community structure within the ridge-forest, but did not find sufficient support for the presence of nested communities despite the COA on the 171 plots and field observations; the COA ordination of the 15 species produced a group of points with no discrete separations and neither COA component significantly' correlated with elevation (2nd component: r = -0.25; P = 0.38; 3rd component: r = 0.28; P = 0.32).
Penfound and Howard (1940) reported the results of a phytosociological study in a ridge forest at Bayou Sauvage, roughly 4 km southwest of our site. At the time of their study the ridge vegetation was approximately 50-y-old (regrowth after clear-cutting for sugar cultivation) and Sapium sebiferum had not yet been introduced. Among native trees, they listed Quercus virginiana and Q. nigra as dominants and Celtis laevigata ("mississippiensis") as the only important subdominant species. They did not find several species common at our study site (e.g., Salix nigra, Ilex decidua, Carya aquatica, Liquidambar styraciflua, and Diospyros virginiana) and, conversely, Crataegus spp. appeared regularly within their quadrats, but was absent at BSR. Differences in native species composition between the two sites may reflect differences in elevation, successional changes after clear-cutting or competition with the introduced exotic (i.e., Sapium sebiferum at BSR). Penfound and Howard, however, did not compare species distributions with ridge elevation and, unfortunately, the forest they studied has since been cleared for housing development.
White et al. (1983) sampled a similar bottomland hardwood forest near Lafitte, Louisiana, approximately 48 km southwest of the BSR study site. Most species found at BSR were also reported from Lafitte, although with different relative dominance. At Lafitte, the dominant native hardwood was Quercus nigra. Subdominants were Celtis laevigata, Quercus virginiana and Liquidambar styraciflua. Sapium sebiferum was not listed in the White et al. inventory, but was mentioned as a pernicious weed of spoil banks outside their study site; individuals were scattered in the sampled forest but considered too small to be counted or measured (pers. obs., White et al., 1983).
It remains to be seen what affect Sapium sebiferum will have on the structure of the Lafitte forest, although personal observations suggest that this introduced exotic continues to have high seed output and increased establishment there. As such, it seems likely that the Lafitte site will come to resemble BSR more closely in its S. sebiferum abundance and dominance in the near future. Additional direct comparisons should prove useful.
Species distributions and elevational gradients. - Elevation is a complex gradient affecting various environmental factors (e.g., soil, light availability, salinity and hydrology). At BSR, as in other floodplain forests, hydroperiod (flood frequency and duration) is the most important factor determined by elevational gradients, and largely defines the distributions of plant species that have different water tolerances (Hosner and Minkler, 1963; Bell, 1974; Bell and del Moral, 1977; Nixon et al., 1977; White, 1983; Hupp and Osterkamp, 1985). Data from Hosner (1958, 1960) and Hosner and Boyce (1962) indicate that saplings of species such as Salix nigra, Acer rubrum and Fraxinus spp. will tolerate saturated soils for extended periods, whereas species of Quercus, Celtis, Liquidambar and Ulmus suffer high mortality; Salix species are the most tolerant of inundation. Those findings generally agree with our data from BSR: Salix nigra and Daubentonia drummondii tolerate the wettest conditions, whereas Acer rubrum, Fraxinus americana and Diospyros virginiana tolerate slightly drier soils and intergrade with species such as Quercus virginiana, Celtis laevigata, Liquidambar styraciflua and Ulmus rubra (more common at "high-ridge" elevations). As expected, increases in elevation to drier soils on the BSR parallel an increase in number of species, as found at other riparian sites (e.g., Bell, 1974; Bell and del Moral, 1977; Hupp, 1982). Also, such specific elevational requirements among species at BSR (Table 3) may be useful in predicting future forest species composition based on a known hydrological regime, or alternatively, predictions about a site's hydrological regime given the species composition (see "Wetland indicator species" this section). Likewise, such strong elevational/hydrological requirements may be at least partially responsible for the formation of plant communities.
Our COA indicated some influence of elevation on community formation; the ordinations revealed at least three communities: (1) Salix nigra, (2) Daubentonia drummondii and a larger community (3) the ridge forest (15 species). However, we expected that such a plot-based COA would give strong support for the presence of other species associations within the ridge-forest community at BSR. Field observation suggested that some tree species (e.g., Ilex vomitoria, Celtis laevigata, Ulmus rubra, Quercus virginiana, Q. nigra, Liquidambar styraciflua) occur more frequently at higher ridge elevations ("high-ridge" species), while others (e.g., Acer rubrum, Diospyros virginiana, Fraxinus americana, Ilex decidua, Baccharis halimifolia, Gleditsia triacanthos and Carya aquatica) are more frequent at lower elevations toward the marsh ends of our transects ("low-ridge" species). Prior studies (e.g., White et al., 1983) recognized (perhaps intuitively) such high- and low-ridge communities in other Louisiana bottomland forests. This differentiation was noticeable in part from the COA ordination on our sampling plots, but insufficient evidence existed to delimit those communities formally. The shape of our sampling plots (long axis parallel to the elevational gradient), and the fact that plots were adjacent to one another, may have contributed to our inability to distinguish species associations within the ridge-forest community. Also, the large overlap in elevational ranges among many species of bottomland hardwood forests (including BSR) suggests that such communities may not be reliably defined (Braun, 1950; Bell, 1974). Furthermore, human disturbance cannot be ruled out as a potential confounding element in the community structure analysis at BSR, especially when considering the disturbance history of this forest and its surrounding areas.
As mentioned, Salix nigra and Daubentonia drummondii dominate fringing communities at the boundaries of the ridge forest. However, differences in hydroperiod do not adequately explain the confined distributions of S. nigra on the levee fore-ridge and D. drummondii at the marsh fringes, since they occur at comparably low elevations. The local distributions of those low-elevation species may instead reflect soil differences: Sharkey silty clay loam characterizes much of the fore-ridge while Gentilly muck characterizes the marsh. The ordination [ILLUSTRATION FOR FIGURE 2 OMITTED] does not contradict such a hypothesis since the D. drummondii community separated more distinctly along the vertical (axis II) rather than the horizontal (axis III) axis. Only the third axis correlated significantly with elevation; the second may correlate with soil variables, but more precise data are needed before firm conclusions can be reached.
Forest regeneration. - Since its introduction into the southeastern U.S. earlier in the 20th Century Sapium sebiferum has become a prolific weed, particularly of disturbed habitats. A long history of disturbance by humans and feral pigs may have encouraged its recent establishment at BSR. where S. sebiferum individuals are no more than 35-y-old, based on tree core data and measured DBH (Harper, 1995). The species is now naturalized in wet prairies and bottomland forests throughout the southeastern coastal plain (Jones and McLeod, 1989). Harper (1995) found S. sebiferum growing in a wide range of habitats at Lafitte, Louisiana. In her study, seedling growth and survival were strongly inhibited under palmetto cover, yet saplings and adults were not affected by understory shading (Harper, 1995). These data support previous claims that individuals of S. sebiferum are relatively shade tolerant throughout most of their development (Jones and McLeod, 1990). At BSR, the distribution of S. sebiferum saplings was not correlated with palmetto cover and seemed unaffected by shading. Such shade and flood tolerance coupled with adult and sapling densities suggest that although S. sebiferum individuals are short-lived (Jones and McLeod, 1989) the species will continue to be a dominant and potentially disruptive component of BSR and similar forests.
At BSR, saplings of native overstory species were rare, but such apparent underrepresentation may be typical of these floodplain communities. Harcombe and Marks (1978) found that saplings of overstory tree species were significantly underrepresented in floodplain forests of eastern Texas compared with sapling abundances of the same species in drier habitats. They attributed this underrepresentation to the denser canopy and understory in bottomland forests and the inability of saplings to withstand low light conditions. Flooding may also affect sapling mortality. However, it remains to be seen what affect the population explosion of Sapium sebiferum will have on the regeneration of native canopy species.
Wetland indicator species. - Plant species are often reliable indicators of local hydrology (Hupp and Osterkamp, 1985). In southern Louisiana, for example, early settlers frequently used the distributions of tree species to identify land suitable for agriculture (Brown, 1944). Today, some plant species ("wetland indicators") are used by governmental agencies to identify, and delineate wetland habitats (see Reed, 1988). Comparing our findings with Reed's wetland indicators (Table 3), wetland species designated "obligate" by Reed are generally those that have the lowest mean elevation at BSR, whereas species with wide elevational ranges at BSR (e.g., Sapium sebiferum, Sabal minor) are labeled "facultative" (likely to occur in wetlands as well as in nonwetlands).
Some notable discrepancies occur between Reed's classification and our findings at BSR. For example, Reed classified Daubentonia drummondii as a "facultative" wetland species, but at BSR this species is confined to low elevations along the fringes of the freshwater marsh and has an elevational distribution indicative of an "obligate" wetland species. Also somewhat surprising is the listing of Carya aquatica as an "obligate" wetland species, since it is not uncommon at higher elevations at BSR and appears to have a "facultative" hydrological tolerance. Reed's classification scheme, based on species ecology over a regional area, does not appear to be strictly applicable at the local level. We suggest caution when using indicator species to delineate local wetlands. Continued research along these lines should serve to complement our data and better define elevational amplitudes of bottomland forest species in general, making delineation of wetlands via indicator species a more reliable practice.
Acknowledgments. - We thank Thomas W. Sherry, Terry E. Christenson and two anonymous reviewers for their invaluable advice on early versions of the manuscript. Christoph A. Walser, Brett J. Falterman, Anne S. Bradburn, David A. White (Loyola University) and Paul A. Harcombe (Rice University) assisted with data analyses and field methods. We also thank James Harris and Charlotte Parker (U.S. Fish and Wildlife Service, Bayou Sauvage National Wildlife Refuge) for suggesting the study site and for logistical support. John Bruza and Glen N. Montz (U.S. Army Corps of Engineers, New Orleans Division) and Angie Friloux (U.S. Department of Agriculture, Natural Resources Conservation Service) directed us to pertinent documents and literature. Finally, to all those who assisted in the field: Darron A. Collins, Maria G, Fadiman, Alexander M. Gleason, Catherine M. Downs, Adin R. Murray and Samara Alpern, we extend our sincere thanks,
BARBOUR, M. G., J. H. BURK AND W. D. PITTS. 1987. Terrestrial plant ecology. The Benjamin/Cummings Publishing Co., Menlo Park. xii + 634 p.
BRAUN, E. L. 1950. Deciduous forests of eastern North America, The Free Press, New York. xiv + 596 p.
BROWN, C. A. 1944. Historical commentary on the distribution of vegetation in Louisiana and some recent observations. Proc. of the Louisiana Acad. of Sci., 8: 35-47.
-----. 1945. Louisiana trees and shrubs. Louisiana Forestry Commission Bulletin Number 1. x + 262 p.
BELL, D. T. 1974. Tree stratum composition and distribution in the streamside forest. Am. Midl. Nat., 92: 35-46.
----- AND R. DEL MORAL. 1977. Vegetation gradients in the streamside forest of Hickory Creek, Will County, Illinois. Bull. Torrey Bot. Club, 104: 127-135.
CAUSTON, D. R. 1988. An introduction to vegetation analysis. Unwin Hyman, Boston. xxvii + 342 p.
CLEVELAND, W. S. 1985. The elements of graphing data. Wadsworth Advanced Books and Software, Monterey, California. xii + 323 p.
COTTAM, G. AND J. T. CURTIS. 1956. Use of distant measures in phytosociological sampling. Ecology, 37: 451-460.
COX, G. W. 1976. Laboratory manual of general ecology. Wm. C. Brown Co., Dubuque, Iowa. vii + 237 p.
HARCOMBE, P. A. AND P. L. MARKS. 1978. Tree diameter distributions and replacement processes in southeast Texas forests. For. Sci., 24: 153-166.
HARPER, M. G. 1995. A model of the invasion of Chinese tallow tree (Sapium sebiferum (L.) Roxb.) into a bottomland-hardwood forest in Louisiana. M.S. Thesis, Tulane University, New Orleans, Louisiana. x + 116 p.
HOSNER, J. F. 1958. The effects of complete inundation upon seedlings of six bottomland tree species. Ecology, 39: 371-373.
-----. 1960. Relative tolerance to complete inundation of fourteen bottomland tree species. For. Sci., 6: 246-251.
----- AND S. G. BOYCE. 1962. Tolerance to water saturated soil of various bottomland hardwoods. For. Sci., 8: 180-186.
----- AND L. S. MINKLER. 1963. Bottomland hardwood forests of southern Illinois - regeneration and succession. Ecology. 44: 29-41.
HUPP, C. R. 1982. Stream-grade variation and riparian-forest ecology along Passage Creek, Virginia. Bull Torrey Bot. Club, 109: 488-499.
-----. 1983. Vegetation patterns on channel features in the Passage Creek Gorge, Virginia. Castanea, 48: 62-72.
----- AND W. R. OSTERKAMP. 1985. Bottomland vegetation distribution along Passage Creek, Virginia, in relation to fluvial landforms. Ecology, 66: 670-681.
JONES, R. H. AND K. W. McLEOD. 1989. Shade tolerance in seedlings of Chinese tallow tree, American sycamore, and cherrybark oak. Bull. Torrey Bot. Club, 116: 371-377.
----- AND -----. 1990. Growth and photosynthetic responses to a range of light environments in Chinese tallowtree and Carolina ash seedlings. For. Sci., 36: 851-865.
LUDWIG, J. A. AND J. F. REYNOLDS 1988. Statistical ecology. John Wiley and Sons, New York. xvii + 337 p.
MONTZ, G. N. 1972. A seasonal study of the vegetation on levees. Castanea, 37: 140-146.
NIXON, E. S., R. L. WILLETT AND P. W. COX. 1977. Woody vegetation of a virgin forest in an eastern Texas river bottom. Castanea, 42: 227-236.
PENFOUND, W. T. 1952. Southern swamps and marshes. Bot. Rev., 18: 413-446.
----- AND E. S. HATHAWAY. 1938. Plant communities in the marshlands of southeastern Louisiana. Ecol. Monogr., 8: 1-56.
----- AND J. A. HOWARD. 1940. A phytosociological study of an evergreen oak forest in the vicinity of New Orleans, Louisiana. Am. Midl. Nat., 23: 165-174.
RADFORD, A. E., H. E. AHLES AND C. R. BELL. 1968. Manual of the vascular flora of the Carolinas. University of North Carolina Press, Chapel Hill. xi + 1183 p.
REED. P. B. 1988. National list of plant species that occur in wetlands: Southeast (region 2). U.S. Fish and Wildlife Service (Biological Report), U.S. Department of the Interior. 71 p.
SHELFORD, V. E. 1954. Some lower Mississippi valley flood plain biotic communities: their age and elevation. Ecology, 35: 126-142.
SMITH, R. D. 1996. Composition, structure, and distribution of woody vegetation on the Cache River floodplain, Arkansas. Wetlands 16: 264-278.
SPEARING, D. 1995. Roadside geology of Louisiana. Mountain Press Publishing Co., Missoula, Montana. xlii, 225 p.
TRAHAN, L. J. 1989. Soil survey of Orleans Parish, Louisiana. U.S. Department of Agriculture, Soil Conservation Service. vii + 90 p.
U.S. FISH AND WILDLIFE SERVICE. 1994. Bayou Sauvage National Wildlife Refuge master plan report [Cashio, Cochran, Torre/Design Consortium, Ltd.]. iv + 98 p.
WHITE, D. A. 1983. Plant communities of the lower Pearl River basin, Louisiana. Am. Midl. Nat., 110: 381-396.
-----. 1993. Vascular plant community development on mudflats in the Mississippi River delta, Louisiana, USA. Aquatic Bot., 45: 171-194.
-----, S. P. DARWIN AND L. B. THIEN. 1983. Plants and plant communities of Jean Lafitte National Historic Park, Louisiana. Tulane Studies Zool. and Bot., 24: 100-129.
|Printer friendly Cite/link Email Feedback|
|Author:||Wall, Dennis P.; Darwin, Steven P.|
|Publication:||The American Midland Naturalist|
|Date:||Jul 1, 1999|
|Previous Article:||Composition and diversity of ground-layer vegetation in silvicultural openings of Southern Indiana forests.|
|Next Article:||Differences between seed bank composition and field recruitment in a temperate zone deciduous forest.|