Printer Friendly

Ecology and Vegetation of LeFleur's Bluff State Park, Jackson, Mississippi.

Mississippi Natural Heritage Program, Mississippi Museum of Natural Science Department of Wildlife, Fisheries and Parks, 2148 Riverside Drive, Jackson, MS 39202-1353

An ecological evaluation of the landscape, soils, hydrology, and vegetation of LeFleur's Bluff State Park, Jackson, Hinds County, Mississippi, was conducted during a 14 month period from September 1995 to November 1996. Of the park's 492 ac (200 ha) area, studies included a 390 ac (158 ha) natural area but excluded the golf course. Plot sampling and analysis with the ordination programs, TWINSPAN and CANOCO, demonstrated that vegetation composition is associated with geological strata, soil texture, chemistry and class, springtime ground water level, and flooding frequency and duration of the Pearl River. Applying a Monte Carlo permutation test, average spring ground water depth (p =.005), days of flooding (p=.005), pH (p=.005), and base saturation (p=.05) were shown to be highly correlated to vegetation composition. Preliminary records indicate over 460 species of plants including 69 exotics inhabit the study area. Vegetation classification and mapping indicated bottomland hardwood forest of several moistur e zones forms a major part of the vegetation mantle. The most extensive forest types include ridge bottom (109 ac), low terrace (66 ac), wet bottom (27 ac), and swamp (22 ac), comprising 57 % of the study area. The remaining area includes several oxbow lakes, freshwater marshes, shrub wetlands, special use areas, a mussel bed, a segment of the Pearl River, and calcareous upland forest (13 ac), a rare community in Mississippi. A number of species of special concern inhabit the park environs: a coneflower, an orchid, three fish, one reptile, and two mussels. The natural area will provide opportunities for field studies implemented through the Mississippi Museum of Natural Science, which opened at its new location in LeFleur's Bluff State Park on 3 March, 2000.

The new Mississippi Museum of Natural Science (MMNS), Department of Wildlife, Fisheries and Parks, commands a high point overlooking the natural bluff and forested floodplain at its new location in LeFleur's Bluff State Park (LBSP), in Jackson, Hinds County, Mississippi (Figure 1). As part of its mission to conserve the natural resources of Mississippi, a theme of the Museum of Natural Science is to present flora and fauna in attractive dioramas and aquaria depicting various landscapes and water bodies of the state. In addition, the museum is intended to be a "living museum," not only archiving and displaying preserved specimens, but also taking advantage of the adjacent natural areas and gardens for viewing native plants and animals of the region. The approach will encourage people to discover more about the local biota and ecoregion in which they live. An abundance of life and the intricacies of natural patterns and processes, whether environmental or biological, are dynamically intertwined and unfolding i n the natural area behind the new museum facility. Remarkable viewing is available to those who venture into the natural world.

The park was chosen as the site for the MMNS for several reasons. The site contains representative examples of natural habitats and forested ecological communities and can be used as an outdoor classroom for ecological studies of the natural environment and biota found in central Mississippi. The park offers access to over 390 ac (158 ha) of natural bluff and bottomland forest, several oxbow lakes, a mile of frontage along the Pearl River, Eubanks Creek, and various other wetlands. The forest and wetlands will serve as an outdoor laboratory for natural science education projects and can provide areas to conduct additional ecological research supported by the MMNS. Examples of such research are provided by Skeen (1974, 1976) and Funderburk and Skeen (1976).

The park's natural area could serve as a site for studies to understand and quantify services provided by natural ecosystems. For example, forests and wetlands provide essential services to society by purifying air and water, protecting soil, mitigating floods, maintaining biodiversity, moderating local temperatures and winds, and providing aesthetics and intellectual stimulation (Daily, 1997). On a global scale they help to contain global warming through the sequestration of carbon stocks in trees and soils (Woodwell and MacKenzie, 1995).

The objectives of this project were to conduct environmental and vegetative investigations on the ecological interconnectivity of the landscape, geology, hydrology, soil, and vegetation of LeFleur's Bluff State Park. The study is to provide baseline data for monitoring Wends in forest succession, tree growth rates, and species composition, and enhance the educational opportunities available to park managers, naturalists, museum educators, and visitors. Furthermore, the studies will help to determine the best options available for maintaining natural areas and restoring degraded communities and habitats, especially concerning invasive exotic species.

History of LeFleur's Bluff State Park

The City of Jackson purchased land from the State Asylum Property in 1944 to create Riverside Park, which officially opened in 1949. At that time Riverside Park included uplands on the west side of the present day LBSP and baseball fields adjacent to Lakeland Drive. Later additional land purchases increased the size of the park to include a large area of the Pearl River floodplain bottomlands. Around 1955, several oxbow swamps were converted into lakes by building levees across their lower sides. Several lakeside cottages were established to take advantage of the placid setting and fishing opportunities on Mayes Lakes, named after a former property owner, Mayes Gordon. In the early 1970's, a golf course was established on the uplands situated on the northwestern side of the park. The park was transferred to the state and designated LeFleur's Bluff State Park in the 1980's.

The park was named after LeFleur's Bluff, a small Wading post established in 1821 south of the present day Old Capitol Museum, Jackson. The LeFleur's Bluff site was selected as the location for the new state capitol because of its central location within the state, access to a navigable river, and position near the Natchez Trace, a major transportation route at that time (Brinson, 1980). Although LBSP is only two miles northeast of the Old Capitol Museum, the lands surrounding the park remained relatively rural, as a 1938 photo indicates. Only Jackson's water treatment system and Belhaven subdivision adjoined the park's boundary. Lakeland Drive (Highway 25) did not yet extend as far as the Pearl River. Land use within the watershed of Eubanks Creek, which flows through the park, was largely agricultural. After over a century of agrarian settlements, only a small portion of the watershed remained forested. Over the next quarter century, Jackson grew to such a size that it enveloped the park on its western and northern sides and on its eastern side, Flowood began to expand. In 1963 the completion of a three-mile earthen dam across the Pearl River 10 river miles north of the park near Madison created the Ross Barnett Reservoir which was built to provide a permanent water supply for the metropolitan area. At that time parts of Jackson flooded almost annually (Brinson, 1980). To alleviate the threat of flooding of the city's eastern lowlands and fairground area, additional flood control levees were completed in 1965 (Carroon, 1982). During that year, with the establishment of the Interstate 55 corridor and channelization of the Pearl River below Lowhead Dam, development of the surrounding commercial and residential areas intensified. Channelization reduced the river segment length from 7.7 miles to 3.6 miles and consequently more than doubled the slope of that portion of the river. The levees and channelization projects were not sufficient in preventing flooding that ensued shortly thereafter. Major spring floods on the Pearl River in 1979, and to a lesser extent in 1983, caused extensive damage to low lying businesses and residential areas. In 1979, the floods ranged from 13 feet to 33 feet above the bottomlands of LBSP, which lie below the 270 foot contour. Due to its lowland character, flooding occurs across a significant portion of the park bottomlands on an annual basis. Exemplified by the construction boom along Lakeland Drive beginning in the mid-1980's, development continues to urbanize the landscape surrounding LBSP.

The studies focused on six categories of investigation: 1) regional setting and geology; 2) area, landform, soils, and ground water; 3) hydrology of Pearl River at Jackson; 4) flora and vegetation composition; 5) ordination analysis of vegetation and environmental parameters; and 6) ecological community characterization. Methods used to accomplish this study are subdivided according to these categories and described per category at the beginning of each of the sections that follow. A brief introduction to the field studies will be presented here. The project involved the preliminary reconnaissance of vegetation and landscape features in the fall of 1995, the establishment and selective placement of eighteen permanent plots (macroplots) during the spring and summer of 1996, and subsequent analysis of species-environmental relationships using several ordination programs. The plots were classified and grouped according to the analytical findings and within these groupings data were summarized (Table 1). ARC-INF O GIS software was used to map site boundaries, topography, macroplot location, trails, plant collections, and ecological communities (Belokon, 1996).

Regional Setting and Geological Review

Regional ecological land classification systems can help orient one to a region and point out the important environmental features that dictate the character of that region. The systems of classification provide a means of comparison and contrast of environmental conditions among regions and can be used for planning and monitoring programs. Bailey (1980) developed a national hierarchical classification of ecological units to provide a framework for stratifying the Earth into progressively smaller areas of increasingly uniform ecological potentials. The U.S. Forest Service adopted the framework for instituting ecosystem management on lands under their jurisdiction. According to the classification, there are four major hierarchical levels: the global level is called the Domain; the continental level, Division; the regional level, Province; and subregional level, Section (also called Ecoregion by The Nature Conservancy, 1996). Within this framework, LBSP lies within the Upper Gulf Coastal Plain Ecoregion of the Southeastern Mixed Forest Province of the Subtropical Division of the Humid Temperate Domain (McNab and Avers, 1994; Keys et al., 1995). There are an additional four subdivisions--Subsection, Landtype Association, Landtype, and Landtype Phase--each of which exhibit increasing ecological and landscape specificity. The park lies on the boundary of the Deep Loess Plains Subsection and the Jackson Prairie Subsection, both of which occur within the Jackson Prairie Topographic Region (Lowe 1921). The lower three subdivisions of the classification have not been mapped for the state of Mississippi.

Geological strata that underlay the LBSP landscape were evaluated to help interpret vegetation patterns found in the park. One often physiographic regions of Mississippi (Lowe, 1921), the Jackson Prairie Belt extends across central Mississippi from the edge of the Loess Bluff Region, at Yazoo City, across most of Hinds County, including LBSP, to the eastern boarder between Alabama and Mississippi, and continues into Washington County, Alabama (Moore, 1969) (Figure 1). In the Gulf Coast stratigraphy, the Eocene Epoch is divided into Wilcox, Claiborne, and Jackson Stages. The Jackson Stage, a time-rock unit of approximately 38 million years (Schiebout, 1986), contains two formations that outcrop along the LBSP bluff line. From upper to lower, they are the Yazoo Clay and the Moody's Branch Formations. The Cockfield Formation, which is part of the Claiborne Stage, is not exposed in the park but is positioned underneath the Pearl River floodplain. Surficial alluvial materials of the bottomlands are of Pleistocene Era or Recent Era origin.

Yazoo and Moodys Branch sediments were deposited in a neritic, or deep marine environment of the Mississippi embayment (Smith and Zumwalt, 1987). Yazoo strata is composed of green and gray calcareous clays containing some sand and marl. The Moodys Branch Formation, a sixteen to thirty feet thick stratum, is a mixture of shells embedded in very limey, fossiliferous clays and glauconitic sands (Moore et al., 1965). These strata are responsible for the genesis of circumneutral to basic soils that are found on the bluff slopes of the park, located just behind the new Museum of Natural Science. An outcrop of this strata, found along the base of the bluff, has been recognized as an alternate geologic type locality of the Moodys Branch (Moore et al., 1965). Being rich in well preserved Eocene-aged fossils (Dockery, 1977), the bluff has served as a venue for numerous geology field trips and will continue to provide educational opportunities as an outdoor earth science exhibit for the new museum. The Cockfield Format ion, a stratum consisting of clays and fine-grained silty sands, functions as an aquifer that potentially receives recharge from the Pearl River wetlands.

Area, Landform, Soil, and Ground Water

The study area, included 79% (390 ac) of the park, comprising the park's natural area and several special use areas. The golf course, an associated access area, a driving range, and a storage site were excluded from the study area. The study area consisted of 22% open water (86 ac) and 6% (25 ac) special use areas, including a day use area, campground, several parking lots, and access roads.

Landscape features were subjectively determined for each macroplot by noting its juxtaposition along the landscape catena as referenced in Table 1 (Taylor et al., 1990) (Table 2). A total of five soil pits were dug and the soil profiles were described using standard soil survey methodology (Soil Survey Staff, 1975). Four soil pits, seven feet deep, were dug on a range of low, medium and high floodplain terrace levels. One pit represented each of the low and high terrace landforms and two represented the medium level terrace. An additional pit was dug in the uplands along the bluff shoulder. Soil samples taken at each macroplot were obtained by inserting a probe to a depth of 12 inches at five points along the edge and at the center of each macroplot. Standard mechanical and chemical analysis documented the following physical and chemical attributes of the soil: texture, pH, cation exchange capacity, base saturation, and amounts of the extractable nutrients, potassium, calcium, and magnesium.

A qualified representation of major environmental factors--landscape position, spring flooding frequency and duration, spring ground water level, and soil texture type--provides an overview of their interrelationship and influence on the composition of the vegetation, as represented by the listed indicator species (Table 2). The most important parameter for bottomland species is wetness, manifested in the frequency and duration of riverine flooding, while parent material is an important parameter influencing upland species. Siwell soils found along the bluff slopes have circumneutral to basic reactions because of the calcareous constituents of the Yazoo clay parent materials. Soil differences among landforms were in most cases distinctive enough to merit a different soil series designation for each landform. Investigations revealed that Rosebloom soils underlay swamplands; Chenneby soils underlay lower terraces and wet bottoms; Cascilla soils underlay the middle terraces; Bigbee soils underlay high terraces; and Siwell soils underlay the upper slopes of the bluff uplands (Soil Survey Staff, 1975). Substrate characteristics of plot soil samples were averaged within habitat type groupings as listed in Table 1 (Table 3). Organic matter of mesic upland soils was 58% higher on the colluvial lower slopes as compared to the submesic shoulder slopes. Floodplain soils averaged 1.4% organic matter, somewhat lower than values determined for mesic uplands (2.2%). pH and base saturation averaged 5.1, and 51% for floodplain soils. In contrast, upland soils had an average pH of 7.5 and base saturation of 97%, a difference extremely important from the standpoint of soil fertility (Buckman and Brady, 1969). The indications are that the upland soils are more fertile and have a greater nutrient availability than the floodplain soils. Cation exchange capacity for the mesic upland habitat type was mostly attributable to calcium ions (91%), indicating calcareous soils and the likelihood of finding calciphiles on these soils. However, measurements of floodplain soils averaged 44% or 10 milliequivalents (me.) per 100 gm less than upland soils (17.9 me./100 gm).

While the bottomland periodically are transformed by flooding, the bluff lands are being shaped by precipitation and gravitational forces. The bluff slopes are steep and unstable, ranging from five to ten percent on the shoulders of the bluff and from 35 to 45% along the steep escarpment. Water erosion occurs during flash flood events, mostly in areas where the soil mantle is unprotected by vegetation. Trampling by pedestrians or uprooting of trees by strong winds are additional factors that move soil downslope.

The soil profile descriptions of floodplain soils provide clues for interpreting the historical character of the landscape and land use changes that have affected this part of the alluvial basin. The soil profile described for a low terrace, located along the west bank of Eubanks Creek (Plot No. 430), provided several leads towards understanding the historical context of this area. Although signs are often cryptic, soil profile characteristics suggested the landform has changed considerably from the land features observed at this location today.

Deep horizons of the profile indicated the area was formerly part of an old river channel. It is conceptualized that after the channel was abandoned by the river, it became plugged, and developed into an oxbow lake or backwater slough. Later Eubanks Creek came to occupy the channel. The area continued to actively build up sediments. The major interpretive features of the soil profile are a series of platy lamellae that contain diverse coloration, stains, and colored leaf fragments. In the lower horizons the darker series of lamellae suggest formation under low, wet swampy conditions in which fine silt and organic sediments settled in a slackwater pond. The upper sequence of lamellae, which contain a higher percentage of coarse particles and exhibit lighter brown colors, suggests a different. scenario to their formation. Their origin is from the rapid deposition of silty sediments during flood episodes of Eubanks Creek. The sediments, which eroded from the silty cap of the uplands within the creek's watershed , were deposited during post-settlement times when intensive forest clearing, agricultural row cropping, and urbanization impacted the area (Robert Hinton, pers.com.).

Additional interpretation of the middle terrace soil profiles provide clues for understanding perched water tables and engineering problems that have hampered construction activities on these soils. A temporary perched water table was observed in the Bw2 horizon that is situated above the IIBx horizon. The IIBx horizon, positioned from 43 inches to 57 inches deep, is compact and restricts water movement through the soil profile. Water remains perched above the restrictive layer after rainstorms but gradually dissipates through percolation, lateral flow or moisture uptake by roots. The IIBx layer, which exhibits strong, large angular blocky structure, is a former mid-level (B) horizon that was exposed by truncation of its surface (A) horizon. Later alluvial sediments reburied the horizon to the depth it is presently located. The significance of the strong blocky structure of the IIBx layer is that engineering problems arise when excavating channels through these soils. When exposed in cut banks, the large str uctural units of the IIBx horizon tend to slump during heavy rains or flooding, a condition that is exacerbated by a rapid fall of water levels in the affected channel. Mass wasting caused by the exposure of the IIBx layer can be seen along a lower section of Eubanks Creek (Robert Hinton, pers. com.).

Four inch diameter cased wells were dug to a depth of eight feet for each floodplain macroplot that was accessible in April, 1996. Free water level of the wells were monitored during a 60 day period starting 1 April. Ground water levels for each of the swamp, wet bottom, and terrace landforms were summarized (Figure 2). Water table levels at the beginning of April, 1996, were elevated when flooding occurred in depressions and other low lying areas. The water table draw down beginning in mid-April is coincident with springtime releafing of trees, when a massive amount of water is used by bottomland trees in photosynthesis and transpiration. By the end of May, ground water levels of the floodplain had dropped an average of 3 to 4 feet below ground level. Several low terrace and wet bottom wells had perched water tables that quickly dried up in April, thus indicating a higher rate of drop in ground water levels for these landforms (Figure 2).

Hydrology of the Pearl River at LeFleur's Bluff State Park

Hydrological studies provide insight into understanding the interrelationship of bottomland forest composition to the flood frequency and duration of southeastern rivers (Conner et al., 1981; Wharton et al., 1982; Shankman, 1996). Taylor et at. (1990) categorized hydrological zones and species associations for bottomland hardwood forests of the southeastern United States into six water regimes as follows: I) continuously flooded, II) intermittently exposed, III) semipermanently flooded, IV) seasonally flooded, V) temporarily flooded, and VI) intermittently flooded. Species associated with each of LBSP's floodplain forest communities have been matched to the zones in which they have been categorized (Larson et al., 1981; Taylor et at., 1990) (Table 2).

River flow volume, water level, and flooding duration data are available from permanent gages positioned along the Pearl River. U.S. Geological Survey operates two gaging stations in close proximity to the park: 1) the Pearl River at the Jackson Water Works gage (WW) (Station No. 02485740), situated at the south boundary of the park, and 2) the Pearl River at the U.S. Highway 80 Bridge gage (H80B) (Station No. 02486000), 3.6 miles south of the park. Additional gaging stations are positioned throughout the Pearl River Basin. In Jackson there are also gages established along Eubanks Creek and several other tributary creeks to measure local flash flooding. Because of the significant confusion that may result from converting river stage data to metric units, these data are treated in English Units (U.S. Geological Survey, 2000a, 2000b), and to maintain consistency, all other data, except those referring to soil chemistry, are processed in a similar fashion. Graphical representations of duration, flood frequency, and peak discharge rates were developed.

The Pearl River extends for a maximum length of 240 miles over a watershed of 8,760 [mi.sup.2] that covers about 14.7% of the state of Mississippi and three counties of Louisiana. The portion of the watershed above Highway 80 at Jackson accounts for about 37% of the total watershed area. Water flow rates reflect the humid climate of the area and the large land area within the upper Pearl River watershed.

Annual peak discharge of the Pearl River at Jackson (H80B) is projected to reach 106,000 cubic feet per second during a flood of one hundred year magnitude (Figure 3)(Landers and Wilson, 1991). A volume of this magnitude would inundate the park bottomlands under more than twenty feet of water in most places. The range in elevation of the park's floodplain is from 250 feet at the Lowhead Dam (Jackson Water Works Gage) to 270 feet along the floodplain rim. Daily mean gage heights were used to estimate flooding on the adjacent park bottomlands. For a period extending from 1976 through 1983, the river reached the 270 foot flood level 22% of the years. From 1985 to 1994, the 270 foot level was reached only once (Figure 4). In 65% of the years, the annual maximum was reached during the months of February, March, and April (Figure 5). Based on probability statistics of the U.S. Geological Survey, gage heights for the JWW Station are projected to reach or exceed the 260 foot level for 23% of the winter season (January, February, and March) and 7% of the spring season (Figure 6).

An estimate of the number of days of flooding per elevation level helps quantify flood duration in a more meaningful way (Figure 7). The values were determined by converting monthly river stage duration percentage tables to actual time values (U.S. Geological Survey, 2000a). The expected number of days per year that the permanent plots flood (listed in Table 1) was calculated by applying the formula expressed in Figure 7 to the assumed elevation levels of the plots.

An absolute maximum water level of 283 feet was reached at the park during the 1979 flood (Carroon, 1982). The second highest water level of 277 feet was registered for the area in 1983 (Stone and Bingham, 1991). A probability curve of the percent of years in which water levels at the park will reach a specific elevation has been constructed (Figure 8). The analysis predicts river stages at JWW will reach the following levels in the listed ratio of years: 265 ft (4:5 years), 268 ft (1:5 years), and 270 ft (1:6 years).

The H80B gage has been functional for about 100 years, while the JWW gage was activated during the 1985 water year. Annual maximum gage heights for the H8OB location can be extrapolated to nearby river stretches if the slope of the river is known. Extrapolation of these values is useful because it increases the number of years in which the annual maximum gage heights are available for the JWW Station. River slope between the JWW and H80B location was determined by comparing gage heights for a 4 year period from October, 1994, to October, 1998, when both gages were functional (Figure 9). As the figure indicates, the Lowhead Dam has a major effect on river slope readings during low flow periods. However, once gage heights reach about 256 feet elevation, the weir effect of the dam is undetectable. Slope values range widely from 0.6 to 3.3 feet per mile when the water level at the JWW is below 256 feet. Above that point calculated river slope ratios generally ranged from 0.4 to 0.8 feet per mile and averaged 2.1 9 feet for the 3.6 mile river segment between gages. The 2.19 correction factor was applied to the annual maximum gage heights of the H80B gage to project JWW gage heights for a longer period (Figure 10). Maximum gage heights were averaged over a 20 year period to determine whether any trend was evident. The value represents a running average maximum height for a twenty year interval (ten preceding years plus ten succeeding years) for each year a value is listed in the graph. The averages showed a gradual increase from 264 foot level in 1911 to 268 foot level in 1987. This increase is partially due to the low annual maximum river stages in early years and the extremely high ones in 1979 and 1983. To indicate the accuracy of the readings, extrapolated readings were compared with actual flood levels for the flood years of 1979 and 1983 (Carroon 1982; Stone and Bingham 1991) and JWW gage readings during the range of years from 1985 to 1994, when both gages were active. With an average difference of 0.3 feet and a range of from 0.1 to 0.8 feet, the extrapolated and actual annual maximum readings showed a strong similarity (Figure 10). The extreme water years of 1979 and 1983 were not part of the accuracy comparison.

Flora and Vegetation Composition

A preliminary flora of LBSP was assembled by documenting known plants and collecting and identifying unknown ones. The nomenclature for species listed in this report follows Kartesz (1994). Plant collection records provided species presence and abundance data for the ecological communities that were not sampled. Special plants and animals known to range within the bounds of LeFleur's Bluff State Park were tabulated from lists obtained from the Mississippi Natural Heritage Program (2000). The species are listed as "special" when it is deemed that they are of such rarity or in such a degree of imperilment that they face serious risks of population decline or extirpation. The lists are periodically generated from data accumulated on each species as filed within a tracking system applied across the North American continent, the Biological Conservation Database System (BCD) (ABI, 2000). The Biological Conservation Database is the most complete source of information about Mississippi's rare, threatened, endangered or otherwise significant animals, plants, ecological communities, and natural features. The BCD is a product of ongoing research that has been conducted by many individuals and organizations over the past 25 years. Special plants and animals are discussed along with their local habitat in the following section in which the ecological communities are characterized.

Reconnaissance sampling and the preliminary classification of vegetation types during the fall of 1995 provided insight towards the best placement of permanent plots. During the reconnaissance phase of the project, vegetation cover per layer was estimated and general landscape features were documented at 20 locations using the Braun-Blanquet method, a rapid ecological sampling technique (Mueller-Dombois and Ellenberg, 1974). The next spring eighteen permanent macroplots, 4306 ft[2] (0.1 ac, 400 in[2]) in size, were measured and staked out at locations representing major landforms and vegetation types.

The forest cover was classified into four layers: canopy (plants with diameter at breast height (dbh) greater than 3.9 in [9.99 cm]); subcanopy (plants ranging from 1.2 in [3 cm] to 3.9 in); shrub and saplings (plants with stems less than 1.2 in dbh), and ground cover (herbs and woody plants less than three feet tall). Vegetation composition was determined by layer using measurements suitable for quantifying representative parameters of each layer. Plants in each of the four layers were documented as follows. Canopy and subcanopy plants were identified, mapped, and their dbh measurements were taken. Canopy plants were also numbered and marked with an aluminum tag. For the shrub and sapling layer, all stems were counted by sub-quadrat (quarter of plot), providing four samples per plot. Cover for the herbaceous layer was estimated to the nearest percent interval in 20, 3.28 feet (one meter) diameter rings per macroplot. The cover intervals were as follows: 1 (0-0.9), 2 (1-4.9), 3 (5-9.9), 4 (10-24.9), 5 (25-49 .9), 6 (50-74.9), 7 (75-94.9), and 8 (95-100). Plots were placed along two transects positioned lengthwise, just inside the line marking the perimeter of each macroplot. Basal area, density, and importance percentage were tabulated and summarized by canopy and subcanopy layer for each ecological community in which at least one plot was located. Importance percentage is the average of each species' relative density and relative cover.

When grouped into one matrix, the size of the species data set used in the ordination analysis was 224 species by 18 plots. The number of species encountered per layer per 0.1 ac for the canopy, subcanopy, shrub and sapling, and ground cover layers, averaged 6, 6, 11, and 39 species, respectively. The number of native woody species averaged 17 species per 0.1 ac. Total number of species encountered in upland plots averaged 64 species per 0.1 ac, but the average number of species documented on floodplain was considerably lower, only 44 species per 0.1 ac.

Based on herbarium collection records and sightings of known plants, at least 460 species, of which 69 are exotic, are known to exist in the park. For comparison, Camp McCain, an area 18 mi [2] in size, located near Grenada, Mississippi, harbors over 900 species of plants (MacDonald 1996). Typically countywide floral surveys, ranging in size from 208 to 769 mi [2], have tallies that range from 656 to 1281 species (MacDonald, 1996; Alford 1999). These findings demonstrate that a significant portion of Hinds County's floral diversity is found in LBSP. Of the park's total listing, 112 species are native woody species and 12 are exotics. By comparison, Smith (1985) identified 204 native woody species for an area of 2625 mi [2] of the upper Pearl River watershed, which includes parts or all of the counties of Choctaw, Attala, Winston, Noxubee, Leake, Neshoba, and Kemper. McDaniel (1982) identified 296 native woody species for the entire Pearl River watershed (8500 [mi.sup.2]). LBSP (less the golf course), at a mere 0.62 [mi.sup.2], contains 55% and 38% of the woody species tallied by these authors, respectively.

Ground cover averaged 79% on the uplands and 58% on the floodplains. Toxicodendron radicans (eastern poison ivy) contributed 59% of the total ground cover on the uplands, 49% on the upper terraces, but only 4% on the wetter terraces of the floodplain (Table 4). The average stem counts of tree saplings and shrubs (less than 1.2 in dbh) varied widely among ecological communities, averaging 2570 stems [ac.sup.-1] on the uplands, 1580 stems [ac.sup.-1] on wet bottom and upper terrace types, and 530 stems [ac.sup.-1] on low terrace, riverfront and swamp types (Table 4).

Using the classification scheme exhibited in Table 1, canopy and subcanopy species' basal area, density, and importance percentage were organized by ecological community/habitat type (Table 5). Trees were separated into two layers, canopy and subcanopy, to help evaluate which species are reproducing in the stand (Table 5). For the Mesic Calcareous Bluff Forest, Liriodendron tulipifera (tuliptree) is most important in the canopy but is not present in the subcanopy, indicating the species is not regenerating within the stand. Table 5 provides a complete profile of canopy and subcanopy species, their dominance, and their composition. The species mixes differ enough among the gradient of habitats that they have come to be recognized as separate ecological communities. Tree density of subcanopy and canopy layers for bottomland plots average 2160 and 1010 trees [ac.sup.-1]; for upland plots, 4260 and 1360 trees [ac.sup.-1]; and for all plots, 1440 and 1070 trees [ac.sup.-1], respectively. The higher subcanopy tree density in the uplands was partly due to the edge effect on the submesic upland plot, which was positioned adjacent to the golf course and hiking trail. Basal area ([ft.sup.2] [ac.sup.-1]) for the following group of plots are: bottomland, 181, upland, 203, and total, 184, of which 96% is attributable to the canopy layer. The swamp type exhibited an extremely large basal area of 619 [ft.sup.2] [ac.sup.-1], while the lower terrace community had the lowest basal area of any community sampled (112 [ft.sup.2] [ac.sup.-1]). Stress due to spring flooding may be a reason for the lower basal area of this ecological community.

Ordination Analysis of Vegetation and Environmental Parameters

In the ordination programs, for each sample point, a single value is needed to represent the abundance of each species. When a species is present in several of the layers, additional conversion factors are needed to generate a single value for the species. For this paper, importance value is the term applied to this representative species value. The measurements obtained for each layer were converged into a single importance value by relativizing the data sets for each layer. First step in the process was to determine species' importance values for each layer. For species of the subcanopy and canopy layers, species importance values were equal to importance percentage as defined previously. For species of the shrub and sapling layer, density of stems per acre was relativized per plot. For plants of the herbaceous layer, relative percent ground cover served as the importance value.

A coefficient of solar exposure was applied to weight the importance values according to the solar exposure each layer is expected to receive during the growing season. Importance values were weighted for the respective layers as follows: canopy (1.00), subcanopy (0.75), shrub and sapling (0.35), and ground cover (0.25). After weighting the species values inthismanner, the importance values of all layers were summarized and relativized to a single representative value per species per plot. The conversion factors employed in the process caused some values to be reduced to a magnitude in the one-one hundredths range. The ordination analysis programs tended to misinterpret values of this magnitude. To remedy this problem, a square root transformation was applied to the species data prior to executing the analysis.

Three ordinations of plot data were conducted using the following analysis packages: Two-Way INdicator Species ANalysis (TWINSPAN) (Hill, 1979), Detrended Correspondence Analysis (DCA) (Hill and Gauch, 1980), and CANOnical Community Ordination (CANOCO) (Ter Braak, 1987, 1998) (see Appendix). As an initial effort to classify the stands, TWINSPAN was used to order the plots and vegetation into groups of similarity. Secondly, DCA was applied and the results were graphically plotted to exhibit the variation of species composition among plot samples. The resulting pattern identified the environmental variables most influential in the genesis of gradients of the eighteen sample scores, i.e., those that best explain the variance of the species data. The magnitude of eigenvalues, ranging from zero to one, indicate how well species composition is correlated to the plot axes.

The third ordination involved direct gradient analysis using Canonical Correspondence Analysis of the CANOCO package. The evaluation directly ordinates species data to environmental variables. Initial tests isolated the most important environmental variables. Later only selected variables were included in the analysis. Graphical representations of the data are possible using the CanoDraw and CanoPost Programs (Ter Braak, 1998).

A TWINSPAN ordination analysis of the reconnaissance samples taken during the fall, 1995, helped to complete a preliminary vegetation classification of the park. Permanent plots were subsequently established and sampled during the spring and summer of 1996. The TWINSPAN ordination program was again used to verify the classification of the vegetation. The results of the ordination validated observations in the field. Additional tests proved equally informative when only subcanopy and canopy data were considered. TWINSPAN isolated four groupings that corresponded quite well to the following classes: wet bottom (4), low terrace (6), mid and high terrace (5), and uplands (3), for a total of eighteen plots. After further scrutiny of the TWINSPAN and DCA results, it was determined that three additional splits were necessary to separate communities that showed obvious differences not explained by the TWINSPAN ordination: a swamps type from the wet bottoms; a riverfront type from low terraces, and a submesic uplands type from mesic uplands. Some floodplain gradients were imperceptible, making the determination of the habitat type boundaries difficult to isolate. Several landform categorizations were marginal and were treated as such in the direct gradient analysis by listing them under both groupings.

Additional DCA and CCA evaluations confirmed the TWINSPAN analysis and the validity of the subjective readjustments in the classification of the plots (Figure 11 and 12, respectively). In the DCA analysis, only the first ordination axis, with eigenvalue of 0.76, proved useful for interpretation. The second, third, and forth ordination axes had eigenvalues of 0.27, 0.14, 0.05, all suggesting minor importance. Degree of wetness, which is partially correlated to elevation, fit the gradient represented by the first ordination axis. The diagrammatic representation of sample points, listed in a series from left to right along the axis, places landforms along a wetness gradient from swamp land to bluff ridge top (Figure 11).

The CCA analysis directly ordinates sample scores to the environmental variables collected at each site (Figure 12). The plot clustering and separation among groups was enhanced by direct gradient analysis because the measured environmental variables were highly correlated to species composition of the plot samples. In testing the significance of the first canonical axis of the CCA analysis, which had a relatively high eigenvalue, 0.74, the Monte Carlo permutation test results produced an F-ratio of 2.333 and p-value of 0.005. The most highly correlated environmental variables were (significance level in parenthesis): 1) average depth of water table (p .005), 2) days of flooding (p =.005), 3) soil pH (p .005), and 4) base saturation (p .05). Other arrows on the figure represent nominal environmental variables, which define the relative position of the sample groups by graphically displaying the approximate mean values of their species scores (Figure 12).

Ecological Communities of LeFleur's Bluff State Park

Recently a framework has been established for a standard national vegetation classification system in the United States (FGDC, 1997). The Nature Conservancy's National Vegetation Classification (Grossman et al., 1998) was adopted as the standard classification of ecological communities for the United States; as part of this effort, Weakley et al. (1998) developed the classification for the Southeastern Region of the United States. State Heritage Programs, including the Mississippi Natural Heritage Program, assisted in developing the Southeastern Regional classification. The Mississippi Natural Heritage Program maintains a listing of ecological communities found in Mississippi. The classification strategy employed for this study is compatible with these regionally and nationally recognized systems of classification.

The habitat types and ecological communities are mapped and listed along with their conservation ranking and acreage measurements (Figure 1 3)(Table 6). The listing of ecological communities follows the nomenclature established for the State List of Ecological Communities (Mississippi Natural Heritage Program 2000). Of the 24 general statewide cover types representing over 100 ecological communities in the state of Mississippi, 13 are present in the park. The statewide classification generally organizes communities along a moisture gradient, the most important factor influencing vegetation differentiation in Mississippi. The list progresses from driest to wettest types.

The ecological communities map was hand drawn on an enlarged digitally generated aerial photograph. The investigator traversed the park observing landscape characteristics, vegetation composition, and forest structure. During subsequent visits new findings helped to refine the vegetation mapping units.

Conservation priority ranks established for state listed communities have been applied to the vegetation units of the park. Conservation ranks provide basic information on the relative imperilment or risk of deterioration or extinction of an element within specified geographic ranges, statewide, nation wide, and global (Grossman et al., 1998). First, communities are categorized into natural types, those with little or no modification by human activity, and seminatural ones, those extensively modified by humans. Natural communities are ranked according to the degree of imperilment, of which there are five categories ranging from critically imperiled (S1) to secure (S5). Semi-natural or altered vegetation, which receive the state ranks of SM (managed) and SW (invasive or weedy), indicate a lower conservation priority for these types.

Four percent of the natural area contains imperiled or critically imperiled communities (S1 or S2), due to their rarity or factors that increase their vulnerability to statewide degradation. The types listed as imperiled are the Calcareous Bluff forest types, Freshwater Mussel Bed, and Sand Bar. Sand Bars are highly ranked because they serve as important nesting sites for several endangered herptiles. Another 37% of the area includes communities that are vulnerable to degradation (S3). Other communities (25%) are considered secure because they are more common or widespread (S4). The remaining areas (33%) have a low conservation priority because they contain disturbed (SW) or managed (SM) vegetation of a degraded condition, such as vegetation found in Special Use Areas.

Ecological community names, distinguished by the suffix "EC," are constructed with the use of diagnostic species' names in most instances. In some instances, habitat type features are written into the name to help differentiate similar types from each other. The following ecological community descriptions are listed in order of general degree of wetness. Two of the mapped ecological communities were of minor extent and need not be included in the discussion below: 1) Young Hardwood Forest of uplands and 2) Mesic Grassland (Mowed).

Uplands

Although of relatively minor size (16.8 ac), upland forests represent a significant assemblage of plants, many of which associate with calcareous soils. Ordination results revealed that these areas are the most distinctive among all stands sampled. The array of plants specifically adapted to the calcareous soils of the escarpment were by enlarge absent from other sample points. The soils maintain a high pH due to the gradual movement of colluvial material downslope along the escarpment, action which continually exposes additional calcareous materials on site. Two forest communities occupy the park uplands, mesic and submesic upland forests. A diagrammatic cross-section exhibits landscape features of the bluff and floodplain catena and demonstrates the relationship of hydrology, geology, and soils to several ecological communities (Poore, 1995) (Figure 14).

Submesic Upland Forests

Submesic Mixed Hardwood Calcareous Forest. The Quercus (shumardii, stellata)--Fraxinus americana-Carya (glabra, alba) EC occupies mid and upper level slopes (shoulder slopes) of the upland bluff area of the park. The soils are shallow, some-what mottled, submesic, and have been significantly eroded. The C horizon is reached at a shallow, 15 inch depth. A deep profile is prevented from forming because of active water erosion and gravitational sloughing of colluvial materials down slope. Additional trampling by visitors venturing off marked trails reduces vegetation cover and increase erosion problems along the bluff.

The soils are classified as an eroded phase of the Siwell soils. The Siwell soils consists of moderately well drained soils that formed in a silty mantle over underlying calcareous clay strata of the Yazoo Formation, permeability is very low. With the clay strata very near the soil surface, soil pH is circumneutral. The colors and ped faces of these soils exhibit distinct characteristics of vertic soil features. The mottling indicates periodic wetness, which is typical of soils with a high content of shrink-swell clays. The slickensides along the ped faces show shiny polished surfaces caused by the rubbing that occurs during wetting and drying sequences. As desiccation progresses, the soils will form deep, wide cracks. Upon rehydration the cracks swell shut, causing reduced infiltration and increased surface runoff (Pettry and Switzer, 1993). The soil's cation exchange capacity and the site index is moderate. The submesic nature of this landform causes a reduction in abundance of arboreal associates of wetter types.

The absence of Liriodendron tulipifera (tuliptree), Aesculus pavia (red buckeye), and Tilia americana (American basswood) and presence of Quercus stellata (post oak), Vaccinium arboreum (farkle-berry), Quercus falcata (southern red oak) and Carya alba (mockernut hickory) help to distinguish this submesic type from the mesic upland forest type described below. Several species are common across both upland forest types: Fraxinus americana (white ash), Crataegus calpodendron (pear hawthorn), Quercus michauxii (swamp chestnut oak), Quercus shumardii (Shumard's oak), Viburnum rufidulum (rusty blackhaw), and Cercis canadensis (eastern redbud).

New Museum Woodland. This mapping unit represents the woodland that is situated above the bluff escarpment edge, at the 300 foot contour interval. The 3.6 acre woodland is located in the backyard of the MMNS. Aerial photos from 1936 indicate this area was not forested at that time. It probably remained cleared as pasture or cropland, since a dairy farm occupied the park land west of the floodplain until the city purchased the property in 1944. On 1955 photos, small trees were evident on 20% of the area, specifically along the drainage that passes through the center of the woodland. By 1986, trees covered the whole area. Although some ornamental trees and shrubs were planted in the area, most of the plants are offspring of trees of adjacent forest lands. Several exotics including Elaeagnus umbellata (autumn elaeagnus), Ligustrum sinense (Chinese privet), Lonicera japonica (Japanese honey-suckle), and Nandina domestica (nandina) were planted or have invaded the area. Restoration efforts currently under-way are beginning the process of removing exotics and planting a mixture of native plants adapted to drier conditions. Additional efforts to control erosion along the drainage have been implemented. With the history of the forest known, there are opportunities available for interpreting forest succession.

Mesic Upland Forests

Mesic Calcareous Bluff Forest. The Quercus (michauxii, shumardii)--Liriodendron tulipifera EC flourishes where underlying calcareous parent materials have weathered to circumneutral or basic soils on rich mesic slopes and bottoms. In this state, the habitat type is generally confined to regions where calcareous soils occur, including the Jackson Prairie, the Blackbelt Prairie and the Pontotoc Ridge Physiographic Regions of Mississippi. A similar mesic calcareous bluff community in the Blackbelt Prairie Region was described for an area near Starkville (Morris et al., 1993). The lower bluff slopes are enriched by the movement of soils from upper slopes. Slopes are similar in steepness, ranging from 35% to 45%. Soils on the lower slope have not been classified but are likely similar to Brooksville soils. The loss of soils from the upper slopes has lead to an enrichment of soils on the lower slopes, which in turn has augmented the diversity and productivity of the associated species.

The lower bluff slopes possess deeper organic soils and have an eastern aspect, thus increasing moisture and nutrient availability. Within the park, Liriodendron tulipifera, Tilia americana, and Juglans nigra (black walnut) are essentially confined to these mesic sites. Several oaks, especially Quercus shumardii (Shumard's oak) and Quercus michauxii (swamp chestnut oak), are common. Smaller subcanopy trees are densely spaced along the slopes: Ostrya virginiana (eastern hophornbeam), Ulmus rubra (slippery elm), and Viburnum rufidulum (rusty blackhaw). Among the other less important species are Fraxinus americana (white ash), Celtis laevigata (sugarberry), and Morus rubra (red mulberry).

Both submesic and mesic vegetation types of the bluff lands contain an astonishing diversity of under-story plants, making this community especially worthy of conservation. The riot of species occupying this site include many species more commonly found in the rich woodlands of the loess bluff region to the west. Their diminution is of concern because of widespread forest disturbances and loss of these specialized rich habitats. Other species found in the calcareous uplands require the specialized, fertile, calcareous habitats. Such species are considered calciphiles. Echinacea purpurea (eastern purple coneflower), a rare calciphile found in LBSP, is listed as a species of special concern in Mississippi.

The coneflower is one of an estimated 45 species in LBSP that prefer circumneutral to basic soils of rich mesic woodlands or prairies. Of this group, several are considered true calciphiles: Aster ontarionis (Ontario aster), Cornus drummondii (roughleaf dogwood), Crataegus calpodendron (pear hawthorn), Dasistoma macrophylla (mullein foxglove), Desmanthus illinoensis (prairie bundleflower), Lithospermum tuberosum (tuberous gromwell), Quercus shumardii (Shumard's oak), Ruellia strepens (limestone wild petunia), and Smallanthus uvedalia (hairy leafcup). Additional species are associated with fertile, mesic woodlands where soils are circumneutral or slightly acidic: Aesculus pavia (red buckeye), Agrimonia pubescens (soft agrimony), Arisaema dracontium (greendragon), A. triphyllum (Jack-in-the-pulpit), Aster cordifolius (common blue wood aster), Brachyelytrum erectum (bearded shorthusk), Bromus pubescens (hairy woodland brome), Cryptotaenia canadensis (Canadian honewort), Erythrina herbacea (redcardinal), Fraxinu s americana (white ash), Gaura angustifolia (southern beeblossom), Geranium maculatum (spotted geranium), Geum canadense (white avens), Lindera benzoin (northern spicebush), Lythrum alatum (winged lythrum), Maianthemum racemosum (feathery false Solomon's seal), Monarda fistulosa (wild bergamot beebalm), Penstemon laevigatus (eastern smooth beardtongue), Phlox divaricata (wild blue phlox), Phryma leptostachya (American lopseed), Podophyllum peltatum (mayapple), Polygonatum biflorum (King Solomon's seal), Prunus mexicana (Mexican plum), Silene virginica (fire pink), Smilax lasioneura (hairy nerved greenbrier), Spigelia marilandica (woodland pinkroot), Thaspium trifoliatum (purple meadowparsnip), Tilia americana (American basswood), Ulmus rubra (slippery elm,) Uvularia perfoliata (perfoliate bellwort), Verbesina virgin ica (white crownbeard), and Viburnum rufidulum (rusty blackhaw), among others.

Bottomlands (Floodplain)

Mesic Palustrine Forests

Oak--Mixed Hardwood Ridge Bottom Forest. Quercus (nigra, pagoda, alba)--Carya cordiformis--Asimina triloba EC occupies natural mid and high terraces of the Pearl River floodplain. Infrequent flooding of short duration (five to ten days [yr.sup.-1]), a relatively deep springtime water table (three to six feet), and loamy soils distinguish this habitat type from other floodplain types. The mid-level terraces are small levees created by ancient meander scrolls of the river. The high terraces may have formed from the accumulation of coarse sediments in zones of accretion at river bends. The landforms on which this community is most commonly found include toe slopes, mesic lowlands and second terraces. The habitat type occurs more extensively along larger rivers of the southeast.

Cascilla soils are most commonly associated with this habitat type. The soils, which formed in silty alluvium, are deep, well drained, moderately permeable, and loamy with above average fertility (NRCS, 2000). Cascilla soils have a thick, dark, silt-loam surface horizon, a subsurface gray mottled zone, and a firm, weakly cemented subhorizon with relatively strong structural units. The thick A horizon with dark colors indicates soils containing moderate amounts of organic matter. The gray mottled zone between 28 and 43 inches indicates that the water table fluctuates at that level for part of the year.

Bigbee soils occur on upper terraces and consist of deep, excessively drained, highly permeable, coarse-textured soils. These soils formed in thick sandy sediments on low terraces along stream flood plains (NRCS, 2000). Bigbee soils are composed of very friable, structureless, yellowish brown soils with textures ranging from loamy sand and sandy loam to silt loam. The soils examined on the upper terrace showed little horizonal development, typical for coarse sedimentary soils. The coarseness also provides little restriction for root development. The soils are more droughty than other terrace soils. Ground water depth of the upper terrace was not reached by aten foot deep water well. Evidence of this condition is borne out in the ordination diagrams which show Plot 426 partially separated from the others in the group (Figures 11 and 12).

Diversity of arborescent lifeforms is higher on the middle and high terraces of the floodplain than on all other communities of the floodplain. Species of trees encountered in plots within this habitat totaled nineteen, about equal to the total number encountered in all other alluvial forested habitats combined. The habitat represents a transition zone from uplands to alluvial wetlands and support a mixing of species from both the wetter floodplain and drier upland habitats. Furthermore, this habitat type includes species less tolerant of flooding.

Herbaceous species associated with this habitat type are Arisaema dracontium (greendragon), A. triphyllum (Jack-in-the-pulpit), Botrychium biternatum (sparselobe grapefern), and Elephantopus carolinianus (Carolina elephantsfoot). Common small-sized shrubs and vines include: Bignonia capreolata (crossvine), Dioclea multiflora (Boykin's clusterpea), Rubus cuneifolius (sand blackberry), Sambucus canadensis (American elder), Sebastiania fruticosa (gulf sebastiana) (a good indicator of this type), Smilax glauca (cat greenbrier), S. smallii (lanceleaf greenbrier), S. rotundifolia (roundleaf greenbrier), Toxicodendron radicans (eastern poison ivy), and Vitis rotundifolia (muscadine).

Vine Wall Forest. The Aristolochia tomentosa--Smilax sp.--Vitis sp. vine wall forest EC is found where trees cannot exploit the full exposure of sunlight in areas such as along the river channel. This minor community is mapped at one site located on the southeastern side of the park. The community forms a blanket of long trailing vines, leaves, and twisting stems that shroud the trees inhabiting the area. Among many woody vines, Aristolochia tomentosa (woolly dutchman's pipe) grows luxuriantly. Pipevine swallowtail caterpillars (Battus philenor), can be seen on the vines during the summer. On a sunny summer day, the large and colorful swallowtails are often flitting about in abundance. Other vines encountered are Ampelopsis arborea (peppervine), Berchemia scandens (Alabama supplejack), Campsis radicans (trumpet creeper), Cocculus carolinus (Carolina coralbead), Lonicera japonica (Japanese honeysuckle), Parthenocissus quinquefolia (Virginia creeper), Smilax bona-nox (saw greenbrier), Smilax glauca (cat greenbr ier), Smilax rotundifolia (roundleaf greenbrier), Vitis rotundifolia (muscadine), and the ubiquitous Toxicodendron radicans (eastern poison ivy). The area covers about three acres, but other small patches are occasionally found elsewhere along the river or on the edges of forested areas. Trees occupying the site are represented in the Oak--Mixed Hardwood Ridge Bottom Forest.

Wet Riverfront Forests/Herbland

River Birch--Sycamore Riverfront Forest. Dynamic hydrologic forces of southeastern rivers occasionally cause dramatic changes across their floodplains by cutting new channels, blocking old ones, accreting new lands, taking others, or, in a less invasive way, etching meander scrolls across the floodplain. The one mappable area of active accretion found in the park is located just north of the Lowhead Dam. The Betula nigra--Platanus occidentalis EC is found along the riverfront where sediment deposition is actively occurring. This ecological community represents vegetation in an early stage of forest succession. The riverfront landform has variable wetness, depending on the level of the ground water table and stage of the river. The areas are commonly flooded during spring high water events and support a shallow water table as determined by their proximity to the river channel.

Terraces are natural levee formations built up during flooding. They are formed by the deposition of silty and sandy particles that are dropped at the river channel rim. Alluvial materials settle on levees where a reduction in floodwater velocity occurs. When water is no longer confined within the channel, the increased roughness of the flow path and a reduction in water depth at the levee induce sediments to fall from the water column. Both of the opportunistic species, Betula nigra (river birch) and Platanus occidentalis (American sycamore), are quickly able to recolonize these exposed soils. Additional important trees preferring the moist alluvial sites are: Acer saccharinum (silver maple), Ulmus americana (American elm), and Salix nigra (black willow). Seedlings of Fraxinus pennsylvanica (green ash) are especially abundant. Chasmanthium latifolium (Indian woodoats) is frequently found growing on sandy levees that skirt the river channel.

In later years as the river migrates away from such riverfront habitats, other tree species become established in the understory of these stands and eventually grow to dominate the habitat once alluvial soils are stabilized. An example of an ancient riverfront terrace is the high terrace east of the oxbow lake that lies just below the bluff line (Plot 426, Figure 13). A soil profile completed for this site indicated that substrates are composed of deep sands and loamy sands that were built up during riverfront accretion several hundreds to thousands of years ago. The vegetation found on the old terrace no longer harbors riverfront trees, such as Betula nigra (river birch) and Platanus occidentalis (American sycamore), that probably first colonized the area.

Silver Maple--Mixed Floodplain Forest. Acer saccharinum--Acer negundo mixed hardwoods EC is also uncommon in the park. The ecological community is representative of early successional forests on alluvial wetlands, along oxbows, and along Eubanks Creek at the walking bridge. Only a small portion of the park is occupied by this vegetation type. One stand that is large enough to be mapped occurs south of the park campground. Possibly, clearing of vegetation at the site stimulated the regeneration of mostly Acer saccharinum (silver maple).

Young Bottomland Forest. Clearing efforts related to recent maintenance activities near the west end of the driving range has created the only stand of young bottomland forest of sufficient size for mapping. Additional smaller gaps, openings created by windthrow, natural senescence, diseases, etc., regularly occur within larger bottomland forest stands. An average of 20 snags (standing dead trees) per acre are present in the park woodlands. Openings created by the dead trees allow sunlight to reach the forest floor and enable young trees to regenerate.

Bottomland Herbland and Bottomland Lawn or Mowed Field. Bottomland grassy areas managed by mowing and herbaceous patches left unmowed are considered as a separate cover type. There is only one significant area of each of these types located south of the golf driving range. The area is an old buried landfill site that has been capped with topsoil. Festuca pratensis (meadow fescue) and Lespedeza cuneata (Chinese lespedeza), common on the unmowed portion, were likely planted as cover for stabilizing the landfill soil cap. Trees unable to root deep enough in the shallow soil cap are inhibited from becoming established. The ephemeral annual herbs, many of which are introduced weeds are abundant on these areas during the spring.

A diverse assortment of ephemeral annuals are found here, making this easily accessible area a good site for low impact botanizing during the spring. Herbs include Cerastium dubium (chickweed), C. arvense (chickweed), Galium aparine (stickywilly), Juncus tenuis (Poverty rush), Krigia dandelion (potato dwarf dandelion), K. cespitosa (weedy dwarf dandelion), Nuttallanthus canadensis (Canada toadflax), Plantago virginica (Virginia plantain), Ranunculus sardous (hairy buttercup), and Triodanis perfoliata (clasping Venus' lookingglass). Grasses include Aristida oligantha (prairie threeawn), Brachiaria platyphylla (broadleaf signalgrass), Cynodon dactylon (bermudagrass), Paspalum dilatatum (Dallasgrass), Setaria parviflora (yellow bristlegrass), Sorghum halepense (Johnsongrass), and Tridens strictus (longspike tridens).

In late summer, the unmowed portion is dominated by Ambrosia trifida (great ragweed), Solidago canadensis (Canada goldenrod), and some areas are covered with a blanket of vine stringers of Passiflora incarnata (purple passionflower). Other plants occupying the site are Amaranthus palmeri (careless-weed), Aster pilosus (white oldfield aster), Croton capitatus (hogwort), Dichanthelium scoparium (velvet panicum), Euthamia tenuifolia (bushy goldentop), Pyrrhopappus carolinianus (Carolina desert-chicory), and Verbena brasiliensis (Brazilian vervain).

Wet Palustrine Forests

Sugarberry--American Elm--Green Ash Bottomland Forest. The Celtis laevigata--Ulmus amencana--Fraxinus pennsylvanica EC is found on low terraces where annual flooding is regular and of moderate duration (15 to 30 days [yr.sup.-1]). Slight elevations provide some relief from excessive wetness, but annual flooding is likely. Positioned within the 257--263 foot elevation range (Table 1), the low terraces experience increased flooding, about three times as long as the duration expected for middle and high terraces. Ponding occurs for a shorter duration than that found on wet bottoms. The water table depth during the spring season is shallow, from two to four feet, but drops rapidly as forest green-up occurs (Figure 2). The compact layers in the subsoil of some horizons restrict downward movement of water, resulting in soils that contain perched water tables or intermittent standing water. The Chenneby soils found on low terraces, wet bottoms, and backwater areas, are deep, somewhat poorly drained, moderately permeab le, grayish brown silt loam and silty clay loam soils (SCS, 1979; NRCS, 2000).

A distinctive suite of trees occupy this alluvial zone, most importantly, Celtis laevigata (sugarberry) and Fraxinus pennsylvanica (green ash), with additional common trees including: Acer negundo (boxelder), Acer saccharinum (silver maple), Liquidambar styraciflua (sweetgum), and Quercus lyrata (overcup oak). The tolerance of the most common species, Celtis laevigata, to long periods of inundation enables it to dominate this habitat type. Campsis radicans (trumpet creeper), Cocculus carolinus (Carolina coralbead), and hex decidua (possumhaw) are common understory constituents. Aster lateriflorus (calico aster), Rubus trivialis (southern dewberry), and Viola palmata (early blue violet) exhibit a high constancy on the low terraces. Additional species associated with this habitat type are Acalypha virginica (Virginia threeseed mercury), Carex icuisianica (Louisiana sedge), Commelina virginica (Virginia dayflower), and Spiranthes ovalis (lessor ladies'-tresses), a species of special concern.

Wet Hardwood Bottom Forest. Quercus lyrata-Carya aquatica EC occupies wet depressions, bottoms of chutes, slough edges, and backswamp areas. The wet bottom habitat type encompasses landforms consisting of low elevation, shallow depressions and flats or chutes having outlets to the main river channel. The wet bottom is considered a separate landform from the low terrace because of the increased frequency and duration of flooding and the tendency of water to pond for longer periods during the winter and spring. Either due to high river levels or ponding, this habitat type is often inundated for more than two months during the late winter and spring. Significant wetness is guaranteed by frequent springtime flooding with an annual duration of roughly 30-80 days. The water table is elevated during the spring, often remaining at or just below ground level into late spring, and averages about 0.2 feet below ground level. The areas dry out quicker than swamps because the depressions are shallow and better drained. T he Chenneby series is also characteristic of wet bottom habitats.

Fraxinus pennsylvanica (green ash) is dominant in this forest type as in the low terrace habitats described above, but Celtis laevigata (sugarberry) seldom occurs. Three indicator species for this community are Carya aquatica (water hickory), Planera aquatica (planertree), and Quercus lyrata (overcup oak), while Styrax americanus (American snowbell) and Cephalanthus occidentalis (common buttonbush) are constant shrubby associates. The exotic Sapium sebiferum (tallowtree) is most commonly found in marshes, swamps, and wet bottoms. Intermingled in the understory are the following frequently encountered vines and herbs: Acmella oppositifolia (oppositeleaf spotfiower), Boehmeria cylindrica (smalispike false nettle), Brunnichia ovata (American buckwheatvine), Justicia ovata (looseflower waterwillow), Leersia lenticularis (catchfly grass), Leersia oryzoides (rice cutgrass), Lycopus rubellus (taperleaf waterhorehound), Mikania scandens (climbing hempvine), Pilea pumila (Canadian clearweed), Pluchea camp horata (cam phor pluchea), Saururus cernuus (lizards tail), Triadenum walteri (greater marsh St. Johnswort), and Wisteria frutescens (American wisteria).

Shrub Wetlands, Pocosin, Herb Bogs

Buttonbush Shrub Wetland. Cephalanthus occidentalis-Itea virginica-Forestiera acuminata EC occupies intermittently exposed to semipermanently flooded lowland areas within and along the edges of creeks, rivers, sloughs, and oxbow lakes, often in zones surrounding deeper water. The lands remain flooded or generally saturated throughout the year Though the conditions of this wetland habitats resemble those of swamps, the sparsity of swamp trees distinguishes this type. Fluctuating water levels cause herb density to vary widely in this ecological community. During the spring and early summer the herbaceous vegetation is often sparse due to deep water conditions. However, the marshy edges of sloughs and lakes are often crowded with herbs and vines by late summer. Herbs of this community are similar to those found in marshes. Cephalanthus occidentalis (common buttonbush) is the indicator species for this habitat, while Itea virginica (Virginia sweetspire) and Forestiera acuminata (eastern swampprivet) are less com mon.

Inland Freshwater Marshes or Spring Marsh

Freshwater Marsh. Eleocharis sp.-Scirpus sp.-Ludwigia sp.-Penthorum sedoides-Cyperus BC covers a broad range of herbaceous vegetation types occupying semi-permanently flooded wetlands on lowlands adjacent to rivers, lakes, and ponds. Soils usually remain saturated for most of the growing season. By late summer or early fall, water level on the marsh are shallow or soils become exposed, but usually remain saturated. Flooding of shrub wetlands, wet bottoms, and marsh habitats is of similar frequency and duration. Marshes tend to be present where deep springtime flooding inhibits trees from establishing, or occur in situations where trees have been taken out naturally by winds or artificially by logging. Marsh soils resemble those of swamps that are described in the following section.

Herbs that dominate marshes found in the park include the following species: Alternanthera philoxeroides (alligatorweed), Hydrolea uniflora (oneflower false fiddleleaf), Polygonum hydropiperoides (swamp smartweed), and P. punctatum (dotted smartweed). The grasses, Leersia lenticularis (catchfly grass), Panicum rigidulum (redtop panicgrass), and sedges, Carex joorii (cypress swamp sedge), Cyperus odoratus (fragrant flatsedge), and Eleocharis obtusa (blunt spikesedge), are commonly encountered in marshes. Numerous other species, including Commelina diffusa (climbing dayflower), Heliotropium indicum (Indian heliotrope), Hibiscus moscheutos (crimsoneyed rosemallow), Ludwigia peploides (floating primrosewillow), Ludwigia palustris (marsh seedbox), Lycopus rubellus (taperleaf waterhorehound), Micranthemum umbrosum (shade mudflower), Penthorum sedoides (ditch stonecrop), Pluchea camphorata (Camphor pluchea), and Saururus cernuus (lizards tail) are common.

Swamp Forests

Bald Cypress--Water Tupelo Swamp. The oxbow lake immediately below the bluff and those that form the Mayes Lakes represent channels of the ancient river that were carved during periods of higher stream flow. Exactly when the main channel of the Pearl River passed below the base of the western bluff line of LeFleur's Bluff State Park is unknown, but it likely occurred during the Pleistocene era. The channel segments were subsequently plugged by levees and became isolated depressions. The sedimentation within these low wetlands allows a buildup of fine silty clay soils. In this area of Mississippi, soils of swamps are commonly classified as Rosebloom soils, a very strongly acid deep silty soil formed in poorly drained areas. A key distinguishing feature of these soils is the significant gleying at or near the soil surface, identified by color mottling in an assortment of pale, dark, grayish and yellowish brown colors.

Taxodium distichum--Nyssa aquatica BC occupies semi-permanently flooded or seasonally flooded wetlands on landforms including back bays, depressions, oxbow lakes, and bottomland flats. Flooding ranges from 70 days per year to a semi-permanent duration. Compared to the Wet Hardwood Bottom Habitat Type, soils remain inundated or saturated for longer periods (Figure 14). According to the ordination axis representing wetness in the DC analysis (Figure 11), of all species associations, this community exemplified the greatest degree of wetness. Taxodium distichum (baldcypress) and Nyssa aquatica (water tupelo) dominate the swamp habitat in LBSP, together accounting for 94% of total canopy importance. Swamps range from young to old stands and density ranges from low to high, forming open to closed canopies. Other less important but common woody species are Carya aquatica (water hickory), Planera aquatica (planertree), flex decidua (possumhaw) and Styrax americanus (American snowbell).

Wetland shrubs and herbs common in marshes and bottoms are also found in swamp habitats. Woodwardia areolata (netted chainfern) occurred in abundance, contributing 83% of the total relative cover. Other species with moderate to high sample frequency only provided minor ground cover: Bidens aristosa (bearded beggarticks), Boehmeria cylindrica (smallspike false nettle), Brunnichia ovata (American buckwheatvine), and Mikania scandens (climbing hempvine). Marsh plants such as Echinodorus cordifolius (creeping burrhead), Lobelia cardinalis (cardinalflower), and Lysimachia radicans (trailing yellow loosestrife), which tolerate shady conditions and frequent flooding, appear well suited for swamp habitats.

Lacustrine Communities

Constructed Pond. Mayes Lakes, originally oxbow lake swamps, were deepened by constructing levees across their outlet points by the mid-1950's. On a 1955 aerial photo, west Mayes Lakes appeared to be in the process of being cleared of trees. The lakes became fishing lakes and several cabins were built around them. Once the city began administering the site, the cabins were removed and the lakes became public recreation areas. The lakes provide habitat for fish, reptiles, beaver, aquatic birds, and other aquatic animals. The shallow swampy zones of the lakes support a variety of wetland herbs, including the showy Hibiscus moscheutos (crimsoneyed rosemallow) and provide ideal habitat for wood ducks.

Rivers and Riverine Communities

Sand Bar. Leptochloa mucronata--Ammannia coccinea-Strophostyles helvula EC occurs on fine and medium textured sandy channel-sides, shorelines, and bars adjacent to streams and rivers. The habitat type extends in a narrow linear band along the river wherever sand is exposed. Wide exposed platforms of unvegetated sands are formed on isolated points at bends on rivers. These expanses create an opportunity for recolonization. After spring flooding, the areas are barren or sparsely vegetated, but by mid-summer, many annuals, deeply rooted perennials, or vines have recolonized the areas. Their success is dependent on access to adequate moisture. The deep, coarse sandy zones may remain sparsely vegetated but a few drought tolerant, ephemeral plants are able to take hold. The moist sands found along the river channel segment within the park are attractive to over 50 species of herbs. Perennials will persist over years of moderate flooding and most will tolerate sand burial to some extent.

Five general zones of river channel sideslope and sand bars are recognized: forest edge, dry deep sands, mesic sands, river edge, and areas that remain barren. The forest edge is occupied by trees and shrubs that lean out from the forest canopy and may become overwhelmed by vines (see vine wall forest). The forest edge may be occupied by a strip to trees tolerant to flooding such as Salix nigra (black willow). The loose, coarse, desiccated sands support species such as Diodia teres (poorjoe), Mollugo verticillata (green carpetweed), Paspalum laeve (field paspalum), Fimbristylis vahIii (Vahl's fimbry), Oldenlandia boscii (Bosc's mille graines) and Xanthium strumarium (rough cockleburr). The riot of plants embracing the moist sandy areas include a mixture of herbs and grasses. Grasses that come to dominate the mesic sands are: Digitaria sanguinalis (hairy crabgrass), Eragrostis hypnoides (teal lovegrass), Leptochloa mucronata (mucronate sprangletop), Panicum rigidulum (redtop panicgrass), and P. virgatum (swit ch-grass). Other grasses constituting the herbland are Eragrostis glomerata (pond lovegrass), E. pectinata (tufted lovegrass), Leersia oryzoides (rice cutgrass), and Panicum dichotomiflorum (fall panicgrass). Amongst the rush of herbs are Ammannia coccinea (valley redstem), Chenopodium ambrosioides (Mexican tea), Cyperus erythrorhizos (redroot flatsedge), Eupatorium serotinum (lateflowering boneset), Gamochaeta purpurea (spoonleaf purple everlasting), Hypericum mutilum (dwarf St. Johnswort), Mimulus alatus (sharpwing monkeyflower), and Phyllanthus caroliniensis (Carolina leafflower). Several legumes are able to take over some areas of the sandbar. Strop ho styles helvula (trailing fuzzybean) sprawls for tens of yards, and Glottidium vesicarium (bagpod) rapidly grows to heights over six feet by late summer. Plants preferring very moist or wet habitats at the water's edge include Paspalum fluitans (horsetail paspalum), Sphenoclea zeylanica (sphenoclea), Ludwigia leptocarpa (anglestem primrosewillow), L. decurre ns (wingleaf primrosewillow), and L. alternifolia (seedbox). Graptemys oculitfera, a threatened ringed map turtle (U.S. Fish and Wildlife, Endangered Species Act), prefers the barren, unvegetated portion of sandbars for nesting sites.

Alluvial Creek/River. Channels, watercourses, and streams serve as conduits for run-off from wetlands and adjacent uplands. The watercourses have a variety of substrates and vary greatly in flow volume and turbidity. Although documentation of the riverine diversity is beyond the scope of this study, it is worthy of notice that several fishes of special concern occur during part of their life cycle or in migration on this stretch of the river: Polyodon spathula (paddlefish), Alosa alabamae (Alabama shad), and Acipenser oxyrinchus desotoi (gulf sturgeon), a threatened species according to the U.S. Fish and Wildlife Service.

Eubanks Creek. The lower reach of Eubanks Creek flows through LBSP. The Eubanks Creek watershed was heavily impacted by agriculture since settlement in the early 1800's until the area was urbanized in the late 1940's. The watershed continues to be impacted by the effects of urbanization. When rural lands were annexed to the city of Jackson, most of the watershed area was converted to residential and commercial plots and gridded with transportation and utility corridors. A large portion of the watershed has been paved and many side drainages, which receive runoff from the city street, have been channelized or cemented. Water quality within the creek has been degraded due to the high amounts of development within the watershed. The stream is deeply entrenched into the floodplain and rapid changes in flood levels of the Pearl River on unstable subsoils have caused significant bank sloughing. The creek provides habitat for several turtle species that tolerate the poor water quality of the stream.

Freshwater Mussel Bed. Lampsilis sp.--Quadrula sp.--Obovaria sp.--Pleurobema sp. BC contains numerous mussels that prefer firm and stable (armored), variably textured, fine mud or coarse sand/gravel substrates, occasionally below sand bars, mostly along flats/bottoms of river channels, in which mussels can embed without being dislodged by river currents. The only mussel bed known to occur within the LBSP section of the Pearl River is located just north of Lowhead Dam at the south end of the park, where firm silty/sandy beds provide suitable habitat for numerous mussel species. The bed has a diverse compliment of mussels totaling nearly 20 species. Some of the species found in the mussel bed are: Amblema plicata, Elliptio crassidens, Fusconaia cerina, F. ebena, Lampsilis straminea claibornensis, L. teres, Leptodea fragilis, Megalonaias nervosa, Obliquaria reflexa, Plectomerus dombeyanus, Potamilis purpuratus, Quadrula refulgens, and Utterbackia imbecillis. In addition, several rare mussel species are found: Lasmigona complanata (white heelsplitter) and Pleurobema beadleianum (Mississippi pigtoe) (Tom Mann pers. com.).

Urban/industrial/Residential Habitat Conversion

Mapping units coded t1, t2 through t11 are special use areas that were not a focus of this study. However, plants observed in the special use areas were added to the park's plant list. The special use areas include a day use area, a campground, a storage site, a building site, a golf course facility, parking lots, and access roads.

Conclusion

The Jackson metropolitan area has grown to encompass a large portion of three counties in central Mississippi. In a large swath of urbanization from the Ross Barnett Reservoir to the hanging bridge at Byram, subdivisions, industrial parks, and commercial zones have enveloped the Pearl River. In the process, the ecological integrity of the river, floodplains, and adjacent uplands has been significantly compromised. With the annual threat of high water, the Pearl River has served as defacto protector of the remaining ribbon of floodplain forest, including most of LeFleur's Bluff State Park. This study identified the patterns and processes of the riverine system and the diverse assemblage of interrelating lifeforms that comprise its ecological communities: Principal physical variables are significantly related to the species associations of the park's upland and floodplain forests and wetlands. Over 15% of the plants of Mississippi are found in the park. The park contains nearly 40% of all native woody species that occur within the Pearl River watershed, which covers 14% of the state of Mississippi. Other ongoing research at the Mississippi Museum of Natural Science is resulting in better understanding of the herptiles, birds, fish, and other groups of animals that inhabit the park. The new Museum of Natural Science, functioning as a living museum, can serve as gateway to rediscovering the wealth of biodiversity embraced in the bluffs and floodplains of the Pearl River in LeFleur's Bluff State Park.

Acknowledgments

Appreciation is extended to participants in this study, especially the Urban Forest Council and Mississippi Forestry Commission for project funding (Rick Olson, coordinator); LeFleur's Bluff State Park staff for helpful assistance and recommendations; Robert Hinton, Soil Scientist (retired) for supervising soil profile descriptions; Mike Tagert, Nathan Tircuit, and Rudy Poglich for field assistance; Phil Tumipseed, U. S. Geological Survey, for providing gage height data for the Pearl River; Wally Belokon for GIS technical assistance in mapping of communities; Robert Poore for providing the vegetation diagram; Al Schotz, Alabama Natural Heritage Program for plant habitat information; the Millsaps Botany and Ecology Classes (Dr. Mann, instructor), and Belhaven Ecology Class (Dr. Chestnut, instructor) for their participation in field sampling; museum staff for editorial comments; and to the staff of the city of Jackson GIS lab for supplying the aerial photos used in mapping the site.

Literature Cited

Association for Biodiversity Information (ABI). 2000. Home page. http://www.abi.org/

Alford, Mac H. 1999. The vascular flora of Amite County Mississippi. Masters Thesis, Dept. of Botany, Duke University, Durham, NC. 176 pp.

Bailey, R.G. 1980. Description of the ecoregions of the United States Forest Service, USDA, Ogden, Utah. 77 pp.

Belokon, W 1996. LeFleur's Bluff State Park Geographic Information System maps and database. Mississippi Automated Resource Information System. Institute for Higher Learning, Jackson, MS.

Brinson, Carroll. 1980. Jackson, a special kind of place. City of Jackson, Mississippi. 381 pp.

Buckman, H.O., and N.C. Brady. 1969. The nature and properties of soils. The MacMillan Co, New York, 653 pp.

Carroon, Lamar E. 1982. Flood of April 1979 on Pearl River in Jackson, Mississippi and vicinity. Hydrologic Investigations Atlas HA-665, U.S. Geological Survey 1 pp. + maps.

Conner, W.H., J.G. Gosselink, and R.T. Parrondo. 1981. Comparison of the vegetation of three Louisiana swamp sites with different flooding regimes. Amer J. Bot. 68:320-331.

Daily, G.C. 1997. Nature's services: Societal dependence on natural ecosystems. Island Press, Washington, D.C. 392 pp.

Dockery, David. T. 1977. Mollusca of the Moodys Branch Formation, Mississippi. Miss. Geologic, Economic, and Topographical Survey Bull. 120, 212 pp.

FGDC [Federal Geographic Data Committee]. 1997. Vegetation classification standard, FGDC-STD-005. Web address: http://www.fgdc.gov/Standards/Documents/Standards/Vegetation. DC 1996.

Funderburk, D.O., and J.N. Skeen. 1976. Spring phenology in a mature piedmont forest. Castanea 41:20-30.

Furnas, D. 1992. Software catalog including the Cornell Ecology Programs and CANOCO. Microcomputer Power Ithaca, NY. 19 pp.

Grossman, D.H., D. Faber-Langendoen, A.S. Weakley, M. Anderson, P. Bourgeron, R. Crawford, K. Goodin, S. Landaal, K. Metzler, K.D. Patterson, M. Pyne, M. Reid, and L. Sneddon. 1998. International classification of ecological communities: Terrestrial vegetation of the United States. Vol I: The National Vegetation Classification System: Development, status, and applications. The Nature Conservancy, Arlington, Virginia.

Hill, M.O., and H.G. Gauch. 1980. Detrended correspondence analysis, an improved ordination technique. Vegetatio 42:47-58.

Hill, M.O. 1979. TWLNSPAN-A FORTRAN program for arranging multivariate data in an ordered two-way table by classification of individuals and attributes. Cornell University, Ithaca, N.Y 90 pp.

Hinton, Robert. 1996. Soil Scientist, retired, Jackson, MS.

Kartesz, John. 1994. A synonymized checklist of the vascular flora of the United States, Canada, and Greenland. Vol 1-Checklist. 2nd edition. Timber Press, Portland, OR.

Keys, Jr., J.E., C. A. Carpenter, S.L. Hooks, F.G. Koenig, W.H. McNab, W.E. Russel, and M. Smith (compilers). 1995. Ecological units of the eastern United States: First approximation. CD-Rom in ARCINFO, map and descriptions. U.S.D.A. Forest Service, Washington, DC.

Landers, M.L., and K.V Wilson, Jr. 1991. Flood characteristics of Mississippi streams. U.S. Geological Survey Water Resources Investigations Report 91-4037, 82 pp.

Larson, J.S., M.S. Bedinger, C.F. Bryan, S. Brown, R.T. Huffinan, E.L. Miller, D.G. Rhodes, and B.A. Touchet. 1981. Transition from wetlands to uplands in southeastern bottomland hardwood forests. Pages 225-273 in J.R. Clark and J. Benforado, eds. Wetlands of bottomland hardwood forests, Elsevier Science Publ. Co., Amsterdam.

Lowe, E.N. 1921. Plants of Mississippi; a list of flowering plants and ferns. Miss. State Geological Surv. Bull. No. 17. Bureau of Geology, Jackson. 292 pp.

MacDonald, John. 1996. A survey of the flora of Monroe County, Mississippi. Masters Thesis. Dept. of Biological Science, Mississippi State Univ., Starkville, MS. 163 pp.

Mann, T. 2000. Zoologist, Mississippi Natural Heritage Program, Museum of Natural Science.

McDaniel, S. 1982. Checklist of the native woody plants of the Pearl River Basin Area. Crosby Arboretum Foundation, Picayune, Mississippi. 17 pp.

McNab, W.H., and P.E. Avers (compilers). 1994. Ecological subregions of the United States: Section descriptions. Administrative Publication Wo-WSA-5. Washington, DC: U.S. Dept. of Agric., Forest Service. 267 p.

Mississippi Natural Heritage Program. 2000. Biological Conservation Database of ecological communities, special plants, and special animals, listed with conservation priority ranking. Mississippi Museum of Natural Science, Department of Wildl., Fisheries and Parks, Jackson. Mimeographed Lists.

Mississippi Cooperative Extension Service. 1996. Mechanical and chemical analysis reports. Plant and Soil Science Soil Testing Laboratory, Mississippi State Univ., MS.

Moore, W.H., A. Bicker, Jr., T. McCutcheon, and W Parks. 1965. Hinds County geology and mineral resources. Mississippi Geol., Econ., and Topog. Survey Bull 105, Jackson, MS. 244 pp.

Moore, W.H. (Ed.) 1969. Geologic map of Mississippi. Mississippi Geological Survey, Bureau of Geology, Jackson, MS. (1:500,000 Scale).

Morris, M.W., C.T. Bryson, and R.C. Warren. 1993. Rare vascular plants and associate plant communities from the Sand Creek Chalk Bluffs, Oktibbeha County, Mississippi. Castanea 58:250-259.

Mueller-Dombois, D., and H. Ellenbeig 1974. Aims and methods of vegetation ecology John Wiley & Sons, New York. 547 pp.

National Soil Survey Center (NRCS) 2000. Official Soil Series Descriptions: Bigbee, Cascilla, Cheuneby, and Rosebloom and Siwell soil series. Natural Resource Conservation Service, Lincoln, NE.

Pettry, D.E., and R.E. Switzer. 1993. Expansive soils of Mississippi. Miss. Agr. & For. Exp. Stn., Bull 986. Miss. State, MS 29 pp.

Poore, R. 1995. Diagrammatic cross-sections of ecological communities. In Barlow and Plunkett, Ltd., The Potomac Group, P.C., and Resource Management Associates.

Comprehensive development plan, Mississippi Museum of Natural Science, Jackson. Department of Wildlife, Fisheries and Parks, Jackson, MS.

Schiebout, J.A. 1986. Montgomery Landing and the Montgomery Landing Project (1978-1982). pp 5-34. In J.A. Schiebout and W. van den Bold (ed.) Montgomery Landing Site, Marine Eocene (Jackson) of Central Mississippi. Proc. of a Symposium 1986 Annual Meeting Gulf Coast Assoc. of Geol. Soc., Oct. 22, 238 pp.

Shankman, David. 1996. Stream channelization and changing vegetation patterns in the U.S. Coastal Plain. Geogr. Review 86:216-233.

Skeen, J.N. 1974. Composition and biomass of tree species and maturity estimates of a suburban forest in Georgia. Bull. of the Torrey Botanical Club. 101:160-165.

Skeen, J.N. 1976. Regeneration and survival of woody species in a naturally-created forest opening. Bull. of the Torrey Botanical Club. 103:259-265.

Smith, S.M., and G.S. Zumwalt. 1987. Gravity flow introduction of shallow water microfauna into deep water depositional environments. Mississippi Geology 8:1-8.

Smith, T.E. 1985. Distribution of woody plants in the upper Pearl River Basin of Mississippi. M.S. Thesis, Mississippi State Univ. 94 pp.

Soil Conservation Service. 1979. Soil Survey of Hinds County Mississippi. Soils Conservation Service, Jackson, Mississippi. 112 pp. + maps.

Soil Survey Staff. 1975. Soil taxonomy: a basic system of soil classification and for making and interpreting soil surveys. USDA Soil Conservation Service, Agric. Hdbk. No. 436, Robert E. Kierger Publishing Co., Melbourne, FL. 754 pp.

Stone, R.B., and R.H. Bingham. 1991. Floods of December 1982 to May 1983 in the Central and Southern Mississippi River and the Gulf of Mexico basins. U.S. Geological Survey water-supply paper no. 2362, U.S. Geological Survey, Denver, CO.

Taylor, J.R., M.A. Cardamone, and W.J. Mitsch. 1990. Ecological processes and cumulative impacts illustrated by bottomland hardwood wetland ecosystems. pp. 13-86 In J.G. Gosselink, L.C. Lee, and T.A. Muir (Ed.). Ecological processes and cumulative impacts illustrated by bottomland hardwood wetland ecosystems. Lewis Publishers, Chelsea, MI.

Ter Braak, C.J.F. 1987. CANOCO-A FORTRAN program for canonical community ordination by [partial][detrended][canonical] correspondence analysis, principal components analysis and redundancy analysis (version 2.1). DLO-Agricultural Mathematics Group, Wageningen.

Ter Braak, C.J.F. 1998. CANOCO reference manual and user's guide to CANOCO for Windows. Centre for Biometry, Wageningen. 351 pp.

Teskey, R.O., and T.M. Hinckley. 1977. Impact of water level changes on woody riparian and wetland communities, Vol. II: The Southern Forest Region. USDI, Fish and Wildlife Service, Biological Ser. Progr. Report No. FWS/OBS-77/59. 46 pp.

The Nature Conservancy. 1996. Conservation by design: A framework for mission success. The Nature Conservancy Arlington, VA.

U.S. Geological Survey. 2000a. Daily mean gage height values, daily mean values statistical program, duration curve statistical characteristics, and duration table of daily values by month for period 1985-1999 for the Pearl River at Jackson Water Works (Station No. 02485740) and at Highway 80 Bridge (Station No. 02486000). Miscellaneous data sets. U.S. Geological Survey, Water Resources Division, Jackson, MS.

U.S. Geological Survey. 2000b. Peak flow data for the Pearl River at Highway 80 Bridge, Jackson (Station No. 02486000) for the period 1874-1996. U.S. Geological Survey, Water Resources Division, Jackson, MS.

Weakley, A.S., K.D. Patterson, S. Landaal, M. Pyne, and others (compilers). 1999. International classification of ecological communities: Terrestrial vegetation of the Southeastern United States. Working Draft of March 1998. The Nature Conservancy, Southeast Regional Office, Southern Conservation Science Department. Community Ecology Group. Chapel Hill, North Carolina.

Wharton, C.H., W.M. Kitchens, E.C. Pendleton, and T.W. Sipe. 1982. The ecology of bottomland hardwood swamps of the Southeast: a community profile. US Fish and Wildlife Service, Biological Service Program. Washington, D.C. FWS/OBS-81/37 133 pp

Woodwell, G.M., and F.T. Mackenzie, eds. 1995. Biotic feedbacks in the global climatic system: Will the warming feed the warming? Oxford University Press, New York.

Appendix

TWINSPAN is:

"a FORTRAN program for two-way indicator species analysis in which species, and samples are analyzed. The result is similar to that obtained by Braun-Blanquet table arrangement (Furnas, 1992)."

DECORANA is:

"a FORTRAN program for detrended correspondence analysis. DECORANA results are superior to reciprocal averaging, multidimensional scaling, and principal components analysis, making it the best general purpose ordination, particularly for very heterogeneous data sets (Furnas, 1992)."

CANOCO is:

"a program for canonical community ordination by [partial] [detrended] [canonical] correspondence analysis. A common problem in community ecology is to discover how a multitude of species respond to external factors such as environmental variables. Data are collected on species composition and the external variables at a number of points in an area. Five years ago, regression and ordination were integrated into techniques of multivariate direct gradient analysis, called canonical (or constrained) ordination. The use of canonical ordination greatly improves the power to detect the specific effects one is interested in. CANOCO escapes the assumption of linearity and is able to detect unimodal relationships between species and external variables (Fumas, 1992)."
COPYRIGHT 2000 Mississippi Academy of Sciences
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2000, Gale Group. All rights reserved. Gale Group is a Thomson Corporation Company.

Article Details
Printer friendly Cite/link Email Feedback
Author:Wieland, Ronald G.
Publication:Journal of the Mississippi Academy of Sciences
Geographic Code:1U6MS
Date:Jul 1, 2000
Words:14549
Previous Article:New Museum Of Natural Science Opens.
Next Article:The Terrestrial Hemiptera and Auchenorrhynchous Homoptera of Point Clear Island and Surrounding Marshlands, Hancock County, Mississippi.
Topics:


Related Articles
Clustering effects of lone star ticks in nature: implications for control.
Editorial.
New Museum Of Natural Science Opens.
Where the wild Things Are.
Calender of events.
Calendar of events.
Alligator adventure.
The historical distribution of prairies in the Jackson Prairie Belt and in western Mississippi.
November & December.
Where in Mississippi is ... Monticello? From the peak of prosperity to the depths of devastation, this riverside town has weathered an incredible...

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters