Library Digitization Homepage
OAS Homepage
Volume 76—1996

{Page 23}

Woody Species Composition of Floodplain Forests of the Little River, McCurtain and LeFlore Counties, Oklahoma

Bruce W. Hoagland, Lisa R. Sorrels and Susan M. Glenn1
Oklahoma Biological Survey, 111 East Chesapeake Street, University of Oklahoma, Norman, OK 73019

1Present address: Centre for Applied Conservation Biology, Department of Forest Sciences, Faculty of Forestry, University of British Columbia, Vancouver, B.C. Canada V6T 1Z4

Received: 1996 Apr 09; Revised: 1996 Aug 10

Species composition and structure of bottomland hardwood forests were studied in the coastal plain region of southeastern Oklahoma. The objectives of this study were to develop a quantitative vegetation classification and analysis of species diversity patterns in bottomland forests of the Little River. Fourteen bottomland sites were sampled with 10–m2 circular plots. Data were compiled into a species-by-site matrix for analysis using detrended correspondence analysis and two-way indicator. The following plant communities types were identified: 1) Quercus phellos, 2) Carpinus caroliniana and 3) Taxodium distichum. Sites were separated in ordination space by the apparent flooding tolerance of dominant tree species. The diversity of woody species was higher in southeastern Oklahoma bottomland forests than in central Oklahoma bottomland forests.


Bottomland hardwood forests are the characteristic vegetation of coastal plain and Piedmont rivers with broad floodplains (1,2). The vegetation of the Oklahoma coastal plain was mapped by Küchler (1) as Southern Floodplain Forest, primarily along the Little and Red Rivers (3-5). These forested wetlands are highly productive ecosystems with well documented ecological and economic values (3,6). Dominant woody genera include Quercus, Nyssa, and Taxodium (2,4,6,7). Vegetation surveys specific to Oklahoma list Fraxinus pennsylvanica, Liquidambar styraciflua, Pinus taeda, Quercus phellos, Quercus nigra, and Quercus lyrata as dominant species (5,8,9). Despite the ecological and economic importance of these natural communities, there is a paucity of information regarding bottomland hardwood forest composition and structure in Oklahoma. In addition, approximately 84% of bottomland hardwood forests in LeFlore and McCurtain counties have been destroyed (9). In this study, we attempt to broaden the knowledge of Oklahoma bottomland hardwood forest communities by developing a quantitative vegetation classification and analyzing patterns of species diversity.


Fourteen bottomland forest sites were sampled in the Little River drainage of McCurtain and LeFlore counties, Oklahoma. Eleven sites were located at the Little River National Wildlife Refuge (referred to as Refuge), one on Cypress Creek, and two on Cucumber Creek, a tributary of the Mountain Fork River (Figure 1).

The Refuge and Cypress Creek sites were situated on the dissected coastal plain physiographic province (10). Two geologic formations, both of Cretaceous age, flank the Little River floodplain: the Antlers sand to the north and Goodland limestone to the south (11). The floodplain is composed of deep Quaternary alluvial deposits with remnant, heavily eroded terrace deposits (11). The predominant soil association at the Refuge and on Cypress Creek is the Guyton silt loam (12). No sites were located on the Cahaba fine sand loam soils, which represent remnant terrace deposits (12). The two Cucumber Creek sites were located in the Ouachita Mountain physiographic province (10). Steep hills composed of Pennsylvanian sandstone and shale flank Cucumber Creek (13). The Ceda-Rubble Land complex, described as shallow with moderate to large stones just below the surface, was the predominant soil along Cucumber Creek (14).

Annual precipitation for southeast Oklahoma averages 107 cm, 62% of which falls

{Page 24}

during the growing season (15). Average relative humidity at mid-aftemoon is 50% (14). Mean annual temperature is 17.3 °C. The annual high temperature occurs in July (38.0 °C) and the low in January (27.5 °C). The growing season extends from April to September (14).


Data were collected during the summers of 1989 and 1990. Since the study area was located in a timber-producing region, sites which had been recently harvested, high-graded, or were current or abandoned pine plantations, were not sampled. Selected forest stands are referred to as sites and each site was quantitatively sampled using 25 randomly placed 10-m2 circular plots. The diameter-at-breast-height (DBH) for all woody species greater than 2.5 cm was recorded within each plot. Relative frequency, relative density, and relative basal area were calculated and summed to derive an Importance Value (IV) for each species at a site (16). Importance values were compiled into a species-by-site matrix for analysis using Two-Way Indicator Species Analysis (TWINSPAN) and Detrended Correspondence Analysis (DCA, 17). Rare species were downweighted for these analyses.

Species importance values were averaged for each TWINSPAN cluster in order to describe the vegetation types they represented. Species richness, evenness, and diversity were then calculated for each TWINSPAN cluster. Species richness was reported as the number of species encountered. Evenness, a measure of the distribution of individuals within a species among the community of species, was calculated according to Pielou (18). Evenness is maximal when there is the same number of individuals among all the species in a community. The Shannon-Weiner Index, a measure of species richness weighted by species evenness, was used to calculate species diversity (19). Woody species nomenclature follows Little (20).


A total of 47 woody species were encountered at the 14 sites sampled. The genus Quercus (nine species total) was commonly encountered during sampling. Quercus phellos and Carpinus caroliniana were the most prominent species in this study, though C. caroliniana was more broadly distributed (Table 1 [Page 25, Page 26]).

Classification. TWINSPAN analysis of the species-by-site matrix produced three clusters. These clusters were named as community types according to the species with the highest average IV. These community types were: (1) Taxodium distichum, (2) Quercus phellos, and (3) Carpinus caroliniana. Of the fourteen sites sampled, seven were assigned to the Quercus phellos community type, four to the Carpinus caroliniana community type, and three to the Taxodium distichum community type. In the first TWINSPAN division, the Quercus phellos community type was separated from the Taxodium distichum and Carpinus caroliniana community types by the presence of Acer saccharum. The Taxodium distichum

{Page 25}

{Page 26}

[Page 25 and Page 26 consist of Table 1.]

community type was then separated from the Carpinus caroliniana community types in the second TWINSPAN division by the Betula nigra, which was present only at the three Taxodium distichum sites.

Carpinus caroliniana and Liquidambar styraciflua were common overstory constituents in the Quercus phellos community type. Understory species with the highest total importance values were Crataegus marshalii, C. viridis, and Cornus florida. Canopy composition was most variable in the Carpinus caroliniana community type. Several tree species were important in this community type, including Acer rubrum, Nyssa sylvatica, and Quercus alba. Crataegus marshallii was a commonly encountered shrub. The Taxodium distichum community type was prevalent in sloughs and along stream margins. Common woody species in this community type included Acer saccharum, Carpinus caroliniana, and Ilex decidua.

Ordination. Sites were separated in ordination space by the apparent flooding tolerance of dominant woody species (Fig. 2). Sites in the Quercus phellos community type had the highest proportion of flood-tolerant Quercus spp. (i.e., Q. falcata, Q. lyrata, Q. nigra, Q phellos) and the lowest first axis DCA scores. The high IV for Liquidambar styraciflua separated the Caney and Refuge-North sites from others in the Quercus phellos community type. Overall, importance values for water tolerant Quercus spp. were low in the Carpinus caroliniana community type. Second axis DCA scores were high for sites with Quercus alba, regardless of community type. The high axis 1 and low axis 2 scores for the Taxodium distichum community type are most likely due to the singular presence of Taxodium distichum, Betula nigra, and Acer saccharum at those sites.

Species Diversity. Species richness was greatest in the Quercus phellos community type, but species diversity and evenness were low (Table 2). Species diversity and evenness scores were highest for the Taxodium distichum community type, but this community type had the smallest number of species. Richness, evenness and diversity values were not significantly different between communities.


In this study, we identified three bottomland forest vegetation types in the Little River drainage. Previous vegetation surveys had listed Liquidambar styraciflua, Quercus alba, and Acer rubrum as bottomland hardwood forest dominant species in Oklahoma (5,8,9,20). These species were constituents of all three community types, but Carpinus caroliniana and Quercus phellos were most abundant. The Quercus phellos community type reported here is closely allied with the SAF forest type 92, sweetgum. - willow oak (21). Although Liquidambar styraciflua was not a dominant species, it was a secondary species in the Q. phellos community type. Liquidambar styraciflua is an early successional species in bottomland hardwood forests and its abundance may be indicative of successional status at a site (8,20,21).

The Carpinus caroliniana community type was distributed throughout the study area. Variation in canopy composition was most likely effected by the difference in elevation between the Refuge and Cucumber Creek sites. In general, the

{Page 28}

Cucumber Creek stream gradient is greater and the floodplain more confined than the lower Little River. Carpinus caroliniana is a common tree on high elevations within floodplain sites (2). These differences were apparent in the ordination diagram. Many of the woody species at the Cucumber Creek sites had a lower flood tolerance than those at Refuge sites. For example, Quercus rubra is a weakly flood-tolerant species (22) that was encountered only on Cucumber Creek. Quercus spp. with greater flood tolerance (i.e., Q. nigra, Q. nuttallii, Q. phellos, Q. lyrata; 22) were prominent in Refuge C. caroliniana communities.

The Taxodium distichum community type (equivalent to SAF type 101; 21) was of limited extent in this study. Nonetheless, we noted vigorous regeneration and several mature stands at the Refuge. Taxodium distichum was an important commercial tree throughout the coastal plain, including Oklahoma, at the turn of the century (23). Taxodium distichum is the only tree species present in Oklahoma deepwater habitats (5,7). Interestingly, Acer saccharum was found to be a prominent member of the T. distichum community type. Although A. saccharum is common on poorly drained flats (24), it is not typically considered an associate of T. distichum. Sampling with randomly placed quadrats may be responsible for this apparent anomaly. The sampling method could have obscured vegetation patterns produced by floodplain microtopography, an important factor in seedling regeneration and vegetation structure in bottomland hardwood forests (25,26).

At the continental scale, species diversity in bottomland forests has been shown to decrease from east to west (27). Bottomland forest composition also changes from eastwest in Oklahoma (28,29,30). In southcentral Oklahoma, species diversity (H'=2.99) was higher for a riparian forest than the community types reported here, although species richness (n=18) was considerably lower (31). Species diversity in north-central Oklahoma bottomland forest (H'=2.49) did not exceed those reported from the Little River (30). At the community level, though, species richness (n=29, x=21.2) was comparable to the Carpinus caroliniana community type (30). Interestingly, species diversity in bottomland forests of LeFlore county (H'=2.22; 32) was lower than the community types on the Little River. Species diversity for the LeFlore county sites was comparable to that at Cucumber Creek sites.


We thank Newell McCarty, Carter Miller, and Danny Sorrels for field assistance. Useful comments were provided by an anonymous reviewer, Ian Butler, Julianne Hoagland, Forest Johnson, and Ernest Steinauer. Funding for this research was received from the U. S. Fish and Wildlife Service, The Nature Conservancy, and the University of Oklahoma.


1.   Küchler, A.W., The Potential Natural Vegetation of the Conterminous United States. American Geographical Society Special Pub. No. 36, New York (1964).

2.   Wharton, C.H., Kitchens, W.M., Pendleton, E.C., and Sipe, T.W., The ecology of bottomland hardwood swamps of the southeast: a community profile. U.S. Fish and Wildl. Serv., Biol. Serv. Program FWS/O.S.-81/37 (1982).

3.   Wharton, C.H., The Southern River Swamp: a Multiple Use Environment. Georgia State University, Atlanta, GA (1970).

4.   Wilkinson, D.L., Schneller-McDonald K., Olson, R.W., and Auble, G.T. Synopsis of wetland functions and values bottomland hardwoods with special emphasis on eastern Texas and Oklahoma. US Fish Wildl. Serv. Biol. Rep. 87(12) (1987).

5.   Blair, W.F., and Hubbell, T.H., The biotic districts of Oklahoma. Am. Midl. Nat. 20, 425-454 (1938).

6.   Mitch, W., and Gosselink, J., Wetlands Van Nostrand Reinhold, New York (1986).

7.   Penfound, W.T., Southern swamps and marshes. Bot. Rev. 18, 413-446 (1952).

8.   Bruner, W.E., The vegetation of Oklahoma. Ecol. Monogr. 1, 99-188 (1931).

9.   Brabander, J.J., Masters, R.E., and Short, R.M., Bottomland hardwoods of eastern Oklahoma. U.S. Fish and Wildl Serv., Tulsa, OK (1985).

10.   Curtis, N.M. Jr., and Ham, W.E.,

{Page 29}

Geomorphic Provinces of Oklahoma (1:2,000,000 scale map). Oklahoma Geological Survey, Norman, OK (1972).

11.   Davis, L.V., Geology and ground-water resources of southern McCurtain county, Oklahoma. Oklahoma Geological Survey Bulletin 86 (1960).

12.   Reasoner, R.C., Soil survey of McCurtain County, Oklahoma. U.S. Dept. of Agriculture, Soil Conservation Service. Washington, DC (1974).

13.   Briggs, G., Geology of the part of the Lynn Mountain syncline, LeFlore county, Oklahoma. Oklahoma Geol. Survey, Circular 75 (1973).

14.   Abernathy, E.J., Olszewski, K.M., and Peters, R., Soil survey of LeFlore County, Oklahoma. U.S. Dept. of Agriculture, Soil Conservation Service, Washington, DC (1983).

15.   Wickham, P., (ed), Statistical Abstract of Oklahoma: 1993. Center for Economic and Management Research, College of Business Administration, University of Oklahoma and Oklahoma Dept. Commerce, Oklahoma City, OK (1994).

16.   Curtis, J.T., The vegetation of Wisconsin. University of Wisconsin Press, Madison, WI (1959).

17.   Gauch, H.G. Jr., Multivariate Analysis in Community Ecology. Cambridge University Press, New York, NY (1982).

18.   Pielou, E.C., An Introduction to Mathematical Ecology. John Wiley and Sons, New York, NY (1969).

19.   Magurran, A.E., Ecological Diversity and Its Measurement. Princeton University Press, Princeton, NJ (1988).

20.   Little, E.L., Forest trees of Oklahoma. Oklahoma Forestry Division, Pub. No. 1, Oklahoma City, OK (1981).

21.   Erye, F.H., Forest cover types of the United States and Canada. Society of American Foresters, Washington, DC (1980).

22.   Teskey, R.O., and Hinckley, T.M., Impact of water level changes on woody riparian and wetland communities. Vol. II the southern forest region. U.S. Fish and Wildl. Serv. Biol. Serv. Program FWS/OBS-77/59 (1977).

23.   Little, E. L. Baldcypress Taxodium distichum) in Oklahoma. Proc. Okla. Acad Sci. 60, 105-107 (1980).

24.   Teskey, R.O., and Hinckley, T.M., Impact of water level changes on woody, riparian and wetland communities. Vol, IV: eastern deciduous forest region. U.S. Fish and Wildl. Serv. Biol. Serv. Program FWS/OBS-78/87 (1977).

25.   Strang, D.R., Glitzentein, J.S., and Harcombe, P.A., Woody seedling dynamics in an east Texas floodplain forest. Ecol. Monogr. 59,177-204 (1989).

26.   Jones, R.H., Scharitz, R.R., Dixon, P.M., Segal, D.S., and Schneider, R.L., Woody plant regeneration in four floodplain forests. Ecol. Monogr. 64, 345-367 (1994).

27.   Marks, P.L., and Harcombe, P.A., Community diversity of coastal plain forests in southern east Texas. Ecology 56, 1004-1008 (1975).

28.   Hefley, H.M., Ecological studies on the Canadian River floodplain in Cleveland County, Oklahoma. Ecol. Monogr. 7, 346-402 (1937).

29.   Rice, E.L., Bottomland forests of northcentral Oklahoma. Ecology 46, 708-714 (1965).

30.   Collins, S.L., Risser, P.G., and Rice, E.L., Ordination and classification of mature bottomland forests in north central Oklahoma. Bull. Torrey Bot. Club 108, 152-165 (1981).

31.   Petranka, J.W., and Holland, R., A quantitative analysis of bottomland communities in south-central Oklahoma. Southw. Nat. 25, 207-214 (1980).

32.   Johnson, F.L., Woody vegetation of southeastern LeFlore county, Oklahoma, in relation to topography. Proc. Okla Acad. Sci. 66, 1-6 (1986).