The source of this uncorrected OCR text may be viewed in the DjVu format at: http://fax.libs.uga.edu/J84xIx49x89x82x04/ or http://purl.galileo.usg.edu/ugafax/J84xIx49x89x82x04/ FWS/OBS-82/04 November 1982 THE ECOLOGY OF SOUTHEASTERN SHRUB BOGS (POCOSINS) AND CAROLINA BAYS: A COMMUNITY PROFILE by Rebecca R. Sharitz and J. Whitfleld Gibbons Savannah River Ecology Laboratory Alken. SC 29801 Project Officer Edward C. Pendleton National Coastal Ecosystems Team U.S. Fish and Wildlife Service 1010 Gause Boulevard Slldell. LA 70458 Performed for National Coastal Ecosystems Team Division of Biological Services Fish and Wildlife Service U.S. Department of the Interior Washington. DC 20240 and Region 4 Fish and Wildlife Service U.S. Department of the Interior 75 Spr'ng Street, SW Atlanta. GA 30303 The findings 1n this report are not to be construed as an official U.S. Fish and Wildlife Service position unless so designated by other authorized documents. PREFACE Library of Congress Card Number : 82-600630 This report should be cited as: Sharitz, R.R., and J.W.'Gibbons. 1982. The ecology of southeastern shrub bogs (poco- sins) and Carolina bays: a community profile. U.S. F1sh and Wildlife Service, Division of Biological Services, Washington, D.C. FWS/OBS-82/04. 93 pp. This report Is one of a series of U. S. F1sh and Wildlife Service Community Profiles synthesizing the available literature for selected critical eco systems Into comprehensive and definitive reference sources. The objective of this particular account is to review the infor mation available on the shrub bog com munities, primarily pocosins but also including many Carolina bays, of the mid- Atlantic Coastal Plain. Both of these ecosystems are Included under the same community profile because they frequently exhibit similar vegetatlonal character istics, have overlapping geographical ranges, and usually have low nutrient, peaty soil conditions. The combination of similar environ mental conditions and the regional overlap of the two ecosystems results in their having similar, species compositions. They are generally distinguishable, however, because of the distinctive morphology of Carolina bays (oval-shaped depressions with identifiable margins) and clearly different geologic origins, although dis agreement exists about how both pocosins and Carolina bays were formed. Nonethe less, they share 1n being the major sites of shrub bog communities on the mid- Atlantic Coastal Plain. s These shrub bog communities are under severe environmental impact throughout the Southeast although the major threats vary in different areas. Most man-related environmental alterations have occurred because pocosins and Carolina bays can be drained and the land used for agricul tural, forestry, or urban purposes and because they often have peaty soils, which may be commercially valuable. Because of previous and continuing efforts to use or alter these areas for commercial purposes, the amount of remaining natural habitat diminishes annually. The shrub bog community of pocosin and Carolina bay ecosystems Is one of the least-studied, most poorly understood of the natural wetlands in the Eastern United States, a fact that makes the rapid dis appearance of these habitats due to the environmental alterations associated with land use and development of critical importance. This report will provide a synthesis of what 1s currently known about the geologic origins, hydrology, soil con ditions, water quality, biota and general ecology of these ecosystems. A considera tion of the community as a whole reveals major gaps in our knowledge of its struc ture and dynamics at all levels. The limited information already available clearly documents the close environmental tie between coastal pocosins and contigu ous estuaries. A strong Interaction between Carolina bays and mobile animals from surrounding terrestrial habitats can also be demonstrated. This community profile addresses Information gaps which still exist about the significant environ mental Interactions between shrub bog communities and surrounding habitats. The information in this community profile should be useful to all environ mental planning groups, ecosystem mana gers, and concerned laymen 1n the Atlantic Coastal Plain, as well as to students and professional wetland ecologists. The quantitative presentations in figures, tables, and text will permit easy access to detailed Information about specific qualities and characteristics of the habi tats. Thus, the document should be valu able for report writing, planning environ-, mental assessment studies, preliminary determinations of potential Impacts of N land management, or general Information for formal courses on the topic of natural wetlands. Such Information on environ mental impact, rate of habitat loss, and gaps In our knowledge about these systems should Interest all ecologlsts, environ mental planners, and environmentally aware citizens. An extensive, thorough and cur rent bibliography is provided for anyone wishing to pursue particular topics In depth. The authors accept full responsibil ity for all statements, Interpretations, citations of other investigators, and original data presented. Any questions or comments about or requests for this publication may be directed to: Information Transfer Specialist National Coastal Ecosystems Team U. S. F1sh and Wildlife Service NASA/SIidel Computer Complex 1010 Cause Boulevard SI 1dell, LA 70458 CONTENTS PREFACE ..................................... 111 FIGURES ..................................... vi TABLES ...................................... viii ACKNOWLEDGMENTS ................................. 1x INTRODUCTION COMMUNITY PROFILE BACKGROUND AND OBJECTIVES: SHRUB BOGS OF THE ATLANTIC COASTAL PLAIN ............. 1 CHAPTER 1. DESCRIPTION OF SHRUB BOGS OF .THE ATLANTIC COASTAL PLAIN ..... 5 1.1 Definitions of Shrub Bogs, Pocosins, Carolina Bays, and Similar Ecosystems ...................... 5 1.2 Geographic Distribution of Pocosins and Carolina Bays ...... 10 1.3 Theories of Geologic Origin ................... 12 CHAPTER 2. PHYSICAL AND CHEMICAL CHARACTERISTICS. .............. 16 i 2.1 Size and Morphology of Community Types .............. 16 i 2.2 Substrate Conditions ....................... 16 2.3 Hydrologie Characteristics .................... 20 CHAPTER 3. BIOLOGICAL FEATURES ....................... 27 3.1 Plant Community Composition ................... 27 3.2 Animal Community Composition ................... 51 3.3 Endangered and Threatened Species ................ 59 3.4 Pocosins and Carolina Bays as Refuge Areas ............ 59 3.5 Trophic Relationships in Pocosin and Carolina Bay Ecosystems .......................... 60 CHAPTER 4. INFLUENCE OF HUMAN ACTIVITIES ON POCOSINS AND CAROLINA BAYS. .......................... 62 4.1 Historical Perspectives ..................... 62 4.2 Current Land Management Practices and Perturbations ....... 63 4.3 Current Ownership of Pocosins and Carolina Bays ......... 70 4.4 Recommendations for Research ................... 70 CHAPTER 5. RECOMMENDATIONS FOR CONSERVATION, PRESERVATION AND MANAGEMENT .......................... 74 REFERENCES ......'.............................. 76 FIGURES Number Page 1 Evergreen shrub bog pocosin 1n.Dare County, NC ............ 2 2 Two aspects of a typical Carolina bay, showing gradation In vegetation types .......................... 3 3 The relation of shrub bogs, pocoslns, and Carolina bays to each other and to the general class of palustrine wetlands ...... 5 4 Habitats In eastern North Carolina that are Identified as pocosin wetlands ........................... 7 5 South Atlantic Coastal Plain showing distribution of Carolina bays ............................. 9 6 Distribution of blackland soils in North Carolina ........... 11 7 Aerial comparison of two Carolina bay area densities ......... 12 8 Conceptual model of one proposed means of pocosin formation and development in flat, Inter^stream areas of lower Coastal Plain terraces ....'................ 15 9 Section and plan views of a typical Carolina bay, indicating key morphological features, soil profiles, and vegetation types .... 17 10 Soil types In North Carolina's Green Swamp pocosin .......... 19 11 Relationships among precipitation, evapotranspiration, and streamflow in a pocosin habitat .................. 21 12 Water level changes in two Carolina bays as influenced by precipitation and temperature ................... 25 (a) Mean monthly changes 1n water level, 1975-1981 .......... 25 (b) Changes in water level at Ellenton Bay .............. 25 (c) Seasonal drops in water level due to evapotranspiration at Ellenton Bay ......................... 25 (d) Changes in monthly mean water levels due to temperature ..... 25 13 Representative cross section through the Croatan National Forest pocosin substrates ....................... 28 14 Comparison of two types of pocosin habitats .............. 30 15 Vegetation types In Green Swamp, NC .................. 31 16 Dominant woody plant species of pocosin communities .......... 36 17 Major woody plant community cover classes in the Green Swamp, NC . . . 38 18 Proposed relationships among vegetation types, hydroperiod, and fire 1n pocosin habitats ....................... 45 19 The pitcher plant (Sarracenia purpurea) is adapted to nutrient- poor pocosin soils .......................... 47 20 Aerial view of vegetational zone pattern around a Carolina bay .... 48 21 Vegetation types of a Carolina bay that are characteristic of many undisturbed sites ........................ 49 22 The pine barrens treefrog (Hyla andersoni), an amphibian species indigenous to pocosin h~a~l>1tatsT ............... ^52 -23 The black bear (Ursus americanus), a major game species dependent on pocosin habitat in the lower Coastal Plain of North Carolina .... 53 24 A drift fence with pitfall traps at Ellenton Bay, SC ......... 57 25 A "gâter hole," indicating how a single species, the American v alligator (Alligator mississipplensis). can influence the ecology of a Carolina bay ...........'............ 58 x vl Number 26 27 28 29 30 31 Page Species adapted to fluctuating water levels characteristic 5g of Carolina bays ••••:•••'••: l-l,! '„* •••••••••' General pattern of ownership of pocosin habitats ot ^ Pocos1naSrainage'for*agricultural or forestry management ....... 64 Drought conditions in a ditched Carolina Bay . . . . . . . . - • • • • JD Aerial view of Carolina bay habitats in an agricultural region .... 66 Peat mining in a pocosin habitat ..........••'•••••• vti TABLES ACKNOWLEDGMENTS 4 5 6 7 8 9 Substrate characteristics of 13 pocosins in the Coastal Plain of South Carolina ............. Substrate characteristics of five Carolina bays in northeastern South Carolina .......... (a) (b) Comparison of water quality data from a natural pocosin stream with data from three drainage ditches in the same area .... Comparison of mean annual concentrations of constituents of pocosin seepage and farm ditch water during the clearing and drainage phase of wetlands development .......... Water quality parameters of seven Carolina bays in South Carolina .' (a) Classification of pocosin communities in the Green Swamp . . . (b) Associated communities peripheral to pocosins ........ Density and basal area of trees in a pocosin in South Carolina ! '. Density and transect cover of shrubs, woody vines, and tree seedlings in a pocosin in South Carolina ............. Comparison of various geological and ecological parameters'of pocosins, pine savannas, and bay forests in South Carolina . . . Secondary productivity associated with Carolina bay habitats . . . Page 18 20 23 23 24 34 35 ,39 40 44 55 We are grateful to numerous individ uals for their contributions to the com pletion of this project. We particularly thank C. J. Richardson (Duke University), C. B. McDonald, A. N. Ash, and M. M. Brinson (East Carolina University), and B. J. Cope!and (North Carolina State Univer sity) for instructive comments and infor mation about the shrub bogs in North Carolina. Resource material was provided by R. W. Skaggs (North Carolina State Uni versity), C. C. Daniel, III (Geological Survey, Raleigh), and T. M. Williams (Belle W. Baruch Forest Science Insti tute). C. B. McDonald and A. N. Ash spent time with us in the field visiting a vari ety of pocosin habitats in North Carolina and graciously provided illustrative mate rial for use in the report. A number of reviewers, including A. N. Ash, C. B. McDonald, and L. J. Otte (East Carolina University), C. J. Richard son and N. L. Christensen (Duke Univer sity), E. C. Pendleton and J. M. Hefner (U. S. Fish and Wildlife Service), C. A. Gresham (Clemson University) and J. F. Schalles (Creighton University), provided useful comments on an early draft. Members of the National Coastal Ecosystems Team (NCET) staff, especially W. M. Kitchens and E. C. Pendleton, were extremely help ful in offering guidance and suggestions. G. Farris, S. Lauritzen, and E. Krebs of NCET provided editorial, layout, and word- processing assistance, repectively; G. Golden assisted in drafting. We thank members of the University of Georgia's Savannah River Ecology Labora tory (SREL) staff who participated in re search cited in this document, T. Willing- ham who typed many copies of the manu script, J. B. Coleman who drew some of the figures, and L. R. Bost who helped with proofreading and preparation of the bibli ography. Data collection and manuscript preparation were aided by contract EY-76- C-09-0819 between the U. S. Department of Energy and the University of Georgia. fx INTRODUCTION COMMUNITY PROFILE BACKGROUND AND OBJECTIVES: SHRUB BOGS OF THE ATLANTIC COASTAL PLAIN The recognition and appreciation that natural wetlands are a valuable, non- renewable resource that can provide criti cal or essential habitat for natural vege tation and wildlife, including non-game and endangered species, are becoming wide spread among government, academic, and even commercial organizations. This con cern has led to establishment of Federal and State regulations (for example Execu tive Order 11990, Protection of Wetlands) to protect these natural systems and their associated flora and fauna. Understandably, initial attention by most interested groups focused on large bodies of permanent water such as estua ries, major lake systems, and rivers or streams which supported obvious game or commercial enterprises. When industrial, agricultural, or urban impacts on these large aquatic areas affected commercial and sports fisheries interests or the hunting of waterfowl, environmental con cerns quickly became an issue. Resolution of the complex of environmental, legal, ethical, and other issues surrounding the best approaches of preserving, managing, or utilizing natural wetlands is not likely to be attained in the near future. It is agreed, however, that our basic knowledge of the structure and function of most natural systems is woefully inade quate for us to make prudent decisions about the optimal management (or non- management) approach. Far too little is known about the life histories of most plants and animals inhabiting aquatic areas. Biological surprises about even the most thoroughly studied species emerge monthly in professional journals (e.g., Echelle and Mosier 1981; Fairchild 1981), reinforcing the awareness that we know far too little about the ecology of most natu ral systems to be able to exercise rea sonable judgment about how these systems should interface with our agro-industrial society. Among the natural wetlands that are most poorly studied are two types re stricted to the Atlantic Coastal Plain and abundant from southern Virginia to north ern Florida. These wetland ecosystems— pocosins (Figure 1) and Carolina bays (Figure 2)~share a common feature of be ing systems of low-nutrient status that in most instances do not have permanent standing water, but are strongly influ enced by the hydrologie regime. The vege tation of pocosins and of many extant Carolina bays is dominated by pines and broadleaved evergreen shrubs or low trees that are generally but not always growing on highly organic or peat soils. Man's activities have imposed a vari ety of environmental modifications and ecological changes in pocosin and Carolina bay ecosystems. Subsequent to regional timber removal in the early development of the Southeast, pocosin ecosystems have generally been considered of low economic value for agriculture, although they have recently become recognized as a source of peat. In regions of excessively drained soils, Carolina bays have commonly been farmed because they are areas of higher soil moisture. Few studies, however, have been carried out to assess the regional importance of either ecosystem in terms of faunal and floral dependence and utiliza tion. It is not known how critical a pocosin or a Carolina bay is to regional wildlife. For example, black bears are characteristically encountered in pocosins of eastern North Carolina, primarily because these ecosystems contain some of the few natural areas left in the region. Are present-day land management practices jeopardizing the future of cer tain plants and animals that rely on such natural habitats for feeding or breeding sites or other purposes? The extent of man's alteration of these wetlands cannot be determined. Dur ing colonization, widespread timbering occurred throughout the Coastal Plain re gion. Although such logging was not exclu sively in pocosins or Carolina bays, these habitats were not spared from considerable tree removal. Thus, major losses of Atlan tic white cedar, cypress, and black gum probably occurred. As a result, the orig inal vegetationa characteristics of many localities are unknown today. The shrub bog community may be second-growth in some instances, following logging of the orig inal forest species. In other areas, par ticularly in smaller Carolina bays, an original evergreen shrub vegetation may have been altered or removed by farming operations so that the shrub community is sparse or absent today. Because of the peat soil feature characteristic of many pocosins, large expanses of pocosin vegetation may have been burned during dry periods in early historical and recent times. A major modern impact is the removal of peat for commercial purposes. The highly organic soil can be processed to produce methanol or sold as a mulching agent for horti cultural operations or home use. Although at the present time peat has been mined only experimentally in the major pocosin areas of North Carolina, peat-methanol gasification operations are scheduled for future development. Thus, there is poten tial that thousands of acres of natural pocosin habitat could succumb to the commercial efforts of peat mining in the Carolinas. Major changes have also occurred in the hydrologie regimes of many pocosins and Carolina bays, primarily because wet f ' t '. • • /•» " •— v •* - J " - +- •j . : .*; vt • v -.-. Figure I.Evergreen shrub bog pocosin in Dare County, NC. co CO T- C = C O _. o < .C "5 3 01 (D « « «n «£ *x o 0) O S CO 3 c DC CO o « » " jS'V ri ^-'/S'i/â^^^ 'x-^ ;s«i^ r I 'T-/MW/) "'T ,.•'* • V "- -.^^V a ,-< of tfflL» . •/ * r.f ^\ *• xvx: s: ,-~ t -t?*5^ - w •*• A. • Il «E .E O « c (D W . » oo =- 6 broadleaved evergreen; water regime: saturated; water chemistry: fresh-acid; and soil: medisaprist.2 The definition of pocosin is con founded by various terminologies that have arisen among different professions. For example, pine-dominated flatwoods occur ring in areas with prolonged hydroperiods may be included in a forester's defini tion, whereas a hydrologist might consider only those shrub bogs occurring in broad, undrained interstream areas to be true pocosins. Because of the various defini tions and the similarities between pocosin vegetation and other plant communities such as bay forests, some of the material in this review may be taken from or ap plied to Coastal Plain wetlands that are closely related to pocosins but do not fit the ecological definition in all aspects. Possible successional relationships among these various wetland communities are dis cussed in Chapter 3. In summary, for the purposes of this community profile, pocosins are defined as freshwater wetland ecosystems character ized by broadleaved evergreen shrubs or low trees, commonly including pond pine, and commonly growing on highly organic soils that have developed in areas of poor drainage. Their present range of occur rence is the Atlantic Coastal Plain from southern Virginia to northern Florida. Carolina Bays Carolina bay ecosystems are formed in elliptical depressions which occur abun dantly in a broad band across the Coastal Plain province of the Southeastern United States (Figure 5). Some bays are less than 50 m (162 ft) in length; an unusually large one (Lake Waccamaw in North Caro lina) is more than 8 km (5 mi) long. These puzzling physiographic features of the landscape present a remarkable consis tency of shape and degree of parallelism in compass orientation along their axes. Called "bays" by early European pioneers who observed the evergreen shrubs and bay Medisaprist soils are well-decomposed organic soils (histosols) with an organic depth greater than 40 cm (16 inches). trees typically growing on their margins (Figure 2), these ovate wetlands afford suitable habitat for a vast assortment of plants and animals. Colloquial names such as "highland pond" or "wet weather lake" indicate the dependency of Carolina bay ecosystems upon rainfall and their lack of association with other lentic (still water) or lotie (flowing water) habitats. They character istically have no tributary streams, are not spring-fed, and rely on direct precip itation and run-off to maintain water volume. Groundwater recharge has been suggested as an additional source in some situations (Schalles 1979). Evaporative water loss, a temperature-dependent phe nomenon, reduces water volume and can result in the complete drying of shallow bays. Thus, many smaller Carolina bay depressions contain temporary aquatic hab itats that may dry up seasonally or for longer intervals under local conditions of low precipitation or regional drought. Because of extensive variability in size, depth, substrate conditions, and geographic location, Carolina bays do not have a single characteristic vegetation type. Many, including Jerome Bog (Buell 1946a) and Rockyhock Bay (Whitehead 1981) in North Carolina, contain evergreen shrub bogs or pocosins. Others contain lakes (Frey 1949), herbaceous marshes (Wharton 1978), or swamps (Porcher 1966). Still others, such as Ellenton Bay in South Car olina (Sharitz and Gibbons 1982) may have once supported forest or shrub communi ties, but have been severely modified by agricultural practices. Both the natural unaltered plant communities and the direc tion and rate of successional change in Carolina bays following disturbance are highly dependent upon,environmental var iables such as^ water level and fire. Therefore, a particular community type does not necessarily represent a particu lar stage of succession. For more ' than 300 years, natural habitats in the-1 region encompassing Caro lina bays have been influenced by human alteration as a consequence of agricul ture, forestry, industry, and 'other land management. These impacts, in addition to any burning by Indians (Wells and Boyce 1953b), have left unanswerable the CHOWAN ROANOKE PAMLICO NEUSE CAPE FEAR LAKE WACCAMAW COOPER COMBAHEE BROAD SAVANNAH OGEECHEE <$> ^ -. A^ . • if f. '•' '* :" :lt.\ * '-A» •• -* X to \ a • ' ^ t* - •*-.£ '» ^fc^ j« ^'l| GREAT DISMAL SWAMP -*-^V-ALBEMARLE SOUND - \ ™ LAKE PHELPS i*. >!' ! \ LAKE ^-^ " ' MATTAMUSKEET -*—-W- PAMLICO SOUND ^^'CAPE HATTERAS - GREAT DOVER SWAMP OPEN GROUND SWAMP A •. CROATAN NATIONAL FOREST CAPE LOOKOUT HOFMANN FOREST ANGOLA BAY HOLLY SHELTER SWAMP GREEN SWAMP CAPE FEAR Figure 6. Distribution of blackland soils (organic soils and dark-surfaced mineral soils) in North Carolina (Lilly 1981b in Pocosin Wetlands: An Integrated Analvaia oj Coastal Plain Freshwater Bogs in North Carolina, ed. C. J. Richardson. Copyright 1981 by Hutchinson Ross Publ. Co., Stroudsburg, PA. Reprinted by permission of the publisher). 10 11 (1952), are found in the Carolinas (North Carolina Natural Heritage Program unpub lished data, cited by Richardson 1982). It is difficult to estimate the total area covered by Carolina bays because they are non-contiguous in distribution, and many may have boundaries that are hard to rec ognize, may be obscured by farming or other land management or development practices, or may partly overlap .each other. The geographic range of Carolina bays encompasses approximately 23,310 km2 (9000 mi2). High density area, Bladen County, NC (Photograph by Charles E.Roe, North Carolina Natural Heritage Program.) Low density area. Barnwell County, SC Figure 7. Aerial comparison of two Carolina bay area densities. Neither pocosins nor Carolina bays demonstrate clear gradients in geomorphol- ogy or in species composition throughout their range of occurrence, although cer tain regional differences nay be recogniz able. For example, directional orienta tion of Carolina bays may show a slight geographic variation (Johnson 1942a), as described in Chapter 2. The total acreage of land occupied by pocosin wetlands on the Coastal Plain decreases from north to south in accordance with changes in the topography and geology. Although organic soils supporting pocosin vegetation may develop under several geologic situations, flat interstream areas"of the lower Coast al Plain terraces in North Carolina sup port the greatest expanse of pocosin vege tation. In South Carolina where pocosin ecosystems are more restricted in area, they occur chiefly in depressions associ ated with ridge and swale topography and in Carolina bays. Likewise, certain dif ferences in the characteristic species may be recognizable in pocosin or Carolina bay communities throughout their range. Wood- well (1956) classed pocosins according to their dominant shrub species and reported a shift in species dominance along a north-south gradient and following fire (see Chapter 3). The North Carolina poco sin communities represent the northern range extensions for certain species, including loblolly bay. Similar geo graphic patterns of species distribution may have become obscured by disturbance from natural causes or from land manage ment practices. 1.3 THEORIES OF GEOLOGIC ORIGIN Pocosins On the basis of the most thoroughly studied pocosin habitats—those in the Coastal Plain of North Carolina (Richard son 1981a)—these wetland communities be gan to develop following the Wisconsin Ice Age, about 15,000 years ago. Daniel (1981) described the geologic processes that led to the development of these areas. More than 75,000 years ago prior to the last expansion of the polar ice cap, the level of the Atlantic Ocean in the Southeast was 13 to 15 m (approximately 45 to-50 ft) higher than at present. During the Wis consin Ice Age, however, the sea level dropped to as much as 120 m (about 400 ft) below its present level and exposed large areas of the Continental Shelf. Because sea levels during the Wisconsin Age were much lower than at present, fast-flowing rivers cut through the Coastal Plain ter races to the Atlantic Ocean. During the next several thousand years as the ice receded, sea levels gradually rose ( - 30 cm/century, Milliman and Emery 1968) and brought about major changes in the hydrology of the Coastal Plain. One theory of pocosin formation relies on the assumption that during this period river flows were slowed and organic sediment loads were deposited in the interstream areas as the lotie (flowing) systems shifted to near-lentic (or slow- moving) aquatic systems (Daniel 1981). Aquatic plants began to grow in these shallow bodies of water, adding to the accumulation of sediments and the buildup of aquatic debris. Concomitant with the buildup in organic sediments and the cli matic warming trend that accompanied the end of the Wisconsin Ice Age, the cooler- climate, boreal forest communities were gradually eliminated (Whitehead 1972) and replaced with hardwood forests, swamps, bogs, and marshes. In the nutrient-poor areas often associated with the buildup of deep peat, evergreen shrub forests became predominant. It is assumed that wet habitats with a propensity for buildup of organic soils developed as a consequence of a shallow water table, large distance between streams (both resulting in slow runoff and subsequent accumulation of water), and rainfall that has exceeded evapotranspira tion (Daniels et al. 1977). The accumula tions of peat that would ultimately form the Coastal Plain pocosins generally began between 8,000 and 10,000 years before pre sent (B. P.) (Daniel 1981). For example, radiocarbon dating of the earliest organic sediments in the Great Dismal Swamp of Virginia indicates that these peats were formed about 8,900 ± 160 years ago (Oaks and Coch 1973). However, a mantle of organic deposits in the Hofmann Forest of North Carolina may have begun as long ago as 220,000 years. Radiocarbon dates of sediment profiles from other sites (e.g., Chesapeake Bay, Harrison et al. 1965) generally indicate ages of 8,000 to 15,000 years B. P. Additional evidence gained by matching pollen profiles from pocosin peat with known floral succession patterns associated with the glacial retreat sup ports the assumption that the pocosin wet lands developed between 10,000 and 12,000 years ago (Otte 1981). Otte (1981) challenged the theory of pocosin development in stream valleys and flat interstream areas as a consequence of blocked drainage associated with sea level rise. He pointed out that 10,000 to 12,000 years ago the sea level was approximately 25 m (81 ft) below the present level and many miles eastward of the present shore line. He contended that the position of the sea at that time was so far removed from the sites of development of the blocked drainage systems that it is un likely that sea level could have been the major controlling factor in the develop ment of these wetlands. Furthermore, Otte noted that pocosins have developed in only a portion of the drainage systems on the Coastal Plain, not in all. This evidence also strongly suggests a non-sea-level cause of pocosin formation. If Otte's theory is correct, the origin of the drainage blockages that resulted in poco sin formation in certain stream systems is as yet unknown. Four different types of geologic sit uations are considered to support pocosin communities in the southeastern Coastal Plain. These are (1) flat areas associated with blocked stream drainage on the lower terraces (Figure 8), (2) Carolina bays, (3) areas of ridge and swale topography between relict beaches and dune ridges (Woodwell 1956), and (4) springs and stream heads of the upper Coastal Plain area (Wells 1932; Christensen et al. 1981). In addition, this evergreen shrub bog association may develop, although not usually as extensively, in other areas such as along floodplains of streams (Daniel 1981). The major pocosin areas of the Carolinas and Virginia have formed in stream valleys or on broad interstream areas with restricted drainage. Paludification (bog expansion caused by a gradual rising of the water table as peat accumulates) has proceeded for sev eral thousand years in these broad inter stream areas in which natural drainage is' 12 13 blocked (Daniel 1981; Richardson et al. 1981). Richardson (1982) described the process by which these pocosins may have developed from a primary mire system (formed in a basin or depression) through a secondary mire stage in which they expanded beyond the physical boundaries of the depression because the peat acted as a water reservoir (Figure 8). Many of the North Carolina pocosins are tertiary mire systems which have developed above the physical limits of the ground waters (Richardson et al. 1981) under conditions of rainfall in excess of evaporation for several thousand years. In these wetland systems, the peat continues to act as a reservoir holding water by capillarity above the level of the groundwater. Hy- drologically, the pocosin has a perched water table elevated above that of the surrounding areas and receives water pri marily from rainfall (Richardson et al. 1981). Water moves slowly out of these raised bogs into adjacent areas. Carolina Bays No theory on the origin of Carolina bay depressions is universally accepted by scientists or laymen. The earliest pub lished theory (Tourney 1848) was that springs rising to the surface of a sandy plain formed circular lakes. Glenn (1895) suggested that the depressions were formed when coastal sandbars built up across the mouths of shallow embayments. Although the rounded contours of Carolina bays were noted in these early descriptive accounts, their abundance, elliptical shape, and parallel alignment were not recognized until aerial photographic surveys were made of the southeastern Coastal Plain in the 1930's. Melton and Schriever (1933a,b) noted the uniform shape and consistent direc tional alignments of the depressions and proposed the captivating hypothesis that they are the result of an ancient meteor shower striking the earth at an oblique angle. Prouty (1952) supported this hypothesis with explanations that the shallowness and shape of such depressions could be accounted for by shock waves from a swarm of meteors crashing into sandy, uncompacted coastal plain soils. An argument that meteorite fragments have not been found associated with any Caro lina bay depression (Johnson 1942b) was not considered as compelling negative evidence since a lack of fragments could be due to the low probability of finding pieces of an exploded meteorite (MacCarthy 1937; Prouty 1952) or due to a non-metal lic meteor swarm (Sharitz and Gibbons 1982). In rebuttal to this early extrater restrial theory of origin, other theories followed which implicated wind and wave action (Cooke 1940, 1943a,b; Johnson 1944a, b; Odum 1952; Thorn 1970; Whitehead 1973; Bliley and Pettry 1979; Kaczorowski 1977); dissolution of substrate minerals in solution pits (Smith 1931; Johnson 1942b); or a combination of factors (John son 1942b, 1944a,b; Whitehead 1981). More imaginative hypotheses regarding the ori gin of Carolina bays include the formation of pools by large schools of spawning fish (Grant 1945b), melting icebergs (Kelly and Dachille 1953), collision of the earth with anti-matter (Baxter and Atkins 1976), or launching sites for extraterrestrial spacecraft (see Justis 1974). In support of the wind and wave action theory, perhaps the strongest chal lenge to the meteor theory, Kaczorowski (1977) compared the Carolina bays with modern analogues found in Alaska, Chile, and Texas. He proposed that oriented lakes develop in topographic depressions created by any one of several processes including those coastal, fluvial, aeolian, solutioning, glacial, or tectonic in nature. He argued that once these depres sions are formed, orientation could be a function of wind regime, perhaps accompan ied by wave action. He stated that no evidence existed to support a meteor ori gin for the initial Carolina bay depres sions. Much disagreement also exists regard ing the age of Carolina bays, although they are generally accepted as being pre- Holocene by most authorities (Schalles 1979). Wells and Boyce (1953) estimated the age of Carolina bays as 250,000 years B.P. Other estimates range from 10,000 to 100,000 years B.P. From a study of pollen profiles, Whitehead (1973) suggested that the oldest sediments were formed about 100,000 years ago. He later stated that there were at least two time periods when bays were formed (Whitehead 1981). Some bays that occur on the lower Coastal Plain terraces are considered by Thorn (1970) to be younger than those on the middle Coastal Plain, also an hypothesis requir ing an explanation that allows for multi- temporal formation. BOREAL FOREST In conclusion, authorities do not agree on the age (within the 10,000 to 250,000 year limits), concurrency of formation, or mechanism of formation of Carolina bays. Disagreement also still exists about whether one of the processes described by Kaczorowski (1977) or a meteor shower created the initial depres sions. STREAM DRAINAGES DEVELOPMENT OF BOTTOMLAND FLOODPLAIN FOREST PEAT FORMATION "SHORT" POCOSIN SUBMERGED AQUATIC VEGETATION "TALL" POCOSIN 15,000-10.000 years B.P. Late Wisconsin Ice Age; low sea levels and complex stream drainage patterns on the Coastal Plain and the exposed Continental Shelf. 10,000-5,000 years B.P. Continued rise in sea levels; raised water tables; decline in stream velocities; blocked (debris clogged) stream drainage; growth of aquatic vegetation; buildup of sediments and organic material to form peat Up to Present Continued buildup of peat with increased water retention and perched water table; expansion of bogs beyond physical limits of stream depression; development of shrub bog community; periodic fires PEAT POND FORMED FROM PEAT FIRE PERCHED WATER TABLE Figure 8. Conceptual model of one proposed means of pocosin formation and development in flat interstream areas of lower Coastal Plain terraces. 14 15 CHAPTER 2 PHYSICAL AND CHEMICAL CHARACTERISTICS 2.1 SIZE AND MORPHOLOGY OF COMMUNITY TYPES Pocosins Pocosins have no characteristic size or shape. In many instances well-defined boundaries are not apparent as the pocosin shrub bog habitat grades into other vege- tationally related communities. A pocosin that would be a recognizable soil and veg etation community unit could conceivably be smaller than an acre or large enough to encompass thousands of contiguous acres. The largest identifiable pocosins occur in eastern North Carolina where 281,000 ha (695,000 acres) were recognized as undis turbed in 1979 (Richardson et al. 1981). Pocosins show no orientational aspects in regard to other topographic features other than location between old or active stream systems in many instances. Carolina Bays Carolina bays may range in length from 50 m (164 ft) to 8 km (5 mi). Lake Waccamaw in North Carolina, perhaps the largest, is 4.8 x 8 km (3 x 5 mi ; Frey 1949). Carolina bays can be readily dis tinguished from other types of depressions on the southeastern Coastal Plain by a unique suite of characteristics (Fig ure 9). These include generally shallow depth, an ovate shape with the narrow end pointed in a southeasterly direction, and an orientation in which the long axis is in a northwest to southeast direction so that Carolina bays are essentially paral lel (Prouty 1952). In most instances a low sandy rim borders each bay and is usu ally most pronounced on the southeastern margin of the ellipse. Although some authors (e.g., Melton and Schriever 1933a, b) emphasized the consistent paral lelism of the long axis of Carolina bays, Johnson (1942a) strongly challenged this concept. He reported that although the predominant long axis orientation was northwest-southeast, the range of direc tional deviation was high and varied geo graphically. Bays that he measured in Georgia and southern South Carolina were oriented on the average along a S 20°E (= N 20°W - S 20°E) direction. The majority ranged from S 10°E to S 30°E, but a few extreme examples were oriented as westerly as S 26°W and as easterly as S 56°E. A 'large sample of Carolina bays measured in North Carolina and northern South Carolina averaged S 45°E and ranged from N-S to S 67°E (Johnson 1942a). Deviation from pre cise parallelism is a factor to be consid ered in a discussion of the geological ori gin of Carolina bays, but is of no conse quence from an ecological standpoint. The basic ovate shape and shallow basin are always maintained and are the dominant physical features of ecological impor tance. 2.2 SUBSTRATE CONDITIONS Pocosins Both very poorly drained mineral soils and organic mucks and peats charac terize areas of pocosin vegetation (Barnes 1981; Gilliam and Skaggs 1981). Barnes (1981) described the mineral subsoil of these wet coastal plain soils as consist ing of layers of marine sediments varying in texture from heavy clays to sands. Clay layers permit the ponding of water at early stages ( of pocosin development and therefore can be an important soil fea ture. The shallow water table has pre vented the development of a conventional soil profile typically formed by weather ing and^ leaching processes. The pH of these pocosin soils is low. The pH of the organic horizons is commonly between 3.5 and 4.1, surface mineral soils average 0.3 watei ion« lttent egi 0 r u £ •> o S I« ,f 5S. 1 11 m •o i cd 5 u Ü H! ^ ? *• SÎI "5 S to o o> o •o CO o o. o (0 « o o CO u 'S» o o & o O» •o o CÔ o "5 _o a o . *s ~ dt *ö m E i* nu s*- l Illl o u >Zl 16 17 to 0.4 pH units higher than organic soils, and mineral subsoils frequently range between 4.1 and 4.7 in pH (Barnes 1981). Although as many as 40 different soils may be recognized in the region of the northern North Carolina pocosins (Gil lian and Skaggs 1981), three major edaphic (soil) groups generally characterize these systems. These are (1) mineral soil-with an organic epipedon (or high organic mat ter surface horizon) that does not extend deep enough ( < 40 cm or 16 inches) for the soils to be classified as organic, (2) shallow organic soils with a high organic surface extending downward 40 to 130 cm (approximately 16 to 50 inches), and (3) deep colloidal organic soils in which the organic matter horizons extend deep into the soil profile ( > 130 cm or 51 inches). These deep organic soils frequently con tain large amounts of buried wood. Jones (1981) characterized pocosin soils from 13 sites in South Carolina as having an ex tremely high organic content that gener ally extended deeper than 40 cm or 16 inches (Table 1). Although the water table is high and the soils may frequently be saturated, pocosins occasionally become dry enough to burn and some of the organic surface may be lost in combustion. Soils information for pocosins in North Carolina is provided by Lilly (1981a). In a study of the distribution and relationships of plant communities of the Green Swamp in North Carolina, a major portion of which is characterized by poco sin vegetation (see Chapter 3), Kologiski (1977) described the soils of this area (Figure 10) and provided maps derived from Table 1. Substrate characteristics of 13 pocosins in the Coastal Plain of South Carolina (Jones 1981). Substrate characteristics Range Organic matter (Ï) Organic surface depth (cm) Texture": Sand (t) S1lt (I) Clay (%) "Measured only 1n nan-organic soils 39.2 >40.0 80.6 10.4 9.0 (6 of the 8.0 - 25.3 - 63.8 - 7.6 - 2.7 - 13 sites). 94.9 >100.0 89.8 17.6 18.6 various sources. He reported the follow ing Coastal Plain soil series to be the major ones underlying the pocosin and re lated habitats. Organic Soils: Four major organic soils series are represented, grouped into two mapping units (Figure 10). They differ chiefly in depth, permeability, and drainage charac teristics. Dare Dorovan Series. These are very deep, very poorly drained and slowly per meable soils formed from the remains of swanp vegetation. The water table is at or near the surface most of the year, and the soils are extremely acid. Dorovan soils appear to frequently extend deeper than those of the Dare Series. Dare and Dorovan soils characterize the major areas of pocosin and bay forest vegetation in the Green Swanp (Kologiski 1977). Pamlico-Ponzer Series. These are moderatelydeep,very poorly drained, slowly or moderately permeable soils formed by organic deposits over marine sediments. They frequently occur in de pressions, and the water table is at or near the surface 6 to 12 months of the year. Panlico and Ponzer soils also under lie pocosin shrub communities in the Green Swamp. Mineral Soils: Kologiski identified seven minera soils in the Green Swamp (Pantego, Rains, Lynchburg, Wrightsboro, Foreston, Leon, and Torhunta), all of which are deep, poorly to moderately drained, and usually moderately permeable. However, only the Torhunta series is mapped as occurring ex tensively in the pocosin areas of the Green Swamp. Like the organic soils, Tor hunta soils are1 very poorly drained, with the water table at or near the surface 2 to 6 months of the year. ^ Despite high levels of organic mat ter, pocosin soils are deficient in avail able nutrients (Richardson 1982). The anaerobic conditions resulting from the long hydroperiods and shallow water table and the acidity of the peats preserve the organic constituents (Maki 1974). For DARE-DOROVAN SERIES m® PANTEGO SERIES | LYNCHBURG SERIES LEON SERIES I RAINS SERIES ! PAMLICO-PONZER SERIES t^^ TORHUNTA SERIES i^FORESTON-WRIGHTSBORO SERIES 18 Figure 10. Soil types in North Carolina's Qreen Swamp pocosin (based on Kologiski 1977X as redrawn by the U.S. Department of the Interior, National Park Service 1979). i 19 example, Barnes (1981) cited studies which indicate that although nitrogen levels are high, nore than 70% of the total nitrogen may be in forms (such as fulvic or humic acids) unavailable to plants. Likewise, phosphorus, generally present in the or ganic form in pocosin soils, may have low availability. Even copper may be defi cient (Barnes 1981). Jnus, on these raised organic soils precipitation is the chief external source of plant nutrients (Richardson 1982). These wetlands may therefore be considered ombrotrophic3 be cause their waters are low in available nutrients. Carolina Bays Carolina bays apparently have no con sistent relationship to sub-surface strata types. Both Thorn (1970) and Gamble et al. (1977), however, have reported that bays seem to be restricted to sandy surface deposits. The characteristic sand rim is extensive enough in some instances to make sand quarrying a profitable venture. Other than this association with sandy substrates, no consistent relationship with particular geological formations or topography in the Coastal Plain is appar ent (Prouty 1952; Kaczorowski 1977). Schalles (1979) described the upper sediments of Thunder Bay (Savannah River Plant, South Carolina) as sandy loam. The presence of 11% "charcoal fragments" in the upper soil layer results from the burning of a peat-type material. Schalles further reported the presence of an "im pervious lens" of clay of unknown thick ness about 20 to 30 cm (8 to 12 inches) below the upper sediments. This near- surface hardpan, composed of clay or a sand-iron-humate complex, may result in a perched surface water table (Schalles 1979). In perhaps the earliest report of the substrata of Caroline Bays (Glenn 1895), a similar clay layer up to 8 m (25 ft) thick was described beneath the sand layer of the bays around Darlington, South Caro lina. Likewise, Bryant and McCracken In ombrotrophic ecosystems most or all of the water and nutrients come from precipi tation rather than from other sources such as ground water or stream flow. (1964) reported that surficial sands over laid micaceous clays beneath 15 Carolina bays that they examined in Scotland Coun ty, North Carolina. In contrast, Bliley and Pettry (1979) reported sand to be the dominant soil type of surface and substra ta soils of Carolina bays on the Eastern Shore of Virginia. Carolina bays undisturbed by burning or land management practices commonly have organic surface horizons overlying the sand and clay layers. Jones (1981) de scribed the substrate characteristics of five bays in South Carolina (Table 2), each of which had more than 30 cm (12 inches) of peat or highly organic soil. In addition, Ingram and Otte (1981b) indi cated that many of the larger bays in eastern North Carolina contain high qual ity peat up to 4.6 m (15 ft) thick which might be of commercial value. 2.3 HYDROLOGIC CHARACTERISTICS Pocosins: Uater Quality, Storage, andjtel ease Hydrological fluxes in pocosins are influenced by four input-output events. Precipitation is normally the exclusive input source of water and must ultimately be equalled by the three outputs. Of these, evapotranspiration and surface run off are the primary routes of departure of water from pocosin systems (Figure 11), and groundwater discharge (the sub-surface loss of stored water within the pocosin soils) is the least significant. Although the pocosin system is not hydrologically complex, the difficulties inherent in pre cise measurement of the output events in Table 2. Substrate characteristics of five Carolina bays In northeastern South Carolina (Jones 1981). Plant cormunlty type Pocosin Pocosin Pocosin Pocosin . Bay Forest 'Measured only 1n / Organic natter (« 7*.5 74.8 29.5 12.7 14.8 non-organic Organic surface depth (cm) > 100.0 >40.0 > 100.0 33.9 x 38-9 soils. Texture' Sand S1H Clay \ t * / - . ~ \ 63.8 17.6 18.6 73.3 9.1 17.6 such geographically extensive systems have resulted in few thorough studies of the hydrodynamics of natural pocosins. Because pocosins do not drain effec tively, groundwater discharge has the least effect on annual water flux. In the Albemarle-Pamlico peninsula, which com prises extensive pocosin habitat, ground- water discharges averaged less than 1.3 cm/year (Heath 1975), equivalent to 130,000 1/ha (1842 ft3/acre), and were as sumed by Brinson (1980) to be similar to the Croatan National Forest. Daniels etal. (1977) also reported that ground- water discharge was an insignificant pro portion of the water flux in the Hofnann Forest of North Carolina. Overland runoff in the Croatan was calculated by Brinson (1980) to be approximately 52 cm (5.2 x 106 1/ha or 73,680 ft3/acre) on the basis of a regional precipitation of 144 cm/year (57 inches/year) and cumulative evapo transpi ration estimated at 91 cm/year (9.1 x 106 1/ha or 128,940 ft3/acre). The relative contributions of transpira tion and evaporation in pocosins are unknown, although Richardson (1982) re ported that 60% to 70% of the water output may be through evapotranspiration during the summer and fall. Runoff becomes more important during the winter and spring because of increased precipitation. Quan titative information regarding the role of evapotranspiration would be a useful addi tion to our knowledge and understanding of the function of these systems. Under natural conditions the surface runoff from pocosins surrounding Albemarle and Pamlico Sounds and elsewhere is as sheet flow in response to precipitation events. The input into estuaries or other receiving bodies of water occurs several days after a rain and is spread over wide areas rather than being confined to a few point locations (Ash et al., in press; Copeland and Hodson 1982). The widespread installation of drain age ditches for land management has altered the hydrology of pocosins. Drain age ditches lower the water table to some extent, particularly in their immediate vicinity, and significantly affect surface runoff (Ash et al., in press). Runoff is greater than normal especially during periods of heavy precipitation and high water tables, conditions occurring most frequently in pocosin areas during winter and spring (Ash et al., in press). The increased 'runoff rates in drainage ditch 10 i 9 8 7 6 oo 0.6 0.7L • 1 Changes in monthly mean water levels due to temperature (after adjustments for precipitation) over the 5-year period. Figure 12. Water level changes In two Carolina bays (Savannah River Plant. SC) as inf uenced by prec pftation and temperature. obtaining this vital information will be critical to understanding the hydrodynam ics of these systems. The data for Ellenton Bay support a hypothesis that rainfall raises water lev el proportionally to the amount of rain fall, whereas water level is lowered as a consequence of evapotranspiration. Evapo rative water loss is, as expected, greater at high temperatures (and therefore, dur ing warm seasons) than at low ones. Evap oration rates would also be affected by wind and by relative humidity, which are unmeasured variables at Ellenton Bay. In addition, transpiration rates in bays with heavy aquatic vegetation might cause sig nificant changes in water level during the growing season. Soil permeability and the surrounding watershed would also influence water collection and retention. Although groundwater recharge (Schalles 1979) has been considered as a potentially important factor in changing water levels in certain Carolina bays, there is little evidence to support a general hypothesis that the water table is influential in affecting water level fluctuations in Carolina bays. This limited information for Ellenton Bay is the only published data available to address the influence of environmental factors on the fluctuations of water lev els in Carolina bays. Long-term measure ments of such basic phenomena as local precipitation, evapotranspiration rates, groundwater levels, soil permeability, and local topography and their relations to changes in water level would be valuable. Dynamic habitats such as these can support only species with effective adaptations to cope with such fluctuating conditions. CHAPTER 3 BIOLOGICAL FEATURES 3.1 PLANT COMMUNITY COMPOSITION Pocosins Few studies provide more than a brief or superficial description of pocosin veg etation and fauna. The apparent lack of interest in pocosin ecosystems is no doubt due in part to their perceived low commer cial value and the difficulty involved in working in these habitats. As Christensen et al. (1981) stated, "...their limited economic value, impenetrable character, and alleged dense populations of venomous or otherwise malevolent critters have suc cessfully repelled ecologists for the past fifty years." In addition, disagreement regarding what actually qualifies as a pocosin and how the boundaries of »these ecosystems are defined, coupled with the fact that many of their characteristic species occur in other wooded wetland sys tems, has hindered or confused ecological description. There is limited information on species composition, productivity, and successional development of pocosins and their relationships with other Coastal Plain ecosystems. Developmental history of pocosin com munities.Developmental historyofthe vegetation of the coastal area of the Mid- Atlantic States has been described by sev eral investigators (e.g., Buell 1945; Frey 1951b, 1953, 1955b; Whitehead 1963, 1964, 1981), based primarily upon examination of pollen profiles obtained from soi cores. Although there exists no complete descrip tion of the vegetational history specific to pocosins, the development of this flora can be inferred from these studies and a more recent one by Otte (1981), which used samples taken from Carolina bay and poco sin sediments. Frey (1951b, 1953, 1955b) sampled cores from a series of Carolina bays and_ described a pine-spruce forest that exist ed during the period of Wisconsin glacia tion. Whitehead (1964) reported a greater dominance of pine during this period with some boreal species occurring primarily on more mesic sites. Late-glacial (about 10,000 to 15,000 B.P.) forests dominated by oak, hickory, hemlock, and beech re placed the pine and spruce forests as the glaciers retreated and the climate warmed. More recent post-glacial changes until the present time have included the gradual decrease in oak-hickory forests, develop ment of southern pine forests, and an increase in cypress-gum dominated vegeta tion in wetter sites. The development of shrub bog vegeta tion in pocosin areas is difficult to trace. Ingram and Otte (1981b) reported that peat began to build up about 8,000 to 10,000 years ago in the dissected depres sions of the outer Coastal Plain, often building up to form broad, dome-shaped surfaces. As noted earlier, carbon-14 dating of sediments from the' Great Dismal Swamp and the Chesapeake Bay indicates that deposition of organic clays and peat accumulation in these areas began between 10,340 ± 130 and 8,135 ± 160 years B.P. (Daniel 1981). These sediments contained late-glacial pine and spruce pollen or oak-hickory pollen. However, older peat samples (15,280 ± 200 years B.P.) contain ing abundant spruce pollen from the mid- glacial period have been recovered from one area of the Chesapeake Bay (Harrison et al. 1965). Furthermore, Richardson (1982) noted that radiocarbon dating of other areas of the Dismal Swamp indicates that much of the peat may be less than 3,500 years old. He interpreted these various ages along with information from pollen profiles, the presence of charcoal in the peat, and historic climatic changes to indicate a dynamic developmental his tory of pocosin substrates that includes 26 27 r fluctuations in oxidation and in the rates of peat accumulation and the occurrence of extensive fires. Peat accumulation, fire, nutrient levels, hydroperiod, human use, and other ecological and environmental factors have contributed to the development of the present pocosin ecosystems. Opinions of investigators differ about which of these factors are primarily responsible for the establishment and maintenance of pocosin communities. From an extensive survey of North Carolina pocosins, Otte (1981) pro posed a pattern of pocosin community de velopment based upon peat profile analy sis. He believes that most pocosins began as marsh systems dominated by aquatic mac- rophytes and grasses, succeeded to cypress and Atlantic white cedar (Chamaecyparis thyoides) forests, and finaîTyto shrub bog pocosins. He reported that most of the pocosin ecosystems appear to have de veloped the evergreen shrub-dominated com munity relatively recently (in the last few thousand years) in terms of the over all age of the wetland. Only the Croatan National Forest pocosins (Figure 13) and southern Dare County pocosins appear to be relatively old (Otte 1981). The substrates of the Croatan pocosin consist of 0.6 to 1.2 m (2 to 4 ft) of white cedar peat overlain with pocosin peat 1.2 to 1.5 m (4 to 5 ft) thick (Figure 13). Otte noted, however, that within the pocosin peat are patches of marsh peat, indicating local wetter conditions and perhaps the sites of peat burns that left shallow open pools where marsh vegetation developed. CHOATAN NATIONAL FOREST PROFILES THROUGH MAJOR PEAT BOG E3 HUMIC POCOSIN PEAT g FIBROUS WHITE CEDAR PEAT E3 PEATY SAND AND SAND B. W. Wells (Wells and Whitford 1976) described the poorly drained interstream areas of the Coastal Plain as historically being covered by broadleaf swamp forests dominated by black gum (Nyssa). sweet gum (Li'qui'dambar), and naple (Acer). Appar ently, they dried out frequently enough for tree seeds to germinate and seedlings to become established. Wells further re ported that fires by Indians and later by white settlers began to change these hard wood swamp forests into shrub bogs, or po cosins. He believed that the frequency of fire played a major role in the develop ment of pocosin vegetation and in the relationship between pocosins and other plant associations. For example, if the hardwood swamps were burned as frequently as every decade, the deciduous forest dis appeared or became dominated by pond pine (Pinus serotina) and southern cane (Arund- inaria gigantea). When burning was even frequent, fire resistant shrubs or shrubby trees such as sweet bay, (Magnolia yirgiana), red bay (Persea borbom'a), and' leucothoe (Leucothoe spp.) became domi nant, along with greenbrier. All of these species have the' ability to sprout from stumps or roots following burning, and they form virtually impenetrable jungles. If these pocosin communities were burned annually, the shrubs practically disap peared and were replaced by grasses, sedges, and herbs. If fire frequency was reduced, these grassy habitats reverted to shrubby pocosin communities. Wells and Whitford (1976) noted that if fire was eliminated altogether, savannas and shrub bogs returned to swamp forests, although 0.5 1 MILE -OFT -2 -4 Figure 13. Representative cross section through the Croatan National Forest pocosin sub strates (from Otte 1981). 28 Otte (1981) suggested that pocosin on deep peat will remain unless the peat is se verely burned. Fire and other disturbances have played a major role in development of the pocosin vegetation of the Green Swamp in North Carolina (Kologiski 1977). Kologiski noted early reports that much of the area was once dominated by Atlantic white cedar swamps and extensive cypress-gum swamps. Logs buried under several meters of peat are all that remain of the white cedar forests believed to have been destroyed by fire. Recent surveys of the area revealed remnants of the cypress-gum forests that were heavily logged in the 1800's and early 1900's (Kologiski 1977). Characterization of pocosin vegeta tion.Early accounts of pocosin vegeta- tion, such as those of Kerr (1875) and Harper (1907), were primarily lists of plant species and brief descriptions. Wells (1928) included a description of pocosin or bay vegetation in his charac terization of coastal plain plant commu nities and later provided some quantita tive data on the pocosin vegetation of Holly Shelter Wildlife Management Area (Wells 1946). Other early studies (e.g., Buell 1939a, b, 1945, 1946a, b; Penfound 1952) added limited information to the characterization of pocosin habitats. More recently, Woodwell (1956, 1958), Kologiski (1977), Christensen (1979), Christensen et al. (1981), Otte (1981), and Jones (1981) have concentrated exten sive sampling and analysis on this vege tation type. Most of the information regarding pocosin vegetation that is presented in this community profile is taken from the work of these researchers. Throughout this report, the use of plant names follows Radford et al. (1968). The typical pocosin or shrub bog vegetation is characterized by a shrub understory with scattered emergent trees (commonly pond pine). The height of the shrub cover usually ranges from 0.5 m (1.5 ft) to 4 m (13 ft). If the woody vegetation (tree component) is less than 6 m (19.5 ft) high (scrub-shrub), it is generally called short pocosin; if greater than 6 m high (forested), it is considered to be tall pocosin (Figure 14), although these size designations differ among the investigators (e.g., Otte 1981). Pocosin communities are commonly termed evergreen shrub bogs; however, many of the charac teristic pocosin species are partially deciduous (e.g., titi [Cyril!a. racemi- flora]) or wholly deciduoTfs (e.g., sweet fepperbush [Clethra alm'folia], fetterbush Lyoni a lucidaj, zenobia [Zenobia pulver- ulenta], blueberry CVaccinium spp.], huckleberry [Gaylussacia frondosa]). Some pocosin species, such as titi, may tend to be deciduous in more northern pocosin com munities and to retain their leaves in more southern habitats. Deciduous species may even dominate in some of the shrub bogs (as may tree species rather than shrubs). In addition to pond pine, other pine species, especially loblolly (£. taeda) and longleaf (P_. palustris), may occur in the better drained areas. In some pocosins, Atlantic white cedar and also cypress (Taxodium spp.) and gum (Nyssa spp.) may be found. Hardwood trees, especially sweet bay, loblloly bay (Gprdonia lasianthus). and red maple (Acer rubrum) are common in many pocosin areas, particularly on less peaty sites. One of the few comprehensive descrip tions of pocosin and related wetland com munity types has been provided by Kologi ski (1977) in an extensive survey of the Green Swamp in Brunswick County, North Carolina. He described the Green Swamp vegetation as a complex continuum of popu lations arranged according to soil, mois ture, and disturbance factors (Figure 15). By using several ordination techniques designed to demonstrate relationships among samples, Kologiski examined data from 220 stands, including the evergreen shrub pocosin community as well as related communities. Although he reported much overlap between vegetation types, he iden tified discrete vegetational units, which he organized into a hierarchial system of classification adapted from the natural areas classification system of Radford (1977). Using this classification, Kolo giski (1977) described five vegetational systems in the Green Swamp, based upon species composition and growth form, which he arranged in a successional sequence from a marsh grass pioneer community to a lowland forest. Within these systems, he separated nine community classes according to the dominant canopy, shrub and herba ceous understory species. 29 i I [EXPERIMENTAL TREE PLANTING CONIFER-HARDWOOD POCOSIN \%& BAY FOREST fa i i ' (^ATLANTIC WHITE CEDAR FOREST | PINE SAVANNA PINE-ERICALEAN POCOSIN I MIXED WHITE CEDAR AND BAY FOREST Low or short pocosin habitat, near the center of Angola Bay. Pender County, NC \:< v N t'- .»• High or tall pocosin habitat. Holly Shelter Wildlife Management Area. Pender County, NC (with evidence of a fire) * Figure 14. Comparison of two types of pocosin habitats. (Photographs by Charles B. McDonald, East Carolina University, Greenville, NCJ 30 Figure 15. Vegetation types hi Green ^wamp, NC (based on Kologlski 1977. as redrawn by the Department of the Interior. National Park Service 1979). 31 Only one of the five vegetational systems in the Green Swamp (the "wet scrub-shrub system," as described by Kologiski) can be considered true "poco- sin" according to the definition followed in this community profile. The wet scrub- shrub system contains the pocosin commun ity cover class that includes those com munities dominated by evergreen shrub and tree species characteristic of pocosins (Table 5a). Other vegetation systems of the Green Swamp (marsh grass; low wood land, which contains the pine savanna cover class; lowland gymnosperm including cedar bogs; and lowland angiosperm con taining the evergreen bay and deciduous forests) represent peripheral and/or suc- cessionally related habitats (Table 5b). The successional relationships among these vegetation types are still a matter of speculation, and the suitability of Kologiski's classification system outside of the Green Swamp has yet to be deter mined (Christensen et al. 1981). Further more, portions of the Green Swamp have been affected by land management practices since Kologiski's study, so that his map of the vegetation (Figure 15) is now out of date. Nevertheless, because it is one of the most complete vegetational studies, Kologiski's description and community classification are summarized here to pro vide an example of an interstream pocosin and related plant communities in North Carolina. Kologiski described the wet scrub- shrub system (pocosin community cover class) as the dominant vegetation of the Green Swamp. According to Cowardin et al. (1979), who developed the system of class ification of wetlands and deepwater habi tats for the U. S. Fish and Wildlife Ser vice, the class scrub-shrub includes areas of woody vegetation generally less than 6 m (19.5 ft) tall. Vegetation in this class includes shrubs and trees that are small or stunted because of environmental conditions (Figure 16). The pocosin community of the Green Swamp is a complex unit containing two community classes (pine-ericalean and conifer-hardwood), each with two or more community types (Table 5a and Figure 17a). Distinct boundaries between community types are often difficult to determine. Kologiski indicated, however, several com munities that differ primarily because of the length of hydroperiod and time since the last fire. The Pine-Erioalean (pine and heath shrub) Community Class generally develops on deep to intermediate organic soils that are exposed to long hydroperiods and fre quent fire (Table 5a). There are three community types: (1) Pond pine (Pinus serotina) canopy with titi (Cyrilla racemiflora) and zeno- bia (Zenobia pulverulenta)shrubs. This vegetation usually occurs on deep organic soils with water at or near the surface throughout the year. Although the commun ity is dominated by shrubs, widely spaced pond pines may form an open canopy. Growth of the pines is slow, and they frequently exhibit a gnarled, twisted, and stunted -form. Their average height may range from 2 m (6 ft) to 12 m (39 ft). Dominant shrubs are titi and zenobia, both of which may exhibit greater than 75% cover. Fetterbush is abundant, but it usually has a lower cover value. The height of the shrub layer ranges between 0.5 m (1.5 ft) and 2 m (6.5 ft). Other shrubs as well as herbaceous species such as broomsedge (Andropogon virgim'cus) occur, and greenbrier drapes over and twines through the shrubs. (2) Pond pine (Pinus serotina) and loblolly bay (Gordonia lasianthus) canopy with fetterbusü(Lyonia lucida) shrubs. This community occurs on slightly drier elevated (approximately 0.5 m or 1.5 ft) areas within the previous community type. These elevated areas resemble islands up to or greater than 100 m2 (more than 1000 ft2) and are most likely formed by an accumulation of sphagnum and litter around the stumps or bases of pond pine trees. In addition to the pond pine, loblolly bay, and fetterbush, shrubs that may form an almost impenetrable thicketxinclude sweet gall berry (Ilex coriacea). bitter gall berry (K.glabra), and titi. (3) Pond pine (Pinus serotina) canopy with titi (Cyril la racemi flora) and fet terbush Sx- utaa&a Cr * till (Cyrilla racemiflora) Ct - Atlantic white cedar (Chomaecyparis thyoides) Gl = loblolly bay (Cardonia lasianthus) Ic sweet gallberry (Ilex coriacea) LI - fetterbush (Lyonio lucida) Vc = highbush blueberry (Vaccinium corymbosum) Me Pp lg Gf As Aq Ic Mh Ns Mv Cr Ar H I O PIHUSIArislida b. P NE SAVANNA Cr titi (Cyrilla racemiflora) Ct - Atlantic white cedar (Chamaecyparis Ihyoides) Cf - huckleberry (Caylussacia frondosa) Gl = loblolly bay (Gordonia lasianthus) Ic sweet gollberry (Ilex coriacea) lern- dahoon (Ilex cassine var. myrlifolia) Ig = bitter gallberry (Ilex glabra} LI fetterbush (Lyonia lucida) Mv sweet bay (Magnolia virginiana) Pb - red bay (Persea borboniol Ps = pond pine (Pinus serotina} Ta pond cypress (Taxodiutn oscendens) Va = blueberry (Vaccinium atrococcuml Wv = chn'n fern (Hoodwardia virgintca) CHAMAECYPARIS/tyonia-/fcx c. CEDAR BOG Ar red maple (Acer rub rum) Cr titi (Cyrilla racemiflora) Gf - huckleberry (Gaylussacia frondoso) Ic = sweet gallberry (Ilex coriocca) Ig bitter gallberry (Ilex globra) LI - fetterbush (Lyonla lucida) Mv = sweet bay (Magnolia virginiana) Ns - black gum (Hyssa sylvatica var. biflora) Pb - red bay (Persea borbonia) Ps - pond pine (Pinus serotina) Ta pond cypress (Taxodium ascendens) Vc highbush blueberry (Vaccinium corymbosum) Wv chain-fern (Woadwardia virginico) Wv Ct GIVa Ps lern Mv Cr GfTalg Pb LI PERSEA PINUS-GORDONIA CHAMAECYPARIS-MAGNOLIA/Cyr/Ho-Lyon/o Ilex d. EVERGREEN BAY FOREST TAXODIUM-ACER-NYSSA/CyriHo Lyonia-llex e. DECIDUOUS FOREST Figure 17. Major woody plant community cover classes In Green Swamp, NO (based on Kolo- giski 1977). 38 1979; Chrlstensen et al. 1981). Using reciprocal averaging ordination to evalu ate Woodwell's (1956) classification, Christensen determined that although the three categories do not fall into discrete groups, they do tend to dominate specific parts of the ordination. Christensen et al. (1981) further attempted to relate species dominance to known environmental features including geologic type of poco- sin, peat depth, water table depth, sub strate type, and age of the community since burning. They also concluded that zenobia tends to dominate in recently dis turbed areas in which the shrub community has relatively low productivity and higher diversity, whereas titi and fetterbush occur in more productive bogs. They were unable, however, to add much additional insight about the causes of variation among pocosins. Jones (1981) described 52 lowland forests in the northern Coastal Plain of South Carolina, including 13 pocosins. Densities and basal areas of trees from a typical pocosin site (Table 6) are low in comparison with other wetland forests. Pond pine dominated these communities, with loblolly bay and pond cypress being next in density. The chief shrub was fetterbush, with greenbrier often the second most important species (Table 7). The gall berries, blueberries, wax myrtle (Myrlea cerifera). and zenobia also con- tributed significantly to the shrub cover which commonly reached or exceeded 100% ground coverage and was the highest cover age measured by Jones (1981) in any of the wetland communities. Associated communities. In addition to the pocosin community, Kologiski (1977) described four peripheral and perhaps re lated community cover classes in the Green Swamp: sedge bog (representing the marsh grass system), pine savanna (low wood land system), cedar bog (lowland gymno- sperm forest), and bay forest (lowland angiosperm forest; Table 5b). Each of these vegetation systems will be briefly characterized. Table 6. Density and basal area of trees in a pocosin in South Carol'na (from Jones 1981). Species Density (stems/ha) Basal area (n /ha) jMnus serotina (pond pine) 538 Gordonia lasianthus (loblolly bay) 125 Taxodium ascendens (pond cypress) 103 Nyssa sylvatica var. biflora (black gum) 77 Persea borbonia (red bay) 18 Magnolia vlrqini'ana (sweet bay) 7 Myrica cerifera (wax myrtle) 11 Acer rubrum (red maple) 7 _Ilex cassine (dahoon) 4 Total 890 8.70 0.88 1.34 0.42 0.06 0.02 0.04 0.07 0.01 11.54 O 39 Table 7. Density and transect cover of shrubs, woody vines, and tree seedlings In a South Caro lina pocosln (from Jones 1981). Species Lyorn'a lucida (fetterbush) Smilax spp. (greenbrier) II jx glabra (bitter gall berry) Vaccinium spp. (blueberry) Persea borbonia (red bay) Magnolia virginiana (sweet bay) Myrica cerifera (wax myrtle) Gaylussacia spp. (huckleberry) Ilex coriacea (sweet gall berry) Sorbus arbutlfolia (chokeberry) Gordom'a lasianthus (loblolly bay) Pi nus seroti'na (pond pine) Rhus radicans (poison ivy) Ilex cassine (dahoon) Nyssa sylvatica var. biflora , (black gum) Lyoni'a ligustrina (fetterbush) Total Density (stems/ha) 10144 3480 2016 1401 1231 594 446 446 , 403 191 255 255 85 64 42 21 21074 Transect" cover (cm/m) 45.12 11.08 6.67 9.12 7.79 6.25 2.92 0.58 4.50 3.00 2.46 1.12 0.29 1.42 1.42 , <0.01 103.74 Sedge bog. This is the most hydric community type included in Kologiski's classification of the Green Swamp vegeta tion. The community type represented in this area is dominated by the sedge Carex walteriana and occupies shallow depres sions in organic soils that contain stand ing water during most of the year. It is therefore "wetter" than the pocosin shrub community. Such depressions are frequently formed by fires burning into the peat dur ing periods of low water. Sedges, especially £. walteriana. are usually the first plants to invade the wetter sections of a pocosin after fire or other major disturbances. These sedges often form dense stands with almost 100% cover, although other herbaceous as well as shrub species may be found in this com munity type. After the depressions begin to fill, zenobia and then titi invade, and the community becomes shrub dominated. Pine savanna. This vegetative commu nity cover class has a widely spaced can opy of pines with an understory ranging from predominantly grasses to a mixture of grasses, shrubs, and ferns (Figure 17b). According to Kologiski (1977), the savanna communities of the Green Swamp are charac terized by mineral soils, short hydroper- iods, and frequent fires. The soils are usually well drained; however, because of a perched water table, water is within a meter (3 ft) of the surface most of the year. Although the moisture gradient from pocosin to pine savanna is generally grad ual , the vegetational ecotone may be sharp (Christensen et al. 1981), largely because of the higher frequency of fire in savanna areas. Three types of pine savanna occur in the Green Swamp: (1) Longleaf pine (Pinus palustris) canopy with wire grass (Aristida strieta) and sedges (Rhynchospora spp.) as ground cover. This savanna type occurs on min eral soils that are usually well drained, although in the more hydric areas, water roay be at or near the surface during the rainy season. Pond pine is the chief can opy species and wire grass dominates the ground cover. Several species of sedge roay share the dominance with wire grass. (2) Pond pine (Pinus serotina) canopy with wire grass (Aristida stricta) and sedges (Rhynchospora spp.) as ground cover. In the more hydric savannas, pond pine becomes the canopy dominant. A simi lar community with switchcane (Arundinaria gigantea) as the ground cover develops in response to frequent burning, although it is not common in the Green Swamp. (3) Longleaf pine (Pinus palustris) canopy with huckleberry (Gay!us.sacia fron- dosa) and bitter gal 1 berry (IIex~g1 abra) in the shrub layer. This community, which occurs on better drained mineral soils than (1) and (2), has a more closed pine canopy and a low shrub layer. Huckleberry and bitter gall berry are the most impor tant species of the latter. Other shrubs include sweet gall berry, leucothoe (Leuco- thoe axillaris), lyonia (Lyom'a man'ana). and several spp.). rica cerifera wax myrtle ( __ species of "blueberry (Vaccinium As in the Green Swamp, savannas fre quently occur in association with poco- sins. They develop .on ridges or high areas that are elevated sufficiently to prevent peat accumulation (such as the rims of Carolina bays). A difference in elevation of only a few centimeters may be sufficient to separate pine savannas from the surrounding pocosins (Ash et al., in press). Fire maintains the savannas by limit ing invasion by shrubs. Without frequent fire, the savannas will become pine- dominated forests with an evergreen shrub understory. Succession will proceed more rapidly in areas with greater surface drainage. The extent of savannas is usu ally limited. In the Green Swamp, only 93 ha (230 acres) of the 5,609 ha (13,860 acre) nature preserve is savanna (Mclver 1981). Cedav bog. Atlantic white cedar for ests are often associated with pocosins. White cedar may become established in open habitats on organic soil, especially after fires when the water table is high (Figure 17c). It commonly forms dense even-aged stands in areas of highly acidic peat and stagnant water (Dean 1969). Seedlings may become established under shrubs and parent trees, but since they require sunlight, growth is decreased once the canopy begins to close. Kologiski identified two commu nity types: 40 41 III III LUI (1) Atlantic white cedar (Chamaqcypa- ris thyoides) canopy with fetterbush (Lyonia lucida) and sweet gal 1 berry (Ilex coriacea) in the shrub layer. According to Kologiski, in the Green Swamp this type is best developed over deep organic soils that are usually wet throughout the year, although standing water is unusual. The ground is covered with a mat of white cedar leaves and other organic debris, and fallen white cedar logs and stumps are abundant. These stands commonly form a dense canopy. The understory of smaller deformed white cedars and the distinct shrub layer form an almost impenetrable thicket. The shrub layer ranges in height from 1m (3 ft) to several and is domi nated by fetterbush and sweet gall berry. Greenbrier and yellow jessamine (Gelsemium sempervirens) are important vines. The herb layer is almost absent. (2) Atlantic white cedar (Chamaecypa- ris thyoides) and loblolly pTnë(Pinus taeda) canopy with fetterbush (Lyonia lu cida) and sweet gal1 berry (Ilex coriacea) in the shrub layer. Loblolly pine is prominent in the canopy of this community, although it is never actually a codominant with white cedar. While very similar to the previous community (1), this type is usually found on shallow organic soils underlain by sandy clay loam. Successional studies (Buell and Cain 1943) indicated that burned white cedar forests typically become evergreen bay forests. These cedars have no resistance to fire, and burning usually kills all the trees in a stand. Successful regeneration following fire depends upon a store of viable seeds in the upper layers of peat (Ash et al., in press). If fires occur during extremely dry periods and burn the peat, the seed source may be eliminated. In such areas where the cedars have been destroyed completely, pocosin and bay spe cies will become dominant. Bay forest. The bay forests described by Kologiski (1977) are characterized by shallow to deep organic soils, intermedi ate to long hydroperiods, and canopies dominated by combinations of red maple, Atlantic white cedar, titi, loblolly bay, sweet bay, black gum, red bay, pond pine and pond cypress. One or all of the bay trees (sweet bay, red bay, loblolly bay) are present in both the evergreen and deciduous phases of this forest (Figures 17d and e). Community types are: (1) Sweet bay (Magnolia virginiana), red bay (Persea borbonja), and loblolly bay (Ggrdonia lasianthus) canopy with titi (CyriTTäTäcerniflora) and fetterbush (Lyonia lucida) shrubs and Virginia chain- fern(Woodwardia virginica) as ground cover (FigureTTd). This lowland ever green bay forest develops on deep organic soils that are poorly drained and have long hydroperiods. Standing water is typ ically present during part of the year and the water table is always near the sur face. The three bay species (loblolly bay, red bay, and sweet bay) dominate the canopy, which may be shared by Atlantic white cedar, titi, dahoon, pond pine, and deciduous species such as red maple, black gum, and pond cypress. The canopy is 3 to^lO m (10 to 32 ft) high, and the shrub stratum usually blends into it. Major shrubs include titi, huckleberry, fetter- bush, sweet gall berry, bitter gall berry, smooth winterberry (Ilex laevigata). dahoon, bayberry (Myrica heterophylla), swamp azalea (Rhododendron viscosum), highbush blueberry (Vaccinium atrococcum, V^. corymbosum), and zenobia. Greenbrier is abundant and Virginia chain-fern is the most important herb. Kologiski (1977) believes that in the Green Swamp, this evergreen bay forest succeeds the titi-zenobia pocosin and probably the white cedar forest (following logging). He reported that the most com mon seedlings in many of the evergreen bay forest stands were red bay, indicating eventual domination by this species. A major fire may cause this forest to revert to a pocosin or white cedar forest (with a shallow burn), a sedge bog (with a deep burn and high water table), or a deciduous bay forest (with a deep burn and temporar ily low water table). (2) Red maple (Acer rubrum). black gum (Nyssa sylvatica var. biflora), and pond cypress (Taxodium ascendens ) canopy with titi (Cyril!a racemi flora) and fet terbush (Lyonia lucida)shrubs and Vir ginia chain-fern (Woodwardia virginica) as ground cover (Figure 17e). This lowland deciduous forest xdevelops on shallow organic soils with intermediate to long hydroperiods and an absence of major fires. Standing water may be present over the hummocky substrate for at least sev eral months of the year. Most of the same species are present in both the evergreen bay and the deciduous forests and the two types blend into one another. In the Green Swamp, the dominant canopy species in the deciduous forests are red maple, black gum, and pond cypress, with frequent occurrence of red bay, pond pine, and lob lolly pine. The height of the canopy ranges between 10 and 25 m (32 to 82 ft). Although the dense understory contains a number of shrubs, titi and fetterbush are usually the dominants. Several species of greenbrier are abundant, and Virginia chain-fern is the dominant herb. Kologiski reported that the presence of canopy seedlings and saplings indicates that the lowland deciduous forest in the Green Swamp is perpetuating itself. He suggested that under existing environmen tal conditions, it can be assumed that a mature lowland deciduous forest is rela tively stable. Over time, however, espe cially if drainage occurs, the area will become dominated by more upland species. Extensive disturbance such as fire or cut ting would probably revert this type to conifer-hardwood pocosin. Only the Pine-Ericalean and the Conifer-Hardwood Community Classes as described by Kologiski (1977) in the Green Swamp typify the shrub bog definition of pocosin used in this community profile. However, the other marsh, swamp, and for est associations represent related commun ity types. A comparison of various eco logical attributes of several of these communities, based upon the work of Jones (1981), is given in Table 8. Kologiski's ordination of 220 stands sampled in ever green shrub and related communities pro vides a rough reflection of the soil mois ture gradient. Sites with relatively dry mineral soils, dominated by pond pine, wire grass, sedges, huckleberry, and bit ter gall berry, are located at one extreme of the arrangement; and sites with highly organic wet soils, dominated by Carex SPP., zenobia, and titi, are located at the other extreme. Proposed relationships between the vegetation units of the Green Swamp and environmental factors are de picted in Figure 18. General patterns of community classes as influenced by soil/ hydroperiod and fire frequency can be interpreted from this scheme. Kologiski (1977) noted that hydroperiod is probably the most important of the factors since it controls the establishment and the growth of the plants and also, to some extent, the severity of fire and its effect on the vegetation. Around lakes or ponds, or in areas with distinct drainage patterns, pocosin communities may intergrade with swamp for est vegetation. Only at their margins, however, do these upraised shrub bogs ap pear to be succeeding to swamps (Christen- sen et al. 1981). Characteristic bay for est species (cypress, black gum, and white cedar) are usually absent except in these wetter areas. At the other end of the moisture gradient, shrub bog vegetation becomes pine flatwoods and savannas domi nated primarily by loblolly and longleaf pine with a diverse herbaceous understory. Successional relationships. As noted earlier in the discussion of the histori cal development of pocosin communities, evidence from peat profiles suggests that most of these shrub bog communities devel oped in the last few thousand years. Otte (1981) reported that the pocosin peats were typically underlain by organic re mains of marsh vegetation with Atlantic white cedar and cypress common in deeper layers of the peat, especially near ancient stream channels. Paludification, or expansion of these pocosin wetlands, proceeded as a result of peat accumulation and an associated gradual rise in the water table. There are basically two theories of pocosin succession which are in sharp con trast to each other. The first, proposed by Wells (1928), assumes that the fre quency and intensity of fire controls suc- cessional development. According to this theory, short pocosin is a pioneer stage leading to bay forest as a climax. The Successional process could therefore be completed within a few hundred years if disturbance is prevented. The second theory of pocosin succession (Otte 1981K assumes that nutrient levels are the con trolling factor. According to this hypo thesis, the successional sequence is marsh —»-swamp forest——+- bay forest —^ tall pocosin—»-short pocosin. This sequence 42 43 of events may require thousands of years for the modification of the substrate and development of an upraised bog with a perched water table to occur. Relationships of pocosin communities to other wetland plant associations have been mentioned throughout the above de scriptions. A delicate balance of envir onmental factors controls these community relationships. The three major factors directly controlling the distribution of vegetation within the pocosin ecosystem are thickness of the peat, length of the hydroperiod, and frequency and severity of fire (Otte 1981). These factors are interrelated in such a complex way that an alteration of one will affect the others. Therefore, it is impossible to separate their individual effects and relative sig nificance to the maintenance and integrity of pocosins. In general, the evergreen shrub bog pocosin communities occur on the deepest peat substrates. Following an extensive survey of pocosins in North Carolina, Otte (1981) proposed that peat thickness, along with hydroperiod and associated nutrient availability, is primarily responsible for the maintenance of short versus tall Table 8. Comparison of various geological and ecological parameters of pocosins, pine savan - nas. and bay forests in South Carolina (from Jones 1981). Values represent means for all sites of each community sampled unless indicated otherwise. Parameters No. of tree species No. of shrub species Total no. of woody species Tree basal area (m2/ha) Shrub cover (cm/m) Depth of organic substrate (cm) % organic matter Mineral soil: % sand % silt % clay No. of sites sampled Pine savanna 7.7 8.2 15.9 14.5 22.6 18.0 6.5 81.2 12.8 6.0 6 Pocosin 7.9 11.3 19.2 14.3 ' s 117.2 >40.0 39. 2a 4. 80. 6a 10.4a 9.0a 13 ' 1 Bay forest 11.0 13.8 i 24.8 33.3 88.9 >40.0 53.8 59.3 31.2 9.5 6 X Measured only in non-organic s (6 of the 13 sites). z LU O 01 Ö LU ce H Z LU 03 PINE SAVANNA (GRASS UNDERSTORY) / MARSH BOG \ CONIFER-HARDWOOD PINE-ERICALEAN POCOSIN / \ DECIDUOUS BAY FOREST PINE SAVANNA (SHRUB UNDERSTORY) EVERGREEN BAY FOREST WHITE CEDAR FOREST MINERAL SHORT SOIL TYPE HYDROPERIOD ORGANIC LONG Figure 18. Proposed relationships among vegetation types, hydroperiod, and fire In pocosin habitats (based on Kologiski 1977). 45 pocosin vegetation. According_to his theory, the[deeper portions of the thicker peat substrates are always saturated with water because the water table rarely falls lower than 1.2 to 1.5 m (4 to 5 ft) below the surface. Thus the root systems of plants never reach the underlying mineral sediments and are confined to the nutrient-poor organic substrates. Such plants are therefore stunted (short poco sin). In contrast, in the shallower peats the water table frequently drops below the peat-sediment interface (Otte 1981). In this circumstance, downward-growing roots can reach the more nutrient-rich mineral sediments and growth is enhanced (tall pocosin). Furthermore, Otte suggested that the short pocosin is the only pocosin community that can be considered a true climax community. Once peat substrates become so thick as to prevent plant roots from reaching the mineral soil, the short pocosin will remain a permanent feature unless a major fire or other disturbance removes part of the peat. Pocosins typically occur in areas of relatively long hydroperiod with the water table at or near the surface for 6 to 12 months of the year. This prolonged expo sure excludes many of the bottomland hard wood species characteristic of stream floodplains and bay forests as well as the upland pines and hardwoods. Water-tolerant swamp species such as cypress and black gum are occasionally associated with poco sin communities, but seldom become domi nant, presumably because of the low nutri ent availability of the organic pocosin substrates. An important difference be tween swamps and pocosins is in the direc tion of water flow (Otte 1981). Water flows into and through swamp ecosystems, replenishing nutrients. In pocosins, the major direction of water flow is out of the ecosystem via runoff; therefore, pre cipitation is the only major source of water and nutrients. The frequency and intensity of fire also influence the vegetation of pocosins. Obviously, since the highly organic soils burn readily when dry, the potential severity of a fire is related to the depth of the water table. Following shal low peat burns that destroy surface vege tation and the surface layer of peat, but do not damage the root and rhizome mat, the original pocosin type will recover quickly (Otte 1981). A burn that destroys the rhizome and root mat will be followed by recolonization from seeds and vegeta tion growing inward from the edge of the burned area. A fire that burns the peat to a depth such that the roots of the recolonizing vegetation can more readily come in con tact with the mineral soil may permit re covery of the same species as originally present. However, because of increased nutrient availability, their growth may be enhanced (Otte 1981). A severe burn over shallow peat that removes all the organic material above the mineral sediments will probably lead to the development of a non-pocosin community. On the other hand, a severe burn in deep peat substrates during a period of low water table could open up an area that would become a lake upon return of a high water table (Otte 1981). On better-drained soils, commonly associated with slight increases in eleva tion (such as the rims of Carolina bays), frequent occurrence of fire will tend to maintain a pine savanna community. Shrubs and upland hardwoods nay become estab lished if fire is prevented. Adaptations of pocosin vegetation. Many of the characteristic pocosin species are not only tolerant of fire, but