Natural Communities of Virginia

The Natural Communities of Virginia
Classification of Ecological Community Groups
SECOND APPROXIMATION (Version 2.2)

Since the First Approximation of The Natural Communities of Virginia (Fleming et al. 2001) was completed, a great deal of progress has been made in both the field inventory and formal classification of vegetation types. Several large regional data sets have been analyzed, making it possible to evaluate vegetation plot data in an appropriate geographic context and to produce relatively robust classifications for state-wide interpretation. Moreover, work on a customized database platform for permanently archiving digital compositional data, environmental data, and metadata has been completed. Projects involving vegetation classification and, in many cases, mapping are underway for most National Park units in Virginia.

Based on intensive work since 2001, the Second Approximation details further refinements in the classification of ecological community groups, and provides preliminary information about the association-level community types nested within each group. Moreover, photographic illustrations have been added. Conceptual and nomenclatural changes to the ecological community groups described in the First Approximation are listed in a separate section. It has proven impractical to continuously revise a printed document to incorporate ongoing, numerous changes to the classification. Therefore, it is our intention to update the web version of the classification at frequent intervals (version 2.1, version 2.2, etc.) until a consistent level of development is achieved to support a Third Approximation.

Why Classify Communities?
This web page is the second in a planned series of reports that will provide a comprehensive classification of natural communities in Virginia. The purpose of this effort is to construct a broad framework for understanding and defining such communities at several hierarchical levels. This goal is integral to the mission of the Virginia Department of Conservation and Recreation’s (DCR) Natural Heritage Program (VANHP), which is responsible by statutory authority for documenting, protecting, and managing “the habitats of rare, threatened, or endangered plant and animal species, rare or state-significant communities, and other natural features” (section 10.1: 209-217, Code of Virginia). As part of its work, the VANHP maintains database information on the status, distribution, and ecology of rare native species and all natural communities; protects and manages these resources through a system of natural area preserves; and provides information and technical advice to other agencies, organizations, and individuals. Within this context, community inventory and classification represent an important “coarse-filter” approach to biological conservation that ensures the protection of intact ecological systems containing diverse organisms. We believe that by identifying and protecting excellent examples of all natural community types in Virginia, the majority of our native plant and animal species, including many cryptic and poorly known ones, can be protected without redundant individual attention.

What is an Ecological Community?
An ecological community is an assemblage of co-existing, interacting species, considered together with the physical environment and associated ecological processes, that usually recurs on the landscape. This present treatment is restricted to natural communities, those which have experienced only minimal human alteration or have recovered from anthropogenic disturbance under mostly natural regimes of species interaction and disturbance. No portion of Virginia’s landscape, however, has altogether escaped modern human impacts – direct or indirect – and only a few small, isolated habitats support communities essentially unchanged from their condition before European settlement. Most of the communities treated here, while somewhat modified in composition or structure, are in mid- to late-successional stages of recovery from some form of human disturbance, such as agricultural conversion or logging. This document generally does not include early-successional communities that have experienced recent disturbance or highly modified habitats such as fields and plantation forests that are artificially maintained in an arrested stage of succession. Such communities, which cover extensive areas of Virginia, may nevertheless develop into natural systems given sufficient time and freedom from further anthropogenic disturbance. A few communities that are very rare in the state and now represented only by highly degraded examples are included because of their importance to the state’s biodiversity.

The specific assemblage of plants at a site is typically closely related to abiotic factors such as bedrock type, soil chemistry, slope, aspect, and elevation. Photo © Gary P. Fleming.

Classifications of natural communities can be based on various components (e.g., vegetation, fauna, landforms, hydrologic regime), used singly or in combination. Natural community classifications prepared by Natural Heritage ecologists for several other eastern states (e.g., North Carolina [Schafale and Weakley 1990], Pennsylvania [Fike 1999], Vermont [Thompson and Sorensen 2000], and New York [Edinger et al. 2002]) rely on a “multi-factor” approach that incorporates both biotic and abiotic elements in a community concept. Except in deepwater systems, however, plants have proven to be the most useful components for characterizing finer-scale communities on the landscape and providing a basis for comparing classifications covering different geographic areas. Although animals, especially invertebrates, can be very important in natural communities, they are often highly mobile, difficult to document, and found in many different ecological settings. Likewise, environmental conditions and processes encompass a spatially diverse array of factors from regional climate to site-specific moisture conditions that are impossible or excessively time- and labor-intensive to measure directly. On the other hand, the plants that together form the vegetation of a site are essentially immobile, are easy to measure in a variety of ways, and typically reflect (both individually and as an assemblage) specific site conditions. Despite the drawback that it can be dynamic over short time periods,

Vegetation is often chosen as the basis for a single-factor system for classifying terrestrial ecological systems because it generally integrates the ecological processes operating on a site or landscape more measurably than any other factor or set of factors .... Because patterns of vegetation and co-occurring plant species are easily measured, they have received far more attention than those of other components, such as fauna. Vegetation is a critical component of energy flow in ecosystems and provides habitat for many organisms in an ecological community. In addition, vegetation is often used to infer soil and climatic patterns. For these reasons, a classification ... based on vegetation can serve to describe many (though not all) facets of biological and ecological patterns across the landscape (Grossman et al. 1998).

The United States National Vegetation Classification (USNVC; Grossman et al. 1998), currently under development by NatureServe (formerly the Association for Biodiversity Information, ABI), The Nature Conservancy (TNC), and state Natural Heritage programs, in conjunction with the Vegetation Panel of the Ecological Society of America and the Federal Geographic Data Committee, promises to deliver a comprehensive “single-factor” approach to ecological communities based on a hierarchical classification of vegetation. Divisions within the upper levels of the USNVC hierarchy rely on physiognomic criteria such as vegetation structure and predominant leaf phenology. The two lowest divisions, the alliance and the association, are based on floristic criteria. The association, constituting the basic unit of inventory and biodiversity assessment, serves as a surrogate for ecological communities.

Methods and Data Sources
The USNVC proposes to determine the compositional variation, geographic range, and conservation status of each unique vegetation type in the country. In its current state of development, however, this massive project is far from complete; it contains vegetation types that are inconsistent in level of resolution and geographic scale and poorly supported by data. Hence, it does not readily meet the parochial needs of many individual Natural Heritage programs and their clientele. To satisfy the requirement for a system by which natural communities in Virginia can be identified, named, and ranked, VANHP ecologists have embarked on a multi-year program to develop a comprehensive, hierarchical classification of the state’s vegetation and natural communities which can be applied in the field by a range of users. Two central tenets of this endeavor are that (1) the basic systematic units of the classification are based on quantitative, plot-referenced data comprising full floristic composition of vegetation, and (2) these units are identified through rigorous analysis of data using multivariate techniques. This effort will in turn help to inform and refine the USNVC. The qualitative documentation of “significant” natural community occurrences in Virginia was initiated by The Nature Conservancy in 1983, and continued after the establishment of VANHP as a state agency in 1986. Such occurrences were defined as outstanding examples of common community types and all examples of rare community types. The required assessments of rarity and quality were based on subjective evaluations and experiences of staff ecologists, often in consultation with other knowledgeable individuals in Virginia and other states. During the early years of the program, emphasis was placed on inventory and identification of the rarer, more threatened communities in urgent need of protection (e.g., bogs, fens, and barrens). This, in turn, supported the identification and acquisition of a number of natural area preserves. To date, VANHP has compiled database information on more than 1,000 significant community occurrences.

Collecting vegetation plot data at The Nature Conservancy's Piney Grove Preserve in Sussex County. Photo: Gary Fleming.

Staff ecologists began collecting quantitative vegetation data in 1989, and this effort has greatly intensified since 1997. Our program is strongly committed to a “specimen-based” approach to community classification that depends on structural, floristic, and environmental plot data collected from uniform areas supporting homogeneous stands of vegetation. Although the analogy is imperfect, data from each plot could be considered a “specimen” that may be compared with other plot data in order to delimit vegetation “taxa,” in much the same way as a botanist analyzes preserved plant specimens to determine the taxonomic limits of species. As a result, procedures for observing, measuring, describing, and comparing vegetation are standardized to a specific scale, which facilitates more precise and objective characterization of vegetation types than is possible through purely qualitative field observation. In addition, a standard system and format for digitally storing data achieves satisfactory archiving of plot records and ensures the availability of data for further analysis or reanalysis in the future. Standard VANHP protocols for data collection and analysis are summarized in a separate introductory section.

To date, more than 3500 vegetation plots have been sampled by VANHP ecologists and contractors. A substantial proportion of these data has been collected as part of landscape-level classification projects for several discrete areas of the George Washington and Jefferson National Forests (Fleming and Moorhead 2000, Coulling and Rawinski 1999, Fleming and Moorhead 1996, Rawinski et al. 1996); the Grafton Ponds complex in York County (Rawinski 1997), three adjacent watersheds in southeastern Virginia (Fleming and Moorhead 1998); the Fort Belvoir military reservation (McCoy and Fleming 2000); the Pamunkey River Watershed in the northern Coastal Plain (Walton et al. 2001), the Bull Run Mountains (Fleming 2002b), Manassas National Battlefield Park (Fleming and Weber 2003), and several ongoing projects (e.g., Shenandoah National Park and the Potomac River Gorge).

During the past several years, classification initiatives using data collected across a broader geographic range have been undertaken, including:

a classification of forests and woodlands on calcareous substrates in western Virginia for the U.S. Forest Service (Fleming 1999);
classification of 956 plots collected throughout the mountain region of Virginia, for the U.S. Forest Service (Fleming and Coulling 2001).
classification of 415 plots of tidal wetland vegetation collected throughout eastern Virginia (Coulling 2002);
preliminary classification of 477 plots of Piedmont and inner Coastal Plain upland and alluvial floodplain vegetation (Fleming 2002a), and
preliminary classification of 1,300 plots of Piedmont and inner Coastal Plain vegetation in Virginia, Maryland, and the District of Columbia.

These projects have allowed us to re-examine data and results generated by the landscape-scale projects on a regional basis, and have fostered an increased understanding of the relationships between Virginia vegetation units and those proposed by other states or the USNVC.

Beginning in 1997, a provisional classification of natural communities was developed internally through a subjective analysis of plot and other field data. This approach synthesized a top-down natural-group (“multi-factor”) approach with a bottom-up classification of vegetation types consistent with the USNVC and based on rigorous analytical methods. The provisional classification has served as a "place-holder" for the program's significant community records, and has been refined through quantitative data analyses, field testing and, to a lesser extent, a review of Virginia vegetation literature. With sufficient plot data to support various within-state landscape and regional analyses, the First Approximation (Fleming et al. 2001) presented an updated version of the classification which formalized the overall hierarchical framework and conceptually defined and described 120 ecological community groups in Virginia. The Second Approximation continues the iterative refinement of a state-wide community classification.

Structure of the Virginia Natural Community Classification
The divisions of the Virginia classification hierarchy, from the top down, are:

System
Ecological Class
Ecological Community Group
Community Type

At present, our work is limited to all vegetated upland communities and all wetland communities that are vegetated with emergent, floating, or submerged aquatic vascular plants. The System level is based on gross hydrologic regime and presently includes five divisions: the Terrestrial System includes all upland (non-wetland) habitats, while the Palustrine System encompasses all non-tidal wetlands dominated by woody plants and herbaceous emergents. The Estuarine System includes emergent and floating / submergent tidal wetlands, extending to the upstream limits of tidal influence. The Riverine System and the Marine System are each represented by a single ecological group that supports vascular plants. This system-level treatment generally follows Cowardin et al. (1979), except that freshwater tidal wetlands are included in the Estuarine System, and some communities that would be placed in the Lacustrine System of Cowardin et al. (1979) are included in the Palustrine System. Classifications of deepwater Lacustrine, Riverine, Estuarine, and Marine System communities that lack vascular plants, as well as of Subterranean System (cave) communities, are currently under study or development by other groups of specialists.

Except for a few nomenclatural and conceptual modifications, the Ecological Class level follows the comparable level of Schafale and Weakley’s (1990) classification of North Carolina natural communities. Divisions at this level are based primarily on gross climatic, geographic, and edaphic similarities, e.g., High-Elevation Mountain Communities or Non-Alluvial Wetlands of the Coastal Plain and Piedmont. Ecological Classes are primarily useful for grouping physiographically and topographically related community groups, which often co-occur on the landscape.

At the Ecological Community Group level, which this document describes in detail, units are based on combinations of topographic, edaphic, physiognomic, and gross floristic similarities. This level is comparable to the level at which many natural community classifications define their basic units, e.g., Northern Red Oak Forests or Low-Elevation Basic Outcrop Barrens. Ecological community groups are not defined at a single, standard scale. Because community groups differ in their extent on the landscape, some are very broadly defined and have large geographic coverage (e.g., Oak / Heath Forests), while others are very narrow in concept and distribution (e.g., Granitic Flatrocks). A few groups (e.g., Inland Salt Marshes) may have only a single occurrence in Virginia but are known to have representatives in other states.

Community Types are the fundamental units of the classification system and will be nested within the Ecological Community Groups. The Community Type level is equivalent to the “association” level of the USNVC and traditional phytosociological studies. Types within a given ecological community group share definite environmental, structural, and floristic similarities. Occasionally Community Subtypes are employed to describe well-marked variation within community types. This document provides preliminary information about community types that have been classified or proposed for each Ecological Community Group. The methods and criteria used to define community types, as well as the types themselves, will be comprehensively described in future iterations of this classification. Classifications at the community type level, both in Virginia and in the USNVC, are dynamic and subject to frequent change. Maintenance of the web-based version of the classification will facilitate keeping this information up to date. A partial example of a fully nested classification within the framework outlined here is provided in Table 1. The protocol for naming community types is described in the introductory section on Procedures for Collecting and Analyzing Vegetation Data.

Because this classification is based on composition in all layers, not just the tallest, community types differ considerably from “cover types” used in forestry and large-scale vegetation mapping. Each approach is valid for certain purposes. Since our purpose is to classify ecological units, it is very important to consider all plants at a site. In forests, for instance, shrubs and herbs often respond to more subtle environmental gradients and may reveal more about local site conditions and associated animal species than do trees, which tend to be more broadly distributed and exhibit less environmental specificity. Likewise, herbaceous species occurring with low cover may be more restricted to certain site conditions and thus far more diagnostic of a community type than more widespread, dominant shrubs and trees.

Table 1. Partial example of the Virginia hierarchical classification system

System: ESTUARINE
Ecological Class: TIDAL WETLANDS
Ecological Community Group: TIDAL FRESHWATER MARSHES

Community Types: Nuphar advena – (Ceratophyllum demersum) Tidal Herbaceous Vegetation
                                     Tidal Freshwater Marsh (Spatterdock Mud Flat Type)
                                     —
                                     Typha latifolia – Schoenoplectus fluviatilis – Carex comosa Tidal Herbaceous Vegetation Tidal Freshwater Marsh (Broad-leaved Cattail – River Bulrush – Bristly Sedge Type)
                                     —
                                     Zizania aquatica – Polygonum punctatum Tidal Herbaceous Vegetation
                                     Tidal Freshwater Marsh (Wild Rice – Dotted Smartweed Type)

Community Subtypes: Zizania aquatica – Bidens laevis Subtype
                                          (wild rice – smooth bur-marigold subtype)

                                          Zizania aquatica – Polygonum punctatum Subtype
                                         (wild rice – dotted smartweed subtype)

The aggregation of community types defined exclusively with biotic criteria into coarser, broadly defined ecological community groups and classes is controversial and may not satisfy those who adhere strictly to either physiognomic-floristic or natural-group classification systems. Physiognomic-floristic systems such as the USNVC clearly identify the physiognomic and vegetational criteria used to distinguish types at different levels and provide consistent upper-level criteria that can be applied to virtually all vegetation in the world. Nevertheless, the higher physiognomic levels of the USNVC hierarchy pose many uncomfortable constraints on the lower, floristic levels, and units above the association level have limited utility at the regional level, especially for conservation. In response to these problems, developers of the USNVC have recently completed an effort to define an alternative, complementary system of standardized Ecological Systems (most comparable to the Virginia Ecological Class level) that can be applied across the United States (NatureServe 2003).

The ecological community groups treated in this report offer a number of advantages over comparable hierarchical units based strictly on physiognomy or floristics. These include:

concepts and terminology that are more communicable, familiar, accessible, and useful to many potential users
a system that encourages the ecological interpretation of vegetation patterns on the landscape
units that have greater utility for conservation purposes and for naming community records where more detailed classifications of vegetation are lacking.

One of the disadvantages of the approach taken here is the imperfect correspondence of community names with those used in similar classifications. For example, although we have adopted a number of names used in the North Carolina natural community classification (Schafale and Weakley 1990), major biogeographic differences between the two states made it impossible to do so consistently. Additionally, the colloquial application of many terms used in community nomenclature, including “bog,” “fen,” and “swamp,” varies widely. We have addressed the latter problem by including a Glossary of Technical Terms and Abbreviations that clearly articulates our interpretation of terms that may be unfamiliar or ambiguous. Continued crosswalking of Virginia community types to the USNVC, as well as the USNVC’s development of standardized ecological systems, should assist in resolving problems of inconsistent regional nomenclature.

Identifying community types or ecological community groups in the field is sometimes problematic since vegetation and associated site conditions are gradational across the landscape. While the boundaries between vegetation types may be sharp in some areas, more often they are represented by broad zones of compositional and environmental transition. It is important to recognize that the VANHP classification system, like other similar systems, is an attempt to partition an extremely complex set of factors into practical units for conservation, mapping, and management. Such an artificial group structure can never be perfect and there will always be environments and vegetation that are intermediate, anomalous, and/or difficult to classify.

Future Directions and Feedback
A hard-copy version of the Second Approximation will be released in the near future, although information will be periodically updated only on the website version. Detailed descriptions of community types classified by us or attributed to Virginia by others, as well as more detailed information about our plot-sampling and data analysis methodologies, will be posted in the future. We also hope to develop links between our community web pages and the natural community information on the NatureServe website. Plot data and PDF versions of technical reports may also be made available online.

We expect the full development of the state classification to be an iterative process of successive approximations. Progressive classification at the community type level will likely continue to result in modifications to and greater consistency in the circumscription of ecological community groups. Collaborative efforts with NatureServe regional ecologists, as well as our colleagues from other state natural heritage programs, will ensure that vegetation types classified through these efforts are crosswalked or incorporated into the USNVC. Over the next ten years, we anticipate a series of technical reports and peer-reviewed publications that will detail state-wide classifications of various broad ecological/taxonomic groups (e.g., Coastal Plain tidal wetlands). A relatively non-technical version of the classification will continue to be available to the public on the VANHP website, and possibly as a well-illustrated publication in the future.

As always, we appreciate and welcome comments or feedback on our work, as well as information or questions concerning Virginia’s natural communities. Click here for contact information.

Acknowledgments
This classification is the result of work by numerous people over a period of many years. The authors thank them and regret that space and logistics preclude mentioning everyone who has contributed to our knowledge of Virginia vegetation.

A large number of current and former VANHP employees and contractors have conducted or assisted with ecological field work and data analysis, but special thanks are extended to Kathleen McCoy, Dean Walton, Tom Rawinski, Bill Moorhead, Shelly Parrish, Joe Weber, Chris Clampitt, Tad Zebryk, Chris Ludwig, Allen Belden, Nancy Van Alstine, Johnny Townsend, Mark Hall, Tresha White, Gina Pisoni, Rebecca Wilson, Mike Leahy, Jennifer Allen, David Richert, Mike Lipford, and Dorothy Allard for major or sustained contributions.

Cooperators and consultants who have provided invaluable data and assistance include Harold Adams, Mark Anderson, Woodward Bousquet, Pete Bowman, Byron Carmean, Leah Ceperley, Gwynn Crichton, Tom Dierauf, Judy Dunscomb, Don Gowan, Stephanie Flack, Cecil Frost, Jason Harrison, Mike Hayslett, Michael Kieffer, David Knepper, Susan Leopold, Steve Martin, Robert Mueller, Leah Oliver, Ken Metzler, Doug Ogle, Milo Pyne, Rick Rheinhardt, Garrie Rouse, Mike Schafale, Rod Simmons, Rob Simpson, Jocelyn Sladen, Charles Smith, Lesley Sneddon, Scott Southworth, Steve Stephenson, Charles (Mo) Stevens, Donna Ware, Stewart Ware, Alan Weakley, Tom Wieboldt, Hal Wiggins, Gary Williamson, and John Young.

Finally, we gratefully acknowledge our clients who have supported community inventory and documentation in Virginia in substantial ways – National Park Service: Tom Blount, Wendy Cass, Gordon Olson, Bryan Gorsira, Beth Johnson, Melissa Kangas, John Karish, Chris Lea, Diane Pavek, Carol Pollio, Chuck Rafkind, Brent Steury, and many others. U.S. Forest Service: John Bellemore, Steve Croy, Kenneth Hickman, Fred Huber, Lisa Nutt, Jesse Overcash, and many others. Department of Defense: Robert Dennis, Alan Dyke, York Grow, Dorothy Keough, Tim Stamps, and Robert Wheeler. U.S. Environmental Protection Agency: Alva Brunner. U.S. Fish and Wildlife Service: Stephanie Brady, Lloyd Culp, John Gallegos, and John Schroer. NatureServe: Milo Pyne, Lesley Sneddon, Judy Teague, Rickie White, and others. Virginia Outdoors Foundation: Leslie Grayson.

The Natural Communities of Virginia Classification of Ecological Community Groups SECOND APPROXIMATION (Version 2.2)

The Natural Communities of Virginia
Classification of Ecological Community Groups
SECOND APPROXIMATION (Version 2.2)