Abstract. Phragmites australis (common reed)is an invasive marsh plant spreading in many wetlands on and near the tidal Hudson River. Phragmites is generally considered a pest with low value to wildlife and threatening rare plants, but scientific documentation is ambivalent. Some organisms are favored by Phragmites invasion and some are not. Phragmites appears to have considerable value for water quality amelioration and soil stabilization. Ecological functions of Phragmites vary greatly depending on site and stand factors. Important site factors include depth and duration of flooding, salinity, soil organic matter content, and microtopography; important stand factors include Phragmites height, density, dominance, prevalence of inflorescences (tassels), vine loads, presence of trees or shrubs, stand size, and interspersion of Phragmites patches with other plant communities. Phragmites is often encouraged by, and a symptom of, underlying problems, such as siltation, nutrient loading, and hydrological alteration. Yet Phragmites does not necessarily indicate poor habitat quality. Many restoration and management projects seek to remove Phragmites despite poor understanding of its ecology, the nontarget impacts of removal, and the sustainability of alternate species. I conducted a review and synthesis of information pertinent to the ecology and management of Phragmites on the Hudson River estuary and in nearby areas. This synthesis is unique in focusing on the Hudson River, considering a wide taxonomic and functional range of Phragmites impacts, and including extensive published and unpublished data and observations. I describe Phragmites ecology, address management issues on the Hudson, outline management techniques and their nontarget impacts, and suggest how research needs can be defined. Depending on management goals, site and stand factors, the surrounding landscape, and the local biota, it may be appropriate to take no action, remove a Phragmites stand, or alter the stand to change its habitat functions and ecosystem services. An explicit and documented decision-making process should be used to justify decisions and acquire information about management outcomes that can inform subsequent management.
Key words: Biodiversity; Birds; Fishes; Herbicides; Heterogeneity; Hudson River; Invasive plants; Phragmites australis; Restoration; Spartina; Tidal wetlands; Typha; Wetland management
Kiviat, E. 2006. Phragmites Management Sourcebook for the Tidal Hudson River. Report to the Hudson River Foundation, New York, New York. Hudsonia Ltd., Annandale NY 12504 USA.
Table of Contents
Hudson River Study Area
Materials and Methods
Prehistory and History of Phragmites in the Study Area
What Favored Establishment and Spread?
What Comes Before and After Phragmites?
Claims and Data Concerning Impacts
Phragmites, Biodiversity, and Ecosystem Processes
Conflicts of Interpretation
Goals of Management on the Hudson River
Advantages and Disadvantages
Achieving Particular Goals
Constraints on the Tidal Hudson
The Management Decision Process
Discussion and Conclusions
List of Tables Table 1. General characteristics of Phragmites australis in the northeastern United States, page 11
Table 2. Historic and prehistoric Phragmites status information from the Hudson River and Hudson Valley region, page 16
Table 3. Conspectus of available information on Phragmites pertinent to the tidal Hudson River, page 20
Table 4. Findings on tidal marsh fish communities in Phragmites compared to alternate vegetation, page 26
Table 5. Studies of nitrogen and phosphorus biogeochemistry in Phragmites australis stands, page 33
Table 6. Common assumptions of the prevailing approach to managing Phragmites australis in the northeastern United States, page 38
Table 7. Management techniques (treatments) that have been used on Phragmitesaustralis, page 39
Table 8. Suggested Phragmites management according to stand size, page 52
Table 9. Suggested Phragmites management by habitat type on the Hudson River, 53
Table 10. Characteristics of Phragmites stands relevant to habitat functions and management decisions, 55
List of Figures Figure 1. Map of tidal Hudson River showing many of the on-river localities mentioned in the text, page 8
Introduction This paper addresses the ecology, impacts, and management of an invasive marsh plant, common reed Phragmites australis (= P. communis; hereinafter “Phragmites”).Phragmites is spreading rapidly in many Hudson River tidal marshes, and if left alone might constitute the majority of the vegetation cover in at least half the marshes within a few decades. This invasion, if it progresses that far, may change plant and animal communities, ecosystem processes, marsh physiography, and the availability of fishery, wildlife, and recreational resources for human use. Management decisions will be made that will be expensive and have substantial non-target impacts. Although prevailing technical and public opinion considers Phragmites a serious pest in tidal marshes, evidence from recent research shows that many claims about Phragmites impacts are incorrect or exaggerated, and suggests that Phragmites is not all “negative” from a human point of view. Scientists, managers, regulators, and policy-makers need to strengthen the scientific basis for management of Phragmites to avoid errors that may eventually be costly ecologically, politically, and financially. The first step is to thoroughly and objectively review existing information pertinent to Phragmites in the Hudson River. In this paper I present information and ideas that will help biologists, managers, and policy-makers develop management approaches for the tidal Hudson River, individual sites on the river, and most areas of the northeastern states.
Phragmites was evidently an uncommon and local species in both tidal and nontidal wetlands of the Hudson Valley, despite a few pockets of abundance, before the 1960s (Muenscher 1935, 1937, Foley and Taber 1951, McVaugh 1958, Winogrond and Kiviat 1997). In an exhaustive survey of Hudson River wetlands ca. 1950, Foley and Taber (1951) reported Phragmites in only 6 sites (Bronck Island, Stockport-Nutten Hook, Peekskill Bay, Grassy Point, Croton River Marsh, and Piermont Marsh); the amounts of Phragmites increasing from north to south. In the 1960s-1970s, Phragmites was still uncommon and local in the Mid-Hudson region. Since then, Phragmites has proliferated, especially in many Hudson River tidal marshes; Piermont Marsh is saturated with Phragmites, Iona Island Marsh is rapidly becoming Phragmites-dominated (the expansion rate is logarithmic), and small colonies are spreading rapidly in Tivoli North Bay (Winogrond and Kiviat 1997, C. Nieder unpublished data). Some supratidal pools are dominated by Phragmites (e.g., on Sleightsburg Spit south of Kingston, and on the north side of the mouth of Mill Creek). Furthermore, there are many patches of Phragmites on dry or ephemerally flooded, supratidal dredge spoil, e.g., on Steward Island (Inbocht Bay). Phragmites has also appeared in many nontidal sites in roadside ditches, ponds, and marshes. Phragmites is capable of dominating dry infertile mineral soil such as on inactive sanitary landfills or dredge spoil deposits.
Strong concerns have been voiced about impacts of Phragmites expansion on flora, birds, fish, recreation, fire hazards, and other aspects of marsh ecology in North America (summarized in Howard et al. 1978, Kiviat 1987, Cross and Fleming 1989, Marks et al. 1994, Meyerson et al. 2000) and specifically on the Hudson River (Winogrond and Kiviat 1997, Anonymous 1998). Some ecologists, nonetheless, have questioned the scientific basis for this strongly negative assessment of Phragmites, and have produced data indicating that Phragmites is not always detrimental to biological diversity and marsh function (Meyerson et al. 2000, Kiviat 2005). Many ecological restoration proposals and projects for Hudson River marshes and other Hudson Valley
wetlands prominently include removal of Phragmites. A Phragmites removal project was conducted by the National Audubon Society at Constitution Marsh on the Hudson River several years ago, and another project is underway there. Half of about 45 marsh restoration projects suggested by the U.S. Army Corps of Engineers (ACOE) in 1994 for the Hudson River Habitat Restoration and Enhancement Project cited problems with Phragmites or called for Phragmites removal (Anonymous, no date). Two ACOE restoration projects including Phragmites removal are currently underway or anticipated in the near future (E. Blair, personal communication). The NYS DEC began herbicide treatment of 3 of 6 small Phragmites stands in the open marsh in Tivoli North Bay 2006 (W.C. Nieder, pers. comm.). Many other Phragmites removal projects are underway or proposed in tidal and nontidal wetlands in the Hackensack Meadows of New Jersey, New York City, the Long Island Sound drainage of Westchester County, and the Connecticut River estuary; completed, current, and planned projects in the Hackensack Meadows projects total ca. 1,000 acres (New Jersey Meadowlands Commission 2006). Approximately 4,000 ha of Phragmites-dominated marshes in Delaware Bay are being restored either by re-establishing saline tidal flushing in historically diked salt hay impoundments, or by use of herbicide, burning, and bed-lowering (Weinstein et al. 2001). Very few scientific data are available on “before and after” conditions at these restoration sites, impacts of Phragmites control on other plants and animals, and long term sustainability of replacement plant communities. I believe that, before additional large sums of money are spent on Phragmites removal projects that may not be supported by good science and may not be beneficial or sustainable, we must carefully assess the state of knowledge concerning the impacts of Phragmites on Hudson River marsh ecosystems and the ecological benefits and costs of Phragmites management.
Invasive plants have been an important topic of applied and theoretical studies during the past two decades. Plant invasions can not only displace native flora, fauna, and biological communities, but can also alter ecosystem processes such as biogeochemical cycling and fire regimes, and may be responsible for the degradation of many ecosystem services to society (Drake et al. 1989, Luken and Thieret 1997). Ecologists are realizing, however, that many of the concerns about plant invasions are either not supported by scientific data or are exaggerated. Current controversy about the invasive salt-cedar (Tamarix) illustrates this problem. Although salt-cedar has been considered a serious pest and very large efforts have been invested in its removal from riparian habitats of the southwestern U.S. for at least 30 years, recent information indicates that the ecological picture of salt-cedar invasion is not simple. Levels of 22 of 30 ecological functions did not differ between salt-cedar stands and stands of the native cottonwood (Populus fremontii)(Stromberg1998); biomass and diversity of insects in salt-cedar were comparable to those in cottonwood and willow (Salix) (Anderson 1998); the invasion of salt-cedar may have been a result of anthropogenic change including flood control, lowering of groundwater tables, increased salinity, and overgrazing (Everitt 1980, 1998); and salt-cedar dominated areas may be difficult or impossible to restore to functional native plant communities (Anderson 1996). A recent article (Cohn 2005) summarized the controversy regarding salt-cedar.
Although it is not well known to the public and to many managers, invasive plants may have substantial positive as well as negative values for fisheries, other fauna, and ecosystem services (e.g., Williams 1997). Phragmites is a partly native plant that, at least in some environments, supports substantial benthic and terrestrial invertebrate communities, foraging by fish, breeding
and foraging by a variety of water and marsh birds, and removal of nitrogen and phosphorus from eutrophic waters (e.g., Meyerson et al. 2000). Relative values must be assessed in relation to management goals on a site-specific basis in order to plan management that will prove correct in the long term. The first step in developing an appropriate, scientifically robust, regionalized and integrated strategy for management of an invasive plant is to review and synthesize available information on the ecology of the plant.
Most quantitative research on faunal relationships of Phragmites began only in the mid or late 1990s. The scarcity of this work on the Hudson River requires attention to other northeastern estuaries, and the recent and frequently unpublished state of much research requires time-consuming searches for gray literature and unpublished data, in order to complete a comprehensive and effective synthesis document. Prior synthesis documents on Phragmites (Howard et al. 1978, Kiviat 1987, Marks et al. 1994) covered the entire continent and are out of date. Much of the information on Phragmites ecology, however, is still based on preliminary or localized studies, and only a few of these studies have been conducted on the tidal Hudson River. My goal was to prepare a synthesis of scientific and natural history information relevant to the management of Phragmites and the restoration of Phragmites-dominated areas along the tidal Hudson River, and analyze the scientific accuracy of existing Phragmites management policy for Hudson River tidal wetlands. The variation in levels of different ecological functions provided by Phragmites relative to other plant communities, and the spatial and temporal variability of Phragmites ecology, indicate the need for more sophisticated management policy and procedures. Sutherland et al. (2004) have stated concerning nature management in the U.K., "Much of current conservation practice is based upon anecdote and myth rather than upon the systematic approaisal of the evidence..." This conclusion appears to pertain to Phragmites management in the northeastern U.S. In general (not specifically on the Hudson), the current Phragmites management policy, combined with the generally coarse-grained approach to management and restoration methods and the scarcity of good post-management monitoring data, are potentially a recipe for wasted money and unnecessary ecological damage.
Some regulators, managers, funders, and scientists may argue that this Sourcebook contains too much detail and that simpler guidance is needed for making decisions. My answer is that the current approach to Phragmites management costs too much in dollars and ecological side-effects to forego scientific nuances and consideration of missing information in decision-making. Furthermore, without sufficient knowledge the method of applied science will not work - we will not be able to make informed decisions, arrive at a management approach that considers different goals and sites, design management projects to accomplish specific goals at specific sites, determine if management accomplishes goals, and modify procedures to better accomplish our goals. This report is a work in progress, due to the constant flow of new information and the changing availability of old information on Phragmites, as well as the development of new paradigms in invasive plant management.
This report emphasizes goal-directed and site-specific management of Phragmites, with special consideration of management methods that alter rather than remove Phragmites stands. This is not a new idea in North America, and was referred to by, e.g., Ward (1942), Cross and Fleming (1989), and Kane (1978, 2001a, b). The rationale for this "soft" approach to management and an
appropriate decision-making process with broad applicability have not been set out in detail for applications in North America for Phragmites or any other plant to my knowledge (excepting in certain agricultural weed situations).
For the purpose of this report, “invasive plants” are either native or introduced species that spread at the expense of natural native plant communities. I refer to "alternate" communities or vegetation, which means non-Phragmites (e.g., Spartina, Typha) in comparison to Phragmites. By "community" I mean groups of plant species that tend to occur together under similar conditions of substrate elevation, hydrodynamic energy, natural and human disturbance, propagule availability, etc. Chance is also important in structuring communities although tidal marsh communities are simpler than many upland and nontidal marsh communities. A "stand" of Phragmites or alternate vegetation is a patch or bed of Phragmites (i.e., a "reedbed" in European terminology) or other plant(s). The Phragmites stand is a discrete entity that in many cases represents a single clone (genetic individual) of Phragmiteswith many underground and aboveground branches. I use "hyperdominant" or "highly dominant" instead of "monodominant," "monospecific," or similar terms, with reference to Phragmites stands that contain few other plants species and individuals. A Phragmites "marsh" is usually a group of Phragmites stands generally interspersed with other communities or water features. “Graminoid” refers to grass-like plants, including grasses, sedges, and rushes, as distinct from “forbs” which are broad-leaved herbs. Scientific names of vascular plants follow Gleason and Cronquist (1991). I use the term “management” in preference to “control”; management includes a wide range of approaches from no treatment to altering stands to local eradication.
Due to the rapidly changing knowledge base, today’s analysis or recommendations may prove incorrect tomorrow. This report is a state-of-the-art compilation and should not be considered the final word for Phragmites management; the report should be used as a general guide and not as a “cookbook.” Some decisions are simple (e.g., eradicate very small patches of the introduced genotype of Phragmites immediately when first discovered on a site, because this is cheap, effective, and low in risk). Many decisions are complex and nuanced, however, much as are decisions about managing human health or many other aspects of nature.
Hudson River Study Area Every estuary is different, and the tidal Hudson River has high levels of nutrients and suspended sediment, abundant PCB contamination, and is generally steepsided with pervasive and longstanding shoreline alteration due to the railroads and other present and past human land use. This report focuses on the tidal Hudson River, i.e., the nominal Hudson River and its associated tide-affected habitats from the Battery (the southern tip of Manhattan Island) north to the Troy Dam. Tidal wetlands and the tidal lower reaches of tributaries are included; in some cases this includes 1 km or more of tributary (e.g., Rondout Creek at Kingston). Habitats supporting Phragmites that are not regularly flooded by tides but are subject to irregular tidal flooding are included. “Irregular” flooding constitutes flooding by spring tides, storm surges, unusually heavy runoff, and ice jams (e.g., supratidal pools sensu Kiviat and Stevens  and many dredge spoil deposits up to 1 m or more above Mean High Water [MHW]). The upper reaches of the tidal Hudson (about Kingston northward) are always freshwater, and in the lower reaches salinity
varies spatially and temporally from freshwater to more than one-half Atlantic Ocean salinity. Mean tide range varies from 0.8 m in the Hudson Highlands, to about 1.4 m at Manhattan and the Troy Dam (Geyer and Chant 2006). Means, however, do not convey the great variation in high and low tide elevations due to lunar phase, runoff into the river, and wind; "storm" tides can be 1 m or more higher than MHW. There are about 2895 ha of tidal wetlands in the estuary (Kiviat et al. 2006). All shorelines and wetlands have been altered to a greater or lesser extent by railroads, roads, dredging, filling, and dumping, and pollutants include PCBs, herbicides, and metals. Herbicides have been used heavily on the railroads (I have been unable to find documentation of the substances applied) as well as in agriculture and other vegetation management in the watershed. The tidal Hudson River was described in detail in Levinton and Waldman (2006). Figure 1 shows many of the Hudson River localities mentioned in this report.
Figure 1. Map of tidal Hudson River showing many of the on-river localities mentioned in the text. (Adapted from Kiviat et al. 2006.)
The Plant Phragmites is a "habitat-modifying organism" (Rooth et al. 2003) that can modify soil, hydrology, microclimate, and vegetation. Table 1 summarizes morphological and ecological characteristics of Phragmites most relevant to this discussion of management in the tidal Hudson River. Phragmites is a giant, rhizomatous grass that forms dense, high-biomass colonies (clones) capable of covering large areas. Although a few authors have referred to Phragmites as “woody,” it is not woody in the sense of many bamboos, and the perennating buds of Phragmites are located below, or just above, the ground surface. Reproduction is principally vegetative: both local spread of colonies by rhizome or stolon extension, and longer distance dispersal by water, animal, or machinery transport of rhizome fragments. Stolons are fast-growing horizontal stems with long internodes, produced on the ground surface and apparently capable of crossing unfavorable substrates such as pavement (Kiviat, pers. obs.). Although seeds are very small and have been reported to be often inviable, seeds germinate readily and seedling establishment can occur in the wild (Harris and Marshall 1960, Gervais et al. 1993, Ahearn-Meyerson et al. 1997, Baldwin and Derico 1999). Nonetheless, establishment of stands from seed is probably rare. In southern France, Alvarez et al. (2005) concluded that vegetative and the less-common sexual colonization both played roles in the long-term dynamics of reed marshes.
In tidal marshes, Phragmites stands commonly become established either adjoining upland shorelines, or on the margin of a creek or pool (e.g., Winogrond and Kiviat 1997) although stands may become established in marsh interiors as well. Shorelines and banks are probably loci of establishment because elevations are higher and vegetative propagules (rhizome fragments) become stranded there. Areas of fill, such as the spoil along ditch banks, allow Phragmites to establish on aerobic soil and thence extend into wetter marsh habitat (Bart and Hartman 2002). Stands occur in all sizes and shapes. Often, new stands expand in all directions, evincing roughly circular form, until physical or biological barriers (including other Phragmites stands) are reached. Hudson River Phragmites is normally in the supratidal zone or upper intertidal zone, and at some locations in the middle intertidal zone (Kiviat, pers. obs.). Buckley and Ristich (1977) stated that Phragmites appeared restricted to elevations above mean high tide (i.e., MHW) in the marshes from Constitution to Piermont. Phragmites does not normally grow in the lower intertidal zone in the Hudson.
Monospecific stands? Phragmites is commonly stated to form monospecific or pure stands, or monocultures, i.e., patches of Phragmites in which other vascular plants do not grow. I have found this incorrect, although the interpretation depends upon scale. The outermost 1 m wide belt of a Phragmites stand, i.e., the outer stand edge, typically contains an admixture of several other species of plants. In the freshwater tidal marshes of the Hudson River, these species are likely to be spotted jewelweed, arrow arum, purple loosestrife, narrowleaf cattail, and several other species. Occasional woody plants (e.g., false indigo, willow) may be present. Farther into the Phragmites stand, plant species richness declines rapidly. In many stands, there are only scattered and usually stunted individuals of spotted jewelweed and arrow arum. At the scale of a hypothetical 1 m2 quadrat, many quadrats may contain only Phragmites. At the level of a 10 x 10 m quadrat, other plant species are usually present. Should this be called monospecific? No; ecological terminology needs to be accurate and monospecific or pure means “one species.” The
nearly-pure kind of Phragmites stand should be referred to as highly dominated, or hyperdominated, by Phragmites.
Although vascular plants are typically of most concern in discussions of vegetation, it should be remembered that algae and often mosses are also present in Phragmites stands. Mosses are more likely in nontidal wetlands where standing water is very shallow or absent, and water levels are relatively stable. Shallower, nontidal wetlands with stable water levels are also more likely to have mixed stands in which several other vascular plant species share space with Phragmites, although the other species in some cases comprise a small fraction of the aboveground biomass compared to Phragmites. A countervailing observation is that in 2006 mosses (number of species unknown) constituted significant minor cover beneath Phragmites on an eroding Phragmites turf at the water edge of a small marsh on the east side of Middle Ground Island between Hudson and Athens; several other vascular plants were also present in this Phragmites stand edge in the upper intertidal zone (Kiviat, pers. obs.).
Hyperdominant stands are formed by many native as well as introduced plant species (Sipple unpublished), and are not necessarily symptomatic of anthropogenic stress nor bad for biodiversity or ecosystem services. Some of the native species that form hyperdominant stands in or near the Hudson Valley are narrowleaf cattail, smooth cordgrass, spatterdock (in tidal marshes), and buttonbush, leatherleaf, swamp loosestrife, hybrid cattail, and softstem bulrush (in nontidal wetlands).
No herbivory? It has been stated or implied that Phragmites is little-used as food by native animals in North America (e.g., Marks et al. 1994). Rhizomes, culm bases, and young shoots of Phragmites are eaten by common muskrat and probably by American beaver. Phragmites marshes in the Hackensack Meadowlands often support substantial muskrat populations (Kiviat, pers. obs.). Young shoots are eaten by cottontail (Richard Casagrande, Yale University, pers. comm.). Several insects eat Phragmites leaves (including the meadow katydid Orchelimum vulgare, the larva of Henry’s marsh moth Simyra henrici, and the larvae of broad-winged skipper Poanes viator in the East and Yuma skipper Ochlodes yuma in the West). The mealy plum aphid Hyalopterus pruni, a sap-sucking insect, alternates between Phragmites and Prunus (cherries, etc.); the aphid often becomes hyperabundant on Phragmites leaves in summer. Ladybug adults and larvae feed on the aphids. The reed scale Chaetococcus phragmitis sucks sap beneath the leaf sheaths. Reed scale is nearly ubiquitous on Eurasian Phragmites in the northeastern states, may reach a high biomass in Hudson River tidal marshes (e.g., 1 gm-2), and is avidly consumed in winter and spring by black-capped chickadee, Carolina chickadee, red-winged blackbird, and apparently by downy woodpecker. Other insects feed within the culm (Krause et al. 1997, Schwärzlander and Häfliger 1999, Tewksbury et al. 2002). Downy woodpecker and probably other birds peck into culms to feed on overwintering insects, and the holes left by bird foraging can easily be found in the previous year’s culms in many areas. Lewis and Casagrande (1997) reported downy woodpecker and black-capped chickadee extracting insects from within Phragmites culms. An agromyzid fly larva feeds within the leaf blade creating a patch mine (an area of tan translucent tissue of ca. 2-3 cm2). White-tailed deer (Self et al. 1975) and apparently Canada goose (Kiviat, pers. obs.) eat the leaves to a limited extent. Cattle, sheep, goats, and horses eat the young shoots or leafy culms; horses are said to eat culms after they are too mature
for cattle (Duncan 1992, Tesauro 2001a, b). Young shoots are rich in saccharides and proteins and are good food for horses and cattle, but older shoots lignify and lose their value as forage (Zheng et al. 2005). Tree sparrow, swamp sparrow, and song sparrow eat Phragmites seeds (Russak 1956; Marks et al. 1994; Lewis and Casagrande 1997; Jean Bourque, Brooklyn, NY, pers. comm.; Kiviat, pers. obs.).
Table 1. General characteristics of Phragmites australis in the northeastern states.