Studies in Avian Biology No. 32:229–237

THE IMPACT OF INVASIVE PLANTS ON TIDAL-MARSH VERTEBRATE SPECIES: COMMON REED (PHRAGMITES AUSTRALIS) AND SMOOTH CORDGRASS (SPARTINA ALTERNIFLORA) AS CASE STUDIES

GLENN R. GUNTENSPERGEN AND J. CULLY NORDBY

Abstract. Large areas of tidal marsh in the contiguous US and the Maritime Provinces of Canada are threatened by invasive plant species. Our understanding of the impact these invasions have on tidal-marsh vertebrates is sparse. In this paper, we focus on two successful invasive plant taxa that have spread outside their native range—common reed (Phragmites australis) and smooth cordgrass (Spartina alternifl ora). A cryptic haplotype of common reed has expanded its range in Atlantic Coast tidal marshes and smooth cordgrass, a native dominant plant of Atlantic Coast low-marsh habitat, has expanded its range and invaded intertidal-marsh habitats of the Pacifi c Coast. The invasions of common reed in Atlantic Coast tidal marshes and smooth cordgrass in Pacifi c Coast tidal marshes appear to have similar impacts. The structure and composition of these habitats has been altered and invasion and dominance by these two taxa can lead to profound changes in geomorphological processes, altering the vertical relief and potentially affecting invertebrate communities and the entire trophic structure of these systems. Few studies have documented impacts of invasive plant taxa on tidal-marsh vertebrate species in North America. However, habitat specialists that are already con- sidered threatened or endangered are most likely to be affected. Extensive experimental studies are needed to examine the direct impact of invasive plant species on native vertebrate species. Careful monitoring of sites during the initial stages of plant invasion and tracking ecosystem changes through time are essential. Since tidal marshes are the foci for invasion by numerous species, we also need to understand the indirect impacts of invasion of these habitats on the vertebrate community. We also suggest the initiation of studies to determine if vertebrate species can compensate behaviorally for alterations in their habitat caused by invasive plant species, as well as the potential for adaptation via rapid evolution. Finally, we urge natural-resource managers to consider the impact various invasive plant control strategies will have on native vertebrate communities.

Key Words: food webs, geomorphology, invasive plants, marsh birds, North America, Phragmites, Spartina alternifl ora, saltmarsh, vertebrates.

EL IMPACTO DE PLANTAS INVASORAS EN ESPECIES DE VERTEBRADOS EN MARISMA DE MAREA: EL CARRIZO (PHRAGMITES AUSTRALIS) Y EL PASTO (SPARTINA ALTERNIFLORA) COMO CASOS DE ESTUDIO Resumen. Grandes áreas de marisma de marea a lo largo EU así como y en las Provincias de Marisma de Canadá, se encuentran amenazadas por especies de plantas invasoras. Nuestro entendimiento acerca del impacto que estas especies tienen en los vertebrados de marismas de marea es escaso. En este artículo nos enfocamos en dos taxa de especies de plantas exitosas que se han dispersado fuera de su rango nativo—el carrizo (Phragmites australis) y el pasto (Spartina alternifl ora). Un haplotipo críptico de carrizo ha expandido su rango en las marismas de marea en la Costa del Atlántico. El pasto (Spartina alternifl ora), nativa y dominante del hábitat de marisma baja de la Costa del Atlántico, ha expandido su rango e invadido habitats de marisma intermareal en la Costa Pacífi co. Las invasiones del carrizo en las marismas de marea de la Costa Atlántica y el pasto en marismas de marea de la Costa del Pacífi co parece que tienen impactos similares. La estructura y composición de estos habitats ha sido alterada y la invasión y dominancia por estas dos taxa, pueden derivar en cambios profundos en los procesos geomorfológicos, alterando la mitigación vertical y pueden potencialmente afectar las comunidades de invertebrados y toda la estructura trófi ca de estos ecosistemas. Pocos estudios han documentado impactos de taxa de plantas invasoras en especies de vertebrados de marisma de marea en Norte América. Sin embargo, especialistas del hábitat, los cuales ya están considerados como en peligro, son los que están siendo más afectados. Se necesitan estudios extensivos experimentales para examinar el impacto directo de especies de plantas invasoras en especies nativas de vertebrados. El monitoreo cauteloso de los sitios durante los estados iniciales de la invasión de plantas y el rastreo de los cambios del ecosistema en el tiempo son esenciales. Debido a que las marismas de marea son el foci para la invasión por numerosas especies, también necesitamos los impactos indirectos de invasión de estos habitats en la comunidad de vertebrados. También sugerimos el inicio de estudios para determinar si especies de vertebrados pueden compensarse en términos de comportamiento por las alteraciones en su hábitat causado por especies de plantas invasoras, como también el potencial para la adaptación vía evolución rápida. Finalmente, recomendamos a los manejadores de recursos

229 230 STUDIES IN AVIAN BIOLOGY NO. 32

naturales a que consideren el impacto que tendrán las estrategias de control de varias plantas invasoras en comunidades nativas de vertebrados.

Invasive plant taxa are profoundly chang- the Hudson River north to the St. Lawrence ing North American saltmarshes, but this is estuary, the coastal plain of the United States not an isolated phenomenon. The introduction from New Jersey south along the southeast- of non-indigenous plants in diverse habitats ern US Atlantic Coast to the northern Gulf of represents some of the most dramatic examples Mexico, and the western US along the Pacifi c of biological invasions. Their impact on natural coast (Dame et al. 2000, Emmett et al. 2000, habitats, and the biodiversity of those habitats, Roman et al. 2000). In this chapter, we are inter- is a pervasive threat and one of the most daunt- ested in two areas—the Atlantic Coast, with an ing ecological challenges facing twenty-fi rst emphasis on the northeastern Atlantic region, century natural-resource managers. and the Pacifi c Coast. Considerable attention has been devoted Northeast-coast saltmarshes are formed to understanding the attributes of successful largely by reworked marine sediments and in invaders and the characteristics of invaded situ peat formation. These marshes are largely regions and habitats, as well as to documenting limited to small, narrow fringing systems the patterns and history of invasion (Mooney because the physiography of the region and and Drake 1986, Drake et al. 1989, Cronk and broad expanses of rocky coast limit their areal Fuller 1995, Pysek et al. 1995). The consequences extent (Nixon 1982). Farther south, more exten- of invasive taxa on biological communities and sive saltmarshes occur in the drowned valley ecosystem processes have been documented estuaries of Delaware Bay and Chesapeake Bay more recently (D’Antonio and Vitousek 1992, (Teal 1986). Pacifi c Coast saltmarshes occur in Mack et al. 2000). MacDonald et al. (1989) calcu- a geologically young region structured by tec- lated that of the 941 vertebrate species thought tonic and volcanic forces (Emmett et al. 2000). to be in danger of extinction worldwide, 18.4% Because of the rocky and unfeatured wave- are threatened in some way by introduced spe- dominated shoreline along much of the Pacifi c cies. In North America, they calculated that Coast, extensive areas of saltmarsh are restricted >13.3% of the native avifauna is threatened by to large estuaries such as those associated with invasive species. In another study focusing on the San Francisco Bay and the Columbia River threats to biodiversity in the US, Wilcove et al. or behind sheltering bay-mouth bars. (1998) found that 49% of all imperiled species Strong physical gradients of salinity and (plants and animals) were threatened by inva- tidal inundation contribute to the characteris- sive species. tic patterns of tidal-height zonation. The two The study of biological invasions has only main marsh zones include low marsh and recently focused on coastal and estuarine high marsh. Low marsh is lower in elevation habitats (Grosholz 2002). Large numbers of relative to mean low water and is regularly non-indigenous species have been identifi ed fl ooded by tides. High marsh occupies the in U.S. coastal estuaries >200 non-indigenous higher elevations in the intertidal zone, and is species from San Francisco Bay alone (Cohen less infl uenced by tidal forces. and Carlton 1998). Most research has concen- The organization of tidal-marsh vegetation trated on non-native aquatic invasive species communities varies in the different regions. including crustaceans, clams, crabs, and hydro- A short-statured grass, saltmeadow cordgrass zoans (Cordell and Morrison 1996, Crooks 1998, (Spartina patens), dominates the high marsh Bagley and Geller 2000, Byers 2000). In contrast, along the northeast coast—often intermixed efforts to examine the ecological effects of non- with the short form of smooth cordgrass and native emergent wetland plant taxa in salt- black needlerush (Juncus gerardi) at the upland marshes have lagged behind (except Weinstein border of the high marsh. Smooth cordgrass et al. 2003). In this chapter, we focus on two of persists as a dominant species in low-marshes, the more problematic invasive marsh-plant taxa in this region reaching heights as tall as 1.25–2 m in North American saltmarshes—common reed (Teal 1986). Open tidal mudfl ats characterize (Phragmites australis) and smooth cordgrass the lower intertidal zone of Pacifi c Coast estu- (Spartina alternifl ora). aries. The mid-intertidal zone is dominated by California cordgrass (Spartina foliosa), which ATLANTIC AND PACIFIC TIDAL SALTMARSHES forms a narrow band of vegetation along the outer edges of the native terrestrial vegetative In the contiguous US and Maritime Provinces zone and along tidal channels (Mahall and Park of Canada saltmarshes occur in three distinct 1976, Ayers et al. 2003). California cordgrass’s regions: the Northeastern Atlantic Coast from range extends from Baja California north to IMPACT OF INVASIVES—Guntenspergen and Nordby 231

Humboldt Bay (Josselyn 1983). California cord- by seaside arrowgrass and the replacement grass, the only species of Spartina native to the of saltmeadow cordgrass by the short form of Pacifi c Coast, grows sparsely and is relatively smooth cordgrass (Niering and Warren 1980, short (usually <1 m tall; Ayres et al. 2003). The Warren and Niering 1993). Cultural eutrophica- Pacifi c Coast high marshes, with high salinity tion leading to higher loadings of nitrogen to and saturated soils, are dominated by low- northeast tidal marshes is also hypothesized to growing (<0.5 m) cordgrass species (Baye et al. have resulted in changes in tidal-marsh vegeta- 2000). A transition in vegetative composition of tion patterns (Bertness et al. 2002). Nitrogen fer- saline marshes occurs in the Pacifi c Northwest tilization experiments in nitrogen-limited New and these marshes are dominated by Salicornia England tidal marshes resulted in increased spp. and beach salt grass (Distichlis spicata) abundance of smooth cordgrass in high-marsh in the high marsh and seaside arrowgrass plots while marsh hay decreased (Levine et al. (Triglochin maritima) in the low marsh (Seliskar 1998, Emery et al. 2001). and Gallagher 1983) but much of the intertidal area in the Pacifi c Northwest remains un-veg- INVASIVE TIDAL SALTMARSH PLANT SPECIES etated (Simenstad et al. 1997). Over the last 200 yr, marsh and tidal fl ats Atlantic and Pacifi c coast tidal saltmarshes have been lost to or degraded by human devel- are characterized by a few dominant emergent opment activities including diking, draining, plant species organized in characteristic zones dredging, or fi lling for agriculture or urbaniza- resulting from both physical stress and compe- tion, and conversion to salt-production ponds. tition, leading to distinct plant communities at A substantial portion of U.S. tidal wetlands specifi c elevations (Bertness and Ellison 1987). has been destroyed (Tiner 1984) and unaltered But, because of habitat degradation, they may coastal saltmarshes are rare (Roman et al. be among the most susceptible to invasive plant 2000). Over 80% of the saltmarshes that once species. Shoreline development, tidal restric- occurred in New England have already been tion, and habitat destruction result in disturbed lost (Teal 1986). Originally, New England salt- conditions including bare soil, high nutrient marshes had networks of salt ponds, pannes, inputs, altered hydrology, and high light levels potholes, and channels in the high marsh which are thought to be among the conditions where the water was semi-permanent. Roads that promote successful plant invasions. The and other obstacles have cut off or reduced colonization and spread of common reed in tidal fl ow into these habitats (Roman et al. Atlantic Coast marshes and cordgrass in Pacifi c 1984, Burdick et al. 1997). Most saltmarshes Coast marshes has been rapid and follows a pat- along the Atlantic Coast have also been tern often typical of plant invasions. Windham ditched to remove standing water and pools (1999) describes the typical invasion sequence and prevent mosquito breeding, resulting in of reeds in Atlantic Coast saltmarshes initiated lowered water tables, vegetation changes, by the fi rst appearance of isolated small patches, and associated trophic impacts on fi sh and the continued initiation of numerous other iso- waterbirds (Roman et al. 2000). Pacifi c Coast lated patches over time, the coalescence of these marshes have suffered the same fate as those patches and eventual dominance of an area. She along the Atlantic Coast. For example, a 70% cited an average annual rate of spread >20% at a loss of tidal wetlands has occurred in the Puget site in southern New Jersey from 1972–1991. Sound estuary in Washington with localized loss being virtually complete in heavily urban- Common reed ized areas (Washington Division of Natural Resources 1998). At the turn of the 19th cen- Common reed is found worldwide. It toler- tury, the San Francisco Bay estuary included ates a range of abiotic conditions and is found in approximately 76,900 ha of tidal marshes and both freshwater and coastal habitats, although 20,400 ha of open tidal fl at. Today, only about its establishment and growth is limited by 16,300 ha (21%) of tidal marshes and 11,800 fl ooding duration and high salinity and sulfi de ha (58%) of tidal fl ats remain (Goals Project levels (Chambers 1997). Reeds have been shown 1999). to form extensive stands in tidal marshes with Northeast coast saltmarsh vegetation patterns salinities <15 ppt. Small, more recently estab- have changed dramatically over the past 50 yr. lished plants grow well at salinities from 0–5 Surveys of southern New England saltmarshes ppt, exhibit some reduction in growth up to 35 suggest that increases in sea levels leading to ppt, and have diffi culty persisting when salini- increased waterlogging of upland marsh soils ties exceed 35 ppt (Chambers et al. 1999). In and plants has in turn led to the replacement North America, the range of reeds has expanded of black needlerush in the upper high marsh dramatically since the late 19th century, and in 232 STUDIES IN AVIAN BIOLOGY NO. 32 some areas reeds have formed extensive mono- Smooth cordgrass cultures displacing native species (Chambers et al. 1999). Reeds now occupy many tidal Among the invasive plants in Pacifi c Coast habitats in Maritime Canada, New England, the marshes, several cordgrass species have been mid-Atlantic, and the northern Gulf of Mexico. particularly successful because they are among Reeds form dense monocultures following the most abundant and aggressive intertidal establishment (Meyerson et al. 2000) and are plants in North America (Adam 1990). For thought to be a robust competitor relative to instance, four of the 12 non-native plant spe- other saltmarsh species. They grow to 3–5 m tall cies identifi ed as introduced species of concern and can form solid stands with stem densities in the San Francisco Bay estuary are cordgrass ranging from 50–125 shoots/m2 (Meyerson et species—smooth cordgrass, saltmeadow cord- al. 2000). grass, dense-fl owered cordgrass (S. densifl ora) Wrack accumulation, erosion, ice scour that and common cordgrass (S. anglica) (Grossinger promotes bare soil, ditching and other hydro- et al. 1998). The introduction and spread of logic disturbances, and nutrient enrichment smooth cordgrass, however, is arguably the associated with shoreline development provide most devastating of the tidal-marsh-plant reeds with opportunities to become established invasions on the Pacifi c Coast. The dominant (Chambers et al. 1999). Dispersal and burial plant species of low marsh along the Atlantic of large rhizome fragments into well-drained Coast of the US, smooth cordgrass has become and low-salinity sites improve the chances of established in open tidal mudfl ats of the Pacifi c successful establishment (Bart and Hartman Coast and has extended its range up through 2003). Once established, poorly drained areas the high-marsh zone as well. and sites with high salinity and sulfi de levels Multiple intentional and accidental intro- tend to be invaded by clonal spread (Chambers ductions of smooth cordgrass have occurred in et al. 2003). Pacifi c estuaries. In Washington, smooth cord- Many explanations have been invoked for grass was accidentally introduced to Willapa the recent change in the relative abundance Bay sometime before 1911 (Scheffer 1945) and and distribution of reeds in North American by 1988 it had spread to occupy >445 ha of tidal marshes (Chambers et al. 1999, Orson tidal fl at (Aberle 1993). More recently, the rate 1999). Recent advances in genomics, including of spread appears to be accelerating. In 1997, the ability to examine nucleotide sequences in the area solidly covered by smooth cordgrass chloroplast DNA, have shed considerable light was estimated at >1,315 ha, and by 2002 it was on this question. Comprehensive genetic analy- estimated at >2,500 ha equaling nearly 47% of ses of herbarium specimens collected before the tidalfl at habitat in Willapa Bay (Washington and after 1910 reveal signifi cant changes in the Division of Natural Resources 2000, Buchanan haplotype frequency of North American reed 2003). In the 1930s and 1940s, smooth cordgrass populations (Saltonstall 2002, 2003). Today one was intentionally introduced in four areas of distinct haplotype derived from an introduced Puget Sound for duck-habitat enhancement, Eurasian lineage (Type M) is the dominant type but the spread there has been minor compared found in the tidal marshes of the Northeast to that in Willapa Bay (Frenkel 1987). Smooth and mid-Atlantic Coast although populations cordgrass was also introduced in the late 1970s of native haplotypes still persist in the region. into one area in Oregon, the Siuslaw estuary Although native haplotypes still dominate (Aberle 1993). In California, the Army Corps of along the Pacifi c coast, haplotype M has been Engineers brought smooth cordgrass plants into identifi ed in urban areas in the western US the South Bay of San Francisco Bay in 1973 for (Saltonstall 2003). a marsh-restoration project and over the next It is currently not known why haplotype M decade it was transplanted to at least two other has become the dominant reed lineage and has sites, and likely others, within the South Bay increased its distribution throughout Atlantic (Ayres et al. 2004, Grossinger et al. 1998). Over Coast tidal habitats. Type M may be a superior the next 10 yr, smooth cordgrass spread slowly competitor or environmental conditions may to other areas and began to hybridize with the have changed and played a role in the expan- native California cordgrass. This hybrid (smooth sion of its range (Silliman and Bertness 2004). cordgrass × California cordgrass) is highly pro- New experiments evaluating the growth and ductive, out competes both parental species, persistence of native and invasive haplotypes and is the form that is now aggressively spread- along salinity and hydrologic gradients as ing throughout San Francisco Bay (Daehler well as competition experiments with other and Strong 1997, Ayres et al. 2004). The latest saltmarsh dominants are currently underway surveys show that smooth cordgrass/hybrid (Vasquez et al., 2005). now occupies approximately 2,030 ha which IMPACT OF INVASIVES—Guntenspergen and Nordby 233 equals approximately 17% of the tidal fl at and the native marsh plants and has been dubbed an marsh habitat in south San Francisco Bay where ecosystem engineer because of its ability to alter the invasion is concentrated (Ayres et al. 2004). habitat through increased sediment accretion Ayres et al. (2003) predicted that, if unchecked, (Ayres et al. 1999). When invaded by smooth invasive smooth cordgrass has the potential to cordgrass, marshes can ultimately be trans- spread throughout the San Francisco Bay estu- formed into solid non-native cordgrass mead- ary and beyond such that it would cause the ows (Daehler and Strong 1996). In San Francisco global extinction of the native California cord- Bay, this non-native cordgrass colonizes open grass. Indeed, recent surveys have confi rmed intertidal mudfl ats and clogs tidal channels the presence of smooth cordgrass in two other (growing as low as 73 cm above the lower limit California estuaries north of San Francisco Bay of the intertidal zone), and grows throughout (Bolinas Lagoon and Drakes Estero) indicating the marsh plain up to the high marsh (as high as the potential for widespread colonization of 15 cm below the maximum elevation of tidal- other Pacifi c Coast estuaries (Ayres et al., 2003). marsh vegetation) where it appears to be dis- Daehler and Strong (1996) estimate, that along placing native plant species (Ayres et al 1999, the Pacifi c Coast of the US, the fi nal distribution Collins 2002). Based on estimates of smooth for smooth cordgrass will stretch from Puget cordgrass tidal inundation toleration rates, cur- Sound, Washington, south through the Tijuana rent water levels, and tidal regimes, Stralberg et River Estuary, California. al. (2004) predicted that approximately 33% of intertidal mudfl at habitat could be encroached EFFECTS OF INVASIVE REEDS AND CORDGRASS ON upon by smooth cordgrass and its hybrids. In TIDAL-MARSH HABITAT addition, the upward spread of smooth cord- grass could be accelerated by future sea-level The expansion of reeds into high-marsh areas rise (Donnelly and Bertness 2001). along the Atlantic Coast of the US can result in Changes in habitat structure and composi- important changes in plant community struc- tion that accompany the smooth cordgrass ture and potential declines in the vertebrate invasion on the Pacifi c Coast and the common species dependent on these habitats. In New reed invasion on the Atlantic Coast, lead to England marshes, the impacts of human devel- alterations in geomorphological processes in opment and cultural eutrophication are affect- tidal marshes and have implications for many ing the distribution of plant species (Bertness et aspects of the tidal-marsh ecosystem including al. 2002). Shoreline development and enhanced basic hydrologic function (e.g., altering fl ow nitrogen supplies appear to be associated with regimes in marshes by clogging tidal channels). the expansion of common reed populations into Thus the effects of the invasion and dominance the high marsh. Rooth et al. (2003) documented of tidal wetlands by common reed and smooth increased rates of sediment accretion following cordgrass could cascade throughout the tidal- invasion by reeds in oligohaline tidal marshes marsh system and alter the trophic structure of of the Chesapeake Bay. The high productivity the marsh ecosystem as well, although little is of reeds and accumulation of litter on the marsh currently known about these effects. For exam- surface, coupled with high stem density and ple, the increase in sediment accretion (e.g., high inorganic sediment loading, appears to be 1–2 cm/yr in Willapa Bay [Sayce 1998]), coupled the mechanism resulting in the higher rates of with the increase in mass and density of above- sediment accretion. The enhanced rates of sedi- ground biomass of smooth cordgrass invasions ment accumulation in reeds stands can alter the in Pacifi c Coast estuaries, could potentially physical structure of tidal marshes by building change the invertebrate community composi- up the marsh surface and fi lling in topographic tion of intertidal zones, reducing benthic inver- depressions and fi rst order tidal channels, tebrate densities (Capehart and Hackney 1989), resulting in a loss of microtopographic varia- while increasing insects and arachnids of the tion (Lathrop et al. 2003). cordgrass canopy. Similar habitat alteration occurs in Pacifi c Coast estuaries that have been invaded by non- POTENTIAL IMPACTS ON TIDAL-MARSH VERTEBRATES native smooth cordgrass (and/or the hybrid form smooth cordgrass × California cordgrass). The impact of the introduction and spread The non-native cordgrass often grows to heights of non-native reeds and smooth cordgrass on of 2 m or more and the above- and below-ground tidal-marsh vertebrate populations remains biomass is much denser than any of the native largely unstudied. For instance, few data cor- plant species (Callaway and Josselyn 1992). relate the distribution of these invasive plant Smooth cordgrass is able to occupy a much species with the distribution and abundance larger portion of the tidal gradient than any of of native tidal-marsh bird or mammal species. 234 STUDIES IN AVIAN BIOLOGY NO. 32

One of the most widely recognized values of short high-marsh vegetation, had low frequen- saltmarshes is their support of migrant and cies in stands dominated by reeds. The Marsh resident avian species. Fundamental changes Wren, Swamp Sparrow (Melospiza georgiana), in habitat structure, shifts in primary produc- and Red-winged Blackbird (marsh generalists tivity, and the potential modifi cation of trophic that prefer tall reedy vegetation) were found pathways that accompany the invasion will in sites dominated by reeds. Even within these likely have their biggest impacts on resident, examples of species that maintain populations non-migratory species that are dependant in reeds, more detailed study is required. Olsen year-round on tidal marshes. and R. Greenberg (pers. comm.) report that in Delaware, Swamp Sparrows require saltmeadow POTENTIAL IMPACTS OF COMMON REED cordgrass for nest cover. Clumps of this vegeta- tion can be found along the edge of reed beds, Several studies provide evidence that many but not in the interior of large stands. Therefore, species of vertebrates use marshes dominated it is unlikely that Swamp Sparrows can maintain by reeds (Kiviat et al., pers. comm.), which nesting populations in larger stands of reeds. In appears to be more important to wildlife as fact, Benoit and Askins (1999) also found that shelter than as food. Wildlife species tend to homogeneous stands of reeds did not provide use the edges of stands, mixed-reed stands, and sustainable habitat for many wetland bird spe- smaller patchy stands than the dense extensive cies. Wading birds, shorebirds, and waterfowl interiors of larger stands. Interestingly, colo- were absent from surveyed reed stands. By nial-nesting long-legged wading birds may contrast, the high-marsh stands dominated by benefi t from the proximity of reed stands. In short-stature grasses included a wide variety of certain sites in Delaware, reeds provide criti- generalists: waders, shorebirds, ducks, and aerial cal habitat for nesting colonial wading birds insectivores as well as high-marsh specialists. A by offering substrate and material for nesting, phenomenon less well documented was the use and serves as a buffer from human disturbance of reed stands embedded in a larger more het- (Parsons 2003). erogeneous landscape. Benoit (1997) reported Reed-dominated marshes support more Virginia Rails and King Rails (Rallus elegans) species of coastal marsh-breeding birds than using patches of reeds interspersed with areas of commonly believed (Kiviat et al., pers. comm.). open brackish marsh. Kiviat et al.’s literature review documented 24 In Atlantic Coast tidal marshes where reeds species of birds that utilized reed stands located have recently established, the availability of in either estuarine tidal marshes and creeks or prey resources (snails, amphipods, and iso- saltmarsh habitat. Although dense populations pods) to adult mummichogs (Fundulus hetero- of reeds appear to have little value for birds, clitus) may be no different than in non-invaded stands interspersed with tidal creeks and open tidal marshes (Fell et al. 1998). However, as water and mixed stands or habitat on the edge of the hydrology of these sites change and marsh reed stands do support some bird species (Swift surface heterogeneity and topographic depres- 1989, Brawley 1994, Holt and Buchsbaum 2000). sions disappear, there is evidence that fi sh Holt and Buchsbaum (2000) suggested that recruitment and utilization may change in factors other than the dominant plant species also reed-dominated stands (Weinstein and Balletto have a major role in determining the distribution 1999, Osgood et al. 2003). A growing body of bird species in tidal marshes. They found that of research suggests that mummichogs may the presence of reeds in northern Massachusetts’s exhibit reduced feeding and reproduction in coastal marshes appeared to have little effect on response to the structural changes that occur as the numbers of Red-winged Blackbirds (Agelaius tidal-marsh sites naturally dominated by cord- phoeniceus), Marsh Wrens (Cistothorus palustris), grass species become dominated by common Virginia Rails (Rallus limicola), or Saltmarsh reed (Able et al. 2003, Raichel et al. 2003). This Sharp-tailed Sparrows (Ammodramus caudacu- suggests that prey for larger wading birds may tus). Benoit and Askins (1999) conducted one of not be accessible within dense stands of reeds the few direct comparisons of bird use of reeds but these large wading birds may forage on the and unaltered saltmeadow cordgrass habitat in edges of reed stands intermixed with more typi- Atlantic Coast saltmarshes. They found signifi - cal low or high marsh. cantly fewer species of birds in reed-dominated stands than in high-marsh saltmeadow cord- POTENTIAL IMPACTS OF SMOOTH CORDGRASS grass stands. The Seaside Sparrow (Ammodramus maritima), Saltmarsh Sharp-tailed Sparrow, On the Pacifi c Coast, the smooth cordgrass and Willet (Catoptrophorus semipalmatus), three invasion may have negative effects on native ver- tidal-marsh specialists adapted to nesting in the tebrate species, but as yet few data are available. IMPACT OF INVASIVES—Guntenspergen and Nordby 235

The salt marsh harvest mouse (Reithrodontomys clogging tidal channels, smooth cordgrass may raviventris), a federally endangered species reduce the foraging habitat of rails as well endemic to San Francisco Bay saltmarshes, as alter what food items are available. Also, prefers the mid- and upper-tidal areas that are Clapper Rails do occasionally nest in native largely dominated by pickleweed (Salicornia California cordgrass (Zucca 1954) but no stud- virginica; Shellhammer et al. 1982). Shellhammer ies have yet examined the success of Clapper et al. (1982) found very few mice in pure stands Rail nests placed in either exotic or native of salt-marsh bulrush (Schoenoplectus maritimus), cordgrass. a tall, reedy bulrush with structural character- Like the California Clapper Rail, the istics more similar to the non-native smooth Alameda Song Sparrow (Melospiza melodia cordgrass than to the preferred Salicornia spp. pusillula) does occupy marshes that have been Shellhammer et al. speculated that the value invaded by smooth cordgrass (J. C. Nordby and of pickleweed was higher for the saltmarsh A. N. Cohen, unpubl. data). In a native marsh, harvest mouse than was bulrushes because Song Sparrow breeding territories are typically pickleweed provides denser cover and more arrayed in a tight linear fashion in the taller horizontal branching. While smooth cordgrass plants (gumweed [Grindelia stricta] and Virginia may provide fairly dense cover, it provides picklewood) that occur along tidal channels little horizontal structure. Other mammals that (Marshall 1948a, Johnston 1954). Preliminary occur in San Francisco Bay tidal marshes that analyses from an ongoing study of how salt- may be affected by habitat alteration associated marsh Song Sparrows are responding behavior- with smooth cordgrass include the saltmarsh ally to the rapid alteration of habitat by smooth wandering shrew (Sorex vagrans haliocoetes), cordgrass have shown that Song Sparrows do the Suisun shrew (Sorex ornatus sinuosis), and include the non-native cordgrass habitat in the California vole (Microtus californicus). All their territories and use those areas for foraging three of these species are known to occur in as well as for nesting. However, no observed the middle and upper intertidal zones of salt or Song Sparrow territories have been composed brackish marshes (Lidicker 2000, MacKay 2000, entirely of smooth cordgrass, and nests that Shellhammer 2000). were placed in smooth cordgrass were some- Relatively few bird species are restricted what less successful and much more likely to year-round to tidal saltmarshes in San Francisco fail due to tidal fl ooding than were nests placed Bay. Resident bird species that breed in the tidal in native vegetation (J. C. Nordby and A. N. marshes include the federally endangered Cohen, unpubl. data). It is possible that the California Clapper Rail (Rallus longirostris obso- Song Sparrows are drawn to inappropriate nest- letus) and three subspecies of tidal-marsh Song ing sites in smooth cordgrass that are too low in Sparrow (Melospiza melodia pusillula, M. m. sam- elevation relative to the tides. Whether smooth uelis, and M. m. maxillaris) listed as California cordgrass is acting as an ecological trap for Song species of special concern because of habitat Sparrows, where overall reproductive success is loss and because they have extremely restricted reduced, remains to be tested ranges and adaptations for nesting in Pacifi c The impact of the smooth cordgrass inva- Coast saltmarshes (Marshall 1948a, Johnston sion is not restricted to resident species because 1954). While California Black Rails (Laterallus the open tidal fl ats of Pacifi c estuaries provide jamaicensis coturniculus) and Salt Marsh crucial habitat for migrating shorebirds. San Common Yellowthroats (Geothlypis trichas sinu- Francisco Bay is designated as a Western osa), also species listed in California, typically Hemisphere Shorebird Reserve Network that breed in brackish or freshwater marshes, they provides breeding habitat or critical migratory do occur in saltmarshes during winter and so stopover sites for >1,000,000 waterfowl and may also be affected by the smooth cordgrass shorebirds each year (Kjelmyr et al. 1991), more invasion. than any other wetland along the Pacifi c Coast Although California Clapper Rails do occur of the contiguous US (Page et al. 1999). Most of and nest in areas that have been invaded by these bird species forage extensively on benthic non-native smooth cordgrass (S. Bobzien, pers. organisms found in the vast tidal mudfl ats that comm.; J. C. Nordby, pers. obs.), it is unclear rim the bay (Takekawa et al. 2000). In a study of whether these sub-populations are sustain- the affect of the spread of common cordgrass (a able. Clapper Rails forage mainly at low tide close relative of smooth cordgrass) on shorebird when the mud substrate in tidal channels and populations in the British Isles, Goss-Custard tidal fl ats is exposed and their preferred foods and Moser (1988) found the largest reduction (clams, mussels, snails, and crabs) are more in Dunlin (Calidris alpina) in estuaries where the available (Williams 1929, Moffi t 1941, Albertson cordgrass had replaced much of the intertidal and Evens 2000). By colonizing tidal fl ats and mudfl at foraging habitat. Ayres et al. (2004) 236 STUDIES IN AVIAN BIOLOGY NO. 32 predicted that in San Francisco Bay the loss by these taxa can affect the trophic structure and of tidal mudfl at habitat to smooth cordgrass vertebrate species composition of tidal marshes. colonization could be extensive if the invasion However, we know of no experimental studies goes unchecked over the next two centuries. of vertebrate species that provide quantitative Stralberg et al. (2004) estimated that 33% of estimates of these effects. These studies are shorebird habitat value (range 9–80%) could be needed to examine the impact of habitat altera- lost under realistic spread scenarios. tion by invasive plant species on the structure In Willapa Bay, Washington, where smooth and function of tidal-marsh communities in set- cordgrass increased three-fold between 1994 tings that allow for rigorous comparisons with and 2002 (Buchanan 2003), aerial surveys con- appropriate controls. ducted in 2000–2001 suggest a reduction in Additional studies are needed that determine shorebird numbers by approximately 60% and the current distribution, abundance, and popu- foraging time by as much as 50% in the southern lation trends of native vertebrate species and portions of the bay as compared with data from their correlation with the presence of different the 1991–1995 surveys (Jaques 2002). species of invasive plants, as well as the effects In addition to the direct alteration of habi- of invasive species on important demographic tat, invasive plants may be altering competi- parameters such as reproductive success and tive interactions among native species as well. survival. We also need to assess the landscape- Pacifi c Coast Marsh Wrens (Cistothorus palustris scale consequences of plant invasions in tidal paludicola), which normally nest in freshwater or marshes and whether a relationship exists brackish marshes and not in open saltmarshes between vertebrate community structure and (Verner 1965), have begun to establish breeding landscape patchiness. Small isolated stands territories in the newly available smooth cord- characteristic of the early stages of invasion grass habitat in San Francisco Bay as well as in may not negatively impact native vertebrate other smooth cordgrass-invaded marshes such populations and may even provide additional as those in Willapa Bay (Williamson 1994; J. C. edge habitat for certain species. As patches Nordby and A. N. Cohen, unpubl. data). Marsh coalesce, however, and a threshold is reached in Wrens are highly territorial and will defend the invasive cover of an area, we may only then their nesting areas by breaking the eggs of other see detrimental effects as dense interior areas species that attempt to nest nearby (Picman occupy a greater share of the landscape and 1977). They can control the distribution and alter intact native habitat becomes increasingly rare. the behavior and reproductive strategy of much Because the spread of exotic species is an ongo- larger and aggressive birds, such as Red-winged ing process, we often have unique opportunities Blackbirds and Yellow-headed Blackbirds to establish baseline data in areas that are not (Xanthocephalus xanthocephalus; Picman 1980, yet invaded and also to track changes over time Picman and Isabelle 1995). Preliminary analyses in areas where invasions are actively spreading. of birds in smooth cordgrass-invaded marshes The development of predictive theoretical and in San Francisco Bay have shown that Song empirical models that incorporate metapopu- Sparrows and Marsh Wrens have segregated lation dynamics of vertebrate species would territories with little overlap and that marsh enhance our understanding of the potential wren territories are more highly correlated with future impacts of these invasions. the non-native cordgrass habitat than are Song It is also important that we assess the behav- Sparrow territories (J. C. Nordby and A. N. ioral and genetic responses of native species to Cohen, unpubl. data). It is not yet known, how- the exotic-species invasions. Because the altera- ever, whether Marsh Wrens are excluding Song tion of habitat can occur so rapidly, we need to Sparrows from the smooth cordgrass habitat understand whether, or to what extent, native or if song sparrows are selecting against those species can alter their behavior to compensate areas for other reasons (e.g., nesting habitat or for changes in their environment. A high level food resources are limited). of behavioral plasticity would be benefi cial as it could also buy species more time to evolve FUTURE RESEARCH NEEDS adaptations to their rapidly changing habitat. Not only must we examine the direct impact It is clear that the replacement and domi- of non-native, invasive species, we also need nance of tidal-marsh communities in North to expand our understanding of the indirect America by invasive non-native reeds and cord- impact of habitat alteration that can be associ- grasses can have important and perhaps severe ated with invasions such as trophic cascades in consequences. These taxa may alter geomor- which the entire food web is altered, facilitation phological processes, hydrologic regimes, and of further exotic invasions as newly altered habitat structure. It is presumed that invasion habitat attracts additional non-native species IMPACT OF INVASIVES—Guntenspergen and Nordby 237 or even alters interspecifi c interactions among A mechanistic understanding of the impacts of native species. Only by examining both the invasive species on vertebrate communities is direct and indirect effects of habitat alteration an essential step in determining if suitable alter- and the effects of other invasive species on native management strategies are needed. tidal-marsh vertebrates will we be able to deter- mine the full extent of the impacts of invasive ACKNOWLEDGMENTS plant species. Millions of dollars are being spent on con- We thank S. Droege and R. Greenberg for trolling the cordgrass and reed invasions and their encouragement and endless patience and natural-resource managers have been making E. Kiviat and D. Stralberg for valuable com- decisions about invasive-plant control mea- ments that improved the manuscript. JCN was sures without knowing the appropriateness supported by a David H. Smith Conservation of the different control programs currently in Research Fellowship from The Nature place. These control measures (e.g., large-scale Conservancy. GRG received support from the glyphosate spraying, fi re, mechanical removal, U.S. Geological Survey Eastern Region State or tarps), as well as the seasonal timing of appli- Partnership Program and the U.S. Geological cation, may well have unintended consequences Survey-Fish and Wildlife Service Science for native species and should be balanced with Support Program to examine the impact of careful monitoring of vertebrate communities. Phragmites invasion in tidal marshes.