Environmental Threats to Tidal-Marsh Vertebrates of the San Francisco Bay Estuary

Studies in Avian Biology No. 32:176–197

ENVIRONMENTAL THREATS TO TIDAL-MARSH VERTEBRATES OF THE SAN FRANCISCO BAY ESTUARY

JOHN Y. TAKEKAWA, ISA WOO, HILDIE SPAUTZ, NADAV NUR, J. LETITIA GRENIER, KARL MALAMUD-ROAM, J. CULLY NORDBY, ANDREW N. COHEN, FRANCES MALAMUD-ROAM, AND SUSAN E. WAINWRIGHT-DE LA CRUZ

Abstract. The San Francisco Bay and delta system comprises the largest estuary along the Paciﬁ c Coast of the Americas and the largest remaining area for tidal-marsh vertebrates, yet tidal marshes have been dramatically altered since the middle of the 19th century. Although recent efforts to restore ecological functions are notable, numerous threats to both endemic and widespread marsh organisms, including habitat loss, are still present. The historic extent of wetlands in the estuary included 2,200 km2 of tidal marshes, of which only 21% remain, but these tidal marshes comprise >90% of all remaining tidal marshes in California. In this paper, we present the most prominent environmental threats to tidal- marsh vertebrates including habitat loss (fragmentation, reductions in available sediment, and sea-level rise), habitat deterioration (contaminants, water quality, and human disturbance), and competitive interactions (invasive species, predation, mosquito and other vector control, and disease). We discuss these threats in light of the hundreds of proposed and ongoing projects to restore wetlands in the estuary and suggest research needs to support future decisions on restoration planning.

Key Words: Contaminants, disease, fragmentation, San Francisco Bay, sea-level rise, sediment supply, threats, tidal marsh, water quality, wetlands.

AMENAZAS AMBIENTALES PARA VERTEBRADOS DE MARISMA DE MAREA DEL ESTUARIO DE LA BAHÍA DE SAN FRANCISCO Resumen. La Bahía de San Francisco y el sistema delta abarcan el estuario más grande a lo largo de la Costa Pacíﬁ co de las Américas y el área mas larga que aun queda para vertebrados de marisma de mar, a pesar de que los marismas de marea han sido dramáticamente alterados desde mediados del siglo 19. A pesar de que los esfuerzos recientes para restaurar las funciones ecológicas son notables, numerosas amenazas para ambos organismos de marea, endémicos y amplios, incluyendo pérdida del hábitat, están aun presentes. El alcance histórico de humedales en el estuario incluyeron 2,200 km2 de marismas de marea, de los cuales solo el 21% permaneció, pero estos marismas de marea compren- den >90% de todos los marismas de marea que quedan en California. En este artículo, presentamos las amenazas ambientales más prominentes para los vertebrados de marisma de marea, incluyendo pérdida del hábitat (fragmentación, reducciones en el sedimento disponible, y aumento en el nivel del mar), deterioro del hábitat (contaminantes, calidad del agua, y disturbios humanos), e interac- ciones competitivas (especies invasoras, depredación, mosquitos y otro control vector, y enfermedes). Discutimos estas amenazas a luz de cientos de proyectos propuestos y llevados a cabo para restaurar humedales en el estuario, y las necesidades sugeridas por estudios para apoyar futuras decisiones en la planeación para la restauración.

Coastal and estuarine wetlands are resources al. 1986), and supporting 162 km2 of remaining of global importance to humans and wildlife, tidal marshes. but they encompass <3% of the land surface in Saltmarsh plant communities along the the Western Hemisphere and only 0.3% of the California coast often form mosaic patches contiguous US (Tiner 1984). The most extensive and are dominated by common pickleweed regions of tidal marsh in the coterminous US are (Salicornia virginica, syn. Sarocornia pacifi ca) and found along the Gulf Coast (9,880 km2), southern Pacifi c cordgrass (Spartina foliosa). Common Atlantic Coast (2,750 km2), mid-Atlantic Coast pickleweed occurs throughout the East and (1,890 km2), and New England and Maritime West coasts of the US (U.S. Department of Coast (360 km2; Greenberg and Maldonado, Agriculture 2003), but Pacifi c cordgrass is tra- this volume). In contrast, much less tidal marsh ditionally found along the California coast from is located on the West Coast, of which the larg- Bodega Bay (though it has been introduced to est extent is found in the San Francisco Bay and Del Norte County) in the north, to San Diego delta (SFBD; Fig. 1). SFBD is the largest estuary County in the south (Calfl ora 2003), extending (4,140 km2) on the Pacifi c Coast of the Americas, into the Baja Peninsula of Mexico. In SFBD, rela- encompassing <7% of the land surface, drain- tively narrow strips (3–10 m) of Pacifi c cordgrass ing >40% (155,400 km2) of California (Nichols et occur between mean tide level (MTL) and mean

176 SAN FRANCISCO BAY TIDAL-MARSH THREATS—Takekawa et al. 177

FIGURE 1. Area of California drained by the San Francisco Bay estuary and the Sacramento-San Joaquin River watershed (shaded), and distribution of tidal-marsh habitat within the estuary (San Francisco Estuary Institute 1998). 178 STUDIES IN AVIAN BIOLOGY NO. 32 high water (MHW), and a wider band (up to a hillsides were scoured by hydraulic mining few kilometers) of common pickleweed ranges and mercury (Hg) was used to extract gold. from mean high water (MHW) to mean higher Roughly 389,000,000 m3 of sediment, along with high water (MHHW) (Josselyn 1983). In more Hg-laden sediments, was transported down- brackish waters of the Sacramento and San stream into the estuary from 1856–1983 (U.S. Joaquin river delta, bulrushes (Bolboschoenus Geological Survey 2003). and Schoenoplectus spp.) are dominant. With The SFBD was home to over half of the the extensive losses of coastal wetlands in state’s population by 1860, and the popula- California, the SFBD now supports 90% of the tion has steadily increased (Fig. 2). Population remaining tidal wetlands (MacDonald 1977). growth also stimulated rapid development and urbanization (Figs. 3, 4). Legislation for land HUMAN DEVELOPMENT reclamation (federal Arkansas Act of 1850, state Green Act of 1850) was enacted to encourage The abundance of wildlife resources attracted conversion of grasslands and wetlands to farm- the fi rst humans to the estuary roughly 10,000 yr lands, and by the 1870s a network of levees had ago. Hunter-gatherer societies approached been constructed to protect low-lying fi elds. 25,000 inhabitants but likely posed little threat The deepwater harbor became a major ship- to most tidal-marsh wildlife (San Francisco ping center, and by 1869, completion of the Estuary Project 1991). In the last 200 yr, hunters transcontinental railroad increased movement and traders were attracted to the estuary by the of food and goods from the region. Striped bass abundant wildlife from as far away as Russia, (Morone saxatilis) were intentionally introduced causing the fi rst notable decline of fur-bearing for a commercial fi shery in 1879. populations including sea otter (Enhydra lutra) In the early 1900s, urban runoff polluted the and beaver (Castor canadensis) (San Francisco bays, and thousands of hectares of wetlands Estuary Project 1991). Spanish inhabitants set were fi lled for development. Tidal wetlands up missions and began grazing cattle and sheep in the South Bay were replaced by >5,000 ha in the 18th and 19th centuries. A rapid infl ux of salt-evaporation ponds by the 1930s (Siegel of humans occurred in 1848 when gold was and Bachand 2002). Dams and diversions on discovered in the Sierra Nevada Mountains. nearly every tributary prevented fi sh from Within 2 yr, the city of San Francisco grew spawning upstream, limited sediment trans- from 400–25,000 individuals. Sierra Nevada port downstream, and reduced freshwater

FIGURE 2. Increased human population in the San Francisco Bay estuary and delta from 1860–1990 (Bell et al. 1995). SAN FRANCISCO BAY TIDAL-MARSH THREATS—Takekawa et al. 179

FIGURE 3. Expanding urban areas (in square kilometers) of the San Francisco Bay estuary and delta from 1850–1990 (Bell et al. 1995).

FIGURE 4. Distribution and changes in extent of development in the San Francisco Bay estuary from 1850, 1900, 1950, and 1990 (Pereira et al. 1999). 180 STUDIES IN AVIAN BIOLOGY NO. 32

fl ows. The thousands of workers attracted In this paper, we summarize the major to the area during World War II stimulated threats to tidal-marsh vertebrates including a housing and construction boom. Farmers habitat loss (habitat fragmentation, sediment became reliant on chemical practices increas- availability, and sea-level rise), deterioration ing toxic runoff to estuary waters. The human (contaminants, water quality, and human dis- population was nearly 6,500,000 in 1980, but it turbance), and competitive interactions (inva- increased 34% to 8,700,000 by the end of the sive species, predation, mosquito control, and century (U.S. Census Bureau 2000). The his- disease). We describe how these threats affect toric (pre-1850) saltmarshes covered 2,200 km2 tidal-marsh vertebrates, where proposed res- or twice the extent of open water (Atwater et toration projects may ameliorate their effects, al. 1979), but the modern landscape has been and what studies would be helpful to support altered by loss of 79% of the saltmarshes, 42% restoration planning. of the tidal fl ats, and construction of almost 14,000 ha of artifi cial salt-evaporation ponds HABITAT FRAGMENTATION (San Francisco Bay Area Wetlands Ecosystem Goals Project 1999). Habitat fragmentation results from changing large continuous areas to a pattern of smaller, THE MODERN ESTUARY more isolated patches of less total area within a matrix of altered habitats (Wilcove et al. 1986). The diversity of the SFBD wetlands includes Fragmentation usually also involves an increase freshwater marshes in the eastern delta and in mean patch perimeter-to-area ratio and upstream tributaries, brackish marshes in changes in patch confi guration. Though habitat Suisun Bay and western delta, and saltmarshes loss contributes directly to population decline, (Fig. 1) in the South Bay, Central Bay, and San edge effects and habitat isolation may cause Pablo Bay (Harvey et al. 1992). That diversity further reductions by an impact on dispersal supports 120 fi shes, 255 birds, 81 mammals, and altering the ecological function within 30 reptiles, and 14 amphibian species in the patches, especially near habitat edges (Andren estuary, including 51 endangered plants and 1994). These impacts may increase with time animals (Harvey et al. 1992). SFBD is a Western since isolation. Hemisphere Shorebird Reserve Network site In the SFBD, loss of tidal-marsh habitat has of hemispheric importance used by >1,000,000 resulted in many remnant marshes that are shorebirds, supports >50% of wintering div- small and isolated from other marshes (San ing duck species counted on the Pacifi c fl yway Francisco Bay Area Wetlands Ecosystem Goals of the US during the midwinter, and is home Project 1999, San Francisco Estuary Institute to one of the largest wintering populations 2000). Roads, levees, urban development, and of Canvasbacks (Aythya valisineria) in North non-native vegetation have replaced upland America (Accurso 1992). edges and transition zones (Table 1), much to the Wetland conservation in the SFBD has detriment of animals that rely on upland areas evolved from slowing loss to preserving rem- as high-tide refugia. Many modern wetlands are nant wetlands to aggressively restoring areas. By 2002, 29 wetland restoration projects had TABLE 1. HISTORIC (PRE-1820) AND PRESENT DISTRIBUTION OF been completed in the North Bay alone (S. TIDAL-MARSH PATCHES IN THE SAN FRANCISCO BAY ESTUARY Siegel, unpubl. data); however, these restora- (DERIVED FROM THE SAN FRANCISCO BAY ECOATLAS 1998). tion projects were relatively small (≤12 ha). Large restoration projects, hundreds to thou- Number of tidal and muted sands of hectares in size, have been proposed or marsh patches a initiated recently that will signifi cantly change Patch size Historic habitat Present habitat the regional landscape. Nearly 6,500 ha of salt (hectares) (pre-1800) (1990s) ponds in the South Bay and 4,600 ha in the North Bay have been acquired (San Francisco <10 190 370 10–50 46 97 Bay Area Wetlands Ecosystem Goals Project 51–100 21 23 1999, Steere and Schaefer 2001, Siegel and 101–500 33 38 Bachand 2002). Some areas will be managed 501–1,000 18 3 as ponds to provide habitat for thousands of 1,001–3,000 20 1 shorebirds and waterfowl, but many areas are >3,000 4 0 proposed for conversion to tidal marshes for Total area of tidal 77,530 16,996 the benefi t of tidal-marsh species. Despite these marsh wetland restoration efforts, tidal-marsh verte- a Patches derived by merging adjacent tidal (old and new) and muted marsh polygons. Muted marshes include wetlands without fully tidal brates still face many threats in the estuary. fl ows. SAN FRANCISCO BAY TIDAL-MARSH THREATS—Takekawa et al. 181 patchy, linear or irregularly confi gured, confi ned species, respond to marsh size, confi guration, to edges of large tidal creeks or sloughs, or on isolation, and other landscape-scale factors as the bayside edges of levees where new marshes well as to local-scale factors such as vegeta- formed over sediments accreted during the tion composition and structure. However, the 1850s. Levee networks, tide-control structures, underlying processes contributing to these pat- and mosquito ditches not only fragmented wild- terns, along with dispersal patterns across the life habitat but also altered the hydrology and estuary, are not well understood. sediment dispersion patterns of outboard levee areas (Hood 2004). Although regulatory laws RESTORATION CONCERNS may prevent future losses of wetlands, increased urbanization and loss of native vegetation cor- Several large-scale wetland restoration ridors may decrease the viability of populations projects are underway, including the Napa- (Andren 1994). Tidal-marsh species may be Sonoma Marsh on San Pablo Bay (4,050 ha) differentially affected depending on their level and the former Cargill salt ponds in the South of habitat specialization (Andren 1994) and the Bay (6,475 ha; Siegel and Bachand 2002). These scale of landscape heterogeneity to which they projects involve restoration of areas that were are sensitive (i.e., patch sensitivity, sensu Kotliar historically tidal, but were converted to salt- and Wiens 1990, Wiens 1994, Riitters et al. 1997, evaporation ponds. The goals of these resto- Haig et al. 1998). ration projects are to provide large areas of Habitat fragmentation may act as an isolating contiguous habitat, increasing marsh area with mechanism resulting in higher extinction rates minimal fragmentation. Previous restoration and lower colonization rates, lower species rich- projects in the estuary have been relatively ness (MacArthur and Wilson 1963, 1967), higher small and opportunistic with limited study of nest-predation rates (Chalfoun et al. 2002), higher restoration effects at the landscape scale. nest parasitization rates, and changes in ecologi- One of the most important considerations cal processes (Saunders et al. 1991). Numbers in restoration is determining the optimum of avian marsh species in the prairie pothole confi guration and size of the project. A single region have been shown to vary with patch size large expanse of habitat may be preferable and perimeter-to-area ratio as well as with veg- for some species rather than several smaller, etation and other local-scale factors (Brown and isolated habitat patches (particularly for those Dinsmore 1986, Fairbairn and Dinsmore 2001). species requiring large territories). However, The effects of fragmentation were also found to a population in a single large patch may be be variable, species-specifi c, and context-spe- more vulnerable to extinction because of demo- cifi c in a range of other habitat types (Bolger et graphic stochasticity or catastrophic events al. 1997, Bergin et al. 2000, Chalfoun et al. 2002, (such as fi res and disease; Carroll 1992). Highly Tewksbury et al. 2002). vagile species, including most birds, are gener- Few studies have been done to assess the ally better able to disperse between isolated impact of fragmentation on tidal-marsh birds in habitat patches than small mammals, reptiles, the SFBD. Scollon (1993) mapped marsh patches and amphibians. However, radio-telemetry and dispersal corridors for the Suisun Song studies of the endangered California Clapper Sparrow (Melospiza melodia maxillaris) based on Rail (Rallus longirostrus obsoletus) indicate low theoretical estimates of dispersal distance and rates of movement between and within seasons published population densities. Varying frag- (Albertson 1995) and habitat fragmentation mentation effects on population size have also is considered one of the main threats to the been predicted for San Pablo Song Sparrows persistence of this subspecies (Albertson and (M. m. samuelis; Scollon 1993; Takekawa et al. Evens 2000). chapter 16, this volume). The dispersal of San Pablo Song Sparrows from one fragmented REDUCTION IN SEDIMENT AVAILABILITY marsh to another is thought to be rare since dispersal distance averages 180 m (Johnston Sediment deposition and tidal actions are 1956a). Thus, populations in smaller, more iso- the dynamic processes that sustain tidal-marsh lated fragments were more susceptible to local wetlands. Rapid sediment accretion of tidal extinction. Recent empirical studies (Spautz et marshes in the SFBD extended for at least 20 yr al., this volume; Point Reyes Bird Observatory, after the start of hydraulic gold mining in unpubl. data) have found that San Pablo, the Sierra Nevada Mountains. San Pablo Bay Suisun, and Alameda (M. m. pusillula) Song received 300,000,000 m3 of sediment, and by Sparrows, all California species of special con- 1887 created 64.74 km2 of new mudfl ats (Jaffe et cern, and California Black Rails (Laterallus jamai- al. 1998). The concentration of suspended sedi- censis coturniculus), a California state threatened ments in the delta declined 50% by the 1950s 182 STUDIES IN AVIAN BIOLOGY NO. 32 with the cessation of hydraulic mining and Coarser substrates may have a decreased ability the advent of dams on major tributaries from to retain nutrients (Boyer and Zedler 1998) and downstream reaches (Wright and Schoellhamer may fail to support the vegetation structure and 2004). San Pablo Bay lost 7,000,000 m3 of sedi- height required for target species (Zedler 1993). ment from 1951–1983 or an annual loss of 0.36 Despite the concerns of amending soils with km2 of mudfl ats (Jaffe et al. 1998). dredge material, data in relation to its use in Many wetland restoration project sites have restorations are scarce. In the Sonoma Baylands subsided and require substantial sediment input restoration project, dredge materials were used to reach adequate levels for plant establishment. to accelerate the restoration process. However, When sedimentation rates were studied in the development of channels and vegetation has south San Francisco Bay (Patrick and DeLaune been slow, presumably because of limited tidal 1990), one site (Alviso) was found to have sub- exchange. Additional studies would provide sided by >1 m based on records from 1934–1967 more detailed analyses of the benefi ts of dredge because of groundwater extraction (Patrick and materials against the costs. DeLaune 1990). The large number of restoration projects occurring simultaneously may reduce SEA-LEVEL RISE predicted sediment availability (San Francisco Bay Area Wetlands Ecosystem Goals Project Projections for future sea-level rise in the 1999). A shortage of sediment may result in SFBD vary between 30 and 90 cm in the 21st a reduced turbidity, increased erosion, and a century, depending upon which climate pro- greater loss of mudfl at and intertidal habitats jection models are used (Dettinger et al. 2003). (Jaffe et al. 1998). An estimated 10–20 cm of that rise is expected, Dredge material has been proposed for regardless of anthropogenic global-warming projects where natural sediment supply is effects based on the historic rate of 20 cm/100 yr inadequate. Annual yields from dredging seen during the course of the 20th century operations produce an average of 6,120,000 m3 (Ryan et al. 1999). The remainder is due to the of sediment in the estuary (Gahagan and Bryant combined infl uence of thermal expansion as the Associates et al. 1994). At current rates of sedi- ocean warms in response to global warming, ment accretion, it would take 10–15 yr to raise and accelerated melting of glaciers and ice caps. elevations one meter in South Bay salt ponds Galbraith et al. (2002) predicted a conversion of (although actual rates will vary by pond) to a 39% of the intertidal habitat in San Francisco height appropriate for vegetation colonization Bay to subtidal habitat by 2100, as high as 70% (San Francisco Bay Area Wetlands Ecosystem in the South Bay. In addition to the projected Goals Project 1999). In recent years, regula- long-term rise in global sea level, considerable tory agencies have included the potential use short-term variability can be expected due to of dredge material for restoration to accelerate local factors including tides, increased storm the process of restoration (U.S. Army Corps of surges and changes in upwelling along the Engineers 1987). coast of California, all of which act on a range of mechanisms and timescales (Table 2). The RESTORATION CONCERNS SFBD tidal marshes and their fl ora and fauna now face potentially severe threats associated Many contaminants such as Hg and PCBs with sea-level rise, because the magnitude of are tightly bound to sediment particles. As a change and the accelerated rate of rise over the result, the transportation of contaminants is next few decades, and because human activities closely tied to the movement of sediments. around the marshes have probably dramatically Use of dredge material in restoration projects reduced the marshes’ capacity for coping with has the potential to transport and reintroduce sea-level changes. buried contaminants to the soil surface where it may be biologically available (Schoellhamer POTENTIAL IMPACTS et al. 2003). Another concern of dredge material use is the potential incompatibility with Actual sea-level rise can be seen in the surrounding substrate conditions. Dredge tidal data at the Golden Gate from 1897–1999 material may not complement the fi ne particle (Fig. 5), which have changed at different rates size of naturally occurring tidal wetlands, and (Malamud-Roam 2000). For example, while soils may not support vigorous plant growth mean sea level has increased by about 20 cm (Zedler 2001). Dredge spoils may have coarser during the 20th century, the height of mean soils and less clay content (Lindau and Hossner higher high water (MHHW) and the high- 1981), and consequently less soil organic mat- est highs have increased by about 25 cm and ter and microbial activity (Langis et al. 1991). 28 cm per century, respectively (Malamud-Roam SAN FRANCISCO BAY TIDAL-MARSH THREATS—Takekawa et al. 183 s or . ope in ESTUARY

ative rise in the AY B RANCISCO F AN S THE

RISE

ooding, temperature, and light. LEVEL ooding will probably increase. - SEA

n oscillation. ers of each tidal day observed over the National Tidal Datum Epoch); DLQ = high water will increase. hardened. amplitude; impact likely minor. over-marsh ﬂ global warming. HYDRODYNAMICS

VARIATION

EFFECTS

ooding occurs in winter and Photosynthesis, predation, and dispersal may all respond to ooding in the summer, varying this pattern of ﬂ OVERALL

AND , PATTERNS

TEMPORAL , RANGE

VERTICAL , ECHANISMS 2. M : The abbreviations include: DHQ = mean diurnal high water inequality (one-half the average difference between two wat ABLE slowing about about oscillation Secular (multi-year) climatic cycles slowing 100 m increase over Holocene, Eustatic (absolute) rise ~ 1 mm/yr; relative Global warming could increase the rate of rel (local) rise ~ 2 mm/yr 7,000 yr ago. year. Duration of rise in Estuary 6–12 mo accelerates, the frequency, duration, and height of extreme bay to 3–9 mm/yr for the next 100 yr. Probably solar annual cycles, possibly beat frequencies 10–20 cm El Nino/southern 15–30 cm rise MSL has annual cycle, low in April and Global warming induced changes in oceanic circulation may Episodic, occurring about every seventh ENSO frequency apparently increasing. If this continue high in September and MHHW has semi- change seasonal tidal mean patterns, but it is not clear how. annual cycle, high in January and July slightly over a 18.6-yr period T Mechanism Wind waves, swash, and run-up Seiches <1–60 cm Vertical range Tides ﬂ Oscillatory; frequency in seconds Storm surge Temporal pattern Lunar modulation Few centimeters of the tides Large waves may become more frequent if storms increase; Oscillatory; frequency in minutes 1–2 m Annual cycles of solar 10s of centimeters Range ~ 120 cm; radiation night Monthly MSL Episodic; winter Precession of lunar node Dominant (spring/neap) = 14.6 d; Increasing estuary depth may increase the frequency or HW height ~ 60 cm; Effect None Minor (lunar declination) = 13.6 d Annual cycle of rainfall and runoff Oscillatory; dominant frequencies Dredging can alter the fortnightly circulation patterns ~ 10 cm; Monthly DHQ and DLQ ~ 30 cm induced by lunar tidal modulation. Global warming could change amount and timing of If tidal range continues to increase, the height of HW relative (ppt high in winter, runoff spring, precipitation and therefore runoff water surface sl 12.5 and 25 hr Daylight ﬂ MHHW ~20 cm Frequency and amplitude will probably increase due to both low in fall) Note Swash and run-up may increase where the shoreline is to MSL will increase, and the frequency mean depth of the estuary. mean diurnal low-water inequality; HW = high water; MSL sea level; MHHW higher ENSO El Niño souther 184 STUDIES IN AVIAN BIOLOGY NO. 32

FIGURE 5. Rate of change in centemeters/100 yr in tidal data at the Golden Gate of the San Francisco Bay estuary from 1897–1999, including: MHHW = mean higher high water, MHW = mean high water, MLHW= mean lower high water, MSL = mean sea level, MTL = mean tide level, MHLW = mean higher low water, MLW = mean low water, MLLW = mean lower low water.

2000). Thus, for the animals living on the marsh marsh inhabitants to predation as they surface, sea level has risen by an effective 25– seek refuge upland. 28 cm in the last century. Should this pattern 3. Increased salinity intrusion into inland of a faster rate of change for the highest tides areas of the estuary can change plant continue into the next century, projections of sea- assemblages and alter habitat for wildlife, level rise may underestimate the actual rise in sea such as the Common Moorhen (Gallinula level for the animals living in the tidal marshes. chloropus). Furthermore, most global-warming scenarios for this region predict an increased frequency and CHANGING SALINITY severity of storm surges due to global warming (Cubasch and Meehl 2001). As sea level rises at the Golden Gate, salinity Sea-level rise poses three major risks to wild- can be expected to intrude further up-estuary. life that occupy the surrounding tidal marshes In terms of salinity, the effect of rising sea level and mudfl ats. can be analogous to a decrease in fresh-water 1. Higher sea levels may drown the marshes infl ow, as the depth of the bay has remained and mudfl ats or increase storm surges fairly constant throughout the Holocene causing greater shoreline erosion and (Ingram et al. 1996). Vegetation patterns in the habitat loss. marshes have changed signifi cantly during the 2. Increased frequency or duration of last century, refl ecting a signifi cant increase in extreme high tides can to lead to higher salinity (May 1999, Byrne et al. 2001, Malamud- mortality of songbird eggs and young Roam 2002). For example, pollen cores from the birds during extreme high-tide fl oods Petaluma River have shown dramatic shifts (Erwin et al., this volume; Point Reyes Bird from brackish waters (characterized by tules) Observatory, unpubl. data). In addition around 1,800 yr ago to saline conditions (char- amplifi ed high tides may increase the vul- acterized by pickleweed) around 1,400–800 nerability of the salt marsh harvest mouse years ago, and back again to brackish conditions (Reithrodontomys raviventris), California around 750 yr ago (Byrne et al. 2001). Climate Clapper Rails (Rallus longirostris obsol- explains only a part of the change in estuarine etus), California Black Rails, and other salinity; water diversion for agriculture and SAN FRANCISCO BAY TIDAL-MARSH THREATS—Takekawa et al. 185 human consumption explains the majority of associated contaminants including mercury the change (Peterson et al. 1995). (Hg), selenium (Se), and polychlorinated As sea level rises, the tidal-marsh habitats biphenyls (PCBs) are listed as priority pollut- (fresh, brackish, and salt) can be expected to ants in SFBD under section 303(d) of the Clean change as vegetation responds to higher salin- Water Act. Sediment-bound contaminants or ity conditions. For some organisms, the changes those re-suspended by dredging operations may be favorable, with a greater area covered or changes in sedimentation dynamics pose by saltmarsh habitat. Other organisms that rely a potential threat to tidal-marsh vertebrates on fresh water habitat, such as the Common when transported into wetlands. Although Moorhen, may face signifi cant loss of habitat, numerous other pollutants occur in the bay, we and a shift up-estuary in their range as the focus on those suspected to have direct bearing salt wedge approaches the delta. Strategies to on tidal marsh health and food webs. compensate for the increase in marine infl uence One of the most prevalent contaminants is due to sea-level rise could include increasing Hg. Mercury extracted from the Coast Range the area of marsh habitat within the delta or was used to recover gold and silver in Sierra increased fresh water fl ows through the delta Range mining operations during the Gold Rush (requiring a decline in water diversion). era (Alpers and Hunerlach 2000). Between 1955 and 1990, the SFBD received an average RESTORATION CONCERNS of 6,030,000 m3 of Hg-laden sediments annually (Krone 1996). Methylmercury (MeHg), a A critical synergistic threat facing marsh more toxic and readily bioaccumulated form inhabitants is the combined force of a sediment (Marvin-DiPasquale et al. 2003), is magnifi ed by defi cit and sea-level rise. The modern supply of passing through food webs with transformation sediments to the estuary has been signifi cantly initiated in phytoplankton. Acute toxicity in altered in historic times due to human modifi ca- fi sh, birds, and mammals damages the central tions of the hydrologic system. However, future nervous system (Wolfe et al. 1998, Wiener et al. predictions based on historical rates of rise, and 2003), while lower-level exposure affects repro- potential increases resulting from global warm- duction in vertebrates (Wiener and Spry 1996, ing, suggest signifi cant losses of saltmarsh Wolfe et al. 1998). habitat in the South Bay and more signifi cant Human-health advisories for fi sh have been losses of tidal mudfl ats, which are critical for in effect since 1970, and Hg concentrations in shorebird species (Galbraith et al. 2002). striped bass have not changed signifi cantly The North Bay, however, does not have a since that time (Fairey et al. 1997, Davis et history of high subsidence as in the South Bay, al. 2002). Highest liver Hg concentrations for though it is largely unknown whether the rates small mammals have been found in South Bay of sediment supply will be adequate to main- wetlands (Clark et al. 1992). A recent study doc- tain marsh surface elevations in the future. umented that eggs of Forster’s Tern (Sterna for- Studies of long-term rates of sediment accretion steri) and Caspian Tern (Sterna caspia) foraging in marshes in San Pablo Bay, the Carquinez in South Bay salt ponds and adjacent sloughs Strait, and Suisun Bay indicate that for the last contained the highest concentrations (7.3 and 3,000 yr the average rates of sediment accretion 4.7 mg/kg total Hg) of any bird species in the have matched sea-level rise (Goman and Wells estuary (Schwarzbach and Adelsbach 2002). 2000, Byrne et al. 2001, Malamud-Roam 2002). Endangered California Clapper Rails exhibited However, the long-term records also indicate depressed hatchability and embryo deformities that there were periods when sediment sup- with egg concentrations exceeding the lowest ply was clearly inadequate to maintain the observed adverse effect concentration (LOAEC) marshes, resulting in a conversion of some areas of 0.5 µg/g (Schwarzbach et al. 2006; Novak et to subtidal conditions (Goman and Wells 2000, al., this volume). Malamud-Roam 2002). Selenium (Se) is another persistent threat to tidal-wetland vertebrates. Se is substituted for CONTAMINANTS sulfur in enzymes, resulting in reproductive failure and teratogenesis. Sources include oil The SFBD has numerous sources of pollu- refi nery effl uent and agricultural drainwater. tion, many of which are particle-bound, and Selenium loads from refi neries have decreased their concentrations fl uctuate with suspended (6.8 kg/d–1.4 kg/d) with recent regulation sediment concentrations (Schoellhamer et al. (San Francisco Bay Regional Water Quality 2003). Most restoration projects in the estu- Control Board 1992), but agricultural sources ary are dependent on sediment inputs to range between 20.4–53.2 kg/d (Luoma and elevate marsh plains; however, three sediment- Presser 2000). Dissolved Se concentrations are 186 STUDIES IN AVIAN BIOLOGY NO. 32 consistently <1 µg/L (Cutter and San Diego- liver damage (Hoffman et al. 1996). Failed McGlone 1990, San Francisco Estuary Institute California Clapper Rail eggs collected during 2003), below chronic criterion for aquatic life 1992 in the South Bay had total PCB concentra- (2 µg/L). However, Se concentrations may be tions between 0.65–5.01 µg g-1 (Schwarzbach elevated because chemical speciation controls et al. 2001). Rail sensitivity to PCB congeners bioaccumulation (Luoma and Presser 2000). For is not known, but the authors indicate such example, Black-necked Stilts (Himantopus mexi- concentrations may cause reduced hatching canus) nesting at Chevron Marsh in Richmond, success. California, had eggs (20–30 µg/g dw) with simi- Section 303(d) of the Clean Water Act lists lar concentrations found at Kesterson Reservoir elements copper (Cu) and nickel (Ni), as well (25–37 µg/g dw). Se in source water was 10% of as organic contaminants DDT, chlordanes, the concentrations at Kesterson, but uptake was dieldrin, dioxins, and furans as priority pol- enhanced because the form of Se was selenite, lutants. In addition, sediment monitoring the most bioavailable species. showed that arsenic (As) and chromium (Cr) Despite reduction in Se since 1998, concen- consistently exceeded guidelines at several trations in sturgeon and diving ducks remained sites (San Francisco Estuary Institute 2003). elevated, possibly because of invasion of the Sediment toxicity could threaten tidal-marsh Asian clam (Potamacorbula amurensis), a spe- vertebrates by decreasing their invertebrate cies that concentrates Se more than other clams prey base (San Francisco Estuary Institute 2003). (White et al. 1989, Linville et al. 2002, San Polybrominated diphenyl ethers (PBDEs), used Francisco Estuary Institute 2003). In a recent as fl ame retardants, have been increasing in survey of wetland birds, Se concentrations in sediment and biota, and Caspian Tern eggs eggs ranged from 1.5 µg/g dw in Snowy Plover from SFBD contain the highest concentrations (Charadrius alexandrinus) eggs to 4.2 µg/g dw found in any bird species (T. Adelsbach, USDI in Snowy Egret (Egretta thula) eggs, and Fish and Wildlife Service, pers. comm.). Tributyl 1.6 µg/g dw in failed California Clapper Rail tin (TBT), an anti-fouling additive in paint, is an eggs (Schwarzbach and Adelsbach 2002) com- endocrine disrupter that may pose a threat to pared with a threshold for reproductive prob- wetland organisms by increasing through wet- lems of 6–10 µg/g dw (Heinz 1996). In birds, land food webs (Pereira et al. 1999). Hg and Se are toxicologically antagonistic, Oil spills remain a threat, because California where exposure to one of these elements pro- is the fourth largest oil-producing state and tects individuals from the toxic effects of the the third largest crude-oil-refi ning state in the other (El-Begearmi et al. 1997); however, Heinz nation. Six oil refi neries are located in the bay and Hoffman (1998) showed that the most envi- and comprise nearly 40% of the state’s total ronmentally realistic and most toxic forms of oil production capacity (California Energy mercury (methylmercury) and selenium (sele- Commission 2003). More than a thousand nomethionine) combined caused lower hatch- tanker ships pass through the estuary each ing success in Mallards (Anas platyrhynchos) year, along with countless container ships, rec- than either contaminant alone. reational boats, and other vessels. The tanker Polychlorinated biphenyls (PCBs), synthetic Puerto Rican spilled 5,678,118 l of oil in 1984, chlorinated aromatic hydrocarbons with a wide killing approximately 5,000 birds, while a Shell variety of industrial uses, have been banned in Oil storage tank spilled 1,589,873 l in 1988. the US since 1979. However, PCB concentrations Over an extended period of time (from at least in SFBD water remain high, and 83% of samples 1992–2002, though likely for many years ear- taken during 2001 by Regional Monitoring lier), heavy fuel oil leaked out of the freighter Program (San Francisco Estuary Institute 2003) S.S. Jacob Luckenbach that sank southwest of exceeded water quality objectives. Sources the Golden Gate Bridge in 1953 (Hampton et al. to the estuary cannot be pinpointed, but are 2003). Roughly 378,541 l of oil were removed thought to be mainly historical. Runoff from from the wreck, but over 18,000 bird mortali- creeks and tributaries has been identifi ed as ties were estimated in 1997 and 1998 from this one signifi cant source of PCBs (San Francisco one source (Hampton et al. 2003). In 2002, of Estuary Institute 2003). 6,867 oil spills reported in California, 445 were PCBs are accumulated in fat and readily in the SFBD (Offi ce of Spill Prevention and increase from trophic level to trophic level in Response 2003). estuarine food webs. Although PCB toxicity depends on the structure of individual conge- RESTORATION CONCERNS ners, general effects include thymic atrophy, immunotoxic effects, endocrine disruption, Projects in the bay that re-expose, accrete, reproductive impairment, porphyria, and or use dredged Hg-laden sediments may pose SAN FRANCISCO BAY TIDAL-MARSH THREATS—Takekawa et al. 187 risks to fi sh and wildlife. This is especially due to water diversion for agriculture, munici- true in tidal marshes where sulfate-reducing pal use, and local consumption (Nichols et al. bacteria in anoxic sediments can transform 1986). Pollen and carbon isotope data from inorganic Hg to MeHg. Boundary zones of sediment cores from Rush Ranch in Suisun Bay oxic-anoxic areas in marshes have particularly clearly indicate a shift since 1930 in the domi- high MeHg production. Site-specifi c param- nance of marsh vegetation from freshwater to eters such as dissolved or organic carbon salt-tolerant plants as a response to increased content, salinity, sulfate, and redox cycles salinity due to upstream storage and water also infl uence biotransformation rates and diversion. This recent shift in the vegetative subsequent bioaccumulation in tidal wetland community was as extreme as the shift that organisms (Barkay et al. 1997, Kelly et al. 1997, occurred during the most severe drought of Gilmour et al. 1998). the past 3,000 yr (Byrne et al. 2001). When freshwater fl ows are reduced, salt WATER QUALITY intrusion up-estuary can become a problem (Nichols et al. 1986), particularly in high- Water quality in the SFBD has changed dra- marsh areas that already become hyper-saline matically due to human activity during the past during certain times of the year. For example, 150 yr, often to the detriment of estuarine eco- historically large populations of the endan- systems. The impact of water quality change on gered San Francisco garter snake (Thamnophis tidal-marsh terrestrial vertebrates is virtually sirtalis tetrataenia), have declined presumably unstudied but could be severe if left unman- due to the loss of several prey species from aged. In this section, we discuss water-quality saltwater intrusion into less saline marsh threats exclusive of toxic contaminant and sedi- habitats (San Francisco Bay Area Wetlands ment budget issues addressed earlier. Changes Ecosystem Goals Project 2000). Salt intrusion is in water quality probably affect vertebrate a cause for concern particularly for tidal-marsh populations indirectly via long-term changes restoration projects in the Delta. When levees in the vegetative and invertebrate communities are breached, the total tidal prism increases, that inhabit the marsh and short-term cascading which allows saline waters to travel farther trophic effects starting with aquatic organisms up-estuary during high tides. Change in salin- low in the food web. ity regimes may also facilitate the proliferation The SFBD is a highly variable environ- of invasive species. The Asian clam gained ment commonly experiencing both wet and a foothold in Suisun Bay during a period of dry extremes within a year as well as wet and extreme salinity changes, with far-reaching dry years. Humans have changed the timing consequences (Nichols et al. 1990). and extent of these fl uctuations as well as Marsh vertebrates may be adapted to con- the chemistry of waters entering the estuary. ditions that are specifi c to particular salinity Under current management regimes, river ranges. For example, adaptations to kidney infl ow and freshwater diversions can vary structures in the salt marsh harvest mouse may between years by as much as 25 times (Jassby et have allowed this species to use saline environ- al. 1995). While native organisms are adapted ments (MacMillen 1964). Zetterquist (1977) to the seasonal fl uctuations, recent and future trapped the greatest numbers of salt marsh extremes may tax their survival and reproduc- harvest mice in highly saline tidal marshes. tive capabilities. Whether the high density of harvest mice in these marshes was due to an affi nity for high SALINITY salinity or lack of competitors is not known. In addition, Song Sparrow subspecies in San Salinity, a key variable determining tidal- Pablo Bay saltmarshes and Suisun brackish marsh vegetative and invertebrate commu- marshes have different bill sizes and plum- nity composition, has changed signifi cantly age colors that may refl ect adaptation to local in recent decades. A decrease in salinity can conditions (Marshall 1948). Because of the close convert a saltmarsh to fresh-water marsh, as in association between vertebrate subspecies and the tidal wetlands near the San Jose and Santa tidal marshes of a specifi c salinity, long-term Clara Water Pollution Control Plant, where salinity changes could affect both the persis- up to 567,811,800 l of treated fresh wastewa- tence and the evolution of endemic tidal-marsh ter empty into south San Francisco Bay every vertebrates. Changes in salinity can affect the day. Despite fresh-water fl ows from treat- distribution of invertebrates as well, such as the ment plants, overall salinity in the estuary has winter salt marsh mosquito (Aedes squamiger), increased because of a reduction in fresh-water which is found not in freshwater, but in brack- fl ows from the delta to 40% of historical levels, ish and saline habitats. 188 STUDIES IN AVIAN BIOLOGY NO. 32

PRODUCTIVITY productivity at optimal levels in the Suisun Bay. The Environmental Protection Agency Water management can greatly affect pri- suggested guideline maximum salinity levels mary productivity in the estuarine waters, for sensitive areas of the estuary. Investigators which likely impacts marsh vertebrates that for- at the U. S. Geological Survey, California age extensively on invertebrates and fi sh in the Department of Water Resources, Interagency tidal channels. Phytoplankton forms the base of Ecological Program, and Stanford University the food web that includes most of the wildlife continue to monitor and model water quality in in the estuary (Sobczak et al. 2002). Sometimes, the estuary as well as conduct original research productivity is low enough to affect fi sh growth experiments. and mortality, which may suggest that other vertebrates become food-limited as well (Jassby HUMAN DISTURBANCE and Cloern 2000). Benefi cial phytoplankton blooms occur when the null zone (the location Residents and visitors are drawn to the where freshwater outfl ow balances saltwater waters, shorelines, and wetlands in this highly infl ow) is positioned across the broad shallows urbanized estuary for aesthetic and recreational of Suisun Bay (Jassby et al. 1995). Management opportunities. The needs of >8,000,000 people of delta infl ows may alter the position of the in the SFBD estuary encroach upon the many null zone to facilitate optimal phytoplankton wildlife populations dependent on estuarine blooms and may be important for maintaining habitats. For example, wetlands attract large the food supply of vertebrates that forage in numbers of visitors (10,000/yr, >75% of sur- tidal channels. veyed sites) for recreational activities such as bird watching and jogging (Josselyn et al. 1989). EUTROPHICATION Although natural disturbances may create areas used by birds (Brawn et al. 2001), anthropogenic Wastewater and runoff comprise a signifi cant disturbances that elicit a metabolic or behavior percentage of freshwater entering the estuary response (Morton 1995) generally reduce the (Nichols et al. 1986). The main threat from these value of habitats for birds (Josselyn et al. 1989). waters is toxic contaminants, but nutrient load- ing is also a concern. In past decades, summer TYPES OF HUMAN DISTURBANCE die-offs due to eutrophication and ensuing oxygen depletion occurred in the South Bay. These People traverse marshes on vehicles, bicycles, events have not recurred following the institution and on foot—jogging or walking through the of improved sewage treatment methods in 1979, areas along roads or trails. Automobiles on although nitrogen concentrations remain high established roads may not greatly affect the (Kockelman et al. 1982). Oxygen depletion con- behavior of wildlife, but boats and aircraft may tinues to be a problem in areas of the delta when be highly disruptive by causing animals to fl ush, water retention times are long (B. Bergamaschi, exposing them to predators. The activity or noise U.S. Geological Survey, pers. comm.). of boats and personal watercraft may cause Eutrophication in the estuary is controlled avoidance or fl ight behavior (Burger 1998), espe- by two factors. First, high turbidity causes algae cially near waterways, or result in trampled veg- to be light-limited, so nuisance blooms do not etation, while wakes from boats may erode bank occur on the same scale as in other polluted bays habitats. Low-fl ying aircraft may create large (Jassby et al. 2002). Second, benthic bivalves disturbances in estuaries (Koolhaas et al. 1993), consume much of the primary production and especially near small airports or in agricultural reduce the availability of nutrients in the water areas where aerial spraying is used. Longer and column (Cloern 1982). However, turbidity may more extensive use of areas by campers, fi sher- decline in the near future due to retention of men, hunters, and researchers may have greater sediment behind dams and channel armoring individual short-term effects, while cumulative (International Ecological Program 2003). long-term effects of visitors on trails or boardwalks may effectively decrease the size of the RESTORATION CONCERNS marsh, provide pathways for predators, and degrade the value of edge habitats. Many agencies and hundreds of scientists are working to understand and manage water qual- TYPES OF EFFECTS ity, although most of the focus is on open-water habitats rather than tidal wetlands. Delta infl ow Knight and Cole (1991) described hierarchi- is currently managed to allow for movements cal levels of disturbance on wildlife species from of fi sh populations and keep summer primary death or behavior change (altered behavior, SAN FRANCISCO BAY TIDAL-MARSH THREATS—Takekawa et al. 189 altered vigor, altered productivity, and death), monitoring studies have found elevated heart- to population change (abundance, distribution, rate levels in breeding marine birds (Jungius and and demographics), and fi nally community Hirsch 1979) and wintering geese (Ackerman et alteration (species composition, and interac- al. 2004) when approached (<50 m). In response tions). In their studies of shorebirds in estuaries, to these fi ndings, many government agencies Davidson and Rothwell (1993) described effects support regulatory buffer distances of 31 m at local (movement) and estuary (emigration) including the California Coastal Act buffer levels, as well as impacts at estuary (mortality) between developments and wetlands. Studies and population (decline) scales. The adverse have shown that buffers are effective if they are effects of human disturbance for waterbirds large enough; 35% of buffers <15 m had direct include loss of areas for feeding or roosting, ele- human effects (Castelle et al. 1992), but larger vated stress levels, and reduced reproduction buffers up to 100 m were found to be effective including abandonment of nests or nestlings, for waterbirds (Rodgers and Smith 1997). as well as changes in behavior including avoidance of areas, reduction in foraging intensity, or RESTORATION CONCERNS feeding more at night (Pomerantz et al. 1988, Burger and Gochfeld 1991, Pfi ster et al. 1992, Most tidal marshes in the SFBD are adjacent Burger 1993). to urban development or levees. Unfortunately, Human incursions into marshes may cause tidal marshes are often fragmented by levees trampling of vegetation and soil compaction, that have a narrow and steep transition from reducing the quality of the habitats. Obligate high marsh to upland. The salt marsh harvest saltmarsh species may be particularly sensi- mouse may be found in upland areas up to tive to disturbance, especially because many 100 m from the wetland edge during high tides are adapted to avoid predators by hiding (Botti et al. 1986, Bias and Morrison 1999). within the vegetation and may avoid areas Other marsh inhabitants, such as the California with repeated disturbance. Although habitua- Clapper Rail and the California Black Rail, tion occurs for many species, birds displaced use upland transitional areas as refuge from from coastal marshes were observed to fl y to high tides; however the narrow width of these distant marshes rather than return to the same zones makes rails more vulnerable to predation. areas (Burger 1981), and disturbances displaced During prolonged fl ooding from high tides, shorebirds from beaches (Pfi ster et al. 1992). Suisun shrews (Sorex ornatus sinuosus) utilize Boats, especially small recreational boats, may upland habitats for cover and food (Hays and be a particularly large disruption for waterbirds Lidicker 2000). Without consideration of adja- that fl ush at great distances (Dahlgren and cent habitats, tidal-marsh restoration may fail Korschgen 1992). to support target species. In addition, many restoration projects include public access to DISTURBANCE LIMITS AND BUFFERS the marshes. How and where such access is allowed may greatly infl uence the value of the Josselyn et al. (1989) found that long-legged tidal marshes for some species. waders in estuary marshes fl ushed when approached from 18–65 m, while waterfowl INVASIVE SPECIES fl ushed from 5–35 m. Waterbirds avoided areas near paths where people traveled, and their The SFBD is perhaps the most highly invaded behavioral responses were noted at distances of estuary in the world (Cohen and Carlton 1998). <50 m (Klein 1993). Green Heron (Butorides stria- Many of the plants now found in the tidal tus) numbers were inversely proportional to the marshes, most of the invertebrates in the marsh number of people at a site, and individuals that channels and adjacent tidal fl ats, and several remained foraged less frequently (Kaiser and terrestrial animals that forage in tidal marshes Fritzell 1984). Flushing distance was related to are not native to the Pacifi c Coast. Here, we larger size and mixed-composition of fl ocks in focus on the exotic species that are most likely waterfowl (Mori et al. 2001). to harm the native vertebrates in these tidal Stress responses are more diffi cult to detect. marshes. An exotic disease, the West Nile virus Hikers, joggers, and dogs, as well as avian (WNV), which is expected to soon appear in predators, disturb the federally threatened estuary tidal marshes, is discussed below. Snowy Plover, prompting closure or fencing beach areas. Waterbirds avoided areas near PLANT INVASIONS paths where people traveled, and adverse behavioral responses were noted at distances Grossinger et al. (1998) recommended that of <50 m (Klein 1993). In addition, physiological monitoring, research, and control efforts focus 190 STUDIES IN AVIAN BIOLOGY NO. 32 on the three exotic plants that have the widest the California Clapper Rail, the biggest problem distribution in tidal marshes: smooth cordgrass may be the loss of foraging habitat and food (Spartina alternifl ora), dense-fl owered cordgrass resources in invaded marsh channels where, (Spartina densifl ora) and broad-leaved pepper- during low tides, the rails do much of their for- grass (Lepidium latifolium), and that four other aging (Albertson and Evens 2000). exotic plants that as yet have a very limited One ongoing study in the South Bay is inves- distribution in these marshes be monitored: tigating the impacts on Alameda Song Sparrows common cordgrass (Spartina anglica), salt- through changes in fl ooding regimes and inter- meadow cordgrass (Spartina patens), opposite specifi c interactions among native species. leaf Russian thistle (Salsola soda), and oboe Preliminary analysis shows that Song Sparrow cane (Arundo donax). Of these species, smooth nests in exotic cordgrass are much more likely cordgrass and dense-fl owered cordgrass are to fl ood than nests placed in native vegetation currently the most widespread and are likely and so reproductive success may be lower in to negatively impact native tidal-marsh verte- invaded marsh habitat (J. C. Nordby and A. N. brates because they can become very abundant Cohen, unpubl. data). The invasion may also in the marsh plain (mid- to upper-marsh zones) be altering interactions among native species. (Ayres et al. 1999, Faber 2000). The marsh plain, Marsh Wrens (Cistothorus palustris), which are which is naturally dominated by low-growing native to fresh- and brackish-water marshes, native cordgrass species provides key habitat are now occupying invaded saltmarshes (J. C. for most of the resident tidal-marsh birds and Nordby and A. N. Cohen, unpubl. data). An mammals including the salt marsh harvest increase in Marsh Wren density is potentially mouse (Shellhammer et al. 1982). Several native detrimental for Song Sparrows and other salt- bird species including the federally endangered marsh birds because Marsh Wrens are highly California Clapper Rail, the California Black aggressive and are known to break the eggs Rail, and the three tidal-marsh Song Sparrow of other species occupying adjacent nesting subspecies that are state species of special con- territories (Picman 1977, 1980). The addition cern (Alameda, San Pablo, and Suisun), nest of interference competition from Marsh Wrens and forage in cordgrass (Johnston 1956a, b; San could reduce the reproductive success and Francisco Bay Area Wetlands Ecosystem Goals overall distribution of other saltmarsh-nesting Project 2000). Where smooth cordgrass and bird species. dense-fl owered cordgrass become abundant, they can potentially alter marsh habitat by INVERTEBRATE INVASIONS changing the vegetative structure (including canopy height, density, and complexity), sub- Exotic benthic invertebrates far outnumber canopy physical conditions, root density and natives in the marsh channels and mudfl ats, soil texture, sediment deposition and erosion comprising >90% of the number of individuals rates, and perhaps ultimately marsh elevation, and benthic biomass over most of the estuary marsh topography, and channel morphology (Cohen and Carlton 1995). Marsh birds and (Callaway and Josselyn 1992, Daehler and shorebirds must commonly feed on exotic Strong 1996, Faber 2000). However, few data invertebrates including clams, mussels, snails, are available on how these changes would affect and worms, and probably also ostracods, tidal-marsh vertebrates. amphipods, and crabs (Carlton 1979). Moffi tt Most research on the impacts of exotic plants (1941) found the Atlantic mussel (Geukensia on tidal-marsh vertebrates in the SFBD has demissa) abundant (57% of food by volume) focused on smooth cordgrass. This Atlantic and the Atlantic snail (Ilyanassa obsoleta) cordgrass was introduced in the early 1970s uncommon (2% of food) in the stomachs of and over the next decade began to spread and 18 California Clapper Rails from South Bay. hybridize with the native California cordgrass Williams (1929) observed California Clapper (Spartina foliosa; Ayres et al. 1999). This exotic Rails feeding heavily on the western Atlantic cordgrass (Spartina alternifl ora, S. alternifl ora clam (Macoma petalum), while ignoring the x foliosa, or both) is highly productive and Atlantic snail. Overall, though, little specifi c now occurs in >2,000 ha of marsh and tidal- information exists on what the tidal-marsh fl at habitats (Ayres et al. 1999). The dramatic birds eat. Nor is it known what effect, if any, alteration of tidal-marsh habitat by the tall, the replacement of native prey items by exot- thick, exotic cordgrass will likely affect resident ics has had on these birds—whether more or species the most. Native vertebrate species do less food is available, whether it is more or occupy invaded marshes; however, it is unclear less nutritious, or whether it is more or less whether these subpopulations are sustainable contaminated by toxic pollutants than native (Guntenspergen and Nordby, this volume). For prey. SAN FRANCISCO BAY TIDAL-MARSH THREATS—Takekawa et al. 191

De Groot (1927) reported in detail one the last half of the 19th century either released impact of the Atlantic mussel. This mussel was by hunters or escaped from commercial fox fi rst found in the SFBD in 1894, probably intro- farms. A wild population became established duced in oyster shipments (Cohen and Carlton in the Sacramento Valley, and from this and 1995). It quickly became abundant along chan- other centers, red foxes spread to the East Bay nel banks and in the outer portions of cordgrass region by the early 1970s. They were observed marshes where it typically lies partly buried so in the San Francisco Bay National Wildlife that the posterior margin of its shell protrudes Refuge (SFBNWR) in the South Bay by 1986 and just above the mud with the two valves slightly have continued to expand their range (Foerster open. De Groot (1927) reported that the toes or and Takekawa 1991, Harvey et al. 1992, Cohen probing beaks of rails were frequently caught and Carlton 1995). Dens have been found in and clamped between these valves. He esti- tidal saltmarshes and in adjacent levee banks. mated that at least 75% of adult rails had lost Red foxes prey on resident California Clapper toes, others starved from having their beaks Rails, Black-necked Stilts, American Avocets clamped shut or injured, and one–two nest- (Recurvirostra americana), and Snowy Egrets and lings per brood were caught by mussels and on various other marsh and aquatic birds and drowned by the incoming tide. Whether or mammals, including endangered endemic spe- not these injury and mortality estimates were cies (Forester and Takekawa 1991, Harvey et al. valid, more recent observations confi rm that 1992, Albertson 1995). California Clapper Rails in the SFBD are fre- The SFBNWR began a program of trap- quently missing one or more toes (Moffi tt 1941, ping and killing red foxes in 1991 (Foerster Josselyn 1983, Takekawa 1993), and Takekawa and Takekawa 1991, Cohen and Carlton 1995). (1993) reported that a rail captured with a mus- Recent surveys show a strong recovery in local sel clamped onto its bill subsequently lost part populations of California Clapper Rail follow- of its bill. ing implementation of the red fox removal Exotic invertebrates may also have an (Albertson and Evens 2000). In the early 1980s indirect impact on tidal-marsh vertebrates by California Clapper Rail numbers in the South altering habitat. The southwestern Pacifi c iso- Bay were estimated at 400–500. The local popu- pod (Sphaeroma quoyanum) was fi rst collected lation crashed to roughly 50–60 in 1991–1992 in the SFBD in 1893. It burrows abundantly in surveys, roughly 5 yr after the fi rst detection of mud and clay banks along the channels and red foxes at the SFBNWR. In 1997–1998 winter outer edges of saltmarshes, and has been cred- surveys, rail numbers increased to 330. Because ited with eroding substantial areas of marsh, California Clapper Rails are year-long residents though no direct studies or measurements and have strong site tenacity, the variation have been made (Cohen and Carlton 1995). between survey years is not thought to be from Two recently arrived crabs, the European green dispersal or migration. crab (Carcinus maenas, fi rst seen in the estu- Brown rats (Rattus norvegicus) became ary in 1989–1990) and the Chinese mitten crab established in many parts of California by the (Eriocheir sinensis, fi rst collected in the estuary 1880s. In the SFBD, brown rats are common in in 1992), are also known to be common burrow- riparian areas, in fresh, brackish and saltwater ers in saltmarshes or tidal channels (Cohen et al. tidal marshes, and in diked marshes (Josselyn 1995, Cohen and Carlton 1997). If these organ- 1983, Cohen and Carlton 1995). De Groot (1927) isms do in fact contribute to the erosion and considered the brown rat to be the third most loss of marsh habitat, this would clearly have important factor in the decline of California an impact on marsh vertebrates. The impacts Clapper Rail, after habitat destruction and could be greatest near marsh channels, since the hunting. More recent authorities (Harvey 1988, slightly elevated areas alongside these channels Foerster et al. 1990, Foerster and Takekawa are better drained and support taller vegeta- 1991, Cohen and Carlton 1995) have also found tion, providing better nesting sites and habitat substantial predation on California Clapper for saltmarsh Song Sparrows, and possibly for Rail eggs and chicks, with some estimating California Clapper Rails, and the salt marsh that brown rats take as many as a third of the harvest mouse (Marshall 1948; Johnston 1956a, California Clapper Rail eggs laid in the South b; Shellhammer et al. 1982, Collins and Resh Bay (Harvey 1988). Rats also prey on other 1985, Albertson and Evens 2000). marsh-nesting birds and their nest contents. Because brown rats are more likely in areas that VERTEBRATE INVASIONS abut urban development, habitat buffers might reduce their abundance in tidal marsh. Red foxes (Vulpes vulpes) from Iowa or House cats (Felis domesticus) are wide- Minnesota were introduced into California in spread in California both as house pets and 192 STUDIES IN AVIAN BIOLOGY NO. 32 feral individuals. In the SFBD, house cats have increased predation: habitat fragmentation frequently been seen foraging in saltmarshes, (Schneider 2001, Chalfoun et al. 2002), loss of along salt-pond levees, and wading at the edge vegetated upland edges for use as refugia from of tidal sloughs (Foerster and Takekawa 1991). predators during high tides, establishment of House cats are known to have killed adult boardwalks and power lines across marshes Light-footed Clapper Rails (Rallus longirostris (the latter are used as perches by raptors), levipes) in southern California (Foerster and changes in marsh vegetation structure and Takekawa 1991) and at least one California the spread of urban-tolerant native predators Clapper Rail in the SFBD (Takekawa 1993), and (e.g., American Crow [Corvus brachyrynchos] presumably also prey on other marsh birds and Common Raven [Corvus corax], raccoon [Procyon mammals. The SFBNWR began a program of lotor], and striped skunk [Mephitis mephitis]), removing feral cats in 1991. feral animals (house cats) and other non-native predators (especially red fox), many of whom RESTORATION CONCERNS are human subsidized in urban and suburban areas (USDI Fish and Wildlife Service 1992). Exotic cordgrass colonization is a threat for Sanitary landfi lls and riprap shorelines are also most South Bay tidal-marsh restoration projects sources of predators (USDI Fish and Wildlife due to its gross alteration of vegetative structure, Service 1992). sub-canopy physical conditions and root density, In the SFBD, tidal-marsh fragmentation has and potential alteration of soil texture, sediment resulted in an increased mean perimeter to deposition and erosion rates, marsh elevation, edge ratio, and thus more edge per unit area. marsh topography and channel morphology, Predators are hypothesized to be more active which could in turn affect native plant and inver- at habitat edges. Current studies of the rela- tebrate populations, as well as vertebrate popu- tionship between edge habitat and predation lations. Although a regional exotic cordgrass indicate that nest predators vary in activity control program began in 2004, complete control and impact depending on the taxon and the may take many years to achieve. It is likely to surrounding land use (Chalfoun et al. 2002). remain a signifi cant issue in most restoration Studies in estuarine marshes indicate that pat- of this subregion and an imminent threat to the terns of predation vary between sites (PRBO North Bay. In contrast, few areas are invaded by Conservation Science, unpubl. data). This varia- exotic cordgrass in the North Bay; thus, vigilant tion is probably due to differences in the suite monitoring and removal of hybrid populations of predators, which itself may be dependent on would be highly benefi cial, and restoration of variation in land use on adjacent uplands and tidal marshes with native cordgrass may be vegetation type, and on variation in tidal fl ood- more successful in this subregion. Control of ing, channel and levee confi guration, marsh nonnative predators will likely be an essential vegetation structure and human disturbance part of most tidal-marsh restoration projects to patterns. Some changes in vegetation involv- maintain native fauna. Nonnative predators will ing increases in vegetation density, such as that be of most concern in restoration areas adjacent associated with the spread of invasive smooth to urban development. cordgrass may actually result in decreased nest predation, but with ecological trade-offs PREDATION (Guntenspergen and Nordby et al., this volume). Several predators have been documented to Increased rates of predation on tidal-marsh depredate tidal-marsh birds, bird nests, reptiles, vertebrates can result from three types of and mammals. They include upland mammal human-induced changes: (1) introduction of species that forage in marshes such as: raccoon, non-native predators, (2) changes in the distri- red fox, coyote (Canis latrans), striped skunk, bution or abundance of native predators, and house cat, domestic dog (Canis familiaris), house (3) alterations of habitat that infl uence preda- mouse (Mus musculus), brown rat, and black tion effectiveness or avoidance. For birds and rat (Rattus rattus); wetland mammals such as other vertebrates in tidal saltmarshes of the the river otter (Lutra canadensis); snakes such SFBD, as in most other ecosystems, predation as gopher snake (Pituophis melanoleucus) and is generally the dominant cause of adult and garter snake (Thamnophis spp.) which have been juvenile mortality and nest failure (Point Reyes observed swallowing nest contents (Point Reyes Bird Observatory, unpubl. data). Although the Bird Observatory, unpubl. data); and numerous primary cause of signifi cant declines in popula- wetland birds including Great Blue Heron (Ardea tions of tidal-marsh vertebrates is habitat loss herodeus), Great Egret (Casmerodius albus), Snowy and degradation, other factors may be contrib- Egret, Black-crowned Night Heron (Nycticorax uting to further population declines through nicticorax), and gull species (USDI Fish and SAN FRANCISCO BAY TIDAL-MARSH THREATS—Takekawa et al. 193

Wildlife Service 1992, Albertson and Evens 2000; and other animals, they have been extensively Point Reyes Bird Observatory, unpubl. data). studied over the last century (Durso 1996) and Raptors, especially Northern Harrier (Circus have been the subject of control programs in cyaneus), White-tailed Kite (Elanus leucurus), many areas, including SFBD. Mosquitoes are a and Red-tailed Hawk (Buteo jamaicensis) are diverse group of insects that share a common also documented predators as are the Common life history (egg, aquatic larvae, aquatic pupae, Raven and American Crow. And fi nally, the nest and fl ying adult), and a requirement for blood parasite and egg predator, the Brown-headed feeding by the adult females to produce eggs, Cowbird (Moluthrus ater) has been documented with rare exceptions. In addition, all juvenile in SFBD tidal marshes, although rates of parasit- mosquitoes are weak swimmers and require ism vary greatly among marshes (Greenberg et habitats free from strong waves or currents or al., this volume). Some of these species, such as abundant predators. Thus, all mosquito species the Northern Harrier, nest in or near marshes require shallow, still aquatic habitats for at least and have probably always been part of the tidal- a few consecutive days. marsh food web. Other species, such as Common Despite these similarities, mosquitoes vary Ravens, American Crows, and raccoons, have considerably in their specifi c habitat require- adapted well to urban areas, and their large pop- ments (several species may often coexist in ulations have resulted in increased predation in close proximity) and in their potential for trans- adjacent natural areas. mitting pathogens (public health signifi cance of some species is higher than others). Mosquitoes RESTORATION CONCERNS are often distinguished by their specialized juvenile habitats, adult behavior, and vector sta- Control of red foxes and other non-native tus (Table 3 modifi ed from Durso 1996, Maffei predators in the south San Francisco Bay has 2000). For example, Culiseta incidens is found contributed to a rebound in California Clapper in shaded, cool, clear fresh water, while Aedes Rail numbers. However, predator control is melanimon prefers sunny, warm fresh water not a viable option in all parts of the estuary, with dense grasses. and other measures to reduce predation, e.g., Some generalities are possible in this diverse by modifying habitat, may have better long- assemblage. Larval habitat falls primarily along term results (Schneider 2001). More studies temperature (seasonal) and salinity (spatial) are necessary to identify the primary predators gradients, with only two truly salt-adapted of tidal-marsh birds and mammals in various mosquito species (Aedes squamiger and A. dorsa- parts of the estuary so that managers can decide lis) common in SFBD. Unlike freshwater genera which control measures, if any, are necessary. that lay eggs in stagnant water, saltwater mosquitoes (Aedes), require an egg-conditioning MOSQUITOS AND OTHER VECTORS period of at least a few days in which eggs cannot tolerate inundation. Thus, mosquito Although wetlands, including tidal marshes, production is low in saltmarshes where dry support high densities of many desirable spe- periods are too short for egg conditioning (i.e., cies, they can also produce copious mosquitoes few impediments to drainage; Kramer et al. and potentially other disease vectors. These 1995). High fl ooding frequency is also benefi cial disease-carrying organisms can pose threats for mosquito control because it is associated to tidal-marsh ecosystems because they can with currents suffi cient to fl ush the larvae to sicken and kill marsh animals as well as people, unfavorable sites. Large populations of mosqui- and because mosquito-control measures can toes are almost invariably found where drain- have an adverse impact on marsh processes. age is poor, whether the impounded water is Fortunately, neither traditional endemic vec- saline (spring high tides that do not drain) or tor-borne diseases nor current mosquito-control fresh (rain or seeps). activities pose an imminent threat to existing Although mosquito threats to human and marshlands; however, new diseases may have animal health include disturbance, allergies, and dramatic impacts on wildlife, particularly birds. infection secondary to scratching (Durso 1996), The need to protect wildlife, as well as the pub- the most signifi cant problems are the infectious lic from diseases may pose serious challenges pathogens carried between animals by mos- for wetland restoration proposals. quito blood feeding. West Nile virus (WNV) has killed hundreds of people, hundreds of thou- MARSH MOSQUITOES AND MOSQUITO-BORNE DISEASES sands of birds (in almost all taxonomic groups), and smaller numbers of other vertebrate taxa, Because some mosquito species transmit in the US over the last 4 yr (Center for Disease widespread and serious diseases to humans Control 2001, United States Geological Survey 194 STUDIES IN AVIAN BIOLOGY NO. 32

2003). WNV is the greatest immediate disease threat both to wetland organisms and humans. Lab research has demonstrated that it can infect Culex tarsalis, Culex erythrothorax, Aedes dorsalis, A. melanimon, A. vexans, and Culiseta inornata,

a and that all of these species can transmit the virus at some level, although the two Culex species were the most efﬁ cient vectors (Goddard et al. 2002). Field observations in states where Culex tarsalis occurs conﬁ rms that this species

. will probably pose the greatest threat in areas that have shallow fresh-water ponds (<5 ppt) that last until eggs, larvae, and pupae develop MARSHES

(5 d in the summer), but that become dry peri- odically to eliminate aquatic predators (Maffei 2000). While Culex tarsalis and its particular STUARY

E habitat types are the chief problems, other mos-

AY quito species and poorly drained marshes have B been implicated in diseases in the past (Reisen et al. 1995, Durso 1996) and may also contribute to the establishment and spread of future patho- RANCISCO gens (Center for Disease Control 1998, 2001). F AN S ooding. Localized pest. May transmit CE-like virus. ooding. large mammals day and night. Localized pest. ooding. large mammals day and dusk. Localized RESTORATION CONCERNS IN

Because marsh mosquitoes have histori- SPECIES

cally been recognized as a potential threat to animal health and human health and comfort, government agencies have acted to control marsh mosquito populations through a variety MOSQUITO of activities, traditionally divided into physical control (habitat manipulation), biological control (stocking living predators or parasites), COMMON and chemical control (applications of biotic OF or chemical pesticides) (Durso 1996). Some of Egg conditioning on moist soil and plants. Flies 16–32 km. Bites humans day and dusk these strategies, such as widespread drainage STATUS

of wetlands or extensive applications of DDT, Egg conditioning on moist soil and plants. Flies up to 1.6 km. Bites humans and other

Egg conditioning on moist soil and plants. Flies up to 16 km. Bites humans and other clearly had substantial impacts on marshes and surrounding habitats in the past (Daiber 1986). ) Simultaneous hatch following ﬂ VECTOR Eggs laid directly on cold standing Flies up to 8 km. Bites humans and other

) Simultaneous hatch following ﬂ

) Simultaneous hatch following ﬂ Although these examples clearly indicate a Eggs laid directly on warm to hot standing Flies 16–24 km. Bites birds, and humans

AND need for continued monitoring and research, , mosquito control activities have become more c name Eggs, larvae, and pupae Adult behavior and vector status. target-speciﬁ c in recent decades and have not linked to signiﬁ cant adverse impacts on BEHAVIOR

Aedes squamiger Aedes dorsalis Aedes washinoi Ochlerotatus squamiger Ochlerotatus dorsalis Culiseta inornata Ochlerotatus washinoi Culex tarsalis the marshes (Dale and Hulsman 1990; U.S. Environmental Protection Agency 1991, 1998, ADULT

, 2003; Dale et al. 1993, Contra Costa Mosquito and Vector Control District 1997, Center for Disease Control 2001). In addition to mosquitoes, degraded tidal CONDITIONS marshes can also provide habitat for brown and black rats (Breaux 2000), which are signiﬁ - cant vectors of human disease, and for midges and other invertebrate pests (Maffei 2000). This

EPRODUCTIVE combination of health threats and pests associated with marshes often has led to conﬂ icts 3. R between wetland restoration proponents and

ABLE neighbors, and the greatest threat to marshes CE = California encephalitis; WEE western equine SLE St. Louis WNV West Nile virus. T Winter salt marsh mosquito (Maffei 2000) ( a Common name Scientifi Summer salt marsh mosquito Winter Marsh mosquito (Maffei 2000) Washino’s mosquito (Maffei 2000) Western encephalitis mosquito (Durso 1996) Highly salt tolerant. Cold water. water. Low salt tolerance. Highly salt tolerant. Warm to hot water. Secondary vector of CE and WEE. water. Low to moderate salt tolerance. large mammals at night. Localized pest. Cool to cold water. Low salt tolerance. other mammals at night. Primary vector of pest. May transmit CE-like virus. WEE, SLE. High vector competence for WNV in the lab. (Maffei 2000) ( ( associated with disease vectors may be the SAN FRANCISCO BAY TIDAL-MARSH THREATS—Takekawa et al. 195 continuing development of residential areas this species composition may change as the nearby. Reducing this confl ict will depend on virus moves westward. Coupled with the good working relationships between wetland increasing number of corvids in the region, the restoration advocates and mosquito con- threat to tidal-marsh birds is imminent. Other trol and other public health personnel (San birds that inhabit wetlands of the estuary have Francisco Bay Area Wetlands Ecosystem Goals been identifi ed as hosts and vectors including Project 1999). cormorants, shorebirds, and Song Sparrows. Transmission of WNV occurs when adult DISEASE mosquitoes feed on the blood of an infected avian host followed by another vertebrate host. Avian species are faced with the greatest Mammals (humans and horses) do not appear known, or anticipated, threats from disease to serve as intermediate hosts, though they can among all wildlife populations in the estuary. be infected (National Wildlife Health Center The three diseases of greatest concern and 2003a). The threat may be greatest for spe- demonstrated mortality in the SFBD are West cies of conservation concern, such as the three Nile virus, avian cholera, and avian botulism. subspecies of Song Sparrows in the SFBD. The Infectious diseases are currently on the rise for total estimated avian mortality due to the dis- two probable reasons—the earth’s climactic ease is over 100,000 individuals, though species changes (Colwell 2004) and imbalance in bio- breakdown is not available. Because the disease logical systems, probably because of degraded is so new, the mortality impacts, as well as habitat quality and diminished habitat quantity the ecological interactions of the virus with its (Friend 1992). hosts, are virtually unknown (National Wildlife For example, increased temperatures can Health Center 2003a). Yet, if the Corvidae con- boost the survival and growth of infectious dis- tinues to be the principal group affected by eases such as Pasteurella multocida, the bacteria the virus, WNV may have a benefi cial impact that causes avian cholera (Bredy and Botzler on those bird species that compete with or are 1989). Degraded habitat quality can lead to negatively affected by corvids. changes in microbial populations, and subsequently disease outbreaks (Friend 1992). A AVIAN CHOLERA decrease in habitat size often results in a greater density of birds, increasing exposure, transmis- AC is a highly infectious bacterial disease sion, and spread of the disease to other locations with the highest documented mortality rate (Friend 1992). of any disease for wetland birds in the estuary (USDI Fish and Wildlife Service 1992). WEST NILE VIRUS Transmission is often direct and may involve surviving carrier birds; consequently, crowd- WNV is the most recent of the serious dis- ing is thought to increase incidence and mor- ease threats, fi rst identifi ed in the US in 1999. tality. California leads the nation in reported Reports of bird infections began in the east- disease outbreaks, particularly in the delta. ern part of North America and have rapidly Waterfowl have been particularly affected, spread west. The fi rst cases of WNV in the especially when concentrated on wintering country were reported in New York City, New areas or during spring migration. AC is of par- Jersey, and Connecticut. By December 2003, ticular concern because roughly 50% of birds WNV had been identifi ed all states except migrating along the Pacifi c fl yway may pass Oregon, Alaska, and Hawaii with 12,850 through the SFBD. Outbreaks have occurred in cases of human infection and 490 deaths (U.S. a variety of habitats including freshwater wet- Geological Survey 2003). As of December 2003, lands, brackish marshes, and saltwater envi- California had two cases of human infection of ronments (National Wildlife Health Center WNV in Riverside and Imperial counties, and 2003b). Since World War II, thousands of birds WNV was detected in the SFBD in 2004. WNV (mainly waterfowl) have been reported dead in is transmitted by a variety of mosquitoes, but each year. In one year, documented mortality two species in particular appear to be primary was 70,000 birds for the state (USDI Fish and vectors, Culex tarsalis and C. erythrothorax. Wildlife Service 1992). The disease commonly At least 138 species of wild birds have been affects more than 100 species of birds, though infected, with the family Corvidae demon- the Snow Goose (Chen caerulescens) has the strating the highest prevalence of the disease greatest mortality (National Wildlife Health (National Wildlife Health Center 2003a). Center 2003b). Unlike botulism, AC often American Crow and Blue Jay (Cyanocitta cris- affects the same wetlands and the same avian tata) have most often been infected; however, populations year after year. 196 STUDIES IN AVIAN BIOLOGY NO. 32

AVIAN BOTULISM DISCUSSION

Avian botulism results from a neurotoxin We have summarized some of the key produced by the bacterium, Clostridium botuli- threats to tidal-marsh vertebrates and have num type C (Friend 1987). The disease is caused identifi ed specifi c issues that are major con- by a bacterium that forms dormant spores in the cerns in tidal-marsh restoration projects in the presence of oxygen. The spores are resistant to SFBD. Unfortunately, it is diffi cult to compare heating and drying and can remain viable for the modern estuary to a historical period years. Spores are widely distributed in wetland when the tidal marshes functioned naturally, sediments and can also be found in the tis- because major changes to the system occurred sues of most wetland species, such as aquatic before studies documented the importance of invertebrates and many vertebrates, including tidal marshes. For example, estuaries in SFBD healthy birds. The botulism toxin is produced and the arid Southwest are driven by snow- only when the bacterial spores germinate pack conditions (Dettinger and Cayan 2003, (Rocke and Friend 1999). Kruse et al. 2003), and water-user demands Although botulism is a more serious mor- determine how closely the system follows the tality causing factor than AC statewide, in natural pattern of infl ows. The climate results the SFBD, the reverse is the case. Outbreaks in two freshwater pulses—rainfall in the winter of botulism in waterbirds are sporadic and (November–February) and runoff in the spring unpredictable, occurring annually in some (April–June). Most native vertebrate species are wetlands, but not in adjacent ones. In the past, adapted to these wet periods, but changes in mortalities from botulism have ranged from the environment have created a much different 0–1,000 in south San Francisco Bay. Botulism system than in the past. Flood protection and outbreaks caused 950 and 565 mortalities in urbanization have resulted in a less dynamic 1998 and 2000, respectively, mostly ducks estuary. The natural periodic, inter-annual, and gulls (C. Strong, San Francisco Bay Bird and annual fl ooding will be replaced by a static Observatory, pers. comm.). Botulism has been system that leaves little room for change. Under documented in the South Bay in Ruddy Ducks these conditions, the establishment and spread (Oxyura jamaicensis), Mallards, and Northern of exotic species may be facilitated. Exotic spe- Shovelers (Anus clypeata; USDI Fish and cies continue to arrive in the SFBD at a rapid Wildlife Service 1992). rate (Cohen and Carlton 1998), and the mecha- Ecological factors that are thought to play a nisms introducing these species remain poorly critical role in determining outbreaks include regulated (Cohen 1997, Cohen and Foster 2000). conditions that favor spore germination, the Many important effects of exotic species may be presence of a suitable energy source or sub- indirect or subtle, such as ways in which exotic strate for bacterial growth and replication, and plants alter habitat for vertebrates. a means of transfer of toxin to the birds, pre- In the modern estuary, combined threats sumably through invertebrate prey. Botulism may have the most detrimental consequences outbreaks appear to be associated with moder- for many tidal-marsh vertebrate populations. ately high pH (sediment pH 7.0–8.0) and low to Decreased water quality and increased con- moderate salinity (≤5 ppt). Botulism outbreaks taminant loads may exacerbate the effects of are not specifi cally associated with shallow vertebrate diseases. Human disturbance, frag- water and low dissolved oxygen (Rocke and mentation, and predation by species such as Friend 1999). the red fox may reduce the carrying capacity of native vertebrate populations in remnant RESTORATION CONCERNS marshes. The loss of downstream sediments may greatly alter the sediment balance and Tidal-marsh restoration will require care- the rate of marsh plain accretion in the SFBD. ful management to avoid disease outbreaks. Dredge material may provide a benefi cial solu- Increased density of vertebrate species in new tion to sediment defi cits in some wetland res- restoration sites may encourage concentrations torations, but many estuarine contaminants are that result in disease outbreaks. Restoration bound to sediments and the combined effects projects adjacent to urban development may of dredging operations, dredge materials, and introduce potential disease sources. Finally, sediment-bound contaminants on vertebrates degraded environmental conditions may and their food webs are largely unknown. increase the effects of disease, impairing the In response to severe losses to wetland species targeted for recovery under restoration habitats, major efforts aim to regain and estab- efforts. lish wetlands in SFBD. The current wave of SAN FRANCISCO BAY TIDAL-MARSH THREATS—Takekawa et al. 197

restoration projects will alter the character of (b) relationship between foraging ecol- the estuary for the next century; however, with ogy and bioaccumulation in tidal-marsh current rates of human development, the few vertebrates, and (c) prediction of wetland remaining unprotected bay lands will no longer restoration activities on the concentra- be restorable. Sea-level rise may eliminate tidal tions, distribution, and bioavailability of marshes squeezed between open water and contaminants. urban development, rather than a mere relo- 7. Future research also should document cation of marshes to higher elevations. Thus, how changes in water quality (i.e., salin- identifying the critical environmental threats in ity) may affect vertebrate distribution advance may be the best way to guide restora- through a bottom-up control of vertebrate tion actions and management to ensure the con- food webs. Increased salinity and low servation of tidal marshes into the future. primary productivity have the potential to change the distribution and abundance RESEARCH NEEDS of vertebrate species, but the magnitude of these effects and the mechanisms by This paper represents the fi rst step towards which they act on vertebrates (e.g., via corrective action and management by identify- the food web or changes in the vegetative ing the major threats to tidal-marsh vertebrates community) are not well understood. in SFBD. Further efforts to comprehend the 8. The effect of disturbance on secretive mechanisms and processes at work are listed species of saltmarshes is very diffi cult below as the next step in our understanding of to study, because their responses are not tidal-marsh ecosystems. This list is not intended easily observed. However, documenting to be a complete compilation of research needs, human activities near tidal marshes and but to highlight some key issues that require studying marked populations, including immediate attention. bioenergetic studies, may better quantify 1. Conceptual models and data validation costs of disturbance and lead to specifi c of the combined interaction of threats management plans. (i. e., interaction between water quality, 9. Studies that detail the direct and indirect contaminants, and disease) to tidal- effects of invasive species, as well as the marsh vertebrates is needed to under- mechanisms of exotic species arrival, stand system-wide processes. establishment, and spread may lead to 2. Information about the effects of fragmen- better regulation of introductory path- tation on tidal-marsh function and demo- ways and control options. graphic and population-level processes 10. Predation studies would help to clarify (dispersal, gene fl ow, survivorship, and the relationship between mortality and predation) would improve prediction of tidal-marsh fragment size, number, and those species’ responses to habitat resto- distribution, and the potential effect of ration. exotics. 3. A greater understanding of the hydrology 11. Finally, mortality caused by disease, as and transport of bay sediments is needed well as degraded conditions under which to help determine local effects of restora- they have the greatest effect, should be tions (sediment sinks) and their effect on estimated within the context of overall fl ow patterns and sediment movement. annual mortality. 4. Detailed studies on the dredge-ame- liorated wetlands including vegetation ACKNOWLEDGMENTS and structure, contaminant load, and vertebrate food webs may help resolve We thank the editors, R. Greenberg, J. sediment concerns in marsh restoration. Maldonado, S. Droege, and M. McDonald for 5. A sediment-supply model based on stimulating discussions that led to the prepara- recent empirical data would allow for tion of this manuscript. We thank K. Phillips, better assessment of sediment changes D. Stralberg, N. Athearn, and K. Turner for and effects of sea-level rise on marshes helpful comments on the draft manuscript. that have a sharp upland transition zone The U.S. Geological Survey, Western Ecological (i. e., levee fringe marshes). Research Center, Wetland Restoration Program; 6. Research that would greatly improve National Science Foundation (DEB-0083583 to understanding of contaminant threats ANC), The Nature Conservancy, and the David in tidal marshes includes: (a) factors H. Smith Conservation Research Fellowship (to that control contaminant abundance JCN) provided partial support for preparing the and bioavailability within the wetlands, manuscript.