Jug  Bay  Wetlands  Sanctuary 

Abstracts from the Jug Bay Research Conference

Studies on the role of hydrology and animal activity in vegetation dynamics of Jug Bay wetlands

The vegetation of tidal freshwater marshes is temporally dynamic and spatially variable due to a number of factors, including hydrology and animal activity. Since 1998, my graduate students and I have been studying the influence of hydrology and animal activity on plant community structure in Jug Bay tidal freshwater marshes. This research has included four experiments in Billingsley Marsh, including field, greenhouse, and seed bank studies on the effects of hydrology on plant communities, and a field exclosure experiment to examine effects of animals on high and low marsh vegetation. Additionally, my research group has studied vegetation, soil, hydrology, and seed banks of marshes north and south of Maryland Route 4 as reference sites for vegetation studies at restored tidal freshwater marshes along the Anacostia River in Washington, D.C. (Kingman and Kenilworth marshes). The hydrologic studies show that increasing the frequency and duration of flooding reduces the diversity of marsh vegetation, and suggest that the effect of flooding on reducing diversity is greater if it occurs earlier rather than later in the growing season. In the exclosure study, we found that animal disturbance (e.g., grazing or physical disturbance by geese, carp, and muskrat) can strongly reduce seedling establishment by some annual species (Zizania aquatica and Bidens laevis), but that the effect depends on hydrology: a strong animal effect was seen in the low marsh but not in the high marsh. Our research in the Jug Bay wetlands underscores the importance of interactions between biotic and abiotic variables in maintaining spatial and temporal complexity in the vegetation of these systems.

North temperate estuaries are critical habitats for a wide variety of waterbirds, especially during the winter and migratory periods. Birds often exhibit patterns of habitat segregation on the estuary that are related to salinity, water depth, substrate characteristics and food availability. To investigate the distribution and abundance of waterbirds on the Patuxent River estuary, we organized four, one-day censuses in February 1999, 2000, 2001 and 2002. Sixty experienced birders conducted 20 to 30 minute censuses at 115 pre-selected, permanent sites between the river's mouth at Drum Point and the head of tide (Route 214), a 55-mile distance. Salinity varied from about 16 ppt at the mouth to 0 ppt above river-mile 40. A total of 50 species were identified and a mean of about 41,000 (+6,000) waterbirds was counted on the estuary each winter. Total numbers of waterbirds varied from a high of 46,400 in 1999 to a low of 32,100 in 2000. In all years, birds were most concentrated in a 20-mile segment from 5 to 25 miles upriver from the mouth, where the salinity was approximately 6 to 15 ppt. The most abundant species were Ruddy Ducks, Canada Geese, Greater and Lesser Scaup, Bufflehead, Canvasback, Common Goldeneye, Long-tailed Ducks, Tundra Swans and Ring-billed Gulls. Ruddy Ducks, typically the single most abundant species on the estuary every winter, occurred predominantly between river-mile 10 and 20, with the highest concentrations in the vicinity of Chalk Point. Diving ducks were most numerous in higher salinity waters, whereas dabbling ducks were most numerous in lower salinity waters. Bald Eagles, Great Blue Herons, Canada Geese and gulls were fairly evenly distributed throughout the estuary, whereas American Black Ducks and Common Snipe were quite restricted. Our data reveal typical mid-winter patterns of waterbird distribution and abundance, and underscore the importance of the Patuxent River estuary for these species. This information serves as a baseline for examining population changes and shifts in occurrence, and may provide another tool for assessing the overall environmental health of the estuary.

The System-wide Monitoring Program (SWMP) of the National Estuarine Research Reserve (NERR) system collects water quality observations at 30-minute intervals at numerous locations throughout the United States. At the Jug Bay component of the Chesapeake Bay National Estuarine Research Reserve in Maryland, in situ water temperature, depth, pH, dissolved oxygen and salinity measurements were recorded since 1995 at two sites in the tidal fresh headwaters of the Patuxent River. Conditions at the sites depicted by the March through December 2002 time series, for the main stem site at Jackson Landing (Patuxent River Park) and at a nearby shallow tidal creek site (Jug Bay Wetlands Sanctuary), exhibit general patterns and ranges observed at other tributary monitoring sites in the Chesapeake Bay. The synoptic water quality measurements emphasize the variability of short-term patterns and diverse conditions that exist between deeper main stem waters and adjacent shallow waters. Both sites exhibited similar diurnal temperature and dissolved oxygen fluctuations. Salinity maximums and the range of fluctuations over each tidal cycle were greater at the main stem site, indicating greater vertical and longitudinal stratification effects in the main stem. Lower salinities at the shallow creek site suggest a decoupling of inflowing main stem surface waters on the flood tide. This selective partitioning and distribution of main stem surface waters to shallow, densely vegetated mudflats may account in part for the higher divergent dissolved oxygen fluctuations observed at the shallow creek site. Daily net ecosystem metabolism (NEM) estimates, determined from dissolved oxygen observations to estimate daily open water primary production and respiration rates for each site, were highly variable. Although net productivity rates in shallow waters exceeded by a factor of almost two those in the main stem, respiration rates were proportionately higher in shallow waters. Seasonally, NEM was net autotrophic in main stem waters during spring and summer and at the shallow creek site during spring and fall. NEM was net heterotrophic in the main stem during fall, and at the shallow creek site during summer. The ratios of daily net production to respiration (P:R Ratio) exceeded unity at both sites, indicating that on average over spring, summer and fall, more organic matter was produced than consumed at both sites.

The Delmarva Peninsula is one of the largest poultry areas in the U.S., producing some 600 million chickens and 1.6 billion pounds of manure annually. In addition to nutrients, poultry litter contains feed additives such as trace metals and antibiotics, as well as naturally produced hormones. In spring 2000, a two year field study was initiated to evaluate the potential water quality impacts of these Anon-traditional@ contaminants. The study included the deployment of polar organic compound integrative water samplers (POCIS), organic analysis of poultry litter and sediments, and the analysis of trace metals in litter, water, and sediments at several sites on the Delmarva Peninsula and a reference site on the Patuxent River near Jug Bay on Maryland=s western shore. Biological endpoints included analysis of benthic macroinvertebrate surveys as well as vitellogenin (Vtg) in the blood of male fish, as an indicator of exposure and response to estrogenic compounds. Select results will be presented. Compounds that were detected in the POCIS deployed at sites on the Delmarva Peninsula included atrazine (herbicide), tetracycline (antibiotic), and 17 ß-estradiol (estrogen); the latter two compounds were not detected at the Jug Bay reference site. Diazinon (insecticide) was detected in low concentrations at Jug Bay, but not at the Delmarva sites. Results of benthic macroinvertebrate community analysis did not indicate any differences among sites related to water quality impacts from poultry litter. Vtg was measured in resident male carp from all sites, but not in fish caged for 3 weeks at each site. The Vtg concentrations were highest in carp from the Delmarva sites; however, Vtg levels in fish from all sites, including the Jug Bay reference, appear to be above normal concentrations based on a comparison with data from several hundred carp collected in the Mississippi River Basin. The finding of elevated Vtg in male carp from Jug Bay is somewhat surprising and may be related to exposure to agricultural chemicals, some of which are presumed endocrine disruptors.

The loss of submerged aquatic vegetation (SAV) from the Patuxent estuary during the latter part of the 20th century was explored using diverse data sets that included: historic SAV coverage and distribution data, SAV ground truth observations, water clarity and nutrient loading data, and epiphyte light attenuation measurements. Analysis of aerial photography from 1952 showed that SAV was abundant and widely distributed along the entire mesohaline region of the estuary; however, by the late 1960s, rapid declines in SAV took place following large increases in nutrient loading to the estuary. An examination of water clarity and epiphyte data suggest that the processes that led to the loss of SAV varied in strength along the axis of the estuary. In the upper mesohaline region, Secchi depths were consistently less than established mesohaline SAV habitat requirements at 1m water depth, suggesting that water clarity was responsible for SAV decline. In the lower mesohaline region, where water clarity was consistently above SAV requirements, high epiphyte fouling rates significantly reduced light available to SAV. Experimental results show that epiphyte fouling had the capacity to reduce available light to SAV blades from 30% to 70% of surface light within a week, and likely contributed to the local decline and near total loss of SAV during the late 1960s and early 1970s. As a result, the prognosis for near-term SAV recovery within the mesohaline portion of the estuary seems unlikely given existing water quality conditions.

Coastal marshes are particularly rich in ecosystem goods and services. They provide a wide range of highly valued resources including fisheries, open spaces, wildlife habitat, nutrient cycling, and recreational opportunities. Some threats to these marshes on the ocean side are an increase in relative sea level rise, and on the inland side from human-made flood protection structures. There are multiple drivers causing relative sea-level rise, these drivers act over different time scales and they vary in importance across spatial locations. Coastal wetlands are expected to be severely impacted by sea level rising in the next few decades. Tracking relative sea level rise, while measuring the sediment surface elevation relative to water level height, is the single most important factor for assessing coastal marsh vulnerability. We will draw from examples of short- and long-term monitoring sites to assess the efficacy of the Sediment Erosion Table (SET) device in capturing sediment which causes elevation changes over temporal and spatial scales. We will also identify the rate of change in relative sea level in seven coastal systems including the Jug Bay NERRS site. Based on SET data, we constructed a database that provides an unprecedented potential to assess optimal temporal and spatial scales at which to observe the sediment elevation change phenomena in relation to sea level rise, and the effect of restoration on marsh development. The waste assimilation function performed by tidal marshes involves the dynamics that cause suspended sediments to be imported from the estuarine water column and deposited between the marsh vegetation. While sediments are transported on a daily base between the estuary and the tidal creeks, there are favorable severe weather conditions that are needed to transport the sediments from the tidal creeks onto the marsh. There is the potential for unfavorable weather conditions to transport sediments from the marsh, back to the tidal creeks, a process that can be prevented by a dense cover of marsh vegetation. Marsh vegetation is equally important for fixing carbon, important for denitrification and building root structure for soil formation. To investigate the waste assimilation function, we carried out four comprehensive field experiments that monitored total suspended sediments at hourly intervals and water level measurements at five- minute intervals. Two sampling intervals took place at the Jug Bay North Marsh site, north of the railroad bed and two at the South Marsh site just south of the railroad bed. Marsh levels at both sites were at 40 cm above datum. Weather conditions varied for each of the experiments and ranged from calm with only a trace of rain (October 1999 South side) to windy and 1.42 cm of total rain (September 1999 North side).

To monitor the use of shrub habitat by migrating songbirds, 26 mist nets have been used at the "River Farm" area of Jug Bay Wetlands Sanctuary, during the peak of autumn migration (mid August to mid November) since 1997. Three main areas are being sampled: existing shrub habitat (0.2 ha), a planted shrub corridor (0.13 ha), and 50% of a previously planted 1 ha tree buffer. In the buffer, half of the area has been planted in native shrubs while the control areas remain unaltered. A total of 1600 shrubs have been planted at approximately 2 m spacing. Each year we conducted 42 days of netting, opening nets each day for 3.3 h, for an annual effort of approximately 3700 net hours. A mean of about 2400 birds of 89 species were captured, marked and released annually. A total of 115 species has been caught. The top 10 species captured were: Myrtle Warbler, Indigo Bunting, Song Sparrow, White-throated Sparrow, American Goldfinch, Common Yellowthroat, Swamp Sparrow, Chipping Sparrow, Blue Grosbeak, and Slate-colored Junco. As the shrubs mature, we hope to demonstrate that shrub habitat and understory is important, if not critical, to migrant songbirds. To date, shrub maturation has been insufficient for measurable results, but these first six years will serve as a baseline for us to monitor changes over time. In addition to the habitat experiment, the operation has served as a migration population monitoring station. This all-volunteer effort would not have been possible without the help of 8 to 12 individuals each year, in addition to the dozens of volunteers who planted the native shrubs.

Red-bellied turtles inhabit the freshwater portion of the Patuxent River and adjacent wetlands, and they nest in open upland areas at the edge of the river. Ours is the first detailed study of the ecology and reproductive biology of this species. We monitored 17 red-bellied turtle nests during three nesting seasons (1998 - 2000) to determine the effect of nest conditions on hatchlings and to identify factors that influence hatchling emergence patterns. Within the Jug Bay Wetlands Sanctuary females excavated nests in open, sunny meadows and lawns near the Wetlands Center and at the River Farm. Hatchlings from seven of the 14 productive nests overwintered and emerged in the spring, while hatchlings in four nests emerged in the fall. In three nests, hatchlings exhibited a "mixed" emergence pattern, with some emerging in the fall and others from the same nest emerging in the spring. In total, 43% of hatchlings emerged in the fall, and 57% emerged in the spring. Spring emergence typically followed an abrupt rise in maximum daily air temperature and sometimes coincided with rainfall. Hatchlings from some fall-emerging nests emerged after heavy rainfall. The mean mass of hatchlings emerging in the fall was significantly greater than that of hatchlings emerging in the spring. Hatchlings from eggs incubated during a wet summer in 1998 had significantly greater mean mass than those incubated during a dry summer in 2000. An understanding of a species' reproduction biology is crucial for effective management and conservation.

A limited supply of labile organic carbon in wetland soils leads to carbon competition among microbial functional groups. The dominant biogeochemical pathways of anaerobic carbon metabolism are driven by this microbial competition. Our study was designed to determine whether well-understood shifts in biogeochemical activity are reflected by microbial community composition. Eight tidal wetlands were studied along the Patuxent River estuary, from Jug Bay, a tidal freshwater marsh, to Jack Bay, a mesohaline saltmarsh. Duplicate cores were taken at each site and analyzed by Most Probable Number (MPN) microbial culture technique, Amplicon Length Heterogeneity (ALH) molecular fingerprinting technique, and standard biogeochemical techniques for quantifying methanogenesis, Fe(III) reduction, and sulfate reduction. Sulfate reduction was found to account for 94% of total carbon metabolism at Jack Bay (15‰ salinity), with methanogenesis suppressed to less than 1%. At Jug Bay, where sulfate concentrations are low and Fe concentrations are elevated, sulfate reduction accounted for less than 1% of carbon metabolism, while methanogenesis and Fe(III) reduction led to 23% and 77% respectively. The community composition of culturable microorganisms, however, did not reflect these large variations in biogeochemical activity. MPN cell counts showed no obvious variations in sulfate-reducing bacterial communities between sites, and cell counts for Fe-reducing bacteria were below detection in most of the samples from Jug Bay. We anticipate that the more sensitive ALH fingerprinting technique in progress will be able to detect changes in microbial biological diversity.

Interest in the stopover ecology of the sora (Porzana carolina) at the Jug Bay site of the Chesapeake Bay National Estuarine Research Reserve arises from the species' prominence in fresh-estuarine emergent marshes in general and in wild rice marshes in particular during fall migration. Foremost, the relationship is an energetic one in that soras depend on carbohydrate-rich seed sources, such as wild rice (Zizania aquatica) and smartweeds (Polygonum spp.), to build body reserves and fuel their migration. The importance of these particular marshes to soras cannot be understated, for in the Atlantic Coastal Plain they represent but a small percentage of tidal wetlands overall. Our presentation chronicles more than ten year's of research that began with the development of sora trapping techniques and ended in a study of their migration using radio telemetry methods.

Lyme disease is the most frequently reported vector borne disease in the United States. Maryland had the seventh highest number of Lyme disease cases from 1992-1998 with a mean incidence rate of 8.3 per 100,000 population. Borrelia burgdorferi sensu lato, the etiologic agent of Lyme disease is transmitted by the tick, Ixodes scapularis and is maintained in enzootic cycles with the vertebrate host, the White-footed Mouse (Peromyscus leucopus) in this region. Maryland appears to be a transition zone of Lyme disease risk, such that there is a high incidence of Lyme disease in northern Maryland and lower incidence rates in southern Maryland. This observation may be linked to the fluctuating population densities of nymphal I. scapularis which may be affected by abiotic or biotic factors such as weather, soil type, vegetation, dispersal or access to bloodmeals. To better characterize the enzootic cycles of Lyme disease in southern Maryland, we conducted a cross sectional study of 96 locations in state parks and natural areas in five southern counties by trapping rodents, collecting questing ticks and soil samples, and recording abiotic data. From June through October, 2001, 96 sites were evaluated totaling 11,520 trap nights. Eight hundred and three rodents (717 P. leucopus, 86 other captures) were trapped, blood and ear tissue samples were collected. Thirty five percent of all P. leucopus trapped in the five counties were infected with Borrelia based on genus specific PCR amplification. Three sites were examined at Jug Bay Wetlands Sanctuary, representing the three main land cover categories studied: open meadow, deciduous forest, and mixed forest. Twenty eight rodents were captured, including 23 P. leucopus and 5 Mus musculus. Forty percent of all captures at Jug Bay were infected with Borrelia. The majority of infected mice were captured in a forested area with primarily deciduous land cover and sandy loam soil texture. Twelve I. scapularis were collected at two sites (83% infected). Currently, geographic information system (GIS) and regression analysis are being employed to correlate infection prevalence and ecological factors such as land cover and soil texture. We hope to use this information to develop a risk map of the enzootic cycles of Lyme disease in five southern counties in Maryland.

Long known for majestic stands of wild rice, the marshes of the tidal Patuxent River, Maryland, experienced a catastrophic loss of rice during the decade of the 1990s. It was during this period that grazing by increasing numbers of resident Canada geese overwhelmed the rice and reduced it from extensive stands to isolated patches. Dramatic restoration of rice followed two management actions: first, the implementation of a cooperative management plan that effectively reduced the numbers of geese, and second, the proactive re-establishment of rice by use of protective fencing.

West Nile virus (WNV) was first isolated from a woman in the West Nile District of Uganda in 1937. Since its arrival to the Western Hemisphere in 1999, West Nile virus has acutely infected over 100 humans and caused 14 deaths. The outbreak in the Western Hemisphere originated in the New York City borough of Queens and has since spread to 27 states and the District of Columbia and to Ontario, Canada. Migratory birds are considered the primary dispersal agent of West Nile virus, and the virus may have taken a significant toll on their populations. During the three year epiornitic, over 10,000 dead birds (wild and captive), representing 83 species, 33 families, and 16 orders have been documented with evidence of West Nile virus infection. Here, I review aspects of the current epidemic in the United States including the basic biology, biogeography, and evaluate the impact the virus may have on individuals and populations of migratory and resident birds. Finally we will discuss the research we are conducting on West Nile virus at the Jug Bay Wetlands Sanctuary and surrounding areas.

We have conducted fish population surveys in the tidal and nontidal waters of Jug Bay Wetlands Sanctuary since 1987. Two sites are shallow Patuxent River tidal flats (Farm Point and Otter Point), two are nontidal permanent creeks (Two Run and Pindell Creeks), and one site consists of several beaver ponds. In 1993, we began using 10 m and 4 m long seine nets to capture the fish (previously we used minnow traps.) We developed a seining technique in 1997 to standardize the Farm Point data. We collected data on species composition, body length, and habitat use. We have identified a total of 45 species of fish. Seven species were common to all habitats. Species diversity varied by tidal regime as well as by habitat. Striped bass and hogchokers were only found in the tidal sites, while rosysided dace and least brook lamprey were found only in the nontidal creeks. The two creeks had similar species composition, yet dominant species varied. The two tidal sites at the Farm Point and Otter Point had similar species diversity and abundance. The beaver pond, located at the mouth of Two Run Creek at Otter Point, contained a mixture of tidal and nontidal fish species. I have used species diversity information and abundance to determine an Index of Biotic Integrity.

The appropriate timing of life history events is critical to an individual's survival and reproduction. The optimal timing of activities such as breeding may be influenced by predictable patterns in resource availability or risks, and constrained by preceding events. I tested whether nest predation and brood parasitism varied over the breeding season, and examined their influence on reproductive success in the Ovenbird, a migratory songbird. Predation of both real and artificial nests increased during the breeding season. In contrast, brood parasitism by Brown-headed Cowbirds was highest early in the breeding season. In spite of the increased risk of brood parasitism, early nests were more likely to fledge young than later nests. These results suggest that temporal patterns in nest predation may be important in determining annual fecundity for this species, and may select for early initiation of breeding. Additionally, events immediately prior to nesting that control the timing of breeding may be of critical importance, because delays in breeding would reduce reproductive success.

In Maryland, where an endemic cycle of West Nile Virus has been established, potential mosquito vector species are spatially distributed in ecologically structured environments determined by the individual species' bionomics. The biological and behavioral traits of these vectors suggest that existing trapping methods, commonly used to evaluate vector distributions and for surveillance of mosquito-borne pathogens, may be inaccurate or ineffective for collection of target species. Several studies have demonstrated that 1-octen-3-ol (octenol) used in combination with carbon dioxide (CO2) has significant species specific effects which could potentially enrich a sample for important vector species. In addition, vertical placement of a trap and habitat selection may greatly influence what species are captured. A goal of mosquito-borne disease surveillance is to collect a large sample composed primarily of vector species, therefore, developing a method which would enrich a trap sample for vector species is ideal. During the 2002 mosquito trapping season, CDC light traps were operated in two habitats at the Jug Bay Wetlands Sanctuary: freshwater swamp and transiently flooded woodland. Olfactory attractant and height effects on trapping within each habitat were evaluated using a 4x4 Latin square design. The light traps were baited with CO2 using a container filled with dry ice and with octenol using a reaction vial with a wick protruding through the cap. Similar to previous studies, results from this work indicate that the olfactory attractants have species specific effects. That is, some species were most attracted by CO2 alone while other species were most attracted by CO2 + octenol. No species were highly attracted to octenol alone or to no bait. Likewise, while most species were collected at all four heights tested, the height at which the greatest abundance of mosquitoes was collected varied by species. The results from this study will be utilized to develop trapping schemes to target mosquito species implicated in WNV transmission during the 2003 mosquito season.

Earthworms are often considered to be the most important soil invertebrate group because they influence the physical and chemical properties of soil, as well as the distribution of other soil organisms. A significant portion of the North American Oligochaeta fauna consists of non-native, mostly European species. This exotic earthworm invasion has recently received much attention. Maryland lies south of the southern limit of Wisconsinian glaciation, therefore both native and introduced earthworms occur in this state. The Baltimore Ecosystem Study is a recently established Long-Term Ecological Research site, with studies focusing on urban ecosystems. One of the basic questions of this project is the significance of non-indigenous flora and fauna in urban, suburban, and rural areas. As part of this study, we have assessed the Oligochaeta fauna in Baltimore City and surrounding counties. The Jug Bay Wetlands Sanctuary was one component of this survey. We have found a total of 26 oligochaete species, eleven of which occurred in the Sanctuary. Fifty percent of these species are native earthworms, a remarkably high proportion. Two species, Diplocardia patuxentis and Bimastos sp. nov. were determined to be new to science. One introduced species, Lumbricus friendi is new to the fauna of North America and several species were described from Maryland for the first time. Native species were mostly found in wet areas, such as an underground seepage in a ravine behind the Wetlands Center, areas along Two-Run Creek and along the Patuxent River. Given that habitat loss is the main cause of local extinction of invertebrates, preserving these habitats is extremely important for the survival of native earthworm fauna.

Eastern box turtles (Terrapene carolina) are considered to be common reptiles in suitable habitat throughout their range in eastern North America. In recent years, however, herpetologists have raised concerns about the population stability of this species. Suburban development, road mortality, collection for the pet trade and habitat degradation are all factors related to population declines. Taking advantage of the large, protected population of eastern box turtles in the sanctuary, we initiated a study of their ecology and population status in 1995. To date, we have captured and marked 409 adult and juvenile turtles. Through re-location and mapping of marked individuals we estimate population density in our 47 ha study area to be about 8.7 turtles/ha. This estimate is similar to those from other published studies. In 2000, we began conducting weekly censuses in several one-hectare plots. Two plots (#5 and #1) were located adjacent to or within150m of a tidal wetland; the third plot (#2) was about 500m from the tidal wetland. We conducted 77, 40-minute censuses in these plots in summer through the fall (2000-2002). Density in the plots varied from 1.7 - 4.6 turtles/ha. A total of 26 different turtles were observed in plot #1 in either 2000 or 2001. Fifteen of these 26 turtles (58%) were observed in both years, indicating that their home ranges incorporated parts of these plots. The 11 turtles observed only once or twice in the plots (42% of total) were either transients or their main home range may simply not encompass the plots; most of these were males. The sex ratio was 52% - 58% males; 29% - 39% females; and 7%-8% juveniles. Lucille Stickel's well-known, long-term study at the Patuxent Wildlife Research Center in Laurel, Maryland (only 20 miles from our study site), provides an invaluable source of relevant data with which to compare our results. The utility of small plots for estimating density may be limited, but they are useful for understanding other aspects of these turtles' natural history and use of habitat. Techniques that we have employed and data we have collected on population size and structure may be useful to herpetologists and resource managers elsewhere who need to assess the health and ecology of box turtle populations.

More than 90% of the phosphorus (P) carried by rivers to estuaries is associated with suspended solids. Much of this particulate P (PP) may by bound to iron oxides. We hypothesize that iron oxide bound P is retained in freshwater sediments but released from brackish water sediments where the oxidized iron is reduced and precipitated with sulfide. We are testing this hypothesis in an NSF-funded study of the upper Patuxent River estuary where a well-defined salinity gradient extends from the tidal freshwaters of Jug Bay to brackish waters near Benedict, MD. We used serial extractions to quantify the different forms of PP in suspended sediments entering the Patuxent River estuary and in suspended and deposited sediments along the salinity gradient. About 50-90% of the PP was extracted by a citrate, dithionite, bicarbonate solution, which removes mostly phosphate bound to iron oxides. Lesser amounts of PP were released by extractions designed to quantify organic P, loosely sorbed phosphate, and authigenic and detrital apatite P. Deposited sediments in areas where salinity was less than 1 ppt had higher concentrations of total PP and more iron oxide bound phosphate than did sediments where salinity varied from 1-6 ppt seasonally. The decline in iron oxide bound phosphate with increasing salinity coincides with a marked increase in the efflux of dissolved phosphate from the sediment and a peak in dissolved phosphate concentrations in the estuarine water column. Sulfide production supported by sulfate from sea salts may enhance the mobilization of iron oxide bound phosphate from brackish sediments. This salinity induced release of dissolved phosphate may contribute to a shift from P limitation in freshwater to nitrogen limitation in saltwater.