Applying a coupled biophysical model to predict larval dispersal and source/sink relationships in a depleted metapopulation of the eastern oyster Crassostrea virginica.
KEY WORDS: Crassostrea virginica, connectivity, biophysical model, restoration. Gulf of Mexico, Pensacola Bay
Eastern oysters (Crassostrea virginica) and the reefs they construct provide valued ecosystem services to estuaries (Kennedy et al. 1996, MacKenzie et al. 1997, Coen et al. 1999) and have supported an economically and culturally important U.S. fishery for centuries (Kurlansky 2006). Reefs of C. virginica, however, have been declining in distribution and abundance for decades (Beck et al. 2011), resulting from a diverse suite of factors including disease (Powell et al. 2008), alterations of water management strategies (Livingston et al. 1997, Parker et al. 2013), severe weather events (Berrigan 1988, 1990, Livingston et al. 1999), and overfishing (Wilberg et al. 2011). The northern Gulf of Mexico (GOM), where the following study was conducted, has similarly experienced declines (Seavey et al. 2011), although those declines are less severe than is reported for the eastern seaboard of the United States (Hargis & Haven 1999, Beck et al. 2011). Instead, GOM oyster reefs remain relatively abundant and continue to support a viable wild oyster fishery that supplies over 50% of the average global supply of wild oysters (Beck et al. 2011).
Along the GOM coast of Florida, oysters have supported a commercial fishery since at least the late 1800s (Ingersoll 1881. Swift 1897, Danglade 1917). Harvest peaks were recorded in 1908, 1928, 1945, 1962, and 1981 (GSMFC 2012). No similar peaks have been recorded since 1981, reflecting a decline in oyster abundance along the Florida GOM coast that appears to be more severe than that observed in other areas of the Gulf (Seavey et al. 2011, Beck et al. 2011). Historically, Apalachicola Bay in north Florida has provided an exception to this pattern of decline, contributing approximately 90% of the Florida wild oyster harvest at about 1 million kilograms annually (http:// myfwc.com/research/saltwater/fishstats/commercial-fisheries/ landings-in-florida/). Even Apalachicola Bay, however, suffers from episodic (Berrigan 1988) and chronic declines (Beck et al. 2011), culminating in the 2012 collapse of the Apalachicola Bay oyster population (Pine et al. 2015). Because of the importance of the fishery and the predominance of the Apalachicola Bay system to the productivity of the GOM oyster fishery, the Apalachicola Bay oyster population has been a continuing target of research to identify factors regulating oyster distribution and ecology (e.g., Pine et al. 2015). Recent research has largely focused on understanding the relationship between water management practices, water quality, and the health of the oyster resource (Livingston et al. 2000, Livingston 2015), although Pine et al. (2015) focused on the influence of oyster population dynamics.
Oyster production in other bays and estuaries of the Florida panhandle may equal or exceed that of Apalachicola Bay during some years (Arnold & Berrigan 2002), but those locations have been less studied and are less well understood even though they are similarly susceptible to acute and chronic impacts (Livingston 2015). The Pensacola Bay System (PBS), in the northwest corner of Florida, is no exception. The PBS is composed of three sub-basins, including Pensacola Bay, Escambia Bay, and East Bay (Fig. 1), the latter of which includes Blackwater Bay at its northern extent. Oyster reefs historically occurred in Escambia and East Bays (Fig. 1), although scattered oysters occurring in other areas of the bay probably contribute to ecological function (Drexler et al. 2014). The oyster reefs that occupy the PBS are generally representative of Florida panhandle oyster reefs and supply commercial oyster harvest of about 1 % of that historically harvested in Apalachicola Bay (http:// myfwc.com/research/saltwater/fishstats/commercial-fisheries/ landings-in-fiorida/). Although exhibiting considerable inter-annual variability, oyster landings before the 1980s did not exhibit significant upward or downward trends. Landings peaks occurred in the 1960s (Dugas et al. 1997), the 1970s (Collard 1991), and again in the 1980s (Dugas et al. 1997). Since that early 1980s average of 307,000 pounds of meats, oyster landings in Pensacola Bay have steadily declined to an average of 73,000 pounds in the early 1990s, 18,000 pounds in the early 2000s, and 6,000 pounds in the early 2010s. This decline has been attributed to changes in freshwater inflow, loss of suitable settlement substrate, shoreline development, and possible overharvest (USEPA 2004). Intermittent restoration efforts have been implemented in an attempt to reverse this trend, but with limited success. Only 104 ha of the estimated 3,144 ha of original reef structure have been "restored" through shell planting since the 1970s (GSMFC 2012). A more recent effort (Project Greenshores; Ray-Culp 2007) has focused on habitat creation and shoreline stabilization in a region of the bay where the shoreline is highly altered, rather than for development of harvestable reefs. The long-term success of this effort remains to be determined.
Oyster populations within estuaries, including the PBS, likely function as metapopulations (sensu Sale et al. 2006) characterized by patchy habitats demographically connected through the larval life phase. Metapopulation structure has been described for the bay scallop Argopecten irradians in the eastern GOM (Bert et al. 2014), a species exhibiting local and mesoscale distribution similar to that of the oyster. To the extent that oysters in the eastern GOM do exhibit metapopulation structure, that structure is poorly understood and is only recently developing as a consideration in the plight of the eastern oyster in Florida and throughout the species' range (Haase et al. 2012, LeCorre et al. 2012, Munroe et al. 2012, Kjelland et al. 2015). At present, evidence for oyster metapopulation structure is primarily based on genetic studies, which reveal complex larval exchange patterns among estuaries spanning the entire Gulf (Gaffney 2006, Varney et al. 2009). In theory, both small-scale community-based efforts and large-scale fishery restoration projects serve to increase connectivity among local populations (sensu Hanski & Ovaskainen 2003) that comprise a metapopulation by enhancing larval sources and by providing increased continuity of settlement substrate for larval immigrants.
Connectivity among local oyster populations relies on larval dispersal among those populations. The intricacies of dispersal are not fully understood but evidence supports an important behavioral component in the form of larval swimming. Vertical swimming patterns of oyster larvae affect their depth in the water column, which exposes the larvae to variable patterns of tidal and wind-driven currents (Mann 1988, Baker & Mann 2003) that may influence patterns of retention and dispersal. For example, even small vertical movements by oyster larvae in Chesapeake Bay influence their dispersal and resultant source/ sink dynamics (North et al. 2008). In Mobile Bay. AL, however, observations of the distribution of recently settled oysters (Kim et al. 2010, 2013) indicate a west-east gradient in dispersal and settlement. Those authors used a numerical larval transport model to demonstrate that this distribution was only weakly impacted by the vertical movement of larvae. Such complex retention/dispersal patterns of oyster larvae have been observed in other northern GOM estuaries (Hoese et al. 1972, Saoud et al. 2000) and for other coastal marine species (Leming & Johnson 1985, Bert et al. 2014). Thus, many questions regarding oyster metapopulation dynamics remain, particularly with respect to relative retention, connectivity, and repopulation of depauperate habitats.
In this context, the relationship between source and sink populations (Pulliam 1988) is essential to understanding metapopulation function. Local populations may function as larval sources, contributing recruits for the maintenance of other local populations, or as sinks that are dependent for population maintenance upon larval inputs from neighbors or distant populations (Lipcius et al. 2008). Certainly, a single local population (i.e., reef) may function as a source both for itself (self-seeding) and for neighboring populations. This relationship may be time variant, in that a population may serve as a source at one time but as a sink at another time (Bert et al. 2014, Long et al. 2014) dependent on factors such as the timing of spawning, biology, transport factors, and standing stock. These relationships influence the population abundance and structure as well as the genetic composition of the metapopulation (Buroker 1983, Hare & Avise 1996, Varney et al. 2009). Understanding the mechanisms of larval exchange within the context of source/sink dynamics is therefore fundamental to understanding the process of recovery from natural or anthropogenic disturbance. Knowledge of source/ sink dynamics and resultant connectivity patterns among oyster reefs is an important precursor to designing a successful oyster restoration program. Lack of this knowledge likely has contributed to the limited success of many oyster reef restoration programs.
Here, data obtained from May 2007 through July 2008 are applied to calibrate a hydrodynamic model of PBS to describe circulation and salinity patterns. An individual-based model of the oyster's larval biology (Narvaez et al. 2012) is coupled to the hydrodynamic model to hindcast the temporal and spatial patterns of larval dispersal during four discrete 20-day events. The model simulations are then compared against observations of oyster larval abundance and juvenile recruitment to evaluate the utility of the model for predicting larval retention, exchange among reefs, and export from the estuary. Combined results provide guidance to resource managers regarding the choice of locations for oyster restoration efforts in the PBS, restoration strategies throughout the range of the species, and generally to sustainable management of exploited marine resources.
MATERIALS AND METHODS
The northern GOM, where the PBS is located, is composed of a series of shallow estuaries characterized by large surface area to volume ratios, large freshwater inputs, and narrow, complex, and dynamic openings to the Gulf (USEPA 2004, Park et al. 2014). The surface area of the PBS is 372 k[m.sup.2], mean tidal amplitude is 0.5 m, and average depth ranges from 2.5 m in the north to 6.0 m in the south. Circulation is wind and river dominated, both of which are strongly seasonal. The major freshwater inputs emanate from the Escambia (181 [m.sup.3]/sec annual average) and Yellow (34 [m.sup.3]/sec) Rivers (Bass & Cox 1988). Those flows are maximal in October and November and were five times greater than flow rates recorded in February through April during this study (Fig. 2). A detailed overview of the physical setting of the PBS is available in Livingston (2015).
Empirical Sampling und Analysis Water Quality and Larval Abundance
Existing oyster reefs in the PBS were mapped using historical information (Little & Quick 1976), local knowledge, and visual/ poling surveys conducted in May 2007 and spring 2008. During May 2007 and monthly from June 2007 through July 2008, samples for larval quantification and basic water quality information were collected at 20 stations (Fig. 1). Sample stations were distributed throughout the study domain following a depth-stratified random design with Stations 1-9 ("shallow") falling between the 1-m contour and shore, Stations 10-15 ("mid-depth") falling between the 1- and 2.5-m-depth contours, and Stations 16-18 ("deep") distributed in depths greater than 2.5 m. Within each stratum, station locations were assigned using a gridded random allocation approach. On each sampling date, salinity and temperature were measured at 0.5 m vertical intervals at each station using a YSI 85-50 handheld meter. Oyster larvae were collected from each station by pumping water through a 35-[micro]m-mesh plankton net. Water was pumped using a diaphragm bilge pump attached to a 2.5-cm-diameter hose capable of reaching the desired depth (Frischer et al. 2000, Arnold et al. 2005). The pump was run for 1-2 min at the sample depth before sample collection to flush the system. Approximately 250 1 (range = 242-270 1) of water was filtered, except at Stations 4 and 5 in May 2007 when 5001 of water was collected. This volume, adequate to quantify the larvae (Govindarajan et al. 2015) without overwhelming the methodology (see 'Larval Supply' below), was collected at each of one or more depths depending on the depth of the sample station. At shallow stations, a single water sample was collected 0.5 m below the surface. At mid-depth stations, samples were collected 0.5 m below the surface ("near-surface") and 0.5 m above the bottom ("near-bottom"). At deep stations, samples were collected 0.5 m below the surface, 0.5 m above the bottom, and midway between the two ("midwater"). Two additional water quality stations (no larval samples collected) were included near the mouth of Santa Rosa Sound (Station 19) and Pensacola Pass (Station 20) to assess boundary conditions for model calibration, though collecting bottom samples of temperature and salinity was not always possible in Pensacola Pass due to very strong currents.
Data obtained from oyster recruitment arrays are used as a proxy for settlement, allowing a comparison of empirical patterns of recruitment relative to modeled patterns of settlement. Oyster recruitment was monitored from May 2007 through July 2008 by deploying one recruitment monitoring array at each of the nine shallow and six mid-depth stations. Each array consisted of 12 axenic adult oyster shells (shell height = 5-10 cm) with a hole drilled in the center. Two lengths of galvanized wire were strung with six shells each, which were then suspended from the arms of a T-shaped PVC frame. The shells were oriented with their inner surfaces facing downward, and the PVC frame was pushed into the sediment until the bottommost shell was approximately 5 cm above the sediment surface (Parker et al. 2013). The first set of arrays was deployed only at the six mid-depth stations in May 2007 and retrieved 6 wk later. Beginning in June 2007 and monthly until final retrieval in July 2008, arrays were deployed and recovered monthly from all nine shallow and all six mid-depth stations. Collectors were not deployed at the deep stations because there is no record of oyster reefs occurring at such depths in the PBS. Only counts from the bottom of each shell, and from all but the top and bottom shells in each string, were included in the analyses because data from the upper surfaces and exposed top and bottom shells are biased by high levels of fouling (data not shown). The number of oyster recruits was summed within each station on each date and the average number of recruits per shell per day calculated.
Thirty water samples for estimating larval abundance were collected from 18 stations monthly (total n = 390). Concentrated plankton samples (settled volume of 0.5-35.0 ml of filtrate) were immediately frozen on dry ice in a 50-ml centrifuge tube and then stored at -80[degrees]C until extraction. Oyster larvae were identified and semiquantitatively enumerated by real-time quantitative polymerase chain reaction (qPCR) essentially as previously described (Frischer et al. 2000, Arnold et al. 2005). Briefly, the entire sample was resuspended to 50 ml with sterile artificial seawater and total genomic DNA purified from a 0.2ml subsample using the Qiagen DNeasy blood and tissue kit according to manufacturer's specifications following the recommended protocol for optimal yields. The extracted volume (0.2 ml) was roughly equivalent to a 1-1 seawater sample. To estimate the abundance of oyster larvae in seawater, the abundance of oyster 18S rRNA gene fragments were estimated by qPCR using a pair of oyster-specific 18S-rRNA-targeted oligonucleotide primers designed for this study. The primer pair consisted of the forward primer CV 662F (5' GGT CCA CCT CGT TGT GGT TA) and the reverse primer CV933R (5' GGA TCC AAG CCT TTC ACC TC). This primer set consistently and efficiently produced a 271-bp amplicon from Crassostrea virginica. Optimal amplification cycle conditions were 95[degrees]C for 5 min followed by amplification cycles of 95[degrees]C. 15 sec; 60[degrees]C, 30 sec; 72[degrees]C, 60 sec. The qPCR assays were facilitated using an MX3000P real-time PCR instrument (Agilent Technologies) with ROX dye as a reference (585 nm excitation, 610 nm emission) and SYBR green (492 nm, 516 nm) as the fluorescent dye used to mark and detect amplified DNA fragments. In each assay, a "no template" blank and a series of standards of known concentrations of oyster larvae were included. Specificity of the larval "probe" was confirmed in silico by searching the Gen-Bank and Silva databases using Primer-BLAST (Ye et al. 2012) and TestPrime utilities (Klindworth et al. 2012), respectively. Assay specificity was further confirmed empirically by testing against 13 marine bivalve species, 2 freshwater bivalves, 3 crustaceans, 3 algal species, 2 bacteria, and several natural plankton Haliotis samples that did not contain oyster larvae (data not shown for brevity). Similar studies have been successfully used to quantify clam Rudilapesphilippinarum (Quinteiro et al. 2011), abalone kamtschcitkana (Vadopalas et al. 2006), and oyster Ostrea lurida (Wight et al. 2009) larvae.
Model Development and Application
The Numerical Circulation Model
Circulation in the PBS was simulated by the Estuarine Coastal Ocean Model developed from the Princeton Ocean Model (Blumberg & Mellor 1987). The model used a three-dimensional (3-D), primitive equation, time-dependent finite-difference algorithm with vertical sigma coordinates, boundary fitted curvilinear grid, a split time step for the solution of the baroclinic 3-D mode and the barotropic 2-D mode, and an embedded Mellor-Yamada second-order turbulence closure model that provided vertical mixing coefficients. A Smagorinsky diffusivity (Smagorinsky 1963) was used to account for the horizontal diffusion. Since some boundary conditions were not available, the model was run without active thermodynamics. The model computed the currents, salinity, and water levels at all cells in the active model domain using time-dependent equations for the conservation of mass and momentum. A "wet-dry" capability (Oey 2005) allowed for numerically stable simulation of drying cells and subsequent "wetting" as water level rose (Ma et al. 2009, Xue & Du 2010). When the modeled water depth was less than 0.2 m, a grid cell was considered to be inactive until the water depth rose above 0.2 m again.
A 90 X 90 model grid was overlayed on the coastline (Fig. 1), and cells were defined as potentially active only if they were within the coastline. The model defined 2,533 active cells ranging in horizontal dimension from 200 to 975 m (mean horizontal model resolution = 380 m); 72% of cells were less than or equal to 500 m. Cells outside the coastline were considered land and rendered permanently inactive. The area-weighted mean depth was 3.9 m [mean sea level (MSL)]. A 15-sec time step was used to avoid the Courant-Friedrichs-Lewy instability condition (Haidvogel & Beckmann 1999). The 10-layer vertical grid was compressed near the bottom and top for better resolution of the boundary layers. This yielded a layer thickness ranging from 6.2 mm for a surface layer in a shallow region to about 3.75 m for a mid-depth cell in the deepest location. The interface between the bay and the GOM was defined as a forced open boundary with elevation and salinity values taken from observations.
Bathymetry (Fig. 1) was generated from the merged bathy-topo dataset available from NOAA (http://estuarinebathymetry. noaa.gov/) with water depth and land elevation available in a 30-m grid. Most grid cell depths h(i, j) were defined relative to MSL by averaging all the measurements within the cell at grid location (i, j), though a few grid cells were manually altered to avoid numerical instability associated with areas of steep topography (Mellor et al. 1998).
The model was run from January 2007 through July 2008 using observed salinity data (Fig. 3) and hourly water levels from NOAA station tide gauge no. 8729840 as forcing functions for the model open boundary at the bay mouth (Fig. 1). A 1-h delay in water level was assumed to account for propagation time from the mouth of the bay to the gauge. Additional salinity values for the open boundary were obtained from a Florida Department of Environmental Protection water quality station located at the mouth of Pensacola Bay (Fig. 1). Daily mean surface winds and precipitation were also obtained from NOAA station no. 8729840 and modeled winds were applied uniformly across the bay following unpublished work in Tampa Bay, FL (M. Wilson, personal communication). Daily freshwater river discharge was estimated from USGS streamflow gauges on the Yellow River (gauge no. 02369600) and Escambia River (gauge no. 02376033). Note that limited knowledge of additional freshwater discharges may influence model accuracy, but the volume of freshwater contributed from other point and nonpoint sources is generally minimal compared with the two major rivers. Evaporation was approximated as a mean value, nominally 0.25 mm/day (Meyers & Luther 2008). The numerical model yielded hourly estimates of salinity S(i, j, k, t) and velocity u(i, j, k, t) over the three spatial indices (i, j, k) and time t where the velocity was the 3-D velocity vector (m, east-west; v, north-south; w, vertical component).
The model was calibrated by minimizing the difference between the observed salinity [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII], where the sampled location is p, at times [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII], monthly time index mp and depth index [l.sub.p], and the model salinity at the nearest active grid cell [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]. Using the model output time [t.sub.m] closest to [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] and limiting n to the surface and bottom values, the mean errors [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] where N(= 1,732) is the total number of measurements. Model parameters that control mixing in the model were adjusted until a minimum of [bar.[DELTA]S] was found. The minimum [bar.[DELTA]S] was -3.6. Similarly, the root mean square of the salinity error was 5.0 compared with a salinity range of 26.2.
Four numerical simulations were performed to estimate larval dispersal from Pensacola Bay Stations 6 and 7, Escambia Bay Stations 5, 10, and 11, and East Bay Stations 1, 3, 12, 14, and 15. Those stations were chosen based on their proximity to known oyster reefs (Fig. 1). Each start location in the model was seeded with 10,000 particles representing oyster larvae, which were released simultaneously just off the bottom of the bay. Model release dates were August 9, 2007, October 3, 2007, June 9, 2008, and July 8, 2008, ensuring contemporaneity with field sampling efforts. Virtual larvae were followed for 20 days to ensure the longest expected life span of oyster larval in GOM waters was captured (GSMFC 2012).
Larval dispersal was simulated using a Lagrangian scheme (Burwell 2001, Meyers & Luther 2008) composed of a forward Eulerian scheme with a very small time step (15 sec) relative to the dominant diurnal and semidiurnal tidal signals. Each larval "particle" carried a unique identification number allowing the tracking of individual pathways as well as a variety of statistical measures of the particle distribution. The forward scheme was augmented with a 3-D random motion to represent subgrid circulation. The random term's vertical component was 10 4 of the horizontal term, roughly equal to the aspect ratio of the estuary. The hydrodynamic Lagrangian model was coupled to a biological model as described in Narvaez et al. (2012), which yielded an additional vertical displacement of the larvae reflecting larval swimming and which increased with increasing larval size (= age). The initial size of each larva was 50 [micro]m, and the growth rate of each larva was based on current size, food availability, temperature, and salinity. Temperature from the observations was interpolated to the model grid for use in the biological model. Values for those parameters were approximated using values for Crassostrea virginica from Delaware Bay (Narvaez et al. 2012) since no similar data were available for the PBS; growth rate was increased by 40% to reflect warm water physiology and ensure the larvae grew to setting size within 20 days. Food was held constant at 4 mg C/l as in Narvaez et al. (2012) to impose an unlimited food supply to the larvae. When a larva reached 55 [micro]m it began to swim upward at a nominal rate of 1 mm/sec, allowing the larval population to congregate near the surface in about 1 day. When a larva reached 220 |am it began to swim less and was fully settled on the bottom at 330 [micro]m.
It was assumed that any larvae crossing one of the open boundaries were flushed out of the system and did not return. The fraction of these flushed larvae was calculated for each simulation. The fraction of larvae that had reached 330 [micro]m (assumed to have settled) by the end of each 20-day simulation was also calculated, and from that the rates of self-seeding and larval transfer between the model seeding sites were calculated. Transfer was defined as the fraction of larvae released at one station that settled within [+ or -] 2 grid cells around another station.
Salinity and Temperature
On all dates a salinity gradient from the head of the estuary to the mouth was observed (Fig. 3), driven primarily by discharge from the Escambia and Yellow Rivers (Fig. 2). Precipitation generally contributed only a small fraction of the total freshwater discharge except during short-term events, the most pronounced of those events occurring during late fall 2007 through early winter 2008 (Fig. 2). At the northernmost sites near-surface salinities were generally 0-20 and near-bottom salinities generally 0-25 through the sampling period (Fig. 3). Near-bottom salinity at deeper stations varied by as little as 5 but was much more variable at shallow stations. Near-surface salinities varied by 20-25 within many stations, both deep and shallow. Vertical stratification was as high as 20 and relatively steady at deeper stations although intermittent high stratification was found in shallower locations in the northern PBS near river inputs. Salinity was lowest and stratification commonly largest in late 2007 through early 2008 during heightened river discharge events (Fig. 2) though some stations (e.g., Stations 5-8) had reduced salinities throughout the water column at that time.
The temperature was homogenous throughout the study area but exhibited significant seasonal variation (Fig. 4). Temperatures across the bay ranged from a high of ~34[degrees]C in the summer to a low around 10-15[degrees]C in the winter. Vertical variations in temperature were generally less than or equal to 1[degrees]C at most sites but were as large as 3[degrees]C at the deepest locations.
Low levels of recruitment, generally less than 0.1 recruits/ shell/day, were recorded from the initial (July 10-12, 2007) collector recovery at Stations 10 and 11 in Escambia Bay and 12, 14, and 15 in East Bay (Fig. 5). The collector at Station 13 was lost, and no collectors were deployed at Stations 1-9 during this period. Recruitment increased substantially in both number and spatial distribution beginning in August and continuing through October. In August, relatively large numbers of recruits, sometimes exceeding 0.5/shell/day, were recorded in all three basins including along the south shore of the PBS from Station 8 eastward to Station 15. During late September, peak densities were recorded at Stations 6 and 7 in Pensacola Bay, whereas in October only low levels of recruitment were recorded in Pensacola Bay and along the south shore of the PBS. Instead, during October recruitment predominated in East Bay, including very high levels of recruitment at Stations 12 and 14. Although the highest level of recruitment during October was recorded from Station 11 in Escambia Bay, few if any recruits from the other stations in that bay segment or from the Pensacola Bay segment were recorded.
Recruitment decreased substantially in November, and was essentially zero from December through May. Both the June and July recoveries revealed low levels of recruits, generally less than 0.2 recruits/shell/day with highest numbers at Station 8 in Pensacola Bay and Stations 10 and 11 in Escambia Bay during June.
The CV 662F-CV933R primer pair RT-PCR produced a consistent product from adult oyster tissue, first observable during qPCR cycle 16 (threshold cycle, [C.sub.t]) and continuing to at least [C.sub.t] 60 in extreme dilutions. The product expressed a consistent melt temperature of 82.8[degrees]C [+ or -] 1 [degrees]C, indicative of a product length of 271 bp with a sequence exactly matching the GenBank record for Crassostrea virginica. This primer pair did not produce PCR products when amplifying other metazoans but did produce some PCR products when challenged with filtered plankton. Those products typically had melt temperatures of 60-70[degrees]C and their sequences do not match that previously described for C. virginica. These results yielded a high certainty in positive qPCR reactions, but quantification proved difficult due to the aforementioned interference from general plankton that occurred in some samples. Thus, larval abundance results are expressed in terms of high (>~ 1 larvae/1; [C.sub.t] [less than or equal to] 45), low (>0 but <~1 larvae/1; 45 < [C.sub.t] [less than or equal to] 50), and zero (<~[10.sup.-3] larvae/1, no oyster DNA detected or oyster DNA detected when C, >50).
Of the 390 plankton samples collected, 3 were lost during processing, all in November including the near-surface and near-bottom samples at Station 10 and the near-bottom sample at Station 11. There was no DNA amplified in 87 of the samples using the oyster primer pair. Nontarget DNA was amplified in 37 of the samples, typically at the deep water stations of Pensacola Bay proper, nearest to Garcon Point. Of the remainder, 201 of the samples exhibited DNA that amplified after cycle 50 but maintained the characteristic melt curve at ~82.8[degrees]C. Thus 62 samples resulted in oyster larval abundance estimates that were either "high" or "low" based on the estimation scheme. The maximum observed plankton signal in this study was at Cc 37.3, roughly equivalent to 3 larvae/1.
Oyster larvae were detected in one or more near-surface samples during all but the December and January sample events (Fig. 5). Larvae were most abundant and widely distributed beginning in February 2008 and then during May, June, and July especially at shallow Stations 1-9 but also quite frequently at mid-depth Stations 10-15. No near-surface larvae were recorded from deep stations 16-18 located in the main stem of the estuary before February, but near-surface larvae were recorded at one or more of those three stations from the February. May, and July events. Larvae were recorded from the midwater and near-bottom samples collected at Stations 1618 on various sampling dates, albeit sporadically. A similar sporadic larval distribution pattern was observed in near-bottom samples from mid-depth Stations 10-15.
Simulated Larval Dispersal, Settlement, and Export
Exemplary status plots capture modeled larval locations on each of days 5, 10 (day 8 for the July 2008 event to correspond to the last day of field sampling), 15, and 20 following particle release simulations initiated on October 3, 2007 (Fig. 6), and July 8, 2008 (Fig. 7). In each case, wind-driven currents affected larval spatial distributions resulting in time-varying displacement. Modeled larval distributions on each of those 4 days, for each of the two modeling periods, are discussed in the following paragraphs.
October 3, 2007, Modeled Release
On Day 5, larvae dispersed along the western shorelines of all three bay segments but were largely absent along the eastern and southern boundaries, apparently reflecting the influence of 48 h of moderate but sustained winds with strong easterly components driving the larvae westward. On Day 10, the larvae were more broadly distributed, reflecting a directionally variable wind field during the previous 48 h. Day 15 followed a period of strong sustained SE winds and the larvae were again largely confined to the western shorelines, although larvae were more broadly distributed in Escambia Bay and especially the northern reach of that segment. On Day 20 of the simulation, the larvae were mostly distributed across the southern and western portions of the PBS, with a high concentration in the Pensacola Bay segment. Winds 48 h prior were strong from the SE, shifting to moderate from the NW 24 h prior then returning SE and strong on the modeled day.
July 8, 2008, Modeled Release
Larvae dispersed throughout the PBS on Day 5, except in the Blackwater Bay portion of East Bay and in northern and central Escambia Bay (including a band running from Stations 5 to 10), reflecting a diverse wind field in the previous 48 h. Wind was similarly diverse ahead of Day 8, blowing from the SSE 48 h ahead of Day 8 but then turning to the WNW 24 h out and from the NW on Day 8. The latter 2 days of wind apparently pushed modeled larvae into the southern PBS with only scattered patches north of Garcon Point. Winds were then steady from the SSW ahead of Day 15, resulting in movement of larvae into south facing shorelines and into the upper reaches of Escambia and East Bays. By Day 20, the modeled larvae were predicted to have been pushed back into the southern and especially eastern portions of East Bay, the southern reach of Escambia Bay, and in abundance in Pensacola Bay. This pattern appears to have been driven by a variable wind field in the prior 48 h, first emanating from the SW, clocking around to the NW, then blowing from the west on the day of prediction.
The simulations also produced settlement patterns for each of the four runs (Fig. 8). Final day outcomes were strongly dependent on the release date used in the simulation. In August 2007, competent (i.e., ready to settle), larvae were predicted to be abundant in the East/Blackwater Bay segment of PBS, as well as in the far northern reaches of Escambia Bay. Settlers were also pushed against the northwestern shoreline in both East and Pensacola Bays. As discussed above, outcomes were almost opposite in October 2007, with few (Escambia Bay) if any (East/Blackwater Bay) competent larvae predicted for the northern reach of the PBS but high densities of competent larvae predicted along the western shoreline of lower Escambia Bay, in the southern reach of East Bay, and throughout Pensacola Bay. The June 2008 pattern also reflected few competent larvae in the northern reaches of the PBS. This simulation, however, predicted dense concentrations of competent larvae along the eastern shorelines in both Escambia and East Bays and lower densities throughout the southern reaches of both those segments and in Pensacola Bay. The July 2008 simulation reflected a combination of the outcomes from the October 2007 and June 2008 events. Competent larvae were essentially nonexistent in northern Escambia Bay and confined to the eastern shoreline of Blackwater Bay, but otherwise spread throughout the PBS with extremely high densities along the eastern shores of East and Escambia Bays.
The general patterns of larval exchange among "reefs" (= release points) was analyzed by averaging the percent contributions from all points contributing to a single site to evaluate sources of larvae and from all points receiving from a single site to evaluate sinks for contributed larvae. This approach revealed that donor reefs were found in all three basins (Table 1), with all modeled reefs contributing between 4% and 8% of their simulated larvae either to settlement at that same reef or at another reef within the PBS system. This relatively balanced pattern of larval contributions contrasts sharply with the uneven distribution of average larval receipt among reefs, ranging from less than 1% of modeled larvae arriving at Station 15 in East Bay to a high of almost 15% of modeled larvae arriving at Station 6 in Pensacola Bay. Both Pensacola Bay stations received a relatively large total percentage of the released larvae, but larval receipt was substantially lower and more variable among stations within the other two basins.
On a station-by-station basis, relationships between donor and recipient sites were more intricate, both in the case of dispersal and recruitment within the donor basin and in the case of larvae dispersed to neighboring basins (Table 1). As noted above, Stations 6 and 7 in Pensacola Bay received relatively high percentages (commonly >1%) of larvae from all other stations but especially via self-seeding or from the other of the two Pensacola Bay stations. Stations 6 and 7 also contributed sometimes substantial (0.14-1.55) percentages of their larvae to the Escambia Bay stations. In contrast, the Pensacola Bay stations contributed a very low percentage of their modeled larvae to any stations in East Bay, the maximum being 0.24% from Stations 7 to 12 across the mouth of Escambia Bay.
Similarly, donor sites in Escambia Bay contributed to both stations in Pensacola Bay and to other stations within Escambia Bay. The lowest level of exchange within Escambia Bay was 0.13% of larvae emanating from Station 11 and setting at Station 10. Peak exchange within Escambia Bay was from Stations 5 to 11 (1.62%) and from self-seeding at Station 11 (3.30%). Escambia Bay contributions to stations in East Bay were generally low (<0.30%) with the exception of Station 14 that received between 0.56% and 1.59% of the larvae released from Escambia Bay stations.
As with the other two basins, East Bay contributed larvae to both Pensacola Bay stations but especially to Station 6. Only one East Bay station (12) contributed less than 0.50% of its larvae to a Pensacola Bay station. East Bay stations contributed very few larvae (<0.10%) to any station in Escambia Bay with the exception of a small percentage of larvae from Station 15 to Station 5 and consistent contributions (0.17%--0.61%) from all five East Bay stations to Station 11. Instead, the highest contributions of settlers coming from East Bay were to the same or other reefs within East Bay. All East Bay stations contributed more than 1.00% of their larvae to at least one other East Bay station (including self-seeding). The highest contribution of larvae from one station to any other station in the PBS occurred within East Bay, specifically Station 15 contributing over 4% of its larvae to Station 14. Stations 1, 12, and 14 also received large percentages (commonly >0.50%) of larvae from other East Bay stations, whereas Stations 3 and 15 received lesser percentages (always <0.20%).
The highest proportion of estimated larval export from the PBS occurred in the October 2007 simulation (Table 2), which predicted almost 10% of the total larval pool was exported. The age of exported larvae in this simulation exhibited a bimodal pattern, characterized by an initial export of larvae aged 6-13 days followed by a relatively large export of Day 18-19 larvae. During the other three simulations, export rates were less than 2%. In each of those three cases, the larvae were predominately 12 days of age or younger, in some cases substantially younger.
Based on historical information (Beck et al. 2011, GSMFC 2012), landings data, and initial surveys of the system, oyster reefs are depauperate in the PBS relative to other GOM and eastern U.S. estuaries. This appears to be a relatively recent development, reflecting a substantial decline during the last 50 y from over 3,000 ha in 1971 to their present depleted state (Collard 1991, Beck et al. 2011, GSMFC 2012). Despite these low adult densities, larval abundance was consistent with empirically determined estimates for Crassostrea virginica obtained along a salinity discontinuity in Chesapeake Bay (Mann 1988) albeit generally much lower than estimates for C. virginica from Carriker's classic 1951 study in various New Jersey estuaries. The level at which oyster larval abundance becomes limiting is not clear, and the disparity between data from Mann (1988) and Carriker (1951) provides no guidance. Levels of recruitment, however, to the collectors commonly fell within the range 0-0.2 recruits/shell/day, and rarely exceeded 0.5 recruits/ shell/day, an order of magnitude less than rates measured in other Florida and southeastern U.S. estuaries (Manley et al. 2008, Parker et al. 2013, Drexler et al. 2014). Possible explanations for these low observed recruitment rates include: environmental conditions are impeding larval survival and/or recruitment: measured larval densities are inadequate to support recruitment at rates typical of other southeastern oyster reef systems; suitable substrate is limiting; or, some combination of the three.
Regarding environmental conditions, they appear to be suitable to support oysters in most, but not all, areas of the PBS (USEPA 2004). With respect to larval supply, the Project Greenshores initiative along the northeast shoreline of the Pensacola Bay basin (near Stations 6 and 7) has shown promising outcomes from substrate addition (http://www.dep. state.fl.us/northwest/Ecosys/section/greenshores.htm), indicating larval supply sufficient to rebuild and maintain an oyster population. Data reported here suggest Project Greenshores is ideally located in an area where salinity is suitable (Davis 1958). ranging from 5 to 30 (commonly <25); both modeled and empirical data demonstrate relatively high and consistent larval supply; and rates of recruitment (especially self-recruitment) are high. These outcomes, particularly the reported increase in recruitment following substrate addition, suggest that substrate may be limiting in the Pensacola Bay basin. Locations in both Escambia Bay (e.g., Station 11) and East Bay (e.g., Station 12) experience similar conditions and may only require substrate addition to promote reef establishment. Many of the sample locations, however, in the PBS lacked one or more of the aforementioned conditions, suggesting a complex interplay between environment, larval supply, and substrate availability. For example, at the deeper monitoring sites, salinity was commonly above 30 especially near bottom, whereas at bay-head Stations 2 and 4 salinity was typically below 10. These stations, and the areas they represent, are therefore unlikely to support oyster populations in the long term. At the remaining stations, especially those in East Bay, salinity appeared suitable but larvae were less frequently encountered, and both measured and modeled recruitment rates (especially rates of self-recruitment) were much lower compared with Stations 6, 7, 11, and 12. Substrate addition, as has been done to rebuild oyster populations in other regions (e.g., O'Beirn et al. 2000. Powers et al. 2009, Schulte et al. 2009. Dunn et al. 2014), likely will be necessary at those locations. Because larvae also appear to be limiting in some areas of the PBS, substrate addition will not suffice and successful restoration will require both the addition of substrate and the addition of adults or larvae (Arnold 2008) to ensure sufficient delivery of larvae and subsequent recruitment. This is evidenced, for example, along the western shoreline of Escambia Bay (Station 5 from this study), where substrate addition has been conducted with no discernable benefit. Restoration efforts in these areas likely will not provide a suitable return on the restoration investment unless a larger larval pool develops within the PBS.
A marine metapopulation is composed of discrete populations occupying habitat patches that have interpatch dispersal rates high enough to ensure demographic connectivity while being low enough to maintain independence among the local populations (Sale et al. 2006). Various studies indicate Crassostrea virginica should be considered within a metapopulation context (Haase et al. 2012, Munroe et al. 2012, Kjelland et al. 2015). Metapopulations function at scales ranging from local estuaries (Bert et al. 2014) to range-wide (Bert et al. 2011). In the current study, C. virginica dispersal and connectivity were considered within a single estuary, at the lower end of those multiple scales. Larval retention within a reef, exchange of larvae between neighboring reefs, and patterns of oyster larval dispersal throughout the bay exhibited substantial spatial variation. In general, all reefs served as an important source (>1 % of modeled larvae) to at least one other site, but recipient sites were fewer in number. The two stations in the Pensacola Bay subbasin for which larval transfer was modeled exhibited high levels of retention and intersite exchange compared with other sites in the PBS, levels that were similar to that predicted for oysters in Chesapeake Bay (Kjelland et al. 2015) although much lower than predicted levels of exchange in Delaware Bay (Narvaez et al. 2012). But Stations 6 and 7 contributed few simulated larvae to the East Bay basin or even to the adjacent Escambia Bay. Escambia Bay sites could provide larvae to sites in Pensacola Bay and to contribute larvae to other sites within Escambia Bay, but only Station 11 retained more than 1% of its larvae. Escambia Bay contributed more than 1 % of modeled larvae to Station 14 in East Bay, but otherwise larval transfer from Escambia to East Bay was low. East Bay contributed substantially to Pensacola Bay but little to Escambia Bay. Modeled larval supply to East Bay was almost exclusively dependent upon larvae generated within East Bay. Thus, the source/sink relationships among individual oyster reefs in the PBS system are complex, with reefs operating as a source to some reefs but not to others, and contemporaneously serving as a sink for larvae emanating from some other reefs. Substantial variability in the distribution and abundance of oyster larvae, and in recruitment patterns, suggests a temporal component to source/sink relationships. These observations are consistent with patterns of temporal variability in source/sink relationships observed for bay scallops in GOM waters (Bert et al. 2014). Understanding these dynamic source/sink relationships is an essential precursor to rebuilding oyster reefs and ensuring their long-term persistence and health (Munroe et al. 2012), both in the PBS and elsewhere (Kjelland et al. 2015), including at larger spatial scales.
Evaluating model predictions within the context of measured larval abundance and recruitment events allows for validation of model performance. For that purpose, measured versus modeled data were compared for two of the modeled dispersal events, the first beginning October 8, 2007, and the second beginning July 13, 2008. These runs were chosen because they correspond to field samples with relatively high numbers of larvae or recruits. The 2007 event occurred under conditions of high salinity (always >20) with little vertical salinity stratification. The 2008 event provided contrasting conditions characterized by lower salinity (rarely >20 except near the mouth of the bay) and relatively strong vertical salinity stratification. The last day of the October 2007 field sampling corresponded to Day 15 of the October simulation, and the last day of the July 2008 field sampling corresponded to Day 8 of the corresponding simulation. Note that the measure of larval abundance is a point sample, so the relationship between model prediction and sample estimate is direct. In contrast, settlement estimates integrate over each deployment, so comparisons between modeled and measured outcomes are indirect.
Concordance between model predictions and concurrent field data was not complete but was informative. For the October 2007 event, oyster larvae (<1/1) were recorded at Stations 5 and 10 in Escambia Bay and Station 13 in East Bay, corresponding to areas where larvae were contemporaneously predicted to occur. At other stations, simulated larval abundances were generally very low or zero, again consistent with the measured larval abundances. Exceptions include Stations 6 and 7 along the western shoreline of Pensacola Bay, Station 4 in NE Escambia Bay, and Stations 12, 3. and 1 in East Bay, where larvae were predicted but not detected. Nonetheless, modeled and measured larvae are in reasonable agreement for the October 2007 event. Correspondence between simulated and measured recruitment is more difficult to assess. Recruits were recorded at many stations during the October sampling event. At some of those stations, including, e.g., Stations 1, 13, and 14 along the NE shore of East Bay, moderate levels of recruitment were recorded and 20-day (ready to set) simulated larvae were in the vicinity of those stations. But at other stations, e.g., Stations 6 and 7 in Pensacola Bay, few if any recruits were recorded although the simulations showed very high abundance of 20-day larvae.
While an abundance of recruits and few larvae were recorded during the October 2007 event, the opposite occurred during the July 2008 event. Recruits were recorded only at Stations 6-9, and then at very low levels. This despite a prediction of competent recruits spread throughout the bay. Only along the southern shoreline of PBS was any consistency between predicted and realized recruitment detected. Again, correspondence in larval abundance from measured versus modeled estimates was mixed. The southern shoreline was predicted to be densely populated with larvae on July 16. Larvae were detected at Stations 8, 9, and 18, but not at neighboring south shore Stations 17 and 15. Larvae were detected at Stations 6 and 7, in agreement with model simulations, but at Stations 1, 13, and 14 along the NE shore of East Bay no larvae were measured although the model outcome showed dense larval concentrations. At most stations in the northern reaches of East and Escambia Bays, the model predicted and sampling revealed few if any larvae, but at Station 3 in East Bay a relatively low density of larvae was detected although the model simulation showed no larvae.
Validating model performance is an essential step in the process of effectively describing dispersal and resultant connectivity patterns, and quantifying larvae in the field relative to model predictions is the most direct means of validating model predictions. Previous work (Arnold et al. 2005) shows that larval dispersal models remain only a partial representation of reality. Such models are useful, but outcomes can be misleading and may result in flawed management responses. Refining inexpensive and efficient larval identification and quantification methodologies such as those applied in this study therefore remains an important endeavor.
Methodological limitations likely contributed to the inconsistencies between modeled and empirical results. For example, the recruitment monitoring methodology actually measures settlement and subsequent survival until sampling, and does so by integrating over a relatively lengthy time period, thereby providing only a partial window into the relationship between essentially instantaneous predicted delivery of competent larvae and measured, time-averaged settlement at a specific location. Larval sampling and subsequent detection with genetic methods approximates one moment in the protracted larval life span. Additionally, while these larval detection methods are theoretically sound (Frischer et al. 2000). further enhancements are needed to better align laboratory-based outcomes with actual abundances in the field. One approach would include more frequent sampling at higher spatial resolution in both vertical and horizontal dimensions. Additionally, interference with the Crassostrea virginica probe from unknown sources in general plankton can be reduced or eliminated by using a combination of visual counts of actual plankton and genetic verification of any detected larvae, in combination with mixed plankton samples treated with additional internal genetic markers.
Other potential contributors to inconsistency between modeled and measured parameters are not methodological. As noted above, a key difference between the October 2007 and July 2008 events was the physical character of the water column. A high salinity, weakly stratified water column in October contrasted with a relatively low salinity and stratified water column in July. As summarized in North et al. (2008), Crassostrea virginica larvae orient toward the bottom in the absence of a halocline but tend to swim upward in the presence of a halocline. It would therefore be expected in the October 2007 scenario for larvae to be closer to the bottom and therefore more successfully recruiting. This matches the observations. Larval proximity to the bottom, however, renders the sampling scheme, for which the intake hose was positioned at least 0.5 m above the bottom, unlikely to sample those larvae despite model predictions that larvae would be there. In contrast, during the July 2008 sampling event, larvae would be expected to be nearer the surface and therefore susceptible to harvest by the intake hose. These same conditions would limit larval access to suitable settlement substrate on the bottom, thereby contributing to reduced recruitment success as reflected in these samples. Moreover, water column stratification was also identified during the June 2008 sampling event, indicating that recruitment success would be affected throughout the June-July collector deployment period. The model predicted an abundance of competent oyster larvae in various locations throughout the bay, but those competent larvae would not necessarily translate into recruits to the collectors at least at stations where the collectors were located below the halocline. This is validated by the data. At Stations 6-9 where some recruitment did occur, the water column was well mixed with respect to salinity. At Stations 1, 2, 13, and 14, where dense concentrations of competent larvae were predicted to occur but no recruitment was detected, salinity stratification was present. At Station 12, there was no stratification and no recruitment was recorded, but the model predicted essentially zero delivery of competent larvae.
Numerous other factors, including but certainly not limited to nutritional status, competitive interactions, genetic variability within the larval cohort, and predation, also influence larval dispersal patterns and resultant settlement outcomes. As in previous studies (e.g., North et al. 2008, Narvaez et al. 2012), those factors were not considered but they too would influence measured outcomes relative to model predictions. Determining the extent to which each of those parameters influences outcomes provides numerous opportunities for future research.
When freshwater inputs are high, the freshwater at the bayhead creates an unstable density gradient with the high saline GOM water that drives subtidal circulation (Pritchard 1956, Hansen & Rattray 1965). As a result, the bottom water inflow increases when river discharge increases, but when freshwater discharge decreases vertical mixing is more likely to decrease bottom water inflow and resultant salinity (MacCready & Geyer 2010). This classic form of estuarine circulation can drive buoyant objects such as oyster larvae toward the mouth of the PBS and out into the GOM during stratified conditions. Increased freshwater inputs via river discharge and rainfall during winter and spring 2008 resulted in reduced surface and near-bottom salinity at all stations relative to 2007, likely pushing larvae toward the mouth of the bay. Similar outcomes were predicted for Delaware Bay (Narvaez et al. 2012) and lower Chesapeake Bay (Haase et al. 2012). In contrast, relatively higher salinities during summer and fall 2007 resulted in predictions of successful larval delivery to, and settlement at, living oyster reefs within the Escambia and East Bay basins.
During the October 16-18, 2007, sampling, the water column was well mixed and salinity values were high. Under these conditions little push from the salinity front would be expected. Yet the model predicted 10% of virtual larvae export from the PBS. This inconsistency may be explained by the timing of sampling relative to the timing of predicted export. The PBS experienced a substantial storm front on October 18, concurrent with field sampling. This major freshwater input induced stratification and a resultant push of surface water toward the mouth of the bay, coinciding with the larval export event predicted to occur on October 21 (when >4% of modeled larvae were exported from the PBS) and October 22 (when >3% of modeled larvae were exported). Clearly, short-term events such as a major rainfall can have significant implications for oyster larval dispersal and resultant connectivity among near and distant populations.
Wind-induced currents influenced dispersal of early-stage larvae residing near the surface more than older larvae that have a greater sinking rate (Dekshenieks et al. 1996). In the simulations reported here, winds often drove larvae to accumulate near the shoreline and then moved them back into the bay in a concentrated grouping when the wind shifted, impacting the statistical likelihood of transference between reefs. The rate at which a larva swims to the surface remains an important factor in these results, but information about the relationship between vertical swimming speed and larval size, as well as larval growth rates in the GOM, is needed to refine larval dispersal estimates.
In Florida, once abundant oyster reefs occupied estuaries along the entire coastline, including the PBS, but are now substantially depleted or absent (Arnold & Berrigan 2002, Seavey et al. 2011). The collapse of oyster reefs and the subsequent failure to recover has typically been attributed to changes in estuarine water quality, loss of necessary substrate, or changes in flow and bathymetry due to dredge-and-fill activities (Brumbaugh et al. 2006, Beck et al. 2011). Doubtless those and other factors are important contributors, but understanding the interrelation of individual reefs and bays within multiscale metapopulation structure is essential to understanding the broad-scale health of the species (Lipcius et al. 2008). Export of larvae is a key step in this process. Oyster larval export has been predicted in various modeling studies (North et al. 2008, Narvaez et al. 2012, Kim et al. 2013, Park et al. 2014), and export of various larval species from an estuary into neighboring estuaries has been documented (Ellien et al. 2000, Natunewicz et al. 2001, George et al. 2013, Safak et al. 2015). At least in the case of oysters, however, the biological significance of that export has not been considered. Evidence of genetic connectivity among GOM oyster reefs (Varney et al. 2009) indicates oyster larvae must travel between estuaries (Toba et al. 2007) but the extent of genetic exchange may be masked by selective factors operating at the local scale (Burford et al. 2014). The simulations reported here suggest that 1%-10% of larvae were exported from the PBS during each modeled period. The fate of those larvae is unknown, but nearshore currents greater than or equal to 50 cm/sec in the northern GOM (Ohlmann & Niiler 2005) are capable of transporting oyster larvae from Pensacola Bay to Apalachicola Bay in 6-7 days. The implications to metapopulation structure (and genetic diversity) may be critical to the long-term stability (and adaptability) of oysters within individual GOM estuaries and throughout the species' range. Such implications are generally not considered within the context of Crassostrea virginica population depletion at both local and broad scales. Understanding the magnitude of within- and between-estuary migrations, and the ramifications of such exchange, is essential to the effective management of this economically, ecologically, and culturally important estuarine organism.
We thank Samantha Bund. Michael Drexler, Mark Gambordella, Dan Marelli, and Anthony Vasilas for assistance with field sampling and laboratory analyses. Comments from an anonymous reviewer improved the manuscript and are appreciated. Funding for the study was provided through NOAA Grant NA06 NA06NMF4540319 (subaward CR-M-022-2006-04), with additional funds for development of the larval probe provided through USFWS grant no. 401817G050. D. A. Narvaez was funded through NSF-EID Program Grant OCE-0622642, COPAS Sur-Austral (PFB-31) and MINECONNC120086. M. E. Frischer was supported by NSF award no. OCE 1459293.
Arnold, W. S. 2008. Application of larval release for restocking and stock enhancement of coast marine bivalve populations. Rev. Fish. Sei. 16:65-71.
Arnold, W. S. & M. E. Berrigan. 2002. A summary of the oyster (Crassostrea virginica) fishery in Florida. A Report to the Division of Marine Fisheries, Florida Fish and Wildlife Conservation Commission.
Arnold, W. S., G. L. Hitchcock, M. E. Frischer. R. Wanninkhof& Y. P. Sheng. 2005. Dispersal of an introduced larval cohort in a coastal lagoon. Limnot. Oceanogr. 50:587-597.
Baker, P. & R. Mann. 2003. Late stage bivalve larvae in a well-mixed estuary are not inert particles. Estuaries 26:837-845.
Bass, D. G. & D. T. Cox. 1988. River habitat and fishery resources of Florida. In: Seaman, W., editor. Florida aquatic habitat and fishery resources. Kissimmee, FL: American Fisheries Society, pp. 121-187.
Beck, M. W., R. D. Brumbaugh, L. Airoldi. A. Carranza & L. D. Crawford. 2011. Oyster reefs at risk and recommendations for conservation, restoration, and management. Bioscience 61:107-116.
Berrigan, M. E. 1988. Management of oyster resources in Apalachicola Bay following Hurricane Elena. J. Shellfish Res. 7:281-288.
Berrigan, M. E. 1990. Biological and economic assessment of an oyster resource development project in Apalachicola Bay, Florida. J. Shellfish Res. 9:149-158.
Bert, T. M., W. S. Arnold, A. L. McMillen-Jackson, A. E. Wilbur & C. Crawford. 2011. Natural and anthropogenic forces shape the population genetics and recent evolutionary history of eastern United Slates bay scallops (Argopecten irradians). J. Shellfish Res. 30:583-608.
Bert, T. M., W. S. Arnold, A. E. Wilbur, S. Seyoum. A. L. McMillen-Jackson, S. P. Stephenson, R. H. Weisberg & L. A. Yarbro. 2014. Florida Gulf bay scallop (Argopecten irradians spp. concentricas) population genetic structure: form, variation, and influential factors. J. Shellfish Res. 33:99-136.
Blumberg, A. & G. L. Mellor. 1987. A description of a three-dimensional coastal ocean circulation model. In: Heaps. N. S., editor. Three-dimensional coastal ocean models. Washington, DC: American Geophysical Union, pp. 1-16.
Brumbaugh, R. D., M. W. Beck. L. D. Coen. L. Craig & P. Hicks. 2006. A practitioners' guide to the design and monitoring of shellfish restoration projects: an ecosystem services approach. MRD Educational Report No. 22. Arlington. VA: The Nature Conservancy.
Burford, M. O., J. Scarpa. B. J. Cook & M. P. Hare. 2014. Local adaptation of a marine invertebrate with a high dispersal potential: evidence from a reciprocal transplant experiment of the eastern oyster Crassostrea virginica. Mar. Ecol. Prog. Ser. 505:161-175.
Buroker, N. E. 1983. Genetic differentiation and population structure of the American oyster Crassostrea virginica (Gmelin) in Chesapeake Bay. J. Shellfish Res. 3:153-167.
Burwell, D. 2001. Modeling the spatial structure of estuarine residence time: Eulerian and Lagrangian approaches. PhD dissertation. College of Marine Science, University of South Florida, St. Petersburg, FL.
Carriker, M. R. 1951. Ecological observations on the distribution of oyster larvae in New Jersey estuaries. Ecol. Monogr. 21:19-38.
Coen, L. D., M. W. Luckenbach & D. L. Breitburg. 1999. The role of oyster reefs as essential fish habitat: a review of current knowledge and some new perspectives. Amer. Fish. Soc. Symp. 22:438-454.
Collard, S. B. 1991. Management options for the Pensacola Bay system: the potential value of seagrass transplanting and oyster bed refurbishment programs. Water Resources Special Report 91-4. Havana, FL: Northwest Florida Water Management District.
Danglade, E. 1917. Condition and extent of the natural oyster beds and barren bottoms in the vicinity of Apalachicola, Florida. Document No. 841. U.S. Department of Commerce, Bureau of Fisheries.
Davis, H. C. 1958. Survival and growth of clam and oyster larvae at different salinities. Biol. Bull. 114:296-307.
Dekshenieks, M. M., E. E. Hofmann, J. M. Klinck & E. N. Powell. 1996. Modeling the vertical distribution of oyster larvae in response to environmental conditions. Mar. Ecol. Prog. Ser. 136:97-110.
Drexler, M. D., M. L. Parker, S. P. Geiger, W. S. Arnold & P. Hallock. 2014. Biological assessment of eastern oysters (Crassostrea virginica) inhabiting reef, mangrove, seawall, and restoration substrates. Estuaries Coasts 37:962-972.
Dugas, R. J., E. A. Joyce & M. E. Berrigan. 1997. History and status of the oyster, Crassotrea virginica, and other molluscan fisheries of the U.S. Gulf of Mexico. In: MacKenzie, C. L., Jr., V. G. Burrell Jr., A. Rosenfield & W. L. Hobart, editors. The history, present condition, and future of the molluscan fisheries of North and Central America and Europe. Volume 1 : Atlantic and Gulf Coasts. NOAA Technical Report 127. U.S. Department of Commerce, pp. 187-210.
Dunn, R. P., D. B. Eggleston & N. Lindquist. 2014. Effects of substrate type on demographic rates of eastern oyster (Crassostrea virginica). J. Shellfish Res. 33:177-185.
Ellien, C., E. Thiebaut, A. Barnay, J. Dauvin, F. Gentil & J. Salomon. 2000. The influence of variability in larval dispersal on the dynamics of a marine metapopulation in the eastern channel. Oceano!. Acta 23:423-442.
Frischer, M. E., J. M. Danforth, L. C. Tyner, J. R. Leverone, D. C. Marelli, W. S. Arnold & N. J. Blake. 2000. Development of an Argopecten-specific 18s rRNA targeted genetic probe. Mar. Biotechnol. (NY) 2:11-20.
Gaffney, P. M. 2006. The role of genetics in shellfish restoration. Aquat. Living Resour. 19:277-282.
George, G., D. V. Desai, C. A. Gaonkar, V. M. Aboobacker, P. Vethamony & A. C. Anil. 2013. Barnacle larval transport in the Mandovi-Zuari estuarine system, central west coast of India. ./. Oceanogr. 69:451-466.
Govindarajan, A. F., J. Pineda. M. Purcell & J. A. Breier. 2015. Species and stage-specific barnacle larval distributions obtained from AUV sampling and genetic analysis in Buzzards Bay, Massachusetts, USA. J. Exp. Mar. Biol. Ecol. 472:158-165.
Vanderkooy, S., editor. 2012. The oyster fishery of the Gulf of Mexico. United States: a regional management Plan--2012 revision. Ocean Springs, MS: Gulf States Marine Fisheries Commission.
Haase, A. T., D. B. Eggleston. R. A. Luettich. R. J. Weaver & B. J. Puckett. 2012. Estuarine circulation and predicted oyster larval dispersal among a network of reserves. Estuar. Coast. Shelf Sei. 101:33-43.
Haidvogel, D. B. & A. Beckmann. 1999. Numerical ocean circulation modeling. London, England: Imperial College Press.
Hansen, D. V. & M. Rattray. 1965. Gravitational circulation in straits and estuaries. J. Mar. Res. 23:104-122.
Hanski, I. & O. Ovaskainen. 2003. Metapopulation theory for fragmented landscapes. Theor. Popul. Biol. 64:119-127.
Hare, M. P. & J. C. Avise. 1996. Molecular genetic analysis of a stepped multilocus cline in the American oyster (Crassostrea virginica). Evolution 50:2305-2315.
Hargis, W. J., Jr. & D. S. Haven. 1999. Chesapeake oyster reefs, their importance, destruction, and guidelines for restoring them. In: Luckenbach, M. W., R. Mann & J. A. Wesson, editors. Oyster reef habitat restoration: a synopsis, and a synthesis of approaches. Gloucester Point, VA: VIMS Press, pp. 329-358.
Hoese, H. D., W. R. Nelson & H. Beckert. 1972. Seasonal and spatial setting of fouling organisms in Mobile Bay and eastern Mississippi Sound, Alabama. Ala. Mar. Res. Bull. 8:9-17.
Ingersoll, E. 1881. The history and present condition of the fishery industries: the oyster industry. Department of the Interior. Washington. DC: Government Printing Office.
Kennedy, V. S., R. I. E. Newell & A. F. Ebel. 1996. The eastern oyster Crassostrea virginica. College Park. MD: Maryland Sea Grant College, University of Maryland.
Kim, C.-K., K. Park & S. P. Powers. 2013. Establishing restoration strategy of eastern oyster via a coupled biophysical transport model. Restor. Ecol. 21:353-362.
Kim, C.-K., K. Park, S. P. Powers. W. M. Graham & K. M. Bayha. 2010. Oyster larval transport in coastal Alabama: dominance of physical transport over biological behavior in a shallow estuary. J. Geophys. Res. 115:C10019.
Kjelland, M., C. D. Piercy. T. Lackey & T. M. Swannack. 2015. An integrated modeling approach for elucidating the effects of different management strategies on Chesapeake Bay oyster meta-population dynamics. Ecol. Model. 308:45-62.
Klindworth, A., E. Pruesse. T. Schweer, J. Peplies, C. Quast. M. Horn & F. O. Glockner. 2013. Evaluation of general 16S ribosomal RNA gene PCR primers for classical and next-generation sequencing based diversity studies. Nucleic Acids Res. 41 :e 1.
Kurlansky, M. 2006. The big oyster: history on the half shell. New York, NY: Random House Trade Paperbacks.
LeCorre, N., F. Guichard & L. E. Johnson. 2012. Connectivity as a management tool for coastal ecosystems in changing oceans. In: Marcelli, M., editor. Oceanography. ISBN: 978-953-51-0301-1. Available at: http://www.intechopen.com/books/oceanography/ connectivity-as-a-tool-to-manage-coastal-ecosystems-in-changingoceans.
Leming, T. D. & D. R. Johnson. 1985. Application of circulation models to larval dispersment and recruitment. Mar. Technol. Soc. J. 19:34-41.
Linie, E. J. & J. A. Quick, Jr. 1976. Ecology, resource rehabilitation, and fungal parasitology of commercial oysters, Crassostrea virginica (Gmelin) in Pensacola Estuary, Florida. St. Petersburg, FL: Florida Department of Natural Resources, Marine Research Laboratory.
Lipcius, R. M., D. B. Eggleston, S. J. Schreiber, R. D. Seitz, J. Shen, M. Sisson. W. T. Stockhausen & H. V. Wang. 2008. Importance of metapopulation connectivity to restocking and restoration of marine species. Rev. Fish. Sei. 16:101-110.
Livingston, R. J. 2015. Climate change and coastal ecosystems: long-term effects of climate and nutrient loading on trophic organization. Boca Raton, FL: CRC Press.
Livingston, R. J., R. L. Howell. IV. X. Niu, F. G. Lewis, III & G. C. Woodsum. 1999. Recovery of oyster reefs (Crassostrea virginica) in a Gulf estuary following disturbance by two hurricanes. Bull. Mar. Sei. 64:465-483.
Livingston, R. J., F. G. Lewis, G. C. Woodsum, X.-F. Niu, B. Galperin, W. Huange. J. D. Christensen, M. E. Monaco, T. A. Battista, C. J. Klein. R. L. I. Howell & G. L. Ray. 2000. Modelling oyster population response to variation in freshwater input. Estuar. Coast. Shelf Sei. 50:655-672.
Livingston, R. J., X. Niu. F. G. Lewis & G. C. Woodsum. 1997. Freshwater input to a Gulf estuary: long-term control of trophic organization. Ecol. Appl. 7:277-299.
Long, W. C., R. D. Seitz, B. J. Brylawski & R. N. Lipcius. 2014. Individual, population, and ecosystem effects of hypoxia on a dominant benthic bivalve in Chesapeake Bay. Ecol Monogr. 84:303-327.
Ma, F., C. Jiang, W. Rauen & B. Lin. 2009. Modeling sediment transport processes in a macro-tidal estuary. Sei. China Ser. E 52:3368-3375.
MacCready, P. & W. R. Geyer. 2010. Advances in estuarine physics. Annu. Rev. Mar. Sei. 2:35-58.
MacKenzie, C. L., Jr., V. G. Burrell Jr., A. Rosenfield & W. L. Hobart, editors. 1997. The history, present condition, and future of the molluscan fisheries of North and Central America and Europe. NOAA Technical Report NMFS 127. Seattle, WA: U.S. Department of Commerce.
Manley, J., A. Power & R. Walker. 2008. Patterns of eastern oyster. Crassostrea virginica (Gmelin. 1791), recruitment in Sapelo Sound, Georgia: implications for commercial oyster culture. Occasional papers of the University of Georgia Marine Extension Service, vol. 3. Savannah, GA: University of Georgia Marine Extension Service.
Mann, R. 1988. Distribution of bivalve larvae at a frontal system in the James River, Virginia. Mar. Eeol. Prog. Ser. 50:29-14.
Mellor, G. L., L. Y. Oey & T. Ezer. 1998. Sigma coordinate pressure gradient errors and the seamount problem. J. Almos. Ocean. Technol. 15:1122-1131.
Meyers, S. D. & M. E. Luther. 2008. A numerical simulation of residual circulation in Tampa Bay. Part II: Lagrangian residence time. Estuaries Coasts 31:815-827.
Munroe, D. M., J. M. Klinck, E. E. Hoffman & E. N. Powell. 2012. The role of larval dispersal in metapopulation gene flow: local population dynamics matter. J. Mar. Res. 70:441-467.
Narvaez, D. A., J. M. Klinck, E. M. Powell, E. E. Hofmann, J. Wilkin & D. B. Haidvogel. 2012. Modeling the dispersal of eastern oyster (Crassostrea virginica) larvae in Delaware Bay. J. Mar. Res. 70:381409.
Natunewicz, C. C., C. E. Epifanio & R. W. Garvine. 2001. Transport of crab larval patches in the coastal ocean. Mar. Ecol. Prog. Ser. 222:143-154.
North, E. W., Z. Schlag, R. R. Hood. L. Zhong, T. Gross & V. S. Kennedy. 2008. Vertical swimming behavior influences the dispersal of simulated oyster larvae in a coupled particle-tracking and hydrodynamic model of Chesapeake Bay. Mar. Ecol. Prog. Ser. 359:99-115.
O'Beirn, X., M. W. Luckenbach, J. A. Nestlerode & G. M. Coates. 2000. Toward design in constructed oyster reefs: oyster recruitment as a function of substrate type and tidal height. J. Shellfish Res. 19:387 395.
Oey, L.-Y. 2005. A wetting and drying scheme for POM. Ocean Model. 9:133-150.
Ohlmann, J. C. & P. P. Niiler. 2005. Circulation over the continental shelf in the northern Gulf of Mexico. Prog. Oceanogr. 64:45-81.
Park, K., S. P. Powers, G. S. Bosarge & H. S. Jung. 2014. Plugging the leak: barrier island restoration following Hurricane Katrina enhances larval retention and improves salinity regime for oysters in Mobile Bay, Alabama. Mar. Environ. Res. 94:48-55.
Parker, M. L., W. S. Arnold, S. P. Geiger, P. Gorman & E. H. Leone. 2013. Impacts of freshwater management activities on eastern oyster (Crassostrea virginica) density and recruitment: recovery and long-term stability in seven Florida estuaries. J. Shellfish Res. 32:695-708.
Pine, W. E., III. C. J. Walters, E. V. Camp, R. Bouchillon. R. Ahrens. L. Sturmer & M. Berrigan. 2015. The curious case of eastern oyster Crassostrea virginica stock status in Apalachicola Bay, Florida. Ecol. Soc. 20:46.
Powell, E. N., K. A. Ashton-Alcox, J. N. Kraeuter, S. E. Ford & D. Bushek. 2008. Long-term trends in oyster population dynamics in Delaware Bay: regime shifts and response to disease. J. Shellfish Res. 27:729-755.
Powers, S. P., C. H. Peterson. J. H. Grabowski & H. S. Lenihan. 2009. Success of constructed oyster reefs in no-harvest sanctuaries: implications for restoration. Mar. Ecol Prog. Ser. 389:159-170.
Pritchard. D.W.I 956. The dynamic structure of a coastal plain estuary. J. Mar. Res. 15:33-42.
Pulliam, H. R. 1988. Sources, sinks, and population regulation. Am. Nat. 132:652-661.
Quinteiro, J., L. Perez-Dieguez. A. Sanchez, R. I. Perez-Marin, C. G. Sotelo & M. Rey-Mendez. 2011. Quantification of manila clam Ruditapes philippinarum (Adams and Reeve, 1850) larvae based on SYBR Green real-time polymerase chain reaction. J. Shellfish Res. 30:791-796.
Ray-Culp, M. 2007. A living shoreline initiative for the Florida Panhandle: taking a softer approach. Natl. Wetlands Newsl. 29:9-12.
Safak, I., P. L. Wiberg. D. L. Richardson & M. O. Kurum. 2015. Controls on residence lime and exchange in a system of shallow coastal bays. Cont. Shelf Res. 97:7-20.
Sale, P. F., I. Hanski & J. P. Kritzer. 2006. The merging of meta-population theory and marine ecology: establishing the historical context. Chapter 1. In: Kritzer, J. P. & P. F. Sale, editors. Marine metapopulations. Burlington, MA: Elsevier Academic Press.
Saoud, I. G., D. B. Rouse, R. K. Wallace, J. Howe & B. Page. 2000. Oyster Crassostrea virginica spat settlement as it relates to the restoration of Fish River Reef in Mobile Bay, Alabama. J. World Aquacutt. Soc. 31:640-650.
Schulte, D. M., R. P. Burke & R. N. Lipscius. 2009. Unprecedented restoration of a native oyster metapopulation. Science 325:1124-1128.
Seavey, J. R., W. E. Pine. P. Frederick. L. Sturmer & M. Berrigan. 2011. Decadal changes in oyster reefs in the Big Bend of Florida's Gulf Coast. Ecosphere 2:114.
Smagorinsky, J. 1963. General circulation experiments with the primitive equations. 1. The basic experiment. Mon. Weather Rev. 91:99-164.
Swift, F. 1897. Report of a survey of the oyster regions of St. Vincent Sound. Apalachicola Bay and St. George Sound. Florida. Washington. DC: U.S. Commission of Fish and Fisheries, pp. 187-221.
Toba, M., H. Yamakawa, Y. Kobayashi, Y. Sugiura. K. Honma & H. Yamada. 2007. Observations on the maintenance mechanisms of metapopulations, with special reference to the early reproductive process of the Manila clam Ruditapes philppinarum (Adams & Reeve) in Tokyo Bay. J. Shellfish Res. 26:121-130.
USEPA, 2004. The ecological condition of the Pensacola Bay system, northwest Florida (19942001). EPA 620-R-05-002. Gulf Breeze. FL: U.S. Environmental Protection Agency. Office of Research and Development, National Health and Ecological Effects Research Laboratory. Gulf Ecology Division.
Vadopalas, B., J. V. Bouma, C. R. Jackels & C. S. Friedman. 2006. Application of real-time PCR for simultaneous identification and quantification of larval abalone. J. Exp. Mar. Biol. Ecol. 334:219228.
Varney, R. L., C. E. Galindo-Sanchez, R. Cruz & P. M. Gaffney. 2009. Population genetics of the eastern oyster Crassostrea virginica (Gmelin. 1791) in the Gulf of Mexico. J. Shellfish Res. 28:855-864.
Wight, N. A., J. Suzuki, B. Vadopalas & C. S. Friedman. 2009. Development and optimization of quantitative PCR assays to aid Ostrea lurida Carpenter 1984 restoration efforts. J. Shellfish Res. 28:33-41.
Wilberg, M. J., M. E. Livings, J. S. Barkman. B. T. Morris & J. M. Robinson. 2011. Overfishing, disease, habitat loss, and potential extirpation of oysters in upper Chesapeake Bay. Mar. Ecol. Prog. Ser. 436:131-144.
Xue, H. & Y. Du. 2010. Implementation of a wetting-and-drying model in simulating the Kennebec-Androscoggin Plume and the circulation of Casco Bay. Ocean Dvn. 60:341-357.
Ye, J., G. Coulouris, I. Zaretskaya. I. Cutcutache. S. Rozen & T. Madden. 2012. Primer-BLAST: a tool to design target-specific primers for polymerase chain reaction. BMC Bioinformatics 13:134.
WILLIAM S. ARNOLD, (1, 5) * STEVEN D. MEYERS, (2) STEPHEN P. GEIGER, (1) MARK E. LUTHER, (3) DIEGO NARVAEZ, (3, 6) MARC E. FRISCHER (4) AND EILEEN HOFMANN (4)
(1) Florida Fish and Wildlife Research Institute, St. Petersburg, FL 33701; (2) College of Marine Science, University of South Florida, St. Petersburg, FL 33701; (3) Old Dominion University, Norfolk, VA 23529; (4) Skidaway Institute of Oceanography, University of Georgia, Savannah, G A 31411; (5) NOAA National Marine Fisheries Service, Southeast Regional Office, 263 13th Avenue South, Saint Petersburg, FL 33701 ; (6) Department of Oceanography and COPAS Sur-Austral, University of Concepcion, Biobio, Chile
* Corresponding author. E-mail: email@example.com
Caption: Figure 1. Pensacola Bay "ridded model bathymetry and sampling stations, with water depth in meter below MSL. Tide and weather gauge is "N," water quality station is "S," and Garcon Point is "G." The lines indicate the geographic division of PBS into the Pensacola Bay, Escambia Bay, and East Bay (including Blackwater Bay) basins. Left insert shows location of the PBS on the northern shore of the GOM. Right insert shows location of known oyster beds (GSMFC 2012).
Caption: Figure 2. Boundary conditions used for model calibration, including wind speed, wind direction, total freshwater discharge (precipitation and river discharge), and salinity at the surface and bottom. Shaded areas indicate times of simulated larval transport. The dashed vertical line is January 1, 2008.
Caption: Figure 3. Measured salinity, from each station shown in Figure 1, at each of the near-surface (crosses), intermediate-depth (closed circles), and near-bottom (boxes) sample locations. The depth of deepest measurement at each site is indicated.
Caption: Figure 4. Measured temperature, from each station shown in Figure 1, at each of the near-surface (crosses), intermediate-depth (closed circles), and near-bottom (boxes) sample locations. The depth of deepest measurement at each site is indicated.
Caption: Figure 5. Qualitative oyster Crassostrea virginica larval abundance in samples collected from near-surface (top panel), midwater (second panel), and near-bottom (third panel) sampling locations. Dotted lines separate stations located in depths between shore and 1 m from those located in depths between I and 2.5 m and those located at depths above 2.5 m. Small circles represent low larval densities (>0 but <~1 larvae/1; [C.sub.t] [less than or equal to] 45) and large circle represent high larval densities (>~1 larvae/I; 45 < [C.sub.t] [less than or equal to] 50). Mean oyster recruits/shell/day (bottom panel). Note that recruitment was not monitored at Stations 16-18 located in deep water or at Stations 1-9 during the May-July 2007 deployment period. Inset presents circle size relative to reeruits/shell//day. Complete trap loss events are indicated by x.
Caption: Figure 6. Composite temporal fate of modeled oyster Crassostrea virginica larvae released on October 3, 2007, and sampled on simulated Days 5, 10, 15, and 20 following release. Simulation date, percent of larvae exported from the estuary, and percent of larvae settled are shown in the upper left of each panel. The light gray arrows represent wind 48 h prior, the gray arrows represent wind 24 h prior, and the black arrows represent the wind on the modeled sample day. The circle around the wind vectors indicates 5 m/sec.
Caption: Figure 7. Composite temporal fate of modeled oyster Crassostrea virginica larvae released on July 8, 2008, and sampled on simulated Days 5, 8, 15, and 20 following release. Time since release, percent of larvae exported from the estuary, and percent of larvae settled are shown in the upper left of each panel. The light gray arrows represent wind 48 h prior, the gray arrows represent wind 24 h prior, and the hlack arrows represent the wind on the modeled sample day. The circle around the wind vectors indicates 5 m/sec.
Caption: Figure 8. Modeled oyster Crassostrea virginica larval locations 20 days following each of four release dates in the PBS. Simulated release date, percentage of larvae exported from the estuary, and percentage of original cohort that has settled are indicated. Open circles indicate larval release stations.
TABLE 1. Percentage of oyster Crassostrea virgittica larval transfer between sites averaged from all four simulation periods, with target sites being [+ or -] 2 grid cells. Donor site Station Pensacola Bay Escambia Bay 6 7 5 10 11 Recipient site Pensacola Bay 6 2.23# 1.04 0.84 0.97 1.25 7 2.63 3.45# 0.86 1.1 1.11 Escambia Bay 5 0.63 0.52 0.83# 0.53 1.07 10 0.2 0.14 0.66 0.38# 0.13 11 1.55 0.82 1.62 0.15 3.3# East Bay 1 0.02 0.04 0.04 0.29 0.05 3 0.04 0.12 0.04 0.12 0.01 12 0.09 0.24 0.05 0.06 0.02 14 0.01 0.03 1.59 0.56 1.04 15 0.02 0.02 0.04 0.08 0.04 Total donated 7.42 6.24 6.57 4.24 8.02 Donor site Station East Bay 1 3 12 14 15 Recipient site Pensacola Bay 6 2.42 1.7 1.69 2.03 0.59 7 0.88 0.55 0.3 0.89 1.41 Escambia Bay 5 0 0 0 0 0.16 10 0.05 0.03 0.02 0.03 0.08 11 0.61 0.43 0.17 0.59 0.2 East Bay 1 1.33# 0.44 0.19 1.65 0.41 3 0.13 0.03# 0.09 0.19 0.09 12 0.44 0.8 2.19# 0.51 0.33 14 0.64 1.02 0.19 0.23# 4.23 15 0.08 0.1 0.13 0.14 0.16# Total donated 6.58 5.1 4.97 6.26 7.66 Station Total received Recipient site Pensacola Bay 6 14.76 7 13.18 Escambia Bay 5 3.74 10 1.72 11 9.44 East Bay 1 4.46 3 0.86 12 4.73 14 9.54 15 0.81 Total donated Stations 6 and 7 are in Pensacola Bay: Stations 5. 10. and 11 are in Escambia Bay; and Stations 1.3, 12, 14. and 15 are in East Bay. Zeros indicate values less than 0.01%. Donor sites contribute recruits, recipient sites receive recruits. Bolded numbers in diagonal indicate instances of self-seeding. Note: Bolded numbers in diagonal indicate instances of self-seeding are indicated with #. TABLE 2. Percentage of oyster Crassostrea virginica larvae exported from Pensacola Bay by larval life date, for each of four larval release dates. Release date Day of August 9, October 3, June 9, July 8. larval life 2007 2007 2008 2008 1 0 0 0 0 2 0 0 0 0 3 0.42 0 0 0 4 0.76 0 0 0 5 0.07 0 0 0 6 0.02 0.01 0 0.07 7 0 0.02 0.39 0.56 8 0 0.12 0.39 0.2 9 0 1.23 0 0.21 10 0 0.07 0 0.68 11 0 0.5 0 0.07 12 0 0.11 0 0.09 13 0.01 0.02 0 0 14 0 0 0 0 15 0 0 0.01 0 16 0 0.01 0 0 17 0 0 0 0 18 0 4.28 0 0 19 0 3.24 0 0.01 20 0 0.25 0 0 Total 1.28 9.86 0.79 1.89 Column dates represent the start date of each simulation.
|Printer friendly Cite/link Email Feedback|
|Author:||Arnold, William S.; Meyers, Steven D.; Geiger, Stephen P.; Luther, Mark E.; Narvaez, Diego; Frischer|
|Publication:||Journal of Shellfish Research|
|Date:||Apr 1, 2017|
|Previous Article:||Oyster density and demographic rates on natural intertidal reefs and hardened shoreline structures.|
|Next Article:||Performance of diploid and triploid Crassostrea gigas (Thunberg, 1793) grown in tropical versus temperate natural environmental conditions.|