Anchoa Mitchilli) Eggs and Larvae in Chesapeake Bay ) E.W

Estuarine, Coastal and Shelf Science 60 (2004) 409e429 www.elsevier.com/locate/ECSS

Distribution and transport of bay anchovy (Anchoa mitchilli) eggs and larvae in Chesapeake Bay ) E.W. North , E.D. Houde1

University of Maryland Center for Environmental Science, Chesapeake Biological Laboratory, USA

Received 28 February 2003; accepted 8 January 2004

Abstract

Mechanisms and processes that influence small-scale depth distribution and dispersal of bay anchovy (Anchoa mitchilli) early-life stages are linked to physical and biological conditions and to larval developmental stage. A combination of fixed-station sampling, an axial abundance survey, and environmental monitoring data was used to determine how wind, currents, time of day, physics, developmental stage, and prey and predator abundances interacted to affect the distribution and potential transport of eggs and larvae. Wind-forced circulation patterns altered the depth-specific physical conditions at a fixed station and significantly influenced organism distributions and potential transport. The pycnocline was an important physical feature that structured the depth distribution of the planktonic community: most bay anchovy early-life stages (77%), ctenophores (72%), copepod nauplii (O76%), and Acartia tonsa copepodites (69%) occurred above it. In contrast, 90% of sciaenid eggs, tentatively weakfish (Cynoscion regalis), were found below the pycnocline in waters where dissolved oxygen concentrations were !2.0 mg l1. The dayenight cycle also influenced organism abundances and distributions. Observed diel periodicity in concentrations of bay anchovy and sciaenid eggs, and of bay anchovy larvae O6 mm, probably were consequences of nighttime spawning (eggs) and net evasion during the day (larvae). Diel periodicity in bay anchovy swimbladder inflation also was observed, indicating that larvae apparently migrate to surface waters at dusk to fill their swimbladders. Overall results suggest that wind-forced circulation patterns, below-pycnocline dissolved oxygen conditions, and diel changes in vertical distribution of larvae and their copepod prey have important implications for potential transport of bay anchovy early-life stages. 2004 Elsevier Ltd. All rights reserved.

Keywords: biologicalephysical interactions; larval transport; predatoreprey interactions; bay anchovy; zooplankton; Chesapeake Bay

1. Introduction stages, including those of bay anchovy (Anchoa mitchilli) in Chesapeake Bay. The mechanisms by which organisms respond or Bay anchovy is the most abundant fish in Chesapeake react to the biological and physical environment are Bay and in many coastal areas of the western North critical in estuaries, the site of important spawning Atlantic. It is a pelagic, small (!110 mm), short-lived grounds and nursery areas for many fish species. Many (!3 years) fish that plays an important ecological role in factors, including physics, larval development, food Chesapeake Bay as a major prey for piscivores such as abundance, and predation act and interact to affect the striped bass (Morone saxatilis), bluefish (Pomatomus small-scale distributions and dispersal of fish early-life saltatrix), and weakfish (Cynoscion regalis)(Baird and Ulanowicz, 1989; Hartman and Brandt, 1995). It is ) Corresponding author. University of Maryland Center for a pelagic, serial spawner with a reproductive season in Environmental Science, Horn Point Laboratory, P.O. Box 775, Chesapeake Bay that extends from May to September Cambridge, MD 21613, USA. peaking in July (Luo and Musick, 1991; Zastrow et al., 1 Present address: University of Maryland Center for Environmen- 1991). Spawning occurs at salinities from 0 to 32 psu tal Science, Chesapeake Biological Laboratory, P.O. Box 38, Solo- mons, MD 20688, USA. (Dovel, 1971; Olney, 1983; Houde and Zastrow, 1991) E-mail addresses: [email protected] (E.W. North), ehoude@ and peaks at temperatures from 26 to 28 (C(Houde and cbl.umces.edu (E.D. Houde). Zastrow, 1991). Bay anchovy spawns between 1800 and

0100 hours (Luo and Musick, 1991; Zastrow et al., position in the water column in relation to currents 1991), producing daily cohorts of eggs that hatch into rather than by swimming horizontally (Norcross and yolk-sac larvae at 18e27 h after fertilization, depending Shaw, 1984; Miller, 1988). Several transport mecha- upon temperature (Houde and Zastrow, 1991). Larvae nisms have been documented for larval fish including begin feeding about 2 days after hatching (Houde and tidally-timed vertical migration, also referred to as Zastrow, 1991). Larval anchovy first feed on microzoo- selective tidal stream transport (Rowe and Epifanio, plankton such as copepod nauplii, rotifers, and tintin- 1994), migration around the depth of null velocity nids, and shift to larger copepodites and adult copepods (Fortier and Leggett, 1983), and migration in relation to as they grow (Houde and Zastrow, 1991; Auth, 2003). both time of day and tides (Weinstein et al., 1980). In Their growth rate is temperature-dependent (Houde, estuaries with two-layer circulation, such as Chesapeake 1974) and is about 0.5e0.8 mm d1 in Chesapeake Bay Bay, diel vertical migration could result in up-estuary (Rilling and Houde, 1999b; Auth, 2003). transport if larvae move into landward-flowing waters Physical conditions may influence the spatial and during the day in summer when days are longer than vertical distribution of bay anchovy early-life stages. nights. It is also possible that random movements of The occurrence and survival of early-life stages below larvae, coupled with frequent spawning by adults, could the pycnocline may depend on dissolved oxygen (DO) lead to retention of a sizable fraction of larvae in the concentrations because larvae avoid waters with low DO estuary. (Breitburg, 1994; Keister et al., 2000) which can limit the The objective of this research was to identify the viability of bay anchovy early-life stages (Chesney and mechanisms and processes influencing small-scale verti- Houde, 1989; Houde and Zastrow, 1991). Temperature cal distributions and potential transport of bay anchovy also potentially could influence the distribution of eggs and larvae. The study was designed to determine bay anchovy early-life stages, especially in highly how currents, time of day, physics (temperature, sali- stratified conditions. The reported range of occurrence nity, DO), ontogeny (egg and larval stages), food abun- for temperature is 13.0e30.0 (C for eggs and 15.0e dance, predation, and weather act or interact to affect 30.0 (C for larvae in Chesapeake Bay (Houde and egg and larval distributions. The temporal scale of sam- Zastrow, 1991). In addition, circulation patterns such as pling was designed to detect diel and tidally-timed residual eddies (Hood et al., 1999) and plume fronts vertical migrations of fish larvae. The research consisted (Peebles, 2002) can form retention areas that affect of: (1) an initial survey in Chesapeake Bay to determine early-stage distributions. areas of maximum egg and larvae abundance; (2) depth- Prey and predator concentrations may influence the stratified sampling at a fixed location to describe vertical spatial occurrence and vertical distribution of bay anch- distributions of early-life stages in relation to physical ovy eggs and larvae. Adult bay anchovy may spawn and biological factors; (3) length measurements of larvae where food abundance is high, leading to an association to classify larvae by ontogenetic stage; and (4) an anal- between high concentrations of early-life stages and the ysis of environmental data to evaluate factors that in- copepods that serve as prey for adult anchovy (Peebles fluenced physical conditions and organism distributions. et al., 1996; Peebles, 2002). The location of larval prey may also influence larval distributions, and larvae may follow the vertical migrations of copepod prey. Pre- 2. Methods dation may affect the distribution of bay anchovy eggs and larvae by: (1) causing direct mortality; (2) stimulat- Data were collected in Chesapeake Bay from 18e27 ing predator-avoidance movements by larvae; and/or June 1996 on the 120 ft RV Cape Henlopen. The first two (3) influencing adult spawning-site selection because days of the cruise consisted of an ichthyoplankton and adults may avoid spawning in areas of high gelatinous CTD survey along the axis of Chesapeake Bay (Fig. 1). zooplankton abundance (Dorsey et al., 1996). Major After the initial survey, sampling effort was concen- predators include the gelatinous zooplankton scypho- trated at a fixed station located in mid-Bay (37( 45# N) medusan (Chrysaora quinquecirrha) and the lobate from 20 to 23 June and from 26 to 27 June (Fig. 1). ctenophore (Mnemiopsis leidyi)(Purcell et al., 1994). Ontogenetic changes in swimming ability, buoyancy regulation, and larval behavior also may influence larval 2.1. Axial survey distributions (Boehlert and Mundy, 1988). Potential for buoyancy regulation increases when the swimbladder The initial survey along the axis of Chesapeake Bay first inflates, although precise control may not be pos- was conducted to determine the spatial distributions sible until later in the development process (O’Connell, and abundances of fish early-life stages and gelatinous 1981). zooplankton and to locate suitable areas for inten- Bay anchovy larvae could affect their transport sive depth-stratified collections. Twelve stations at 15 within or out of estuaries by regulating their vertical nautical-mile (w26 km) intervals were occupied from E.W. North, E.D. Houde / Estuarine, Coastal and Shelf Science 60 (2004) 409e429 411

zooplankton was preserved in 1% formalin-plus-ethanol solution and transferred to 100% ethanol within six days of collection. Baltimore 39.25 2.2. Fixed station sampling

Samples were collected at a fixed station to determine vertical distribution and abundance of fish early-life stages, microzooplankton, and gelatinous zooplankton 38.75 in relation to each other and water-column hydrography. The site of the fixed station was determined from information on egg and larval abundances in the axial CBOS survey (visual inspection of samples) and with consid- eration of water-column depth, shipping traffic, and 38.25 gelatinous zooplankton concentrations. The first occu- Potomac River pation of the station (20e23 June) was for 81 h and the second (26e27 June) for 28.5 h. Latitude At the fixed station, a suite of biological and physical measurements was obtained on each tide (every w6 h). 37.75 Sampling Station First, current velocity was measured with an Acoustic Doppler Current Profiler (ADCP). Then, a CTD was deployed to measure water-column temperature, salinity, DO, and fluorescence, and to collect microzooplankton. Following the CTD cast, depth-stratified 37.25 sampling for ichthyoplankton and gelatinous zooplankton was conducted at 5-m intervals from the surface to 25-m (near bottom). After two sets of depth-stratified samples were complete, another CTD cast was made CBBT and additional ADCP measurements were obtained. 36.75 The entire suite of measurements (details below) was completed in w2.7 h. 77.2 76.8 76.4 76.0 2.2.1. Current velocity measurements Longitude Instantaneous measures of current velocity were Fig. 1. Chesapeake Bay, USA. Location of fixed-station sampling (¤) determined from 6-min averages of raw data from the e e on 20 23 and 26 27 June 1996, current velocity measurements (C)at RV Cape Henlopen’s hull-mounted 1200 kHz ADCP. Chesapeake Bay Observing System (CBOS) buoy, and water level records (:) at Baltimore, MD, and Chesapeake Bay Bridge Tunnel Raw data were collected in 1-m depth intervals (from (CBBT), VA. Open circles (B) represent axial sampling stations on 4 to 25 m) for 6e10 min with the vessel at constant 18e19 June 1996. speed and heading (parallel to the channel) every w2.7 h during the first fixed-station occupation, and every w1.7 h during the second occupation. Current velocities the Bay mouth (37( 00# N) to the head of the Bay were separated into longitudinal (along-channel) and (39( 26# N). lateral (cross-channel) components. The longitudinal At each station, a CTD cast (water-column temper- direction was set by determining the principal axis, the ature, salinity, DO, and fluorescence) preceded ichthyo- direction in which velocity variance was at a maximum. plankton collections. Ichthyoplankton and gelatinous zooplankton were collected in an openingeclosing 1 m2 2.2.2. Microzooplankton Tucker trawl with 280-mm mesh nets, a flowmeter, and Depth-stratified microzooplankton samples were a depthetemperature recorder. A pair of 2-min collec- collected during the CTD casts in 10-l Niskin bottles tions was made in a single deployment at each station, attached to the CTD rosette. Samples were collected at one from near bottom (within 1.5 m) to the pycnocline, 2.5, 7.5, 12.5, 17.5, and 22.5 m depths. They were filtered and the second from the pycnocline to the surface. On onto a 35-mm mesh and preserved in 5% formalinesea- average, each collection filtered w110 m3 of water. water. In the laboratory, zooplankton were identified Gelatinous zooplankton was separated from the sample and enumerated under a dissecting microscope. All with large strainers (w5 mm pores), identified, and organisms were enumerated in samples from 12.5, 17.5, biovolumes of each species were determined. Remaining and 22.5 m depths (79% of the samples). When the 412 E.W. North, E.D. Houde / Estuarine, Coastal and Shelf Science 60 (2004) 409e429 number of organisms was high (21% of the samples, adjusted for a 1-m overlap between the sampled depth usually from 2.5 and 7.5 m depths), samples were diluted intervals (Ai Z 4ciCðciCcjÞ=2 where Ai Z abundance to 25 ml and three 1-ml aliquots were enumerated. The in ith depth interval, ci Z concentration in ith depth mean number of organisms from the three aliquots was interval, and cj Z concentration in the jth depth in- multiplied by 25 ml to estimate the total number of terval). Abundances in each depth interval were summed organisms in the sample. to determine total water-column abundance (no. m2 or ml m2) at the station. 2.2.3. Ichthyoplankton and gelatinous zooplankton Mean depths of organism occurrence at the fixed Depth-stratified collections of ichthyoplankton and station were calculated for each set of depth-stratified gelatinous zooplankton at the fixed station were made in samples by averaging net tow depths (the midpoint of the 1-m2 Tucker trawl. This gear was deployed as in the each depth interval) weighted by the concentration of initial axial survey except that the trawl was fished for fish eggs and larvae (North, 2001). To investigate 2 min in the 5e0m,10e5m,15e10 m, 20e15 m, and potential relationships in the distribution of organisms, 25e20 m depth intervals. To minimize potential bias due correlation coefficients (SAS 6.12, PROC CORR) were to time and tidal stage effects during a station occupa- calculated for organism abundances as well as for mean tion, the order of depth intervals sampled was randomly depths of occurrence. Data were loge-transformed and assigned in each set of stratified samples, except that Pearson correlation coefficients were calculated for the 10e5 m and 5e0 m depth intervals were always sam- normally distributed data. Spearman rank-order corre- pled during one deployment, and in that order. Catches lation coefficients were calculated for data that were not were processed as at the axial survey stations, except normally distributed (abundances of bay anchovy yolk- that (1) gelatinous zooplankton processing was limited sac and O13 mm larvae). Correlation analyses of mean to total biovolume measurements (species were identi- depths of occurrences also included pycnocline depth fied in three sets of depth-stratified collections) and (2) and the depth of the 3.0 mg l1 DO contour. Pycnocline samples were preserved in a 2% formalin-plus-ethanol depth was estimated as the depth at which the solution when gelatinous zooplankton biovolumes were BrunteVa¨ isa¨ lla¨ frequency was maximum during each especially large. CTD cast (Mann and Lazier, 1996). The depth of the In the laboratory, fish eggs and larvae were identified 3.0 mg l1 DO contour was calculated by linear in- and counted under a dissecting microscope. Although terpolation of CTD DO measurements adjacent to entire samples were sorted for larvae, eggs were counted 3.0 mg l1. The CTD DO measurements were calibrated in a subsample (1/2 to 1/8 of the whole sample, divided with pre- and post-cruise Winkler titrations conducted with a plankton splitter) when numbers were O200. by E.M. Smith during spring (28 Aprile5 May) and Larvae from the fixed-station collections were separated summer (17e26 July) research cruises onboard the RV into two groups based on swimbladder inflation Cape Henlopen in 1996. (inflated or not inflated). All larvae were measured with Potential advection of water (and planktonic organ- a computer-based digitizing system. isms within it) during fixed-station sampling was estimated by calculating displacement (km) using in- 2.2.4. Analysis terpolated ADCP current velocities (Golden Software The time series of depth distributions of physical Surfer v.6.01, applying kriging with an isotropic linear factors, ichthyoplankton (no. m3) and zooplankton variogram model). Initial interpolated ADCP current (no. l1) concentrations, and gelatinous zooplankton velocities (grid-line geometry was 1 m in the depth biovolumes (ml m3) at the fixed station was contoured direction and 1.67 h in the time direction) were re- (Golden Software Surfer v.6.01). In these plots, time was interpolated with half the distance between grid points depicted in hours since midnight of the first day of fixed- in the timeedirection. This process was repeated until station occupation. The gridding method was kriging the distance between grid points in the timeedirection with an isotropic linear variogram model. Grid-line was w720 s. geometry was no finer than half the average distance Potential displacement during each time step (i) was between measurements in the Y-direction (depth), and calculated by multiplying interpolated velocity (v)at no finer than the average distance between measure- each 1-m depth by the time-step duration (t ¼ 720 s). ments in the X-direction (time). Cumulative potential displacement (Dc) at time c was Abundances (per m2) in a depth interval of ichthyo- calculated by summing displacement at t ¼ c with dis- plankton, microzooplankton, and biovolumes of gelat- placement during all previous time steps (Dc ¼ Svit for inous zooplankton in each tow were estimated by t ¼ 0tot ¼ c). Estimates of lower layer water mass multiplying organism concentrations by the sampled movement based on our displacement calculations depth interval. For fixed-station samples, the abundance (18.5 cm s1) were similar to those reported at the mouth estimates from collections during a single net deploy- of the Choptank River, a Bay tributary (20 cm s1) ment (10e5 and 5e0 m depth intervals only) were (Sanford and Boicourt, 1990). E.W. North, E.D. Houde / Estuarine, Coastal and Shelf Science 60 (2004) 409e429 413

2.2.5. Simulated transport analysis velocities (at 2.4 and 18.9 m depths) from the mid-Bay Displacement calculations were used to estimate CBOS buoy in June 1996 were used in conjunction with potential length-specific transport of bay anchovy larvae NOS water level measurements at Baltimore and at the by size class to determine if diurnal vertical migrations Chesapeake Bay Bridge Tunnel (Fig. 1) to identify that increase in range as larvae develop could affect forces that influenced physical conditions during fixed- dispersal up- or down-estuary. All size classes of larvae station sampling. Water level measurements and CBOS were hypothesized to remain at 5 m during the night to current velocities were filtered with a Lanczos low pass isolate the effects of daytime migrations. During the day, filter with a 34-h half-power point to reveal low- larvae 4e5 mm in length were hypothesized to be at 5 m, frequency water level and circulation patterns (courtesy while the mean depth of successive 1-mm size classes was W.C. Boicourt and C. Derry). June salinity (psu) and 1 m deeper than the next smallest size class. Potential DO (mg l1) measurements from channel stations of the transport (km) during the first fixed-station occupation CBP Monitoring Program were used to provide was estimated by calculating cumulative displacement historical context to the field sampling results in June for the 5-m depth interval during night (sunset to 1996. Mean salinity and DO concentrations were sunrise) and adding it to cumulative displacement at calculated at the depths of CTD measurements from depths specific for each length class during the day 1985e1995 June surveys and compared to June 1996 (sunrise to sunset). monitoring results.

2.2.6. Statistical analysis A statistical analysis was conducted to determine if 3. Results advection (displacement), photoperiod (dayenight), or time periods of differing physical conditions described 3.1. Historical context a significant amount of variability in the abundances of organisms. Time periods of differing physical conditions Freshwater flow into the Chesapeake Bay during the were chosen based upon water-column hydrography. spring of 1996 was far above the long-term mean A repeated measures, multiple-regression analysis was (Boynton et al., 1997). The high input of freshwater conducted on log -transformed abundances (no. m2 e strongly affected June salinity structure in mid-Ches- or ml m2) with displacement, dayenight, and time apeake Bay. Upper layer (above-pycnocline) salinities period as fixed effects, and station occupation (first or were lower in June 1996 (Fig. 2a) than the mean of the second) as a random effect (SAS 6.12, PROC MIXED). 10 prior years, as were lower layer (below-pycnocline) Covariance in time was modeled with a first-order auto- DO concentrations. The intersection of the 10 psu iso- regressive covariance structure. The dayenight variable haline with the surface was O80 km further down- was coded day Z 0 and night Z 1 where night repre- estuary in June 1996 compared to its mean location in sented time from sunset to sunrise. Explanatory variables 1985e1995 and, at the fixed station, the June 1996 sur- passed tolerance and condition index multicollinearity face salinity was w5 psu lower than the June long-term tests (SAS 6.12, PROC REG and PRINCOMP). All average. Dissolved oxygen concentrations !1mgl1 regression models passed ShapiroeWilks normality tests extended 50 km further down-estuary in June 1996 (SAS 6.12, PROC UNIVARIATE) as well as tests for compared to their mean locations in June 1985e1995. homogeneity of variance (plots and correlation analyses At the fixed station, lower layer DO concentrations in of KresidualsK versus predicted values). June 1996 were w1mgl1 less than the June long-term Coefficients of determination (R2) from repeated- average. measures, regression models with fixed and random effects are unreliable so are not reported. Instead, F values are included in tables of results to indicate the relative 3.2. Predominant organisms amount of variance accounted for by fixed effects. The random effect ‘station occupation’ was not significant in The most abundant eggs and larvae collected were any of the models, so was excluded from tables of bay anchovy. Mean abundances (pooled axial and fixed- results. station surveys) of anchovy eggs and larvae were 183:0m2 G 33:2 s.e. and 4:04 m2 G 0:69 s.e., respec- 2.3. Environmental monitoring data tively. Sciaenid eggs were the second most abundant fish egg (mean abundance Z 53:4m2 G 8:4 s.e.). Although Physical conditions during the cruise were evaluated the sciaenid eggs potentially were spawned by 4e7 with environmental monitoring data of the Chesapeake different species (Olney, 1983; Daniel and Graves, 1994), Bay Observing System (CBOS), NOAA National Ocean most probably were weakfish (Cynoscion regalis) be- Service (NOS), and the Chesapeake Bay Program cause most positively-identified sciaenid larvae were (CBP) Monitoring Program. Wind (RMP) and current weakfish (50 out of 72 post-yolk-sac sciaenid larvae). 414 E.W. North, E.D. Houde / Estuarine, Coastal and Shelf Science 60 (2004) 409e429

a) June 11-12, 1996 salinity (psu, lines) and dissolved oxygen (shaded contours) 0 (mg l -1 ) 10 10 8 6 20 4 2 Depth (m) 30 0 0 50 100 150 200 250 300

) b) Bay anchovy eggs -2 1200 upper-layer 800 lower-layer 400

0 Abundance (no. m 0 29 58 89 115 144 174 201 230 257 285 309

) c) Bay anchovy larvae -2 12

0 Abundance (no. m 0 29 58 89 115 144 174 201 230 257 285 309

) d) Sciaenid eggs -2 60

0 Abundance (no. m 0 29 58 89 115 144 174 201 230 257 285 309

) e) Gelatinous zooplankton

-2 4000 3000 2000 1000 0 Biovolume (ml m 0 29 58 89 115 144 174 201 230 257 285 309 Distance from Head of Bay (km)

Fig. 2. (a) Salinity (psu) contour lines and dissolved oxygen (mg l1) shaded contours along the axis of Chesapeake Bay for June 11e12, 1996. Data from Chesapeake Bay Monitoring Program. Black dots indicate depths of CTD measurements. Dissolved oxygen data up-estuary of 50 km were not available. Triangles mark the location of axial survey stations and an arrow marks the location of the ﬁxed sampling station. Bar graphs depict axial survey abundances (no. m2) of (b) bay anchovy eggs, (c) bay anchovy larvae, (d) sciaenid eggs, and (e) gelatinous zooplankton biovolume (ml m2) above (upper layer) and below (lower layer) the pycnocline. Stations from 174e257 km were sampled at night.

Biovolumes of gelatinous zooplankton were mostly com- mainder. Mean biovolume of gelatinous zooplankton was posed of the ctenophore, Mnemiopsis leidyi, (99% in axial 1077:6mlm2 G 104:2 s.e. During fixed-station sampling, survey collections, 98% in fixed-station collections) with most copepods (96%) were Acartia tonsa copepodites the hydromedusa Nemopsis bachei constituting the re- (mean abundance Z 1:08 ! 105 m2 G 0:14 ! 105 s.e.) E.W. North, E.D. Houde / Estuarine, Coastal and Shelf Science 60 (2004) 409e429 415 and adults (mean abundance Z 2:23 !104 m2 G 0:31 ! predominantly in below-pycnocline waters up-estuary 104 s.e.). of 257 km, and apparently associated with salinities O18 psu. 3.3. Axial survey 3.4. Fixed-station sampling In axial survey collections, bay anchovy eggs (Fig. 2b) were two orders of magnitude more abundant than Fixed-station sampling was conducted near 201 km anchovy larvae (Fig. 2c), and one order of magnitude where depths were O25 m, anchovy larval abundances more abundant than sciaenid eggs (Fig. 2d). Peak peaked, and gelatinous zooplankton biovolumes were abundances of bay anchovy eggs and larvae, sciaenid moderate (Fig. 2c,e). Maximum abundance of bay eggs, and ctenophores (Fig. 2bee) occurred down- anchovy larvae (17.9 no. m2) during the fixed-station estuary of the Potomac River (174 km), although cteno- occupation was comparable to peak abundance during phores occurred in significant biovolumes as far the axial survey (11.5 no. m2). up-estuary as 58 km. In the upper Bay (0e89 km), anchovy eggs (max: 34.6 m2) were relatively uncommon 3.4.1. Physics and anchovy larvae were absent (Fig. 2b,c). Within lower Contour plots of physical factors at the fixed station Chesapeake Bay (O174 km), bay anchovy eggs were show a well-defined pycnocline, below which temper- most abundant where ctenophore biovolume was mini- atures (Fig. 3a) were lower than those above the pyc- mal (230 and 285 km), and were virtually absent where nocline. Dissolved oxygen concentrations !3.0 mg l1 ctenophore biovolume peaked (257 km). generally occurred just below the pycnocline (Fig. 3b)and Anchovy eggs, larvae, and ctenophores were most decreased with depth, frequently measuring !2mgl1 abundant above the pycnocline at all stations near bottom. Highest fluorescence values (RFU, raw (Fig. 2bee). In contrast, sciaenid eggs were abundant fluorescence units) occurred within and above the pyc- in both layers at the most down-Bay stations but were nocline and peaked in surface waters during afternoons

a) Salinity (psu) contour lines and temperature (ºC) shaded contours ºC

27 10 25 23 20 21 Depth (m) 19 17 10 20 30 40 50 60 70 80 160 170

–1 b) Dissolved oxygen (mg l ) contour lines and fluorescence (RFU) shaded contours RFU

7 6 10 5 4 20 3 Depth (m) 2 1 0 10 20 30 40 50 60 70 80 160 170

–1 Velocity c) Along-channel current velocity (cm s ) (cm s–1)

45 36 10 27 18 9 0 20 -9 Depth (m) -18 -27 -36 -45 10 20 30 40 50 60 70 80 160 170 Time (hrs)

Fig. 3. Time series of fixed station (a,b) CTD and (c) ADCP measurements, 20e23 and 26e27 June 1996. Two variables are plotted in panel (a), salinity (psu) contour lines and temperature ((C) shaded contours, and in panel (b), dissolved oxygen (mg l1) contour lines and fluorescence (RFU) shaded contours. Panel (c) depicts along-channel current velocity (cm s1) for first (2e83 h) and second (150e179 h) station occupations. Measurement locations indicated by black dots. White and gray bars at the top of (b) indicate day and night. Negative current velocity is up-estuary. 416 E.W. North, E.D. Houde / Estuarine, Coastal and Shelf Science 60 (2004) 409e429

(Fig. 3b). Daytime peaks in DO concentrations corre- Just prior to the second station occupation (Fig. 4, sponded to those of fluorescence, a result of photosyn- shaded area, June 26e27), another strong wind event thesis in the upper layer. from the north (Fig. 4a) reduced water levels in the upper During the initial 30 h of the fixed-station occupa- Bay (Baltimore record, Fig. 4b), increased water level in tion, the pycnocline was shallow (w6 m), temperatures the lower Bay (CBBT record, Fig. 4b), and resulted in in the upper layer were the highest of the time series, and enhanced two-layer circulation (Fig. 4c). At the start of those below the pycnocline were the lowest (Fig. 3a). the second occupation of the fixed station, the pycnocline Tidal currents were reversing in the upper and lower had deepened (w15 m) and intensified, probably a result layers (Fig. 3c). After 30 h, conditions changed. The of mixing during the wind event (Fig. 3a). The 3.0 mg l1 pycnocline deepened (w9 m), water temperatures and DO contour also was driven deeper during the wind event salinity increased in the lower layer and decreased in the as was fluorescence (Fig. 3b). The strong up-estuary upper layer. These changes corresponded to strong up- current velocity throughout the water column (Fig. 3c, estuary currents below the pycnocline where tidal cur- 160e170 h) likely was due to a flooding tide combined rents were no longer reversing. This change in physical with a barotropic response to the removal of wind stress, conditions likely was related to a wind event that oc- producing an up-estuary pulse of water (a seiche effect). curred from 30e45 h during the fixed-station occupa- The subsequent gradual elevation in pycnocline depth tion (Fig. 4a, shaded area, June 20e23). Strong winds during the second fixed-station occupation apparently blowing from the north and north-west reduced water resulted from the decrease in wind mixing as well as the levels in the upper Bay (Fig. 4b, Baltimore record), and barotropic response. enhanced two-layer estuarine circulation. This response Potential displacement of water during the two fixed- can be seen in the CBOS filtered current record where station occupations demonstrates the dynamic change in residual surface velocities increased in the down-estuary physical conditions and potential consequences for direction while lower layer velocities increased in the up- organism transport (Fig. 5). For the first 20 h of the estuary direction (Fig. 4c). time series, displacement in both the surface (!9 m) and

Fig. 4. (a) Wind velocity (10 m s1) at mid-Bay CBOS buoy, (b) filtered water level (cm) at Baltimore, MD, and Chesapeake Bay Bridge Tunnel (CBBT), VA, and (c) CBOS filtered along-channel current speed (cm s1) in the upper and lower layers in June, 1996. Vectors in (a) point in the direction to which the wind was blowing. Negative current speed in (c) is up-estuary. Shaded areas indicate timing of fixed-station sampling. E.W. North, E.D. Houde / Estuarine, Coastal and Shelf Science 60 (2004) 409e429 417

40 down-estuary 30 20 4 m 10 4 m 10 m 0 10 -10 25 m Displacement (km) -20 17 m 16 m -30 up-estuary 23 m -40 0 102030405060708090 150 165 180 195 Time (hr) Fig. 5. Potential displacement (km) over time for the first and second fixed-station occupations. Lines represent potential displacement of water at each 1-m depth interval. Displacement at each time point was calculated by multiplying the along-channel current velocity estimate at that time by the duration between velocity estimates (velocity (km s1) ! time (s) = displacement (km)) and then summing with displacement estimates from all previous time points. Specific depths are marked by open circles and labeled. Negative displacement is up-estuary. lower layers (O9 m) was within that expected by tidal station during the first period could have been separated excursion. Between 25 and 45 h, enhanced gravitational by as much as 60 km by the end of the third period circulation related to the northern wind event resulted in (Fig. 6c). potential displacement of the surface layer w10 km down-estuary and the lower layer w10 km up-estuary. 3.4.2. Biology The upper layer was nearly stationary between 45 and All stages of bay anchovy (Figs. 7a, 8), ctenophores 75 h, suggesting that collections in the upper layer (Fig. 7c), copepod nauplii (Fig. 9a), and Acartia tonsa during this time period were sampling the same water copepodites (Fig. 9b) were located in highest concen- mass and its associated organisms. In contrast, lower trations above the pycnocline during both the first and layer displacement was non-stationary after 30 h, in- second fixed-station occupations (Table 1). In contrast, dicating that up-estuary residual currents continually only 10% of sciaenid eggs were found above the transported water and organisms within it past the fixed pycnocline (Table 1, Fig. 7c). Although only about half station. of A. tonsa adults (49%) were found above the pycno- The fixed-station time series was divided into four cline during the first occupation, most (85%) were above periods, each with different physical characteristics, to the pycnocline during the second occupation (Table 1, examine the influence of advecting water masses on Fig. 9c). biological collections. During the first period, 2e20 h, Periods of different physical conditions explained the pycnocline depth was shallow and displacement was a significant amount of the variability in abundances of near zero in both the upper and lower layers. The second anchovy eggs, yolk-sac larvae, 6e9 mm larvae, cteno- period, 20e40 h, was characterized by changing physical phore, and copepod nauplii in multiple-regression conditions and rapid displacement in both the upper models (Table 2). This suggests that the different upper and lower layer. From 40e83 h, the third period, upper layer water masses sampled during periods one, three layer displacement was near zero, pycnocline depth was and four contained different biological communities. In stable, and lower layer salinity continually increased, as multiple-regression models, displacement included a mea- did lower layer displacement up-estuary. Finally, the sure of spatial association within water masses as well as second fixed-station occupation can be considered a an indication of the direction of water mass movement. fourth period characterized by rapidly changing physical Significant parameter estimates for bay anchovy yolk- conditions: decreasing pycnocline depth, a strong baro- sac, 3e6 mm, 6e9 mm, and 9e13 mm larvae indicated tropic up-estuary response throughout the water column, that abundances decreased as upper layer waters moved and a transition in the upper layer from wind-mixed down-estuary past the fixed station, and that sciaenid egg towards restratification. abundance increased as lower layer waters moved up- Fig. 6 summarizes water mass movement during the estuary past the fixed station. first three periods. Waters sampled in the lower and upper layers during the first period (Fig. 6a) were separated 3.4.3. Anchovy and sciaenid eggs by as much as 20 km by the end of the second period Mean abundances of anchovy eggs were highest (Fig. 6b). Organisms that were located at the fixed during the first period (2e20 h) while abundances of 418 E.W. North, E.D. Houde / Estuarine, Coastal and Shelf Science 60 (2004) 409e429

corresponded to peaks in egg concentrations near the a) Period 1 4 pycnocline (Fig. 7b), suggesting that most of these eggs

9 Depth (m) were newly spawned. The significant dayenight effect for sciaenid eggs in the multiple-regression model (Table 14 2) almost certainly represented effects of evening spawning activity. 19

24 3.4.4. Prey and predators Copepod nauplii and Acartia tonsa copepodites, the -40 -30 -20 -10 0 10 20 30 most common prey items in bay anchovy larval guts in Displacement (km) Chesapeake Bay (Auth, 2003), peaked in concentration and abundance in the upper layer during the ﬁrst two b) Period 1 4 periods (Fig. 9a,b, Tables 1 and 3). Nauplii were Period 2 concentrated in the upper 10 m of the water column,

9 Depth (m) while copepodites were found throughout the upper layer, especially at night. It is likely that most nauplii 14 were A. tonsa because most of the copepods (96%) collected during fixed-station sampling were this species. 19 Peak concentrations of adult A. tonsa were generally shallower during night and deeper during day, often 24 occurring within or below the pycnocline during day (Fig. 9c). Like nauplii and copepodites, abundance of -40 -30 -20 -10 0 10 20 30 adult A. tonsa declined between the second and fourth Displacement (km) periods (Table 3). The significant dayenight effect for A. tonsa copepodite and adult abundances in the multiple- 4 c) Period 1 regression model (Table 2) could indicate that these Period 2 stages aggregated at depths not sampled by the Niskin

9 Depth (m) Period 3 bottles or that they were able to evade Niskin-bottles capture during the day. 14 Ctenophores were present throughout the upper layer (Table 1) and often peaked in concentration during 19 afternoon and night (Fig. 7c). Ctenophore biovolumes increased from the first to the fourth period (Table 3). 24 Mean depths of ctenophores, copepod nauplii, and Acartia tonsa copepodites and adults overlapped -40 -30 -20 -10 0 10 20 30 (Fig. 10b). However, mean depths of nauplii tended to Displacement (km) be shallower than ctenophore and copepodite mean Fig. 6. Cumulative displacement (km) of water at the fixed station at depths during the first occupation (2e83 h), while A. the end of (a) period one (2e20 h), (b) period two (20e40 h), and (c) tonsa adults tended to occur deeper, especially during period 3 (40e83 h). Lower and upper layer water masses during period the day. Mean depths of nauplii, copepodites and e 1(2 20 h) potentially were separated by as much as 60 km by the end ctenophores appeared to track pycnocline depth during of period 3 (within 3.5 days) due to wind-induced, enhanced two-layer circulation. the first occupation and were significantly correlated with pycnocline depth for the entire fixed-station sciaenid eggs were highest during the second period sampling period (Table 4a). At the beginning of the (20e40 h) (Table 3). Anchovy egg abundance declined second occupation (150e170 h), mean depths of cope- by the fourth period. pods and ctenophores no longer overlapped pycnocline Bay anchovy egg concentrations peaked near surface depth, possibly because the wind event, that had at sunset and at night (Fig. 7a), presumably because deepened the mixed layer, also dispersed organisms adults spawn above the pycnocline at night. The deeper through it. mean depths of anchovy eggs (Fig. 10a) just before The negative correlations between ctenophore bio- nightfall probably represented unhatched eggs, possibly volumes and both copepod nauplii and Acartia tonsa dead (Iseki and Kiyomoto, 1997), from the previous copepodite abundances (Table 4b) suggest that high night’s spawn that were diffusing and sinking. In concentrations of nauplii and copepodites did not co- contrast, the relatively shallow mean depths of occur- occur with high concentrations of ctenophores, possibly rence of sciaenid eggs (Fig. 10a) just before sunset the result of differing levels of copepod production E.W. North, E.D. Houde / Estuarine, Coastal and Shelf Science 60 (2004) 409e429 419

-3 a) Bay anchovy eggs (no. m ) shaded contours and salinity (psu) contour lines (no. m -3 ) 120 100 10 80 60 20 40 20 0 10 20 30 40 50 60 70 80 160 170

-3 b) Sciaenid eggs (no. m ) shaded contours and salinity (psu) contour lines (no. m-3 ) 24 21 10 18 15 12 20 9 Depth (m) Depth (m) 6 3 0 10 20 30 40 50 60 70 80 160 170

-3 c) Ctenophore biovolume (ml m ) shaded contours and salinity (psu) contour lines (ml m-3 ) 160 140 10 120 100 80 60

Depth (m) 20 40 20 0 10 20 30 40 50 60 70 80 160 170 Time (hrs)

Fig. 7. Time series of (a) bay anchovy and (b) sciaenid egg concentrations (no. m3, shaded contours) and (c) ctenophore biovolume (ml m3, shaded contours) with salinity (psu) contour lines for the first (2e83 h) and second (150e179 h) fixed-station occupations. Black dots indicate mid-point of trawl depth intervals. White and gray bars indicate day and night. and/or mortality from ctenophore predation within tion of diel vertical distribution of large anchovy larvae. these water masses. In fact, 67% of larvae O5 mm, 75% of larvae O9 mm, and 95% of larvae O14 mm were collected at night. The dayenight effect accounted for most of the variability in 3.4.5. Bay anchovy larvae abundances of larvae O9 mm in multiple-regression Abundance (Table 3) and upper layer concentrations models (Table 2) and was significant for larvae O6 mm. (Fig. 8) of bay anchovy yolk-sac and 3e6 mm larvae Abundances of bay anchovy early-life stages were peaked during the first period (2e20 h) and were more positively and most highly correlated with abundances than an order of magnitude lower in the third and fourth in adjacent length classes (Table 4b). Abundances of bay periods. Larvae O13 mm increased in abundance from anchovy larvae !13 mm were positively correlated with the first to the fourth period (Table 3). Eighty-five prey abundance (copepod nauplii and copepodites). percent of larvae collected during the fourth period Abundances of yolk-sac and 3e9 mm larvae were were O9 mm. These large larvae probably were not the negatively correlated with biovolume of ctenophores, survivors from eggs and newly-hatched larvae during a potential predator. the first period, but represented a new group advected In contrast to mean depths of their predators and into the region. Larvae hatched from eggs spawned prey, mean depths of bay anchovy larvae were often during the first period could not have been Ow7.5 mm located near, but did not always track, pycnocline depth by the fourth period because bay anchovy larvae hatch (Fig. 10c). The difference in mean depths of occurrence at w1.9 mm standard length (Houde and Zastrow, for large and small larvae indicated that ontogenetic 1991) and grow at a rate of w0.5e0.8 mm d1 (Rilling, factors influenced the vertical distribution of larvae. 1996; Auth, 2003). During the first 30 h of fixed-station sampling, the Larvae O6 mm showed clear peaks in concentration average mean depths of larger larvae (6e13 mm) were at night (Figs. 8c, 11), suggesting that evasion of the nearer to surface than smaller larvae (!6 mm) during Tucker trawl during the day can confound determina- the night but were deeper during day. 420 E.W. North, E.D. Houde / Estuarine, Coastal and Shelf Science 60 (2004) 409e429

-3 a) Bay anchovy yolk-sac larvae (no. m ) shaded contours and salinity (psu) contour lines (no. m-3 )

1.4 1.2 10 1.0 0.8 20 0.6 0.4 0.2 0.0 10 20 30 40 50 60 70 80 160 170

-3 b) Bay anchovy larvae 3 - 6 mm (no. m ) shaded contours and salinity (psu) contour lines (no. m-3 ) 0.7 0.6 10 0.5 0.4 0.3 20 0.2 0.1 0.0 10 20 30 40 50 60 70 80 160 170

-3 c) Bay anchovy 9 - 13 mm (no. m ) shaded contours and salinity (psu) contour lines (no. m-3 )

0.6 10 0.5 0.4 0.3 20

Depth (m) (m) Depth (m) Depth 0.2 0.1 0.0 10 20 30 40 50 60 70 80 160 170 Time (hrs) Fig. 8. Time series of bay anchovy larvae concentrations (no. m3, shaded contours) with salinity (psu) contour lines. (a) Yolk-sac larvae, (b) larvae 3e6 mm, and (c) larvae 9e13 mm for the ﬁrst (2e83 h) and second (150e179 h) station occupations. Black dots indicate mid-point of trawl depth intervals. Trawls in which no larvae were captured are marked by an ‘x’. White and gray bars indicate day and night.

The occurrence of larval mean depths nearer to surface Table 5 demonstrates the error in potential transport during the early evening (Fig. 10c) may be related to diel estimates that could be induced by net evasion. For periodicity in swimbladder inflation and the physiolog- example, if the majority of 9e10 mm larvae actually were ical requirement of clupeoid larvae to gulp air at the located at 10 m during the day but most larvae were surface to inflate their swimbladders (Blaxter and Hunter, collected at 12 m because of net evasion, then the estimate 1982). When collections were made O45 min after sunset, of potential transport (1.9 km) would be opposite in the mean depths of O6 mm larvae were nearest to surface direction from the actual transport (6.1 km). in early evening (20, 45, 70 h) and deepened through the night (Fig. 10c). During the entire fixed-station occupation, most larvae with inflated swimbladders consistently 4. Discussion occurred during night (Fig. 11). The diel periodicity in swimbladder inflation increased with larval development: Our results demonstrate that complex and interacting 50% of 6e7 mm larvae and O80% of O11 mm larvae biological and physical factors determine the character- had inflated swimbladders at night. Inflated swimblad- istics of larval fish nursery areas in estuaries. A suite of ders were observed in some larvae as small as 4e5 mm. factors influenced the distribution and potential trans- Results of the simulated transport analysis demon- port of bay anchovy early-life stages, including wind- strate the strong effect of shifts in vertical distributions on forced circulation patterns, pycnocline depth, DO potential transport of anchovy larvae (Table 5). A 1-m concentrations, salinity, and time of day. In addition, difference in daytime mean depth resulted in R2km prey and predator distributions and larval developmen- difference in potential transport after 3.5 days. If bay tal stage may be important factors that influence anchovy larvae make diurnal vertical migrations that transport of bay anchovy larvae. became deeper during the day as they develop, then poten- Wind events as short as 12e24 h had both direct and tial transport of larger larvae is up-estuary while trans- indirect effects on distribution of anchovy early-life port of younger larvae is down-estuary. Alternatively, stages. Direct wind mixing deepened the upper layer of E.W. North, E.D. Houde / Estuarine, Coastal and Shelf Science 60 (2004) 409e429 421

-1 a) Copepod nauplii (no. l ) shaded contours and salinity (psu) contour lines (no. l -1 )

300 270 240 10 210 180 150 120 20 90 60 30 0 10 20 30 40 50 60 70 80 160 170

-1 b) Acartia tonsa copepodites (no. l ) shaded contours and salinity (psu) contour lines (no. l -1 ) 40 35 10 30 25 20 20 15 10 5 0 10 20 30 40 50 60 70 80 160 170

-1 c) Acartia tonsa adults (no. l ) shaded contours and salinity (psu) contour lines (no. l -1 )

5.0 10 4.0 3.0 20 2.0 Depth (m) Depth (m) Depth (m) 1.0 0.0 10 20 30 40 50 60 70 80 160 170 Time (hrs) Fig. 9. Time series of copepod concentrations (shaded contours) with salinity (psu) contour lines. (a) Copepod nauplii (no. l1), (b) Acartia tonsa copepodites (no. l1), and (c) A. tonsa adults (no. l1) for the first (2e83 h) and second (150e179 h) station occupations. Black dots indicate depth of Niskin bottle collection. White and gray bars at top of panels indicate day and night. the water column, broadening the depth zone where tured the plankton community and controlled the overlap first-feeding anchovy larvae usually were located. Wind of bay anchovy larvae, their prey and predators in the events may diminish first-feeding larval survival by water column. Previous research in Chesapeake Bay also dispersing prey organisms (Lasker, 1975) or reduce indicated that the pycnocline influenced distribution of anchovy egg abundance by inhibiting adult spawning organisms. For example, Purcell et al. (1994) found (Moser and Pommeranz, 1999). ctenophores to be most abundant above the pycnocline in Wind events also may indirectly, but significantly, mid-Chesapeake Bay. Also in the mid-Bay, bay anchovy impact anchovy early-life stages by forcing a restructur- larvae were in greatest abundance above the pycnocline ing of the biological community. Wind-forced enhanced (MacGregor and Houde, 1996). In contrast, in baywide gravitational circulation that began after the first period surveys, Rilling and Houde (1999a) and Auth (2003) did quickly provided the plankton community above the not find significant differences between early-stage pycnocline with potential access to a different water mass abundances above and below the pycnocline over a broad below it. This two-layer movement can expand or reduce range of physical conditions. In the lower Bay, Govoni suitable habitat for larval fish and may affect prey and and Olney (1991) reported that bay anchovy eggs and predator populations that could vertically migrate into ctenophores were separated by the pycnocline at a well- the surface layer. In addition, small-scale differences in stratified station (eggs above, ctenophore below) but co- residual current velocities (i.e., shear) may have restruc- occurred at a well-mixed station. This suggests that tured plankton communities within the upper and lower degree of stratification can influence the importance of layers. For example, organisms in waters separated in the pycnocline as a frontal boundary as well as control depth by only 2 m at the beginning of the wind event predatoreprey interactions. could have been separated horizontally by 3 km within The importance of the pycnocline as a frontal bound- 20 h (Fig. 5). ary also may depend upon the DO conditions in the lower Results clearly demonstrated that the pycnocline is an layer. The observed low concentrations of organisms important frontal boundary (Largier, 1993) that struc- (except sciaenid eggs) below the pycnocline suggested 422 E.W. North, E.D. Houde / Estuarine, Coastal and Shelf Science 60 (2004) 409e429

Table 1 Longitudinal gradients in salinity also could have Percent organisms above the pycnocline during the entire fixed-station controlled the distribution of sciaenid eggs, as indicated sampling (total), as well as the first and second occupations by the shift from presence of sciaenid eggs throughout the Total First station Second station water column at high-salinity axial stations near the Bay occupation occupation mouth to presence only in the lower layer at stations Bay anchovy upbay of 230 km (Fig. 2e). Olney (1983) reported that the Eggs 81.5 88.6 61.7 distribution of sciaenid eggs had a distinct polyhaline (G 3.1, n Z 38) (G 2.2, n Z 28) (G 7.2, n Z 10) Yolk-sac 90.8 90.3 100.0 peak (most eggs were at stations with mean salinities larvae (G 4.6, n Z 22) (G 4.9, n Z 21) (n Z 1) O26 psu) in lower Chesapeake Bay in 1971e1973. This Larvae 80.9 78.5 89.9 indicates that sciaenid spawning may be restricted to a 3e6mm (G 4.9, n Z 33) (G 5.8, n Z 26) (G 8.3, n Z 7) narrow salinity range that limits spawning to near- and Larvae 78.5 77.0 83.7 sub-pycnocline waters at up-estuary locations. If sciaenid 6e9mm (G 4.6, n Z 36) (G 5.3, n Z 28) (G 9.4, n Z 8) Larvae 77.5 77.1 79.2 spawning is limited to sub-pycnocline depths up-estuary 9e13 mm (G 5.0, n Z 36) (G 5.6, n Z 28) (G 12.1, n Z 8) of Rappahannock Shoals (w225 km, Fig. 2a), then low Larvae 86.8 81.0 99.0 DO conditions may be a bottleneck to their egg and lar- O13 mm (G 4.5, n Z 28) (G 6.3, n Z 19) (G 0.7, n Z 9) val survival and could reduce reproductive success of Other organisms sciaenid fishes in Chesapeake Bay during years of severe Sciaenid eggs 10.0 9.8 10.4 hypoxia. (G 2.4, n Z 38) (G 2.8, n Z 28) (G 4.8, n Z 10) Prey and predator abundances may have had direct Ctenophores 72.4 70.4 77.9 and indirect effects on the distribution of bay anchovy (G 2.8, n Z 38) (G 2.8, n Z 28) (G 7.4, n Z 10) Copepod 75.9 76.6 73.8 eggs and early-stage larvae by influencing their survival nauplii (G 1.9, n Z 38) (G 2.2, n Z 28) (G 3.9, n Z 10) as well as adult spawning behavior. Adults may spawn Acartia tonsa 68.7 64.1 81.7 preferentially in areas of high copepod abundance copepodites (G 2.3, n Z 38) (G 2.2, n Z 28) (G 4.0, n Z 10) (Peebles et al., 1996; Peebles, 2002) as found during A. tonsa adult 58.4 49.0 84.6 the first 18 h of our study, or avoid spawning in areas of copepods (G 4.3, n Z 38) (G 4.1, n Z 28) (G 6.5, n Z 10) high gelatinous zooplankton biovolumes (Dorsey et al., Mean percentages were calculated for each set of depth-stratified 1996) as we found during 150e179 h. In addition, high samples using organism concentrations (no. m3 for eggs and larvae, no. l1 for copepod stages) and biovolume (ml m3 for ctenophores). ctenophore biovolumes may have reduced abundances Standard error (G 1 s.e.) and sample number (n) are reported in of anchovy eggs by direct predation. Mortality rates of parentheses below mean percentages. Samples were excluded from the anchovy eggs during our study (first period: Z ¼ analysis if no organisms were captured throughout the water column in 0:16 h1, third period: Z ¼ 0:17 h1 (North, 2001)) were one set of depth-stratified collections. higher than mean anchovy egg mortality rates (Z ¼ 0:066 h1 G 0:014 s.e.) in July 1991 in the mid to lower that low DO had an important effect on plankton Bay (Dorsey et al., 1996) when gelatinous zooplankton distributions by concentrating organisms within and biovolumes were lower (612.8 ml m2 maximum biovol- above the pycnocline. Roman et al. (1993) found that ume of ctenophores and the scyphomedusa Chrysaora copepod abundances were highest in the pycnocline when quinquecirrha) than those in our study (Table 3). Peak DO concentrations were !1mgl1 in bottom waters in biovolumes of ctenophores in surface waters at night Chesapeake Bay. Keister et al. (2000) demonstrated that during our study may have enhanced predation mortal- naked goby (Gobiosoma bosc) and bay anchovy larvae ity on newly spawned bay anchovy eggs. and copepod concentrations declined when DO levels Diel periodicity in bay anchovy egg abundances was were !2.0 mg l1 in the Patuxent River, a tributary of apparent: peak concentrations of bay anchovy eggs Chesapeake Bay. The co-occurrence of most cteno- occurred at night. Highest anchovy egg abundances at phores, copepods, and fish larvae in waters O2.0 mg l1 night have been reported in Peconic Bay, New York (above the pycnocline) in this study supports the (Ferraro, 1980) and Chesapeake Bay (Luo and Musick, hypothesis that low DO concentrations could be an 1991; Zastrow et al., 1991). Peak concentrations of sci- important factor influencing predatoreprey interactions aenid eggs also occurred in the evening when they were (Breitburg et al., 1994, 1999; Keister et al., 2000). highest in the water column indicating that spawning Salinity was another physical factor that may have occurred at this time, as reported for Cynoscion regalis influenced the distribution of fish early-life stages. Results in Peconic Bays, New York (Ferraro, 1980). Our ob- of the axial survey suggest that salinity may have servation supports the general supposition that dusk structured the longitudinal gradient in organism abun- spawning is characteristic of sciaenid fishes (Holt et al., dance in the ichthyoplankton community. Low salinities 1985). from high freshwater flow in 1996 could have limited bay The light/dark cycle also was a cue for inflation anchovy production to the lower estuary compared to and deflation of swimbladders by bay anchovy larvae. dryer years such as 1993 (Rilling and Houde, 1999a). Although previously unreported for bay anchovy, diel E.W. North, E.D. Houde / Estuarine, Coastal and Shelf Science 60 (2004) 409e429 423

Table 2 2 Fixed-station regression table: results of repeated measures multiple regression analysis on loge-transformed organism abundances (no. m or ml m2 for ctenophores) NDF DDF Displacement Dayenight Period Parameter FP Parameter FP Parameter FP estimate estimate estimate Bay anchovy Anchovy eggs 1 33 0.42 n.s. 0.97 n.s. 0.58 G 0.15 14.76 0.0005 Anchovy YSL 1 33 0.10 G 0.04 7.68 0.0091 0.04 n.s. 1.57 G 0.18 74.87 0.0001 Larvae 3e6mm 1 33 0.36 G 0.12 8.54 0.0062 1.80 n.s. 0.00 n.s. Larvae 6e9mm 1 33 0.41 G 0.06 53.5 0.0001 0.44 G 0.16 7.83 0.0085 1.43 G 0.35 16.72 0.0003 Larvae 9e13 mm 1 33 0.24 G 0.10 6.14 0.0185 1.51 G 0.29 26.51 0.0001 1.70 n.s. Larvae O13 mm 1 33 2.85 n.s. 2.63 G 0.36 49.27 0.0001 2.86 n.s. Other organisms Sciaenid eggs 1 33 0.06 G 0.02 7.96 0.0080 0.58 G 0.25 5.42 0.0261 3.11 n.s. Ctenophores 1 30 0.07 n.s. 0.49 n.s. 0.20 G 0.05 20.49 0.0001 Copepod nauplii 1 31 0.17 n.s. 0.00 n.s. 0.29 G 0.14 4.68 0.0383 Acartia tonsa copepodites 1 31 3.86 n.s. 0.38 G 0.24 10.96 0.0024 0.06 n.s. A. tonsa adults 1 31 0.47 n.s. 0.56 G 0.13 17.99 0.0002 0.52 n.s. The regression analysis was conducted to determine whether advection (displacement), photoperiod (dayenight), or time periods of differing physical conditions explained a significant amount of variability in organism abundance. Parameter estimates and their standard errors are reported for significant effects (a ¼ 0:05). Mean displacement in the lower layer was used for sciaenid eggs; upper layer mean displacement was used for other organisms. The dayenight variable was coded day Z 0, night Z 1 (night: sunrise to sunset). Period represents the four time periods when physical conditions differed (1 Z 2e20 h, 2 Z 20e40 h, 3 Z 40e83 h, 4 Z 150e179 h). NDF Z numerator degrees of freedom, DDF Z denominator degrees of freedom.

Table 3 Mean abundance of organisms (no. m2 for fish early-life stages and copepods, ml m2 for ctenophores) for the entire duration of fixed-station sampling (total) and for four time periods with different physical characteristics Total 2e20 h 20e40 h 40e83 h 150e179 h Bay anchovy 181.7 508.4 130.0 155.7 58.4 Eggs 34.25 83.72 69.15 44.23 8.25 0.97 4.65 1.26 0.06 0.06 Yolk-sac larvae 0.35 1.35 0.69 0.02 0.06 1.03 3.48 2.53 0.12 0.11 e Larvae 3 6mm 0.30 0.79 1.05 0.03 0.05 0.52 1.19 1.05 0.23 0.26 e Larvae 6 9mm 0.10 0.32 0.36 0.05 0.08 0.96 1.78 1.28 0.56 0.91 e Larvae 9 13 mm 0.19 0.78 0.54 0.14 0.37 0.73 0.35 0.52 0.45 1.55 Larvae O13 mm 0.21 0.21 0.23 0.16 0.71

Other organisms 66.5 48.1 164.2 53.4 40.1 Sciaenid eggs 10.04 18.79 21.63 12.99 9.20 1151.9 849.2 760.8 1229.6 1441.5 Ctenophores 67.66 61.74 80.19 88.05 140.28 8.06 ! 105 1.25 ! 106 1.25 ! 106 6.69 ! 105 4.45 ! 105 Copepod nauplii 0.74 ! 105 0.15 ! 106 0.20 ! 106 0.72 ! 105 0.48 ! 105 ! 5 ! 5 ! 5 ! 4 ! 4 Acartia tonsa 1.09 10 1.60 10 2.12 10 7.87 10 5.61 10 ! 5 ! 5 ! 5 copepodites 0.14 10 0.27 10 0.54 10 7650 6160 2.23 ! 104 1.36 ! 104 4.63 ! 104 2.47 ! 104 8050 A. tonsa adult copepods 3060 4890 6880 4160 1660 Sample size (n)35e38 6 5e615e16 9e10 Standard errors (G 1 s.e.) are given below the mean (in italics). Range of sample sizes (n) for means in each time period are indicated at the bottom of the table. 424 E.W. North, E.D. Houde / Estuarine, Coastal and Shelf Science 60 (2004) 409e429

a) Mean depth of bay anchovy and sciaenid eggs 0

5 Anchovy 10 Sciaenid

15 3mg/l DO Depth (m) night 20

25 020406080150 170

b) Mean depth of ctenophores, copepod nauplii, and A. tonsa copepodites and adults 0

5 Cteno Nauplii 10 Acopep

15 Aadult Depth (m) 3mg/l DO 20 night

25 0 20406080150 170

c) Mean depth of bay anchovy larvae by size class 0

ysl 5 3-6 mm 10 6-9 mm 9-13 mm 15 Depth (m) > 13 mm 3mg/l DO 20 night 25 020406080150 170 Time (hrs) Fig. 10. Fixed-station depth distributions of selected organisms. Mean depth (m) of occurrence of (a) bay anchovy and sciaenid eggs, (b) ctenophores, copepod nauplii, and A. tonsa copepodites and adults, and (c) bay anchovy larvae by size class (mm). Shaded area represents water below the pycnocline. The 3.0 mg l1 oxygen contour (heavy gray line) and night (sunset to sunrise: shaded boxes at top of panel) are depicted. periodicity in swimbladder inflation (inflation at night) bladders and 50% of 6e7 mm larvae had inflated has been reported for other clupeoid species such as swimbladders at night. These observations indicate that European anchovy Engraulis encrasicolus (Re´ , 1987, diel periodicity in swimbladder inflation may begin at 1990), Japanese anchovy Engraulis japonicus (Uotani a smaller size in bay anchovy than in other marine clup- et al., 2000), northern anchovy Engraulis mordax (Hunter eoids (7e10 mm for those listed above). Early swim- and Sanchez, 1976), gulf menhaden Brevoortia patronus bladder development may prevent sinking, an adaptive (Hoss and Phonlor, 1984) and Atlantic menhaden advantage in shallow estuaries where low salinities pro- Brevoortia tyrannus (Forward et al., 1993). Some bay vide less buoyancy and sub-pycnocline waters may anchovy larvae as small as 4e5 mm had inflated swim- contain stressful low DO. In addition, inflation of the Table 4 Results of correlation analysis of (a) mean depths of organisms, pycnocline depth, and 3.0 mg l1 dissolved oxygen contour depth, and (b) fixed-station organism abundances (no. m2 or ml m2 ) Anchovy Anchovy Larvae Larvae Larvae Larvae Sciaenid Ctenophores Copepod Acartia Acartia Pycnocline Oxycline

eggs YSL 3e6mm 6e9mm 9e13 mm O13 mm eggs nauplii copepodites adults depth depth 409 (2004) 60 Science Shelf and Coastal Estuarine, / Houde E.D. North, E.W. (a) Mean depths Anchovy eggs 1 0.45* 0.48** 0.41** 0.38* 0.74*** 0.68*** Anchovy YSL 1 Larvae 3e6mm 1 Larvae 6e9mm 1 Larvae 9e13 mm 1 0.74* Larvae O13 mm 1 0.84** 0.87** 0.74* Sciaenid eggs 1 0.48** Ctenophores 1 0.44** 0.48** 0.52*** 0.56*** Copepod nauplii 1 0.43** 0.66*** 0.59*** Acartia copepodites 1 0.44*** 0.42** Acartia adults 1 Pycnocline depth 1 0.86*** Oxycline depth 1 (b) Abundance Anchovy eggs 1 0.38* 0.41* 0.47** 0.39* Anchovy YSL 1 0.74*** 0.75*** 0.44** 0.37* 0.64*** 0.66*** Larvae 3e6 mm 1 0.81*** 0.47** 0.34* 0.60*** 0.67*** Larvae 6e9 mm 1 0.71*** 0.33* 0.45** 0.60*** Larvae 9e13 mm 1 0.66*** 0.35* 0.39* Larvae O13 mm 1 0.32* 0.38* Sciaenid Eggs 1 0.46** 0.43** 0.53*** Ctenophores 1 0.50** 0.40* Copepod nauplii 1 0.69*** Acartia copepodites 1 0.48**

Acartia adults 1 e 429

Only significant correlations are included in the table (*P ! 0:05, **P ! 0:01, ***P ! 0:001). All data were loge-transformed. Pearson correlation coefficients (r) were calculated for all normally distributed variables; Spearman rank-order correlation coefficients were calculated for tests involving non-normal variables (abundances of bay anchovy yolk-sac and O13 mm larvae). Sample sizes for organism abundance correlation analyses ranged from n ¼ 33 to 38. Sample sizes for the mean depth analysis ranged from n ¼ 6 to 13 for bay anchovy larvae O6 mm, and from n ¼ 22 to 38 for other organisms and anchovy size classes. Due to potential confounding related to net evasion during the day by anchovy larvae O6 mm, only mean depths of larvae O6 mm caught at night were included in the mean depth analysis. 425 426 E.W. North, E.D. Houde / Estuarine, Coastal and Shelf Science 60 (2004) 409e429

100 5 (O6.2 mm) were likely a result of gear avoidance during )

-3 the day. Avoidance of plankton nets during the day is 4 well known for many larval ﬁsh (Morse, 1989), in- 75 cluding clupeoid species (McGurk, 1992). 3 50 4.1. Potential transport 2 Swim Bladders Changes in larval depths of occurrence in relation to 25 1 the light/dark cycle, copepod prey concentrations, and Percent Larvae with Inflated

Total Larvae Captured (no. m lower layer DO concentrations may promote either up- 0 0 estuary or down-estuary transport of larval bay anchovy 04812162024 in Chesapeake Bay. Evidence exists for up-estuary Day Day Length Class (mm) transport of bay anchovy larvae. Bay anchovy larvae oc- Night Night cur up-estuary from spawning locations in the Patuxent Fig. 11. Percent of bay anchovy larvae with inflated swimbladders by River, Chesapeake Bay (Dovel, 1971), and Hudson River length class (mm) during day (open circles) and night (closed circles) (Dovel, 1981; Schultz et al., 2000), possibly indicating up- 3 on left Y-axis. Total larvae captured (no. m ) in each size class during estuary dispersal. In addition, larger larvae were found day (light gray line) and night (dark gray heavy line) on right Y-axis. Day defined as 20 min after sunrise and night as 20 min after sunset. further upstream than smaller larvae in the Patuxent Length classes are in 1 mm (larvae !15 mm) and 3 mm (larvae River (Loos and Perry, 1991), Hudson River (Schultz O15 mm) intervals. Sample sizes were 1012 (larvae !4 mm) and 1846 et al., 2000), and Chesapeake Bay (Auth, 2003). In the (larvae O4 mm). Chesapeake Bay and near the mouth of the Patuxent River, larger larvae were found inshore (MacGregor and Houde, 1996), suggesting directional transport from the swimbladder at night could reduce the energetic costs of Bay’s offshore regions towards shore and mouths of active swimming at night when larvae are not feeding tributaries. (Hunter and Sanchez, 1976; Uotani et al., 2000). De- In contrast, there also is evidence for transport of bay flation of the swimbladder during the day could reduce anchovy larvae down-estuary and out of Chesapeake predation by visual predators (inflated swimbladders are Bay. Most anchovy larvae in our study were found in the highly visible) and enhance prey-capture ability by larvae upper layer where net transport is expected to be down- due to better body mechanics (Forward et al., 1994). estuary. In a study at the mouth of the Bay, Olney (1996) The light/dark cycle also influenced catchability of estimated net flux of bay anchovy larvae to be seaward, anchovy larvae. Evasion of the Tucker trawl apparently although net flux of larvae O6.2 mm was less than half O affected day versus night abundances of larvae 6 mm. that of smaller larvae. Otolith microchemistry provides Olney (1996) also concluded that higher nighttime further evidence that net transport of bay anchovy may abundances of postflexion bay anchovy larvae be down-estuary during larval stages (Kimura et al., 2000). Trends in Sr:Ca ratios in otoliths of young-of-the- year collected at low salinity sites (5 psu) suggested initial Table 5 Potential transport of bay anchovy larvae by length class during the down-estuary transport to polyhaline regions followed first station occupation (2e83 h) under the hypothesis that larvae made by up-estuary migration to lower-salinity regions (Fig. 4a diurnal vertical migrations that became deeper during the day as larvae in Kimura et al., 2000). Otoliths of YOY individuals grew collected in the polyhaline (O18 psu) region did not Length class Nighttime Daytime Potential indicate either up- or down-estuary dispersal. (mm) depth (m) depth (m) transport (km) Despite the physical conditions and high freshwater 4e5 5 5 21.7 flow that may have promoted down-estuary transport in 5e6 5 6 19.7 June 1996, results of this study support the hypothesis e 6 7 5 7 16.7 that up-estuary transport is part of bay anchovy early- 7e8 5 8 13.2 8e9 5 9 10.0 life history strategy. During the first 18 h of this study, 9e10 5 10 6.1 larger larvae (O6 mm) were found deeper in the water 10e11 5 11 2.0 column than smaller larvae (!6 mm) during the day. 11e12 5 12 1.9 During the first fixed-station occupation, Acartia tonsa e 13 14 5 13 4.7 copepodite mean depths were deeper than nauplii mean 14e15 5 14 6.9 depths, and A. tonsa adults were deeper than copepo- In this hypothetical analysis, potential transport was calculated by dites. Larger bay anchovy larvae were more abundant at summing the cumulative displacement (km) at the 5-m depth interval during night (sunset to sunrise) with the cumulative displacement at depths where larger prey were located. Acartia tonsa, the depths specific to each length class during the day (sunrise to sunset). dominant copepod in mid and lower Chesapeake Bay Positive potential transport indicated transport down-estuary. during spring and summer, can make diel vertical E.W. North, E.D. Houde / Estuarine, Coastal and Shelf Science 60 (2004) 409e429 427 migrations, moving down in the water column during L. Beaven and J. Stone with plankton sample process- the day and up at night (Roman et al., 1988, 1993). ing, and C. Derry for processing CBOS and water level Because the diet of bay anchovy larvae shifts from data. Special thanks to E.M. Smith for providing nauplii to later-stage copepods as the larvae develop Winkler titration data. This research was supported by (Auth, 2003), the depth distributions of larvae that track the National Science Foundation Division of Ocean prey during the day may deepen as larvae develop, Sciences (Grant No. NSF OCE-9521512), a Chesapeake potentially resulting in no net transport or up-estuary Biological Laboratory Graduate Research Assistant- displacement. Progressive increases in daytime depths as ship, and the EPA Science-To-Achieve-Results (STAR) larvae develop may begin to counteract the propensity Fellowship Program (Fellowship No. U91-5366). This is for down-estuary transport of eggs and early-stage Contribution No. 3747 of the University of Maryland larvae (Table 5) before the strongly directed up-estuary Center for Environmental Science, Chesapeake Biolog- migration begins during metamorphosis (Kimura et al., ical Laboratory. 2000). Other studies provide supporting evidence for the hypothesis that bay anchovy larvae move up-estuary by References tracking the diurnal migrations of their Acartia tonsa prey during day and that potential for up-estuary trans- Auth, T., 2003. Interannual and regional patterns of abundance, port increases with larval development. Supporting evi- growth, and feeding ecology of larval bay anchovy (Anchoa mitchilli) in Chesapeake Bay. M.S. thesis, University of Maryland, dence from other studies suggests that (1) larvae of some College Park. clupeoids move deeper in the water column during the Baird, D., Ulanowicz, R.E., 1989. The seasonal dynamics of the day (Govoni and Pietrafesa, 1994; Olivar and Sabate´ s, Chesapeake Bay ecosystem. Ecological Monographs 59, 329e364. 1997; Olivar et al., 2001), including bay anchovy (Schultz Blaxter, J.H.S., Hunter, J.R., 1982. The biology of clupeoid fishes. e et al., 2003), (2) clupeoid larvae move deeper in the water Advances in Marine Biology 20, 1 223. Boehlert, G.W., Mundy, B.C., 1988. Roles of behavioral and physical column as they grow (Matsuura and Kitahara, 1995), (3) factors in larval and juvenile fish recruitment to estuarine nursery the daytime vertical distribution of clupeoid larvae can be areas. American Fisheries Society Symposium 3, 51e67. determined by prey concentrations within the region Boynton, W.R., Boicourt, W., Brandt, S., Hagy, J., Harding, L., where light was sufficient for larval feeding (Munk et al., Houde, E., Holliday, D.V., Jech, M., Kemp, W.M., Lascara, C., 1989), and (4) bay anchovy postflexion larvae can ex- Leach, S.D., Madden, A.P., Roman, M., Sanford, L., Smith, E.M., 1997. Interactions between physics and biology in the perience no net or weak up-estuary transport (Schultz estuarine turbidity maximum (ETM) of Chesapeake Bay, USA. et al., 2000). International Council for the Exploration of the Sea CM 1997/ Although this hypothesis is plausible, the significant S:11. confounding factor of reduced daytime catchability Breitburg, D.L., 1994. Behavioral response of fish larvae to low dis- makes it difficult to determine the actual depth-distribu- solved oxygen concentrations in a stratified water column. Marine Biology 120, 615e625. tion of larger bay anchovy larvae during the day and to Breitburg, D.L., Steinberg, N., DuBeau, S., Cooksey, C., Houde, E.D., estimate their potential transport, as seen in the simu- 1994. Effects of low dissolved oxygen on predation on estuarine fish lated transport analysis (Table 5). Testing the hypothesis larvae. Marine Ecology Progress Series 104, 235e246. will remain unresolved until better techniques are devel- Breitburg, D.L., Rose, K.A., Cowan Jr., J.H., 1999. Linking water oped to accurately measure the depth distribution of quality to larval survival: predation mortality of fish larvae in an oxygen-stratified water column. Marine Ecology Progress Series large larvae during day. Understanding potential size- 178, 39e54. dependent distribution and transport of bay anchovy Chesney, E.J., Houde, E.D., 1989. Laboratory studies on the effect of larvae is important for addressing regional differences in hypoxic waters on the survival of eggs and yolk-sac larvae of the mortality and growth (Rilling and Houde, 1999b), iden- bay anchovy, Anchoa mitchilli. In: Houde, E.D., Chesney, E.J., tifying optimum nursery areas, and understanding the Newberger, T.A., Vazquez, A.V., Zastrow, C.E., Morin, L.G., Harvey, H.R., Gooch, J.W. (Eds.), Population Biology of Bay impacts of low DO on the life-history strategies of larval Anchovy in Mid-Chesapeake Bay. Center for Environmental fish in estuaries. and Estuarine Studies, Chesapeake Biological Laboratory. Final Report to Maryland Sea Grant. Ref. No. (UMCES)CBL 89-141, pp. 184e191. Acknowledgements Daniel, L.B., Graves, J.E., 1994. Morphometric and genetic identifi- cation of eggs of spring-spawning sciaenids in lower Chesapeake Bay. Fishery Bulletin, U.S. 92, 254e261. We thank W.C. Boicourt for valuable guidance on Dorsey, S.E., Houde, E.D., Gamble, J.C., 1996. Cohort abundances the physics of estuaries and for the CBOS data that he and daily variability in mortality of eggs and yolk-sac larvae of bay provided. Insights of T. Miller, E. Russek-Cohen, and anchovy, Anchoa mitchilli, in Chesapeake Bay. Fishery Bulletin, e M. Roman on larval behavior, sample design, statistics, U.S. 94, 257 267. Dovel, W.L., 1971. Fish eggs and larvae of the upper Chesapeake Bay. and zooplankton are much appreciated. We thank the University of Maryland, NRI Special Report No. 4. crew of RV Cape Henlopen and cruise participants for Dovel, W.L., 1981. Ichthyoplankton of the lower Hudson Estuary, capable field support (esp. S. Leach and G.C. Rilling), New York. New York Fish and Game Journal 28 (1), 21e39. 428 E.W. North, E.D. Houde / Estuarine, Coastal and Shelf Science 60 (2004) 409e429

Ferraro, S.P., 1980. Daily time of spawning of 12 fishes in the Peconic MacGregor, J.M., Houde, E.D., 1996. Onshoreeoffshore pattern and Bays, New York. Fishery Bulletin, U.S. 78 (2), 455e464. variability in distribution and abundance of bay anchovy Anchoa Fortier, L., Leggett, W.C., 1983. Vertical migrations and transport of mitchilli eggs and larvae in Chesapeake Bay. Marine Ecology larval fish in a partially mixed estuary. Canadian Journal of Progress Series 138, 15e25. Fisheries and Aquatic Sciences 40 (10), 1543e1555. Matsuura, Y., Kitahara, E.M., 1995. Horizontal and vertical dis- Forward, R.B., Hettler, W.F., Hoss, D.E., 1994. Swimbladder deflation tribution of anchovy Engraulis anchoita eggs and larvae off Cape in the Atlantic menhaden, Brevoortia tyrannus. Fishery Bulletin, Santa Marta Grande in southern Brazil. Archives of Fishery and U.S. 92, 641e646. Marine Research 42 (3), 239e250. Forward, R.B., McKelvey, L.M., Hettler, W.F., Hoss, D.E., 1993. McGurk, M.D., 1992. Avoidance of towed plankton nets by herring Swimbladder inflation of the Atlantic menhaden Brevoortia larvae: a model of nighteday catch ratios based on larvae length, tyrannus. Fishery Bulletin, U.S. 91, 254e259. net speed and mesh width. Journal of Plankton Research 14 (1), Govoni, J.J., Olney, J.E., 1991. Potential predation on fish eggs by the 173e182. lobate ctenophore Mnemiopsis leidyi within and outside the Miller, J.M., 1988. Physical processes and the mechanisms of coastal Chesapeake Bay plume. Fishery Bulletin, U.S. 89, 181e186. migration of immature marine fishes. American Fisheries Society Govoni, J.J., Pietrafesa, L.J., 1994. Eulerian views of layered water Symposium 3, 68e76. currents, vertical distribution of some larval fishes and inferred Morse, W.W., 1989. Catchability, growth, and mortality of larval advective transport over the continental shelf off North Carolina, fishes. Fishery Bulletin, U.S. 87, 417e446. USA, in winter. Fisheries Oceanography 3, 120e132. Moser, H.G., Pommeranz, T., 1999. Vertical distribution of eggs and Hartman, K.J., Brandt, S.B., 1995. Trophic resource partitioning, larvae of northern anchovy, Engraulis mordax, and of the larvae of diets, and growth of sympatric estuarine predators. Transactions associated fishes at two sites in the Southern California Bight. of the American Fisheries Society 124 (4), 520e537. Fishery Bulletin, U.S. 97, 920e943. Holt, G.J., Holt, S.A., Arnold, C.R., 1985. Diel periodicity of spawn- Munk, P., Kiørboe, T., Christensen, V., 1989. Vertical migrations of ing in sciaenids. Marine Ecology Progress Series 27, 1e7. herring, Clupea harengus, larvae in relation to light and prey Hood, R.R., Wang, H.V., Purcell, J.E., Houde, E.D., Harding Jr., distribution. Environmental Biology of Fishes 26, 87e96. L.W., 1999. Modeling particles and pelagic organisms in Ches- Norcross, B.L., Shaw, R.F., 1984. Oceanic and estuarine transport apeake Bay: convergent features control plankton distributions. of fish eggs and larvae: a review. Transactions of the American Journal of Geophysical Research 104, 1223e1243. Fisheries Society 113, 153e165. Hoss, D.E., Phonlor, G., 1984. Field and laboratory observations North, E.W., 2001. Transport and retention of fish early-life stages in on diurnal swimbladder inflationedeflation in larvae of the Gulf Chesapeake Bay: mechanisms and implications for recruitment. menhaden, Brevoortia patronus. Fishery Bulletin, U.S. 82 (3), 513. PhD. dissertation, University of Maryland, College Park. Houde, E.D., 1974. Effects of temperature and delayed feeding on O’Connell, C.P., 1981. Development of organ systems in the northern growth and survival of larvae of three species of subtropical marine anchovy, Engraulis mordax, and other teleosts. American Zoologist fishes. Marine Biology 26, 271e285. 21, 429e446. Houde, E.D., Zastrow, C.E., 1991. Bay anchovy, Anchoa mitchilli. In: Olivar, M.P., Sabate´ s, A., 1997. Vertical distribution of fish larvae in Funderburk, S.L., Jordan, S.J., Mihursky, J.A., Riley, D. (Eds.), the north-west Mediterranean Sea in spring. Marine Biology 129, Habitat Requirements for Chesapeake Bay Living Resources. 289e300. Chesapeake Bay Program, Annapolis, MD, pp. 8-1e8-14. Olivar, M.P., Salat, J., Palomera, I., 2001. Comparative study of Hunter, J.R., Sanchez, C., 1976. Diel changes in swimbladder inflation spatial distribution patterns of the early stages of anchovy and of the larvae of the northern anchovy, Engraulis mordax. Fishery pilchard in the NW Mediterranean Sea. Marine Ecology Progress Bulletin, U.S. 74 (4), 847e855. Series 217, 111e120. Iseki, K., Kiyomoto, Y., 1997. Distribution and settling of Japanese Olney, J.E., 1983. Eggs and early larvae of the bay anchovy, Anchoa anchovy (Engraulis japonicus) eggs at the spawning ground off mitchilli, and the weakfish, Cynoscion regalis, in lower Chesapeake Changjiang River in the East China Sea. Fisheries Oceanography Bay with notes on associated ichthyoplankton. Estuaries 6 (1), 6 (3), 205e210. 20e35. Keister, J.E., Houde, E.D., Breitburg, D.L., 2000. Effects of bottom- Olney, J.E., 1996. Community structure, small-scale patchiness, and layer hypoxia on abundances and depth distributions of organisms feeding of larval fishes in an estuarine plume. PhD. dissertation. in Patuxent River, Chesapeake Bay. Marine Ecology Progress University of Maryland, College Park. Series 205, 43e59. Peebles, E.B., 2002. Temporal resolution of biological and physical Kimura, R., Secor, D.H., Houde, E.D., Piccoli, P.M., 2000. Up-estuary influences on bay anchovy Anchoa mitchilli egg abundance near dispersal of young-of-the-year bay anchovy Anchoa mitchilli in the a river-plume frontal zone. Marine Ecology Progress Series 237, Chesapeake Bay: inferences from microprobe analysis of strontium 257e269. in otoliths. Marine Ecology Progress Series 208, 217e227. Peebles, E.B., Hall, J.R., Tolley, S.G., 1996. Egg production by the bay Largier, J.L., 1993. Estuarine fronts: how important are they? anchovy Anchoa mitchilli in relation to adult and larval prey fields. Estuaries 16 (1), 1e11. Marine Ecology Progress Series 131, 61e73. Lasker, R., 1975. Field criteria for survival of anchovy larvae: the Purcell, J.E., Nemazie, D.A., Dorsey, S.E., Houde, E.D., Gamble, relation between inshore chlorophyll maximum layers and success- J.C., 1994. Predation mortality of bay anchovy Anchoa mitchilli ful first feeding. Fishery Bulletin, U.S. 73 (3), 453e462. eggs and larvae due to scyphomedusae and ctenophores in Loos, J.J., Perry, E.S., 1991. Larval migration and mortality rates Chesapeake Bay. Marine Ecology Progress Series 114, 47e58. of bay anchovy in the Patuxent River. Thirteenth Annual Re´ , P., 1987. Ecology of the planktonic phase of the anchovy, Larval Fish Conference, NOAA/NMFS Technical Report Series. Engraulis encrasicolus (Linnaeus 1758), within Mira Estuary 95, pp. 65e76. (Portugal). Investigacion Pesquera (Spain) 51 (4), 581e598. Luo, J., Musick, J.A., 1991. Reproductive biology of bay anchovy in Re´ , P., 1990. Tidal transport and retention of anchovy eggs and Chesapeake Bay. Transactions of the American Fisheries Society larvae within Mira Estuary (Portugal). Portugaliae Zoologica 1 (2), 120, 701e710. 7e13. Mann, K.H., Lazier, J.R.N., 1996. Dynamics of Marine Ecosystems: Rilling, G.C., 1996. Regional and temporal variability in growth, mortality, BiologicalePhysical Interactions in the Oceans, 2nd ed. Blackwell and population age structure of bay anchovy (Anchoa mitchilli)eggs Science, Inc., Malden, Massachusetts, 394 pp. and larvae in the Chesapeake Bay. M.S. thesis, University of Maryland. E.W. North, E.D. Houde / Estuarine, Coastal and Shelf Science 60 (2004) 409e429 429

Rilling, G.C., Houde, E.D., 1999a. Regional and temporal variability Schultz, E.T., Cowen, R.K., Lwiza, K.M.M., Gospodarek, A.M., in distribution and abundance of bay anchovy (Anchoa mitchilli) 2000. Explaining advection: do larval bay anchovy (Anchoa eggs, larvae, and adult biomass in the Chesapeake Bay. Estuaries mitchilli) show selective tidal-stream transport? ICES Journal of 22 (4), 1096e1109. Marine Science 57, 360e371. Rilling, G.C., Houde, E.D., 1999b. Regional and temporal variability Schultz, E.T., Lwiza, K.M.M., Fencil, M.C., Martin, J.M., 2003. in growth and mortality of bay anchovy, Anchoa mitchilli, larvae in Mechanisms promoting upriver transport of larvae of two fish Chesapeake Bay. Fishery Bulletin, U.S. 87, 555e569. species in the Hudson River Estuary (USA). Marine Ecology Roman, M.R., Ashton, K.A., Gauzens, A.L., 1988. Day/night differ- Progress Series 251, 263e277. ences in the grazing impact of marine copepods. Hydrobiologia Uotani, I., Fukui, A., Osaki, K., Ozawa, T., 2000. Experimental study of 167/168, 21e30. the inflation and deflation of gas bladder of Japanese anchovy, Roman, M.R., Gauzens, A.L., Rhinehart, W.R., White, J.R., 1993. Engraulis japonicus larvae. Bulletin of Marine Science 66 (1), 97e103. Effects of low oxygen waters on Chesapeake Bay zooplankton. Weinstein, M.P., Weiss, S.L., Hodson, R.G., Gerry, L.R., 1980. Limnology and Oceanography 38 (8), 1603e1614. Retention of three taxa of postlarval fishes in an intensively flushed Rowe, P.M., Epifanio, C.E., 1994. Tidal stream transport of weakfish tidal estuary, Cape Fear River, North Carolina. Fishery Bulletin, larvae in Delaware Bay, USA. Marine Ecology Progress Series 110, U.S. 78 (2), 419e435. 105e114. Zastrow, C.E., Houde, E.D., Morin, L.G., 1991. Spawning, fecundity, Sanford, L.P., Boicourt, W.C., 1990. Wind-forced salt intrusion into hatch-date frequency and young-of-the-year growth of bay a tributary estuary. Journal of Geophysical Research 95 (C8), anchovy Anchoa mitchilli in mid-Chesapeake Bay. Marine Ecology 13357e13371. Progress Series 73, 161e171.