Influence of Hydrographic Conditions and Wind ...

Can. J. Fish. Aquat. Sci. Downloaded from www.nrcresearchpress.com by Shanghai International Studies University on 06/05/13 For personal use only.

Influence of Hydrographic Conditions and Wind Forcing on the Distribution and Abunda,nce of Dungeness crab, Cancer magister, Larvae Roderick C. Hobbshand Louis W. Botsford Department sf Wildlife and Fisheries Bioisgy and Center for Popuiation Biology, University of California, Davis, CA 954 16, USA

and Andrew Thomas' College sf Oceanogsaphg/,Oregon State University, CorvaElis, OR 9733 7-5503, USA

Hobbs, R. C., L. W. Botsford, and A. Thomas. 1992. Influence of hydrographic csndi.tions and wind forcing on the distribution and abundance of Dungeness crab, Cancer rnagister, larvae. Can. 1. Fish. Aquat. Sci. 49: 1379-1 388. The distribution of larval Duwgeness crab (Cancer rnagister) was sampled by joint U.S.A./U.S.S.R. ichthyoplankton surveys off the coasts of Washington, Oregon, and northern California involving more than 1263 stations from 5 to 360 km offshore and between 40 and 48"N latitude in each of the five years 1981-85. Observed cross-shelf distributions of megalopae were consistent with a mechanism by which die9 vertical migratory behavior of the late-stage megalopae in the presence of wind-induced currents results in onshore transport to favorable settlement areas. Total onshore transport for 45 d prior to sampling, estimated as the effects of wind on a passive particle at the surface at night and in the Ekman layer during the day, was correlated with the nearshore megalopal abundance. Mean larval densities for each of the five cruises declined exponentially with time of year of the cruise. This implied (1) an instantaneous mortality rate of 0.066.d-' and ( 2 ) that survival is independent of interannual variation in environment. Abundance of megalopae was not correlated with station hydrographic data (salinity, temperature, and dissolved oxygen) nor chlorophyll levels (satellite data from the Coastal Zone Color Scanner). C)e 4 981 a 1985, la repartition annuelle des larves du crabe dormeur (Cancer magister) au large des c6tes des Etats de Washington, de 1'8regon et de la Californie nord a $te determinee lors de releves de l'ichtyoplancton men& conjointement par des chercheurs arn6ricains et russes. Des Gchantillons ont kt4 preleves 3 plus de 120 stations a partir de 5 krn des c6tes jusqu'a 368 km au large, entre 48 et 48" de latitude nord. La repartitiow sbserv6e des megalopes sur cette &endue de la plate-forme correspondait au ph6nomGne par B'entremise duquel la migration verticale nycth6merale des stades megalopes avances en presence de courants dus au vent entraine le transport vers des fonds &tiers favorables. Il y a correlation entre le transport total vers les eaux c8tiGres pendant 45 d avant 116chantillonnage, etabli comme I'incidence du vent sur une particule passive presente la nuit dans Ies eaux de surface et le jour dans la couche d'Ekrnan, et B'abondance des megalopes dans les eaux cdti$res. La densite moyenne des lames pendant les cinq croisiGres d'echantillonnage a dirninue de fason exponentielle en fonction du moment de t'echantillonnage. Ceci sugggre un taux de mortalit4 instantank de 0,066-d-' et l'independance de la survie sur la variation interannuelle du milieu. L'abndance des m6galopes n'etait pas en corr4latisn avec les donnees hydrographiques relevees 3 chaque station (salinite, temHrature et oxygene dissous), ni avec les teneurs en chlorophyl le (donnees-satellite recueillies par balayeur couleur de zone cdtihe). Received November 26, 1 990 Accepted Banuary 23, 1992

(hi82H )

andings o f the Bungeness crab (Cancer magistet-) f m m northern California to Washington fluctuate by as much as an order o f magnitude in apparent synchronous cycles sf period 10 yr (reviewed in Botsford 1986; Botsford et al. 1989). Although considerable research has focused on the cause o f these cycles, their dynamic basis is still not understood. One o f the proposed causes sf these cycles is cyclic fluctuations in the physical environment. Corre%ationsbetween m u a l popu-

'Current address: National Marine Mmmals Laboratory, National Marine Fisheries Service, 7688 Sand Point Way NE, Seattle, WA 98 1 15-8070,USA. 'Current address: Atlantic Centre for Remote Sensing of the Oceans, Suite 3381, 6155 North St., Halifax, N.S. B3K 5W3, Canada. Can. J . Fish. Aquat. Sci., Vol. 49, I992

lation data and various environmental variables suggest that the physical environment influences egg and larval stages. Annual catch, considered as a proxy for cohort abundance, has been shown to be correlated with wind stress duping the late hrvd period (Johnson et al. 1986) and sea surface temperature during the hatching period (Wild 1980). Mechanisms proposed are wind-driven transport to a suitable settlement environment and poor egg clutch survival due to high temperatures, respectively. Because variation in larval survival, transport, and successful settlement could determine the cycles in Dungeness crab abundance, a broad-scale survey o f the distribution and abundance o f larval crabs would be useful in understanding the population dynamics o f the Bungeness crab. The joint U.S. A./U. S . S .R. ichthyopliarmkton surveys o f f the coasts o f Washington, Oregon,

Can. J. Fish. Aquat. Sci. Downloaded from www.nrcresearchpress.com by Shanghai International Studies University on 06/05/13 For personal use only.

and northern California during the years 1981-85 coincided with the late larval phase of the Dungeness crab. More than 120 stations from 5 to 368 Ian offshore and between 40 and 48"N were sampled in each of the five years. The data from these cruises provide a unique opportunity for c o m p ~ s o nof lava1 distribution to environmental conditions over a broad spatial range. Possible Environmental Influences Dungeness crab larvae in this region are subject to oceanographic influences during the first 5 or 6 mo of the year. They hatch near January from eggs carried on the female's abdomen and then progress through five zoeal stages during the next 3-4 mo. They then metamorphose into a megalopal stage and settle to the bottom about 1-2 mo later (Lough 1976; Reilly 1983; Hatfield 1983; Stevens and Amstrong 1985). They settle nearshore in shallow coastal regions md estuaries (Reilly 1983; strong 1984, 1985), and few Dungeness crab are found at depths greater than 60 m (e.g. Cmasco et al. 1985). Both the potential effects of physical conditions and the interpretation of the plankton samples depend on the timing of metamorphosis from the zoeal to the megalopal stage. In 1970, sampling off Newport, Oregon, Lough (1976) first collected megalopae in mid-April and observed that most of the larvae were megalopae by early May. Megalopae first appeared somewhat earlier in central California, around mid-March (Reilly 1983). Hobbs md Botsford (1992) found that metmorphosis had occurred 3 wk earlier off Newport, Oregon, in 1984 than Lough reported for 1970 and 1971. They concluded from 5 yr of samples that most metamorphosis from zoeae to megalopae off Washington, Oregon, and northern California occurred over 2-3 wk at a given latitude and that this peak of metamorphosis occurs either nearly simulaneouslyat all latitudes or in a manner that proceeds from south to north over less than a month. One possible environmental effect which we explore here is the potential effect sf the local physical and biological environment of the larvae ow survival rates and timing of larval development. Laboratory studies indicate that Dungeness crab larvae are sensitive to temperature and salinity concentration (Reed 1969; Gaumer 1971; Lough 1976; Sulkin and McKeen 1989). A second possible environmental effect explored here is transport of the larval stages to favorable settlement areas. The facts that this study and others report larval distributions extending st considerable distance offshore (Lough 1976; Reilly 1983; Jmieson and Phillips 1988; Jamieson et al. 1989), while successful larvae must settle nearshore ((Reilly 1983; Stevens and h s t r o n g 1984, 1985; Cmasco et al. 1985), imply that cross-shelf transport may be important in returning Dungeness crab larvae to the nearshore environment for successful settlement, The fact that later megalopal stages are found closer to shore than earlier ones (Hatfield 1983; Jamieson and Phillips 1988) also implies onshore transport of this stage. Some investigators have found that in samples taken on cross-shelf transects at a single point on the coast, successive zoed stages occur progressively farther offshore ( h u g h 1976; Reilly 1983). They have interpreted this to mean that successive zoeal stages we transported uniformly offshore; hence the megdopd stage would require transport onshore. However, it is also possible that alongshore transport contributed to these observed patterns (cf. Johnson et al. 1986). Transport of the larvae will depend on the predominant currents in this region. The California Current System off

Washington, Oregon, and northern California consists of the southward Wowing California Current and the northward flowing California Undercurrent which surfaces as the Davidson Current in the fall and winter (Hickey 1979, 1989). Flows during the winter are typically southward in the upper layer offshore and northward inshore and beneath the upper layers offshore. After the spring transition the California Current moves inshore md flows are typically southward at the surface and northward below. Wind-driven variations in the alongshore Wow are significant (Hickey 1979, 1989; Stmb et al- 1987b). Cross-shelf transport at the surface is almost entirely wind driven, being typically onshore during the fall and winter months and offshore during the spring and su Cross-shelf transport beneath the surface is generally opposite to that at the surface (Hickey 1979, 1989; Stmb et al. 1987b). The winds are typically onshore between 35 and 39"N latitude in the winter diverging to the north and south. Above 42"N latitude the winds generally are northward. At some time during March or April the divergence zone below 39"N shifts northward to around 48"N so that the winds are typically onshore and southward above 45"N latitude and increasingly southward below this latitude during the spring and summer. This shift in the divergence zone is referred to as the spring transition and drives the reversal in cross-shelf transport that occurs at this time (Strub et al. 1987a, 1987b). Wind-forced transport of the l m a e would depend on their vertical distribution, for which some observations exist. Researchers have typically found the megalopal stage assmiated with the surface more than the zwal stages (Reilly 1983; Booth et al. 1985; Jamieson and Phillips 1988; Shedcer 1988; Jmieson et al. 1989; Hobbs and Botsford 1992). Megalopae have been seen during the day clinging to the pleustonic hydroid Ve%ellave&e&&a (Wickham 1979;Reilly 1983; Shenker 1988) and other flotsam and aggregating in windrows at the surface (Shedcer 1988) on cloudy days. There is evidence for die1 verticd migration of megalopae with most in the neuston at night (Booth et al. 1985; Jamieson and Phillips 1988; Shenker 1988; Jmieson et al. 1989) but as deep as 60 m on bright days (L. W. Botsford and J. M. Shenker, unpublished plankton data). From the analyses of diel variation in vertical distribution for 5 yr of data from these U.S. A./U. S.S.R. ichthyoplankton cmises, Hobbs and Botsford (1992) concluded that vertical migratory behavior varies with development stage. Late-stage megdopae migrated vertically on a diel vertical basis, being in the neuston at night and below during the day. The fraction of late-stage megalopae in the neuston was 0.62 k 0.09 from the hours sf 19:W to 08:W PST and 8.08 + 0.02 between 08:W and 19.80 PST. These results are consistent with those obtained by Booth et stl. (1985). Fewer early stage megalopae were in the neuston and they were generally below the neuston at all hours in the one year that the ichthyoplankton cruise coincided with the time of metamorphosis from zoeae to megalopae. In all five ichthyopldcton cruises, zoeae were generally below the neuston at all hours of the day except for 2 or 3 h in the evening (Hobbs and Botsford 1992). We here test directly for (1) an influence of local environment on l m d distribution using hydrographic and chlorophyll data a d (2) an influence of wind-induced transport on cross-shelf distribution of crab larvae, using wind data,

Data and Metho& Abundance Data Five joint U.S.A./U.S.S .R.ichthyoplankon cruises were conducted in 1981-85 (Table 1). The 1984 cruise occurred Can. 9. Fish. Aquat. Sci., VQ%.49, 1992

T ~ L 1.E Cruise details for the 5 yr of sampling.

Can. J. Fish. Aquat. Sci. Downloaded from www.nrcresearchpress.com by Shanghai International Studies University on 06/05/13 For personal use only.

Cruise No.

Research Vessel

Perid

Poseydon Poseydon Equator Poseydon Mys Babyshkim-

4 8 OON

4 7 OBN

46 BBN

.

45 0016

4 4 QOM

-

43 OQN

-

42 OOH

-

41 OBN

-

40 OOH

39 OON

129 OOW

127 OQW

125 OOW

123 OOW

FIG. 1. Typical cruise track. 'This is the 1983 cmise which began at

48"Nlatitude in %ate April and ended at 4O0Nlatitude in mid-May (from Clark and Kendall 1985). The tracks of the other cruises vary from this one only in the order s f sampling of nearshore stations at each latitude.

approximately 5 wk earlier than the 1983 and 1985 cruises, which were about 2 wk earlier than the 1981 and 1982 cruises. Each cruise proceeded from worth to south and sampled 124 stations from 5 to 360 km offshore and between 40 md 48"N latitude (e.g. Fig. 1). Samples at most stations consist of a neuston tow and an oblique bongo tow, but at some stations only a neuston sample was taken. Neuston tows were made at 2.0 knots (1.03 m-s- '1 for 10 min using a 0.3-m-high by 0.5-m-wide Smeoto sampler (Sameoto and Jiiroszynski 1969) Can. . I Fish. . Aquar. Sci., Vok. 49, 1992

Published report

Clark 1984 Clark 1986 Clark a d Kendall 1985 Clark md Savage 1988 Savage 1989a

with a 0.505-mm-mesh net. Bongo tows were standard MARMAP tows (Smith and Richardson 1977) using 64)-cm bongo frames and 8.585-mm-mesh nets. Tows were made between the surface and the bottom or a maximum of either 380 or 578 m (depending on the yea) of wire out. Flowmeters in the mouths of the nets were used to detemine the volume of water filtered. The ichthyoplarnleton smples from each cmise were sorted for fish larvae in Poland, and brachyuran larvae were set aside for our use. Samples containing greater than about 60 zoeae were divided using a plankton splitter to generate a subsample of approximately 25-50 zwae. These zoeae were examined under a binocular dissecting microscope and identified as either C. magister, other brachyurms, or unidentifiable (damaged). All megalopae were examined and identified as C. magistea or other brachyurans (Lough 1975). In both zoeae m d megalopae, Dungeness crab was the predominant species. The 1981-83 m d 1985 cruises occurred at a time of y e a when most (95%) of the crab larvae had progressed to the megalopal stage with a few still in late zoeal stages. The 1984 cruise occurred at the time of metamorphosis with most larvae in the north being zoeae and in the south being megalopae. We estimated local I m a l abundance and the fraction of larvae in the neuston from the paired neuston and bongo samples using a bias-corrected maximum likelihood method (Hobbs and Botsford 1992). Estimation of megalopal abundance in the 1981-83 and 1985 cruises was based on an assumed diel pattern of migration with a constant fraction of megalopae in the neuston during the day and at night (Hobbs and Botsford 1982). For the few stations where only neuston samples were taken (75 stations, d l in the 1982 cmise), we used the estimates of fraction of megalopae in the neuston versus time of day derived in Hobbs and Botsford (1992) as "correction factors" (cf. Jmieson and Phillips 1988; Jamieson et al. 1989) to estimate the abundance of megalopae in the entire water column. Fractions in the neuston were 62% of megdopae in the neuston at night (%9:00-08:00PST) and 8% of megalopae in the neuston during the day (08:00-19:W PST) (Hobbs and Botsford 1992). We used a similar estimation method to calculate the abundance of zoeae found in the 1981-83 and 1985 cruises, the only difference being that this estimate was based on a 3-h moving average of zoeae to reduce variability (Hobbs and Botsford 1992). In 1984 (the earliest cruise), nearly all of the larvae were found in the bongo samples, and the average abundances of both zoeae and megdopae were sufficiently l a g e that we used only the bongo samples as measures of local abundance for this cruise. Before conducting statistical tests to detemine spatial or temporal covariability between two data sources, it is- prudent to first remove m y variabilty introduced by the sampling process. To that end, before attempting to detemine whether the spatial distribution of larval data covxied in some way with environmental data, we first removed the effects of the decline in abundance within each year due to natural attrition of the

Can. J. Fish. Aquat. Sci. Downloaded from www.nrcresearchpress.com by Shanghai International Studies University on 06/05/13 For personal use only.

l m a e 4i.e. mortality and settlement). Because each of the cmises spanned about 1 mo and began on a different date during the spring with nearly 3 mo separating the beginning of the earliest cruise and the end of the latest cruise (Table I), the abundance of larvae would be expected to change dramatically over this time due to mortality and settlement. This causes differences in s m p l e abundance that are due only to differences in inkaannual timing of the s a p l i n g . To remove these differences, we first estimated the rate of decline within the year and then used that estimate to remove that effect. To estimate the intraannual rate of larval decline from our interannual data, we regressed the logarithm of average larval density (per 10 m2)over each cmise divided by estimated annual egg production (approximated by total (male) catch in the following yea, c.f. Hobbs m d Botsford 1989) on time of year at the midpoint of the cruise. We divided l m a l density by estimated mnud egg production to remove the effect of the variation im egg production between yews. Since this regression indicated a constant rate of decline though the year* we were able to use the estimated l m a l mortality rate to remove the effects of that attrition from the abundance data and thereby produce a set of quasi-synoptic abundance estimate (QSAEs). Because the last larval stage is probably the most important one for find settlement, we focused primarily on the megalopae in the later cruises (198 183, 1985) and adjusted all data for each of those years to what we estimate it would have been if sampled synoptically on May 10 each yea. This involved removing the effect of l m d decline within each year from the day-night corrected abundance estimates using where N(to) is quasi-synoptic abundance on day to (May 101, N(t) is the abundance observed on day t , and m is the instantaneous larval rate of decline. We chose May 18 because n to d l four of the late cruises (Table 1). For the 1984 cmise, we use the average sample date for the cruise (Day of the y e a (DOY) 83) for to. Physical Data Hydrographic data and information on sea state and cloud cover were also taken at each station in the five cruises. These data were measurements of salinity, temperature? and dissolved oxygen taken near the surface by the ship taking the plankton samples. The salinity and temperature data have been analyzed by Savage (1989b). ChlorophylB abundance at each sample station in each year was from digital images from the Coastal Zone Color Scanner (CZCS). The CZCS ,an instrument on the NIMBUS-7 satellite, measures intensity of light reflected from the ocean surface from which chlorophyll abundance can be estimated. Both cloud cover and the low sun angle at higher latitudes early in the year obscure the surface reflectance, making the chlorophyll readings at those points unreliable or nonexistent. We were able to obtain useable chlorophyll readings for approximately one third of the stations. These stations me predominantly within 100 h of the shore and later in the year due to the extensive cloud cover earlier in the yearo We used the U.S. Navy's Fleet Numerical Oceanographic Center (FNOC) meteorological wind fields to compute possible wind-driven transport of the larvae. This wind product provides surface wind velocities twice daily on a grid with 380-bspatial resolution over the entire study period. These are estimated geostrophic winds based on ship and buoy measurements of

EI

Total Larvae

Mmn Sample Date (DOY) FIG. 2 . E m a l densities averaged over all locations for each cmise, divided by relative egg prdnctian in each yea, rand plotted against the mean smple date for each cruise. Slops: total larvae = -0.066-d-', zaeae = -0.089-d-', and megalspae = -0.054-d--'.

wind velocity and pressure, as well as model-derived pressure fields. Surface components of these winds are produced by a reduction in velocity md a rotation to simulate boundary layer conditions. In computing wind-driven transport, we accounted for the fact that megalopae me primarily at the surface at night m d below during the day. However, because our howledge of vertical migration and the influence of wind on surface transport are uncertain, transport was modeled in thee ways. We calculated larval transport for the 45-d period prior to sampling for the late cruises (1981-83, 1985) as (1) Ekman transport alone, (2) Ekmm transport for 12 h per day and 39% of the wind speed in the direction of the wind for the remaining 12 h, and (39 E h a n transport 12 h per day and 3% of the wind speed with the direction rotated 15"to the right for the remaining 12 h. The first assumes that the megalopae are distributed throughout the upper E h a n layer and does not take into account vertical migration. The latter two account for vertical migration by assuming that the megalopae are distributed throughout the upper E h a w layer during the day but are in the rneuston at night. The direction of transport of the megalopae in the neuston is considered to be much closer to the wind direction at the sea surface than the direction of E b a n mass transport, but the actual direction in relation to wind direction is not well known. The two different surface transport directions were an attempt to bracket this uncertainty. We computed Ekmaw velocity components from

is the density of seawater, TEKis volume Ekman where p,, transport, and BE,is the depth of the wind-driven layer, which was assumed to be

Can. J . Fish. Aquat. Sci., Vol. 49, 1992

where W is the wind velocity. Trmsprt over each 12-h period was computed from the value of FNOC wind from that period, interpolated to each degree of latitude from 39 to 49ON, at a p i n t 25 km offshore.

abundance (coefficient of variation = 0.21, dividing lmaH density for each year by this factor had little effect on the result. 'The slope of the regression of adjusted megalopal density (-0.054.d- I ) is not as steep as the rate of decline for totd l m a e because it includes not just losses due to mortality and settlement, but also additions due to metamorphosis from zoeae to megdopae. The rate of decline of adjusted zoeal density ( - 0.089-d- I ) includes both losses due to mortality and losses due to metamorphosis. The fact that the zoeal and the megalopal rates are close to the rate of decline for total l m a e suggests that both settlement and metamqhosis rates are less than the mortality rate. Because we have information on only two stages, it is not pssible to estimate rates of metamorphosis and

Can. J. Fish. Aquat. Sci. Downloaded from www.nrcresearchpress.com by Shanghai International Studies University on 06/05/13 For personal use only.

Results Interannual Temporal Patterns The estimated overall rate of decline for total larvae is 0.W.d-', or a half-life of about 10.5 d (Fig. 2). This rate includes losses due to mortality of zoeae and megdopae, settlement of megalopae, and advtxtion outside the area surveyed. Because of low variability in estimated female

9

.

.

. . -

m

e

*

. *

. m

39 129

.

I

128

127

126

125

124

429

128

longitude

39

la

128

127

126

$ongitude

. .

I

127

.

.

.m .

1

126

I

125

124

125

124

longitude

125

124

39 129

128

I27

1%

longitude

Fro. 3. Quasi-synopticplots of the corrected megalopal abundances in each sf the later cruises (adjusted to May 10) with the estimated abundance represented by the size of the symbol. Day and night samples are indicated by the open cireles and solid squares, respectively. (a) 1981;(b) 1982; (c) 1983; (d) 1985. Can. J . Fish. Aquat. Scr'.. b l . 49, 1992

Can. J. Fish. Aquat. Sci. Downloaded from www.nrcresearchpress.com by Shanghai International Studies University on 06/05/13 For personal use only.

Distance Offshore (km) 8

533

18843

15a

2m

2%

3m

3543

Distance Offshore g h )

FIG. 5. Average of the bongo tow rnegalopae and zmae abundmces for the 1984 cruise versus (a) latitude and (b) distance offshore.

0

XI

1m

1%

100

2%

3m

Distance Offshore gkm)

FIG.4. Average of the corrected larval abundances versus distamce offshore for each of the cmises 1981-83 and 1985 and all four yeas together. (a) Megdogae; (b) zoeae-

settlement. In addition, metmophosis is most likely not constant with time, but occurs over a small time period so that much of the difference between the rate for zoeae and the rate of megalopae may be due to the fact that the 1984 cruise preceded the time of metamorphosis in the northern latitudes.

nmtly nearshore and to the north (Fig. 3a, 3b). The offshore rnegalopae were also concentrated in the north in 1981 but distributed over all latitudes in 1982. The QSAEs of megalopae in the 1981-83 and 1985 cruises showed no apparent regular pattern in the distribution of megalopae with latitude; however, cross-shelf distributions were more regular (Fig. 4a). The averaged distribution for all four years shows two peaks, at 110 and 200 km offshore, and very few Bungeness crab megalopae beyond 250 krn offshore. Abundance peaked at 75 and 200 h offshore in the 1983 data md both nearshore and at 200 h offshore in the 1985 data. The offshore peak in 1983 resulted from a single large sample in the south (Fig. 3c). Only a single peak nearshore, with very few megalopae beyond 150 krn offshore, occurred when samples were taken later in the year (1981, 1982). Zoeal QSAEs were highest at the extreme latitudes in the earlier cmises (1983, 1985) md at intermediate latitudes in the later cmises (1981, 1982). The cross-shelf distribution of zseae (Fig. 4b) was dominated by the single large peak in 1983 at 200 km offshore. In the remaining years the zoeae were fairly evenly distributed with lower abundances in the nearest onshore and stations furthest offshore. Significant zoeal abundances were found beyond 150 h in all years. Because the March-April cruise (1984) coincided with the timing of metamorphosis from zoeae to megalopae, the larvae were predominantly zoeae at the beginning of the cruise in the north md megalopae at the end of the cruise in the south (Fig. 5). Peak densities for megalopae and zoeae were less than 100 krn offshore in 1984 with few l m a e found beyond 150 km (Fig. 5).

Distribution of Larvae Although the QSAEs of megalopae were corrected for daynight differences (Fig. 3), the daytime smples had a greater variance and a higher probability of a zero estimate when larvae were present. To indicate differences in precision and the pmbability of false zeros, we differentiated between day and night smples in Pig. 3. The distributions of megalopae for the two cmises during late April and early May differed substantially from each other, 1985 had more megalopae in offshore samples though all latitudes than 1983 (Fig. 3c, 3d). The distribution for 1984 is skewed toward the north with a single large sample offshore in the south. The megalopae in the two later cmises were predomi-

Megalopal Abundance and Local Conditions There were no significant relationships between logtransformed QSAEs (natural log of (1 abundance)) and the local hydrographic conditions. Megalopae abundance appeared to decline at the extremes of both sea surface temperature and dissolved oxygen concentration; however, there was no apparent visual relationship between megalopae abundance and salinity. Chlorophyll abundance declined exponentially with distance offshore and did not depend on latitude or day of the year. There was no significant relationship between the QSAEs and chlorophyll abundance.

+

Can. J e Fish. Aqepdaf. Sci., Vol. 49, 199.2

Can. J. Fish. Aquat. Sci. Downloaded from www.nrcresearchpress.com by Shanghai International Studies University on 06/05/13 For personal use only.

FIG. 6 . Wind drift tracks for the 45 d prior to sampling in each of the later cmises ending at the sample p i n t at five latitudes. These were calculated fmm the FNOC wind field data as (1) port, (2) E h a n trmsport for 12 h per day and 3% of the wind velocity for the remaining 12 h, and (3) E h a n transport 12 h per day and 3% of the wind speed, 15" to the right of the wind direction for the remaining 12 h. (a) 1981; (b) 1982; (c) 1983; (d) 1985.

Wind Fields and Cross-Shelf L m a l Transport Wind-induced trmsport computed in the three different ways showed a consistent latitudinal pattern that varied in magnitude from year to year (Fig. 6). Typically, the onshore component increased with latitude, and the southward component decreased with latitude for methods 2 and 3. Ekman transport only (method 1) was generally southwaf.8 in the northern latitudes and offshore in the south. Both I981 and 1982 had almost no north-south transport in the north and transport of 3-4" of latitude in the south. 1983 and 1985 showed a pattern similar to later cruises except that the rapid southward transport in methods 2 and 3 was preceded by a period of onshore transport. The sample dates in these earlier cmises occurred about 3 wk prior to the same latitudes in the 1981 and 1982 cruises, so that an onshore phase at the same time of year would not have shown up in plots for later cruises. To evaluate the effect on transport of l m a e to nearshore habitat, we examined the onshore component of transport by each of the three methods for each year (Fig. 7). The latitudinal pattern of onshore transport remained the same from year to year (with the exception of the far north in 1983), but the magnitude varied. Onshore transport calculated by methods 2 and 3 was generally strong in the north and weak or reversed in the south. E h a n trmsport only (method I) was much smaller in magnitude and generally offshore in the south. In the MayJune cruises (1981, 19821, method 2 was the only one that resulted in net onshore transport at all latitudes. There is some correspondence between these patterns and the QSAEs (Fig. 3). For example, onshore transport in the north calculated by both the second and third methods in 1981 was more than twice the 1982 values (Fig. 6a, 6b). This difference was consistent with the greater offshore densities in the north Can. J . Fish. Aquat. Sci., V Q ~49, . 1992

in 1982 than in 1981 (Fig. 3a, 3b). 1985 had greater onshore transport (calculated by methods 2 and 3) at d l latitudes than 1983 (Fig. 7c, 7d). Although 198%had many more offshore stations with megalopae than 1983, the greatest abundance was found nearshore in 198%and offshore in 1983 (Fig. 3c, 3d and 4). To summaize the broadscale patterns, we compared the onshore transport calculated by method 2 averaged over the coast with the average quasi-synoptic densities of megalopae nearshore (