Govoni et al.: Distribution of larval Xiphias gladius off the southeastern United States 



65 



spawning may occur as far north as Cape Hatteras. 

 Planktonic eggs have not been identified in the west- 

 em North Atlantic. 



Effective fisheries management requires knowledge 

 of both the spatial distribution of larvae and the dis- 

 tribution of spawning adults. Contemporary stock sta- 

 tus ascertained by virtual population analysis can be 

 calibrated with larval abundance estimates, even for 

 species with rare larvae (e.g. McGowan and Richards, 

 1989; Scott et al., 1993). Because this calibration de- 

 pends on accurate estimates of larval abundance, the 

 spatial and temporal distribution of larvae must be 

 known. Knowledge of spawning distribution would be 

 the first step toward protection of spawning habitat 

 and, perhaps, the restriction of fishing within spawn- 

 ing seasons and locales. 



Here we examine the coarse- and fine-scale distri- 

 bution, and the lengths of swordfish larvae off the 

 southeastern United States. Our focus is on the influ- 

 ence of the Gulf Stream in shaping the spatial distri- 

 bution of larvae and in the determination of probable 

 spawning locales. 



Methods 



We examined published records of larval swordfish 

 (Arata, 1954; Arnold, 1955; Tibbo and Lauzier, 1969; 

 Markle, 1974; Post et al, 1997), data from the Ma- 

 rine Resources Monitoring, Assessment and Predic- 

 tion Program (MARMAP), data and specimens from 

 the Southeast Area Monitoring and Assessment Pro- 

 gram (SEAMAP), and new data from three surveys 

 conducted between the Florida Straits and Cape Hat- 

 teras in 1984, 1988, and 1997 (CF8406, CH8807, and 

 CH9703J. 



Spatial distribution of species 



For spatial distribution, we examined exclusively 

 neuston collections (the upper 0.5 m of water), be- 

 cause swordfish are surface-oriented as are larvae. Al- 

 though some swordfish larvae have been collected in 

 nets that fished principally below the surface (Grail 

 et al., 1983), most have been collected at the surface 

 (Taning, 1955; Yabe et al., 1959; Gorbunova, 1969; Ni- 

 shikawa and Ueyanagi, 1974). All swordfish larvae 

 collected by MARMAP, SEAMAP, and CH8807 were 

 collected in the neuston, none in accompanying, sub- 

 surface ichthyoplankton collections. Small larvae were 

 occasionally taken from below the surface in CH9703, 

 but these nets fished obliquely fi-om 20 m and larvae 

 were likely captured when nets were near the surface. 

 Collections of larvae were classified to water mass 

 (or type) — shelf water (including Georgia water, Car- 



olina Capes water, and occasionally Virginia coastal 

 water [Pietrafesa, 1989]) or Gulf stream water — by 

 applying measured hydrographic characteristics to 

 the classifications of Matthews and Pashuk ( 1986) for 

 MARMAP and CF8406 collections, and Pietrafesa et 

 al. (1985) for CH8807 and CH9703 collections. Fron- 

 tal zone water is a mixture of these water masses 

 (Hitchcock et al., 1994). South of Cape Hatteras the 

 Gulf Stream courses north-northeastward in juxta- 

 position with shelf water. Classically, the definition 

 of the Gulf Stream front is a dynamic one: the Gulf 

 Stream front is the point where the pressure gradient 

 between Sargasso Sea water and slope water (north 

 of Cape Hatteras) or shelf water (south of this Cape) 

 is zero ( Stommel, 1966). Practically, observed horizon- 

 tally compressed surface isotherms and isohalines, ac- 

 companied by sharp discontinuities in sea-surface tex- 

 ture and color, define the Gulf Stream frontal zone 

 (Olson et al., 1994). In our study, we used this obser- 

 vational definition to assign the surface position and 

 to classify the water of the Gulf Stream frontal zone. 

 Surface Gulf Stream water at its western fi-ont be- 

 tween the Florida Straits and Cape Hatteras has char- 

 acteristic temperatures that range from 2^ to 24°C 

 in wdnter and from 27° to 29°C in summer, salinities 

 that range from 35.7 to 36.4 psu and vary little sea- 

 sonally, dissolved oxygen values that range from 4.5 to 

 5.0 mL/L, and nitrate values of 1.0 pM/L (Atkinson, 

 1985; Schmitz et al., 1993; Hitchcock et al., 1994; Xie 

 and Pietrafesa, 1995). Shelf water is cooler and less 

 sahne (Pietrafesa et al., 1985). The course of the Gulf 

 Stream along the continental shelf break is unstable; 

 it meanders onshore and offshore and projects intru- 

 sions, filaments, and eddies onto the shelf (Pietrafesa, 

 1989; Lee et al., 1991 ). These processes complicate the po- 

 sition and distort the configuration of the frontal zone. 



Retrospective examination of 1163 collections taken 

 from the Straits of Florida to Cape Hatteras and off- 

 shore from the 9 m to 2000 m isobath (Fig.l) from 

 1973 to 1980 (MARMAP) afforded the determination 

 of coarse-scale distribution. These collections were 

 taken with neuston nets of two different dimensions 

 and meshes: a 1.0 x 0.5 m net with 505-pm mesh and 

 a 2.0 X 1.0 m net with 947-pm mesh. Both meshes col- 

 lect swordfish larvae, which are reported to be at least 

 4 mm at hatching (Sanzo, 1910; Yasuda et al., 1978). 

 Neuston nets were towed for 10 min at 5.6 km/h. Col- 

 lections were made in all four seasons and at all times 

 of day and night. Larvae were preserved in 5% forma- 

 lin solution. The probable location of the Gulf Stream 

 front for these collections was determined from 1) 

 advanced, very high resolution, infra-red radiometer 

 (AVHRR) satellite images of sea-surface temperature 

 (SST) taken fi-om 1976 to 1980; 2) expendable bathy- 

 thermograph (XBT) profiles taken from 1973 to 1980; 



