NOTE Hare and Govoni: Larval fish transport and vertical distributions on the southeast US continental shelf 



731 



Table 1 



Summary of the pairwise comparisons of larval depth distributions between species classified by the a priori outcome of trans- 

 port. In each table cell, the number to the left is the number of significant pairwise differences, the number to the right is the 

 total number of comparisons across the six station occupations, and the number in parentheses is the G-statistic for evaluating 

 the null hypothesis that the number of observed differences is as expected with a 5% error rate. The critical value at «=0.05 is 

 5.99 and significant values are indicated in bold. Values greater than 5.99 indicate that there are more significant differences 

 between species than expected by chance. Exported taxa are Bothus spp., Peprilus triacanthus, Syacium papillosum, Xyrichtys 

 novacula. Estuarine taxa include Leiostomus xanthurus, Micropogonias undulatus, and Paralichthys spp. Shelf resident taxa 

 include Etropus spp. and Etrumeus tei'es. 



A priori classification of the outcome of transport 



Brevoortia tyrannus 



Exported 



Estuarine 



Shelf resident 



Exported 

 Estuarine 

 Shelf resident 



5/ 17(10.29) 



12/15(55.35) 

 9/ 12(39.11) 



2/17(1.17) 

 12/43(23.88) 

 11/34(25.09) 



5/13(12.96) 

 11/30(27.96) 



2/6(4.39) 



and Mundy, 1994; Brodeur and Rugen, 1994). Variability 

 in larval fish vertical distributions (and zooplankton) is 

 related to processes that influence water column mixing 

 (e.g.. Heath et al„ 1988; Incze et al., 2001) and to spe- 

 cies-specific responses to diel cycles and gradients in 

 turbulence, temperature, and salinity (DeVries et al. 

 1995; Olla et al., 1996; Gray and Kingsford, 2003). The 

 approach used in the present study was to average over 

 shorter-scale variability (hours) in larval vertical dis- 

 tributions to examine longer-time-scale patterns (days) 

 in larval vertical distributions. 



Average larval vertical distributions of exported, 

 estuarine-dependent. and shelf-resident taxa and the 

 implied outcomes of their larval transport are consis- 

 tent with the results of physical oceanographic models 

 and observations of shelf circulation in the southeast 

 United States continental shelf. The model of Janowitz 

 and Pietrafesa (1980) (see also Miller et al., 1984) in- 

 dicated a three-layered, cross-shelf flow during winter: 

 surface and near-bottom offshore flow, and intermedi- 

 ate onshore flow. Similarly, the model of Werner et al. 

 (1999) indicated a two-layered, cross-shelf flow during 

 winter: an offshore flow near the surface and onshore 

 flow throughout the rest of the water column. Surface 

 flow in the study area during winter is typically off- 

 shore (Govoni and Pietrafesa, 1994). On the inner and 

 middle shelf (water depths <40 m), average bottom flow 

 is onshore; on the outer shelf (water depth 40-75 m), 

 average intermediate flow is onshore, whereas bottom 

 flow is offshore (Fig. 5b in Lee et al.. 1989). Modeled 

 and observed flow fields may indicate that larvae in the 

 surface water will move offshore (exported taxa), where 

 the probability of entrainment into the Gulf Stream is 

 higher. Larvae that are in the middle or lower portion 

 of the water column will move onshore (i.e., estuarine- 

 dependent and shelf-resident taxa). Thus, the average 

 larval vertical distributions, the general outcome of 

 larval transport, and the generalized observed and 

 modeled vertical flow fields are consistent. 



Differences between vertical distributions of larval 

 B. tyrannus and the other estuarine-dependent taxa 

 (Fig. 2; see also Govoni and Pietrafesa, 1994) imply 

 differences in cross-shelf transport. There are several 

 possibilities, none mutually exclusive. 1) Onshore trans- 

 port of larval B. tyrannus occurs with northeast wind 

 events and onshore transport of other estuarine-depen- 

 dent larvae occurs with southwest or northwest wind 

 events. This possibility is supported by the model simu- 

 lations of Hare et al. (1999). 2) Cross-shelf transport 

 of B. tyrannus larvae occurs in surface Gulf Stream 

 intrusions (Checkley et al., 1988; Stegmann and Yoder, 

 1996), whereas cross-shelf transport of other estuarine- 

 dependent larvae occurs by wind-driven mechanisms. 

 This possibility has not been adequately evaluated. 3) 

 All estuarine-dependent larvae are transported across 

 the shelf by the same mechanisms, but the rate of their 

 transport differs. For example, southwest wind events 

 cause onshore transport rates to be greater for the other 

 estuarine-dependent taxa because B. tyrannus larvae 

 spend less time in the intermediate portion of the wa- 

 ter column. This possibility is also supported by Hare 

 et al. (1999), who found that in modeled larval vertical 

 distributions, the outcome of larval transport was modi- 

 fied by circulation. From these alternative hypotheses, 

 it is clear that our understanding of the cross-shelf 

 transport of larval fishes remains incomplete and that 

 the effective physical and biological mechanisms are 

 complex. 



One approach to resolving the affect of vertical dis- 

 tribution on cross-shelf larval transport is to develop a 

 specific hypothesis regarding supply of larvae to inlets 

 that is based on the above possibilities and then to test 

 these hypotheses using the long time-series of larval 

 ingress collected at Beaufort Inlet (see Warlen, 1994). 

 Three alternative patterns in ingress, based on the 

 three possibilities presented above, could be evaluated 

 by using ingress data collected at Beaufort Inlet: 1) in- 

 gress of B. tyrannus occurs during northeast winds, and 



