66 



Fishery Bulletin 104(1) 



SK MG 



Northwestern stations 



WH PH 



Florida Key stations 



Figure 3 



Mean concentration of pink shrimp (Farfantepenaeus duorarum ) 

 postlarvae over a three-year period at each sampling station: 

 northwestern Florida Bay stations — Middle Ground (MG) and 

 Sandy Key (SK); Florida Key stations— Whale Harbor (WH) 

 and Panhandle (PH). 



Table 2 



Results of a nonparametric two-way ANOVA of the effect 

 of site and month on pink shrimp {Farfantepenaeus duo- 

 rarum) postlarvae entering Florida Bay. P<0.05 is indi- 

 cated with an asterisk. SS = sum of squares; MS=mean 

 of squares. 



Factor 



df SS 



MS 



Month 



Site 



Month X site 



Error 



80 



1 



1 



461 



756.7 



65.9 



1191.5 



772.2 



9.5 



65.9 



1191.5 



1.7 



5.6 0.007* 



39.3 0.002* 



711.3 0.007* 



Table 3 



Correlation coefficients of pink shrimp {Farfantepenaeus 

 duorarum) postlarval concentrations with sea surface 

 temperature (SST), cross-shelf wind (t/), and alongshore 

 wind (V) at the four sampling stations of MG= Middle 

 Ground, SK= Sandy Key, WH= Whale Harbor and PH = 

 Panhandle. Environmental data are from Long Key 

 CMAN station. Significant correlations (P<0.05 ) are indi- 

 cated with an asterisk. 



water displacement of -86.3 x 10^ m (Fig. 4C ). The 

 subtidal currents in this region are very weak as 

 also observed by other investigators (Koczy et al., 

 1960; Rehrer et al., 1967). Correlation coefficients 

 between winds and currents in the alongshore 

 direction iv) were significant for both onshore and 

 offshore currents time series (Table 4). Correlation 

 coefficients between winds and currents in the 

 cross-shelf direction were higher in the onshore 

 than in the offshore data series. This analysis 

 indicated that prevailing currents, overall, were 

 not favorable to passive transport of larvae from 

 the Tortugas to western Florida Bay. 



A harmonic analysis of the 3-year ADCP data 

 showed that semidiurnal tidal constituents (M.2, 

 S.,, K.,, and N„, see Table 5 for explanation of 

 abbreviations) are dominant on the SW Florida 

 shelf. M., was the strongest tidal constituent and 

 its east-west constituent explained 95% of the 

 total current variance. The east-west amplitude 

 of the M., constituent was 0.32 m/s for both moor- 

 ings, and the north-south was 0.07 m/s for moor- 

 ing A and 0.04 m/s for mooring B. The east-west 

 amplitude (0.32 m/s) was much stronger than the 

 long-term averaged subtidal cross-shelf constitu- 

 ent at both stations (0.001 m/s). The east-west 

 tidal excursion of the M, constituent was similar for 

 both moorings, but a few meters larger on the offshore 

 station (Table 5). This result indicates that it is reason- 

 able to consider that there are similar tidal excursions 

 on the SW shelf up to 50 km. 



Ebb versus flood catches 



Concentration of postlarvae collected hourly over a com- 

 plete dark portion of a tidal cycle showed clearly that 

 dark-ebb catches were negligible (<10%) by comparison 

 with dark-flood catches (>90'7f ) (Fig. 5, A-C). Hourly con- 

 centrations of postlarvae collected on the dark flood were 

 92.7% from 9 to 10 July, 90.2% from 23 to 24 July, and 

 86.9% from 8 to 9 August 2002. A nonparametric Krus- 

 kal-Wallis test showed significant differences in catches 

 of postlarvae between dark-ebb and dark-flood peri- 

 ods (Kruskal-Wallis; H=15.5, n = 36, P<0.001; (ANOVA: 

 7^=18.6; P<0.001). Samples taken during the day on 8 

 August confirmed the hypothesis that postlarvae are 

 present mostly at night in the water column (Table 6). 

 A total of 18 postlarvae (31.5 postlarvae/1000 m^) were 

 captured during 10 consecutive daylight hours, against 

 2657 (3702.5 postlarvae/1000 m^) captured during 10 

 consecutive hours of darkness. Although concentrations 

 of postlarvae tend to increase with the tidal current, 

 these differences were not significant (Kruskal-Wallis: 

 //=10.3, /! = 36, P=0.5). From 9 to 10 July, no postlar- 

 vae were captured during the hours of highest current 

 speed (2:00 to 3:00 h), which coincided with high rain 

 and winds (Fig. 5A). It is not clear whether these fac- 

 tors caused a change of behavior in the postlarvae or a 

 malfunction of the net. From 23 to 24 July, concentra- 

 tion of postlarvae on the dark flood followed the cur- 



