Cnales et al : Variability in supply and cross shelf transport of Farfanlepenaeus duororum postlarvae into Florida Bay 



63 



Postlarvae were collected monthly in each channel 

 during two nights around the new moon from January 

 2000 to December 2002. At the PH station sampling 

 began in June 2000 after the original site {Captain Key 

 channel) was abandoned because it had insufficient wa- 

 ter flow for effective sampling. Two moored subsurface 

 channel nets (net 1 and net 2) of 0.75-m- opening, 1-mm 

 mesh size, and SOO-nm mesh in the codend were sus- 

 pended with floats at 0.5 m depth. Nets were deployed 

 each night before dusk and removed shortly after dawn 

 each day. General Oceanic flowmeters (2030R16, Low 

 Speed Rotor, Miami, FL) were mounted in the mouth 

 of the nets and the volume of water filtered through 

 the nets was calculated for each net. Farfanlepenaeus 

 duorarum postlarvae were sorted, identified, and pre- 

 served in 90% ethanol. The raw catch in each sample 

 was standardized to numbers of postlarvae per 1000 m'^ 

 of water filtered. The average number of postlarvae over 

 the two sampling nights was used as the mean month- 

 ly concentration for each net. The average of monthly 

 postlarval concentration for each region (northwestern 

 Florida Bay vs. Florida Keys) was compared by using a 

 nonparametric two-way analysis of variance (ANOVA) 

 (Anderson, 2001). 



Three 12-hour experiments were conducted in sum- 

 mer 2002 in the SK channel to document the behavioral 

 response of pink shrimp postlarvae to ebb and flood 

 tides. Consecutive pairs of night (i.e. dark) flood and 

 dark-ebb plankton samples were taken hourly from 

 19:00 to 07:00 h from 9 to 10 July (new moon), 23 to 

 24 July (full moon), and 8 to 9 August (new moon). In 

 addition, plankton samples were taken for 10 consecu- 

 tive hours daily on 8 August to verify the response of 

 postlarvae to light. The nonparametric Kruskal-Wallis 

 test and analysis of variance (ANOVA) were used to 

 determine differences in concentration of postlarvae 

 between dark-ebb and dark-flood periods. 



To evaluate the possible effect of environmental vari- 

 ables on larval supply to Florida Bay and the pattern 

 of seasonality in postlarval concentrations, available 

 time series of winds and sea surface temperature (SST) 

 on the coastal shelf were examined in relation to our 

 time series of monthly measured larval concentrations. 

 In particular we were looking for a pattern that might 

 help to determine the reason for the marked summer 

 peak in the concentration of postlarvae at the western 

 border of Florida Bay. Time series of hourly winds and 

 sea surface temperature (SST) for the 3-year sampling 

 period were obtained from the Coastal Marine Auto- 

 mated Network (CMAN) station at Long Key, and from 

 the Coastal Ocean Monitoring and Prediction System 

 (COMP) station in NW Florida Bay (Fig. 1). Tempera- 

 ture and wind time series from the Long Key CMAN 

 station and the NW Florida Bay COMP station were 

 highly correlated with each other (7-2 = 0.9; P<0.01). The 

 longer Long Key time series were used for coastal SST 

 and wind analysis. Wind speed and direction over the 

 Keys, as measured at CMAN sites, are highly coherent 

 (Peng et al., 1999) and useful for explaining currents 

 on the SW shelf (Lee and Williams 1999). Monthly av- 



Table 1 



Mean cross section depth, peak tidal flow, and cross sec- 

 tion area of the four channels sampled for pink shrimp 

 iFarfantepenaeus duorarum) postlarvae at the north- 

 western border of Florida Bay, Middle Ground (MG) and 

 Sandy Key (SKl, and at the southeastern edge in the 

 Middle Florida Keys, Whale Harbor (WH), and Panhan- 

 dle Key (PH). Area was calculated at zero (m) of mean 

 sea level. 



Station 



MG 

 SK 

 WH 

 PH 



erages of wind vectors and SST were calculated from 

 hourly CMAN time series data (years 2000 to 2002) to 

 examine the effect of winds and SST on the monthly 

 postlarval collections. 



Time series of current data from two established sta- 

 tions with moored ADCPs and temperature and salinity 

 sensors were used to drive our transport model. Ini- 

 tially, these data were examined to determine whether 

 prevailing currents alone could explain the transport of 

 larvae from the Tortugas spawning grounds to Florida 

 Bay nursery grounds. These stations also were a source 

 of salinity data used in one set of simulations. These 

 stations were located on the inner SW shelf of the Gulf 

 of Mexico, about 30 km from our MG station (Lee et al., 

 2001) (Fig. 1). This array monitored coastal currents as 

 part of the Florida Bay Circulation and Exchange Study 

 (Lee et al., 2001). The ADCP moorings were located at 

 depths of 6.4 m (mooring A=onshore) and 11.6 m (moor- 

 ing B=offshore) and recorded data every 30 minutes for 

 a 3-year period (A=21 September 1997 to 15 October 

 2000; B=22 September 1997 to 17 October 2000). The 

 two ADCP moorirr^s were about 30 km apart. Lee et 

 al. (2001) reported insignificant differences between 

 currents in the vertical for cross-shelf transport in the 

 shallow SW Florida shelf. Wind and current vectors 

 were resolved into cross-shelf (u=east [+] and west [-]) 

 and alongshore ((;=north [-I-] and south [-] constituents). 

 The east-west and north-south displacement of current 

 and wind constituents (half-hour current data and the 

 hourly wind data) was the product of each current and 

 wind constituent by the respective time interval. Cor- 

 relation analysis was conducted on currents and wind 

 time series. 



A harmonic analysis was conducted on the three-year 

 ADCP raw data to define tidal constituents and current 

 magnitude. Period (Pi) and tidal excursion (Ti) were 

 calculated for each constituent from amplitudes (Ai) and 

 frequencies of the constituents as follows: 



Ti = AiPi/n. 



