xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001711502
007 cr mnu|||uuuuu
008 060530s2005 flua sbm s000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001318
Walters, Sarah Lyle.
Mapping Tampa Bay Cynoscion nebulosus spawning habitat using passive acoustic surveys
h [electronic resource] /
by Sarah Walters.
[Tampa, Fla.] :
b University of South Florida,
Thesis (M.S.)--University of South Florida, 2005.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 75 pages.
ABSTRACT: Spotted seatrout, Cynoscion nebulosus, spawning locations as well as associated environmental variables were determined for Tampa Bay, Florida during the 2004 spawning season using a mobile hydrophone survey. Hydrophones, a type of underwater microphone, can be used to detect and record spawning sounds of soniferous fishes. During their spawning season in Tampa Bay which generally occurs between March and September, mature male spotted seatrout generate sounds associated with courtship in the crepuscular and evening periods by vibrating sonic muscles against the swim bladder. Active spawning sites can be located using hydrophones to find these calling males. Using a random stratified sampling method, 760 stations within Tampa Bay (46 % of the sampling universe) were sampled over the 2004 spawning season. Only 8% of sampled stations had large aggregations of spotted seatrout.Spawning, determined by the sound produced by large aggregations, was detected throughout the bay except for Hillsborough Bay and was most common in the lower bay and eastern region of the middle bay. Presence of submerged aquatic vegetation (SAV), proximity to shoreline, as well as high dissolved oxygen values and shallow depth were positively correlated with spawning areas. Courtship calls of sand seatrout, Cynoscion arenarius, and silver perch, Bairdiella chrysoura were also detected during the survey as they share an overlapping spawning season with spotted seatrout. Aggregations of all three species rarely occurred simultaneously. Sand seatrout and silver perch used different habitats within Tampa Bay to spawn and spawned with a much greater frequency than spotted seatrout. Courtship calls of spotted seatrout were analyzed both by ear and by received sound level to determine if signal processing could be used to assess courtship sound recordings.However, there was no clear relationship between the two methods.
Adviser: Dr. David Mann.
x Marine Science
t USF Electronic Theses and Dissertations.
Mapping Tampa Bay Cynoscion nebulosus Spawning Habitat Using Passive Acoustic Surveys by Sarah Lyle Walters A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science College of Marine Science University of South Florida Major Professor: David A. Mann, Ph.D. Susan Lowerre-Barbieri, Ph.D. Joseph J. Torres, Ph.D. Date of Approval: October 19, 2005 Keywords: spotted seatrout, mobile hydrophone courtship sounds, sciaenid reproduction, estuary Copyright 2005, Sarah Lyle Walters
ACKNOWLEDGEMENTS I would like to thank my committee, Dr David Mann, Dr. Sue Lowerre-Barbieri, and Dr. Jose Torres, for all of their s upport and guidance throughout this process. I especially would like to ac knowledge Dr. Sue Lowerre-Barbi eri for her assistance in helping balance my commitments towards this degree with my obligations for work as well as all her advice in both science and life in general. Data collections as well as mapping were a direct result of Joel Bickfo rdÂ’s dedication (and enthusiasm!) for this work. IÂ’d also like to thank Janet Tunnell for a ll of her efforts collect ing data as well as analyzing sounds in the laboratory. The Fi sheries Independent Monitoring Program (FIM) at the Florida Wildlife Research Inst itute provided the expertise and software knowledge responsible for the SAS sampling selection program. Many thanks to my USF lab mates, Jim Locascio, Brandon Casper Mandy Hill-Cook, Randy Hill, and Bryan Nichols, as their humor and a ssistance have facilitated as well as brought joy to this process. And, most importantly, IÂ’d like to thank my family fo r all their love and support and for raising me to value a nd respect nature. Evidently, my simple love of the outdoors evolved into a professional di rection, one which allows (and even encourages) me to get wet, go fishing, and truly appreciate the intricacies of GodÂ’s great world.
NOTE TO READER Note to Reader: The original of this docum ent contains color that is necessary for understanding the data. The orig inal thesis is on file with the USF library in Tampa, Florida.
i TABLE OF CONTENTS LIST OF TABLES ii LIST OF FIGURES iii ABSTRACT vi INTRODUCTION 1 METHODS 3 Sampling Universe 3 Sampling Periodicity 8 Data Collection 10 Statistical Analysis Acoustic Analysis 14 15 RESULTS Seasonality Data Collected Geographic distribution of spotted seatrout courtship sounds Environmenta l variables associated with s potted seatrout courtship sounds Comparison of spotted seatro ut spawning locations with other sciaenids Analysis of spotted seatrout courtship sounds 17 17 20 25 32 37 48 DISCUSSION 52 Geographic distribution and environmental variables associated with spotted seatrout spawning 52 Comparison of spawning locations between spotted seatrout and other sciaenids 57 Methodology review Analysis of spotted seatrout courtship sounds 58 61 LITERATURE CITED 62
ii LIST OF TABLES Table 1 Number of grids in each region characterized by grid type. 6 Table 2 Number of total grids and stations sampled with the percent of grids sampled in each region. 21 Table 3 Number of grids repeat ed and percentage of repeated grids displayi ng differences in spotted seatrout detections by region. 22 Table 4 Repeated grids disp laying differences in spotted seatrout detections between sampling dates. 23 Table 5 Spotted seatrout detec tions by category and distance. Table 6 Substrata at all samp led stations and stations with large sp otted seatrout aggregations categorized as Â“close-byÂ”. Table 7 Sand seatrout detect ions by category and distance. Table 8 Substrata at all sampled stations and stations with sand seatrout aggregations and percent of each substratum used by aggregation stations. Table 9 Environmental variables at stations where all three sciaenid sp ecies (spotted seatrout, sand seatrout, silver perch) were detected at a ggregation levels simultaneously. Table 10 Silver perch detecti ons by category and distance. Table 11 Substrata at all sample d stations and stations with silver perch aggregations and percent of each substratum used by aggregation stations. Table 12 Stations where all thr ee sciaenid species (spotted seatrout, sand seatrout, silver perch) were detected at aggregation levels simultaneously. 27 31 37 41 42 43 47 48
iii LIST OF FIGURES Figure 1 Tampa Bay Sampling Universe 4 Figure 2 Sampling station distributi on when stations determined by availa ble substrata (A) and by area (B). 7 Figure 3 Bunces Pass, Florida. Kn own spotted seatrout aggregation spawning site a nd location of the long-t erm acoustic recording system (LARS). 9 Figure 4 Spectrograph of (A) an i ndividual spotted seat rout call composed of th ree sets of multiple-pulses followed by a long grunt and (B ) a large spotted seatrout aggregation 12 Figure 5 Spectrogram of an individual sand seatrout. Figure 6 Spectrogram of an individual silver perch. 13 14 Figure 7 Temperature (C) at all stations sampled in the 2004 Tampa Bay hydrophone survey during April and May. 18 Figure 8 2004 Bunces Pass spotted s eatrout spawning season start with daily duration, sunset time, and daily average water temperature. 19 Figure 9 2004 Bunces Pass spotted s eatrout end of season daily duration with associated sunset time and temperature data. 20 Figure 10 2004 Tampa Bay hydrophone survey sampled grids. 24 Figure 11 2004 Tampa Bay hydrophone survey spotted seatrout detections. 26 Figure 12 2004 Tampa Bay hydrophone su rvey spotted seatrout large aggregation stations. Figure 13 The number of stations within a region a spotte d seatrout large aggregation wa s detected (black) and the total number of stations sampled within that region (gray). Figure 14 Large aggregation percent by region compared to the expected percen t (7.7%) of large aggr egation detections. 28 29 30
iv Figure 15 Mean depth (m) of stations without large aggregations versus large aggreg ation stations, +/one standard deviation. Figure 16 Mean temperature (C) of st ations without large aggregations versus large aggregation stations, +/one standard deviation. Figure 17 Mean salinity (ppt) of st ations without large aggregations versus large aggregation stations, +/one standard deviation. Figure 18 Mean dissolved oxygen (mg/ L) of stations without large aggregations vers us large aggregation stat ions, +/one standard deviation. Figure 19 2004 Tampa Bay hydrophone su rvey sand seatrout detections. Figure 20 2004 Tampa Bay hydrophone surv ey spotted seatrout and sand seatrout aggregations. Figure 21 The number of stations wi thin a region a sand seatrout aggregation wa s detected (black) and the total number of stations sampled within that region (gray). Figure 22 Mean depth (m) of spo tted seatrout and sand seatrout aggregat ion stations, +/one standard deviation. Figure 23 Tampa Bay hydrophone surv ey silver perch detections. Figure 24 2004 Tampa Bay hydrophone su rvey silver perch, spotted seatrout, and sand seatrout aggr egation detections. Figure 25 The number of stations wi thin a region a silver perch aggregation wa s detected (black) and the total number of stati ons sampled within that region (gray). Figure 26 Decibel level within 200-300 Hz frequency range corresponding to number of spotted seatrout categorized by ear. All of th e plotted recordings do no t exclusively contain spotted seatr out calls and other sciaenid species may be present. Figure 27 Decibel level within 200-300 Hz frequency range corresponding to number of spotted seatrout categorized by ear. All of th e plotted recordings are exclusively spotted seatrout calls. 32 34 35 36 38 39 40 42 44 45 46 49 50
v Figure 28 Decibel level within 200-300 Hz frequency range corresponding to number of spotted seatrout categorized by ear. All of th e plotted recordings are exclusively spotted seatrout calls detected close-by. 51
vi MAPPING TAMPA BAY CYNOSCION NEBULOSUS SPAWNING HABITAT USING PASSIVE ACOUSTIC SURVEYS Sarah Lyle Walters ABSTRACT Spotted seatrout, Cynoscion nebulosus, spawning locations as well as associated environmental variables were determin ed for Tampa Bay, Florida during the 2004 spawning season using a mobile hydrophone su rvey. Hydrophones, a type of underwater microphone, can be used to detect and reco rd spawning sounds of soniferous fishes. During their spawning season in Tampa Bay wh ich generally occurs between March and September, mature male spotted seatrout gene rate sounds associated with courtship in the crepuscular and evening periods by vibrati ng sonic muscles against the swim bladder. Active spawning sites can be located using hydrophones to find these calling males. Using a random stratified sampling me thod, 760 stations within Tampa Bay (46 % of the sampling universe) were sample d over the 2004 spawning season. Only 8% of sampled stations had large aggregations of spotted seatrout. Spawning, determined by the sound produced by large aggreg ations, was detected th roughout the bay except for Hillsborough Bay and was most common in the lower bay and eastern region of the middle bay. Presence of submerged aquatic vege tation (SAV), proximity to shoreline, as well as high dissolved oxygen values and sh allow depth were positively correlated with spawning areas. Courtship calls of sand seatrout, Cynoscion arenarius and silver perch, Bairdiella chrysoura were also detected during the surv ey as they share an overlapping spawning season with spotted s eatrout. Aggregations of all three species rarely occurred simultaneously. Sand seatrout and silver perch used different habitats within Tampa Bay
vii to spawn and spawned with a much greater frequency than spotte d seatrout. Courtship calls of spotted seatrout were analyzed both by ear and by received sound level to determine if signal processing could be us ed to assess courtship sound recordings. However, there was no clear relatio nship between the two methods.
1 INTRODUCTION Spotted seatrout are estu arine-dependent batch spawners, but their preferred spawning habitat, as well as environmenta l parameters corresponding to spawning site selection, has yet to be fully determined (Brown-Peterson et al., 2002). Mature male spotted seatrout generate courtship sounds associated with spawning. These sounds are made in the crepuscular and evening periods by vibrating sonic muscles against the swim bladder (Tavolga, 1969, Fish and Mowbray, 1970). Active spawning sites can be located using underwater microphones, or hydrophones, to find these calling males (Mok and Gilmore, 1983; Saucier and Baltz, 1992; Sauc ier and Baltz, 1993; Luczkovich et al., 1999). Passive acoustic studies in coastal Nort h Carolina, South Carolina, Louisiana, and the Florida east coast have examined spotte d seatrout spawning sites using a non-random approach. Results from these studies, as we ll as traditional reproduc tive biology studies, indicated spotted seatro ut use a wide range of habitats during courtship and reproduction, including bays, lagoons, channels, deep passes adjacent to open water, deep channels adjacent to vegetated shallow areas, and s eagrass in non-tidal areas (Hein and Shepard, 1979; Brown-Peterson et al., 1988; Sauc ier and Baltz, 1993; Gilmore, 2003). Previous work on Tampa Bay spotted seatrout reproduction has focused on spawning activity in the lower and middle portions of the bay. Using data on the distribution and average size of larval spotted seatrout, McMichael and Peters (1989) concluded that spawning occurred in middle and lower Tampa Bay and also in nearshore Gulf waters. Actively spawning spotted seat rout targeted and collected by LowerreBarbieri (2004) were captured primarily in lower Tampa Bay but location of specific
2 spawning aggregations was difficult. Captur e efforts were focused on lower Tampa Bay with some capture in the southern porti on of the middle bay, but the middle and upper Bay were not sampled. McMichael and Peters (1989) also did not include the upper bay for larval collection. Further research inve stigating spotted seatrout spawning site locations is necessary in order to verify spawning sites in the lower bay, determine the extent of spawning in the middle bay, and es tablish if the upper ba y is being used for spawning. Passive acoustics is an ideal tool fo r characterization of spawning habitat in a large area such as Tampa Bay. While traditiona l collection gear typically limits the time and geographic area sampled in a stu dy, passive acoustic methodology permits comprehensive coverage in a fraction of th e time. Additionally, as a noninvasive tool, passive acoustics does not inte rrupt spawning behavior while data are being collected. The objectives of this study were to lo cate spotted seatrout spawning sites as identified by aggregation sounds, identify phys ical and chemical variables associated with these spawning sites, and determine if significant spawning activity differences exist amongst geographic regions within Tampa Bay.
3 METHODS A stratified random sampling design usi ng a mobile hydrophone was developed and tested the last half of the 2003 Tamp a Bay spotted seatrout spawning season and found to be successful. Data from this surv ey was used in constructing the sampling window for the 2004 survey (see Sampling Periodicity in Methods) as well as to determine the amount of sampling possible pe r evening in 2004 (see Sampling Universe in Methods). A few minor altera tions were made to the prot ocol and the design was used for the 2004-spawning season as detailed below. Sampling Universe Tampa Bay, Florida was divided into four zones based on geographic and logistical criteria (Hillsborough Bay, Uppe r Bay, Middle Bay, and Lower Bay). All zones, except Hillsborough Bay were subdivide d into east and west regions (Figure 1): upper west (Region A), upper east (Region B), middle west (Region C), middle east (Region D), lower west (Region E), lo wer east (Region F). Hillsborough Bay was considered both a region (Region G) and a zone. Each region was a stratum, with sampling units composed of 1-nm grids. Grid s with a high percentage of land or very shallow water (95% or more of the area comp rised of land or water < 1.5 m in depth) were excluded from sampling. Grids were categorized as either Â“open waterÂ” or Â“shorelineÂ”. Shoreline grids were those that had more than 5% of their area c onsisting of either land or water 1.5 m or less adjacent to land. An open water grid had 95% of its area covered by water deeper than 1.5 m and it was not adjacent to land.
4 Figure 1 Tampa Bay Sampling Universe
5 Sampling was based on the lunar calenda r, as spotted seatrout spawning frequencies have been attributed to luna r influences (McMichael and Peters, 1989; Gilmore, 2003; Kupschus, 2004). Two nights pe r week were selected for sampling, both nights falling within two days of the quarter full, three-quarter, or new moon. One region was sampled per evening. One zone was samp led per week. Zones were rotated monthly so that each zone was sampled over the various possible lunar phases during the course of the spawning season. Grids were randomly sele cted with replacement in order to account for seasonal variability. During preliminary testing in 2003, it was found that six grids per evening was the maximum number of grid s that could be sampled within the sound sampling window (see Sampling Periodicity in Methods). Regions also varied in the number of Â“s horelineÂ” and Â“open waterÂ” grids sampled per evening. It was necessary to sample the shoreline/open water grids per region proportionally to what was present in each region. This was necessary because spotted seatrout are reported to spawn in shoreline areas, and certain regions contained more of this habitat than others. Sampling was also proportional for the number of grids per region (Table 1). Four stations per grid were sampled to en sure representative sampling of the area within each grid. Stations were as evenly distributed as possible over the grid area as well as the available substrata (Figure 2A). Ta rgeted substrata for both shoreline and open water grids included submerged aquatic vege tation (SAV), structure, channel, and nonchannel. The SAV category included areas either directly on top of seagrass or adjacent to the flats containing seagrass. Structure en compassed any artificial construction in the water including pilings, jetties, bridges, arti ficial reefs/fish have ns, and range markers.
6 Table 1 Number of grids in each region characteri zed by grid type. Total number of grids in a particular region is listed with th e number of grids samp led per evening in parenthesis. Region Number of Open Water Grids Number of Shoreline Grids Total A 9 (1) 31 (4) 40 (5) B 20 (2) 20 (3) 40 (5) C 38 (4) 13 (2) 51 (6) D 29 (3) 22 (3) 51 (6) E 19 (2) 32 (4) 51 (6) F 29 (3) 27 (3) 56 (6) G 10 (1) 29 (4) 39 (5) Total 154 174 328 The channel substratum was defined as a cl early navigable deeper passage of water surrounded by shallower water. The non-cha nnel substratum was a depth independent describer of areas that were not channels and did not have st ructure or SAV. If the four different substrata were not present within a grid, then the four sampling stations were selected based on differences in depth. If de pth was constant throughout the grid, then the four sample stations were di stributed equally based on area of the grid (Figure 2B).
7 Figure 2 Sampling station distribution when stati ons determined by available substrata (A) and by area (B). A B
8 Sampling Periodicity Hydrophone sampling began on April 5, 2004 and continued through the first week of October 2004. April was selected as the start month because Tampa Bay spotted seatrout males and females typically begi n spawning by mid to late March (LowerreBarbieri, 2004) or April (McMichael and Pe ters, 1989). Similarly, October was chosen as the end month as spawning has been reporte d to cease in mid-September (LowerreBarbieri, 2004) or October (McMichael and Peters, 1989). Because of the diel periodicity associated with spotted seatro ut courtship sounds, sampling was conducted during the window of maximum sound production. This window was based on both the 2003 preliminary hydrop hone survey data, indicating aggregations were detected between sunset and 02:30 and previous research repor ting spotted seatrout spawning aggregation sounds from 17:00-01:00 (Saucier and Baltz, 1993), and sunset to 24:00 (Gilmore, 1994). Based on these sources of information, the sampling window for the 2004 hydrophone survey was set to begin at sunset (roughly 20:00 EDT) and continue for five hours (until roughly 01:00 EDT). Seasonal start and stop dates for the 2004spawning season were confirmed based on data from a known spawning site. Lowe rre-Barbieri (2004) reported a very high percentage of the females collected at Bun ces Pass (Figure 3) we re actively spawning (100% in 2001 and 96.5 % in 2002). A long-te rm acoustic recording system (LARS) was deployed in Bunces Pass for the 2004-spaw ning season. Anchored 0.5 m off the bottom at the mouth of the pass, the LARS was progr ammed to sample ten continuous seconds of sound every ten minutes at a 2634 Hz sampli ng rate and record to onboard Compact Flash memory. Data from the LARS were an alyzed both by ear and spectrographically in
9 Figure 3 Bunces Pass, Florida. Known spotted s eatrout aggregation spawning site and location of the long-term acous tic recording system (LARS). Cool Edit. Temperature data recorded 40 km nor th from the LARS during the 2004 spotted seatrout spawning season was obtai ned from NOAA. Although another agency (consulting firm Delta Seven) collected temp erature data 1.0 km from the LARS during the same time period, this data set was inco mplete. However, as the daily temperature averages did not significantly differ between these two sites during the first and last
10 months of the spawning season (Mann-Whitney Rank Sum Test, n=98, p=0.201), the NOAA temperature data set was ap plied to the LARS data. Data Collection Hydrophone recordings and environmental data were taken at all stations. Once at a sampling station within a grid the engine was turned off, GPS and depth measurements were recorded, and a mobile hydrophone (HTI, model 96-mi n, sensitivity Â–164 dBV/Pa) was lowered one meter in the water. Recordi ngs were made after a two-minute period in the event the spotted seatrout calling ceased because of engine noise disturbance. During the two-minute waiting period, a YSI Model 600 QS was lowered into the water middepth on the opposite side of the boat fr om the hydrophone to measure salinity, temperature, and dissolved oxygen. Mid-depth measurements were taken, as Tampa Bay is a well-mixed estuary with little diffe rence in bottom and surface temperature and salinity (Goodwin, 1989). Substrata type (described in the previous section to include submerged aquatic vegetation (SAV), structur e, channel, and non-channel) and times were also recorded. If sciaenid courtship calls were heard, then a recording was made. All recordings were made for a thirty second period in Â“A -timeÂ” on the Sony digital audio tape (dat) recorder model TCD-D8. Recordings on the dat recorder used the Â“line-inÂ” jack with all recordings on level 10 and a sample rate of 44.1 kHz. Record mode was set to Â“manualÂ” and microphone sensitivity was set on Â“lowÂ”. A miniature Marshall Amplifier with the tone set in the middle and the volume on 10 wa s used to listen to all sounds prior to recording. Headphones were worn if a sound was difficult to detect through the amplifier
11 or if there was too much background noise. Tape label name, program number, tape start and end time, and comments accompanied each recording. Spotted seatrout males produce distinct courtship calls. These calls have been classified into four major sound types: dualpulse calls, a long grunt call, multiple-pulse calls, and a stacatto call (Mok and Gilmore 1983, Gilmore, 2003). Their calls can be distinguished from other soniferous fishes based on pulse duration and intensity of sound by frequency range (Figure 4). Estimated nu mber of spotted seatrout producing sound were categorized as (1) 1-2 individuals, (2 ) 3-5 individuals, (3) small aggregation with individuals still dis tinguishable, or (4) large aggregation. Distance to the fish was categorized as: Â“directly on-top ofÂ”, Â“close-byÂ” or Â“in the distanceÂ”. The directly on-top of category was defined as fish sounds audi ble through the bottom of the boat without the aid of the hydrophone. Sounds categorized as Â“i n the distanceÂ” were quiet, and difficult to hear through the amplifier. Â“Close-byÂ” incl uded a wide range of sounds falling between Â“directly on-top ofÂ” and Â“in the distanceÂ” categories. Courtship calls of two other sc iaenid species, sand seatrout, Cynoscion arenarius and silver perch, Bairdiella chrysoura were regularly heard in Tampa Bay. Males in both of these species also make courtship calls a ssociated with spawning. However, their calls are easily distinguished from spotted seatro ut (Figures 5 and 6) Sand seatrout calls resemble a Â“purringÂ” and silver perch have a distinctive high-pitched Â“knockÂ” (Mok and Gilmore, 1983, Locascio and Mann, 2005, Joel Bickford, pers. comm.). Although sand seatrout and silver perch sh are overlapping spawning seasons with spotted seatrout and apparently similar windows of maximum s ound production, there may be species-specific variability that was not accounted for in this sampling design. Calls made by sand
12 Figure 4 Spectrograph of (A) an individual spotted seatrout call composed of three sets of multiple-pulses followed by a long grunt and (B) a large aggregation. Darker shading corresponds to highe r decibel levels. A B
13 Figure 5 Spectrograph of an individual sand seatrout. Darker shading corresponds to higher decibel levels.
14 Figure 6 Spectrograph of an individual silver perch. Darker shading corresponds to higher decibel levels. seatrout and silver perch we re noted on the datasheet a nd the number of fish was estimated as (1) 1-2 individuals (2) 3-5 individuals, or (3) aggregation. Distance from the hydrophone for these two species was categorized identically for spotted seatrout as either Â“directly on-top ofÂ”, Â“close-byÂ”, or Â“i n the distanceÂ”. All recordings made in the field were reviewed in the lab by ear and verified with known recorded fish sounds. Statistical Analysis A program written in Statistical Analysis System (SAS) was used to randomly select sampled grids. Arc View GIS 3.3 was used to map the location of courtship sounds. Differences in water temperature between April and May, differences in temperature, salinity, and di ssolved oxygen between stations with and without large
15 spotted seatrout aggregations, and differen ces in depth between shoreline and offshorecategorized grids were examined usi ng the Mann-Whitney Rank Sum test. The test with Yates correction for continuity was applie d to test differences between expected and actual aggregation presence of all three sci aenids amongst all regions, by grid type, and by substrata. While examinat ion of sciaenid aggregati ons by region and grid type included all aggregations rega rdless of distance, associa tions with bottom type and environmental parameters were examined at stations categorized as Â“directly on-top ofÂ” or Â“close-byÂ”. Aggregation stations classified as Â“in the distanceÂ” were not used, as those aggregations may not have been in proximity to the in situ measurement locations. The analyses were based on the assumption that the ratio of the number of large spotted seatrout aggregations (and aggregations fo r sand seatrout and silver perch) heard throughout Tampa Bay divided by the number of stations sampled was the expectation if no significant regional, grid type or substrata effects existed. This ratio was then used to determine the number of expected large aggreg ations in each of the categories (by region, grid type, and substrata) and those numbers were then compared to the number of large aggregation stations that act ually occurred. When spotted s eatrout aggregations were compared to sand seatrout and silver pe rch, the large and small spotted seatrout aggregation categories were combined in or der to have comparable categories for all species. Acoustic Analysis To determine if signal processing to determ ine sound level could be used to assess courtship sound recordings, comparisons were made between spotted seatrout number as categorized by the human ear and by received sound level. Each recording was read into
16 MATLAB and analyzed by performing a 44100-poi nt Fast Fourier Transform (FFT). Average sound spectrum levels were then calc ulated over 100 Hz wide bins. Signals were calibrated using the hydrophone calibration a nd a calibration of th e DAT recorder. Sound energy for spotted seatrout calls is concentrated in the 200-300 Hz frequency range (Figure 4). Sound pressure levels w ithin the 200-300 Hz frequency range were compared to all spotted seatrout numerical categories assigned by tr ained technicians.
17 RESULTS Seasonality Although hydrophone survey sampling occurr ed from April through October, only data from May 3rd through September 19th were included in the analyses. Spotted seatrout typically begin spawning in midMarch/April in Tampa Bay, however in 2004 the start of the spawning season was delaye d. Data from the hydrophone survey indicated an absence of sounds of large aggregations until May. All seven regi ons were sampled in April and only six stations out of the 144 sa mpled stations had spotted seatrout calling, none of which were large aggr egations. Conversely, thirty stations (n=148) in May had spotted seatrout calling, three of which were large aggregations and three of which were small aggregations. Water temperatures were also quite low in April, ranging from 19.2 C to 26.2 C, and averaging 22.2 C (Figure 7) Monthly water temperature significantly increased to 26.8 C in May (Mann-Whitn ey Rank Sum test, n=292, p<0.001) and sounds of large aggregations were detected. Due to equipment failure, it was not possible to confirm the start of the spawning season based on the LARS at Bunces Pass. However, large aggregation sounds were only de tected on three days in March (19th-21st) but by May, only four days did not have sounds attrib utable to large aggreg ations (Figure 8). Although data were not available from April 1st to May 5th due to equipment failure, additional sampling with a mobile hydrophone at Bunces Pass detected sounds of a large aggregation on April 30th. Combining data from Bunces Pass with information from the hydrophone survey, the start of the spawning season as defi ned by the sounds produced by large aggregations was estimated to begin in early May.
18 Figure 7 Temperature (C) at al l stations sampled in the 2004 Tampa Bay hydrophone survey during April and May. The threshold wate r temperature of 23 C has been cited as the water temperature necessa ry to initiate spawning (B rown-Peterson et. al, 1988). Ap r il 5 April 6 April 9 Ap r il 1 5 A pr il 1 9 A pr il 2 0 A pri l 26 May 3 May 5 May 1 1 M a y 12 M a y 17 M a y 25 May 26Temperature (C) 18 20 22 24 26 28 30 Reported Threshold Spawning Temperature (23 C)
19 Figure 8 2004 Bunces Pass spotted seatrout spaw ning season start with daily duration, sunset time, and daily average water temperat ure. The threshold water temperature of 23 C has been cited as the water temperatur e necessary to initiate spawning (BrownPeterson et. al, 1988). 0 5 10 15 20 25 30 Temperature (C) Eastern Standard TimeBunces Pass 2004 Large Aggregation Level Sound, May 6-June 615:00 05:00 17:00 19:00 21:00 23:00 01:00 03:00 May 10May 15May 20May 25May 30June 4Start/Stop Time of Aggregation Level Sound Water Temperature Sunset Time Reported Threshold Spawning Temperature (23 C) Although data were collected through the first week of October, only data collected through September 19th were used in further analyses. After September 12th, only three large aggregations were det ected by the hydrophone survey and spawning aggregation sound began to decreas e at Bunces Pass by September 13th. Sounds of large aggregations occurred on only one night between September 20th and September 26th (Figure 9). Although large aggregations were heard again from September 27th through October 5th, start times were variable and durat ion steadily decreased. Sounds produced
20 by large aggregations terminated by October 6th and were replaced by individuals calling until complete cessation of spotted se atrout calls occurred on October 8th. Figure 9 2004 Bunces Pass spotted seatrout end of season daily duration with associated sunset time and temperature data. 0 5 10 15 20 25 30 35 Start/Stop Time of Aggregation Level Sound Sunset Time Water Temperature Eastern Standard TimeAugust 22 September 5 September 19October 3Temperature (C)15:00 05:00 17:00 19:00 21:00 23:00 01:00 03:00 Data Collected Over the course of the sampling season, 760 stations were sampled from 190 grids (Table 2). Thirty-four of the grids were samp led at least twice, with five of the grids sampled three times. Eight of the 34 repeat grids displayed differences in the amount of
21 spotted seatrout detected between dates. Re gion E had the highest percentage (60%) of repeat grids with differences between multiple sampling dates (Table 3). There were no clear trends in average temperature, sa linity, dissolved oxygen, sampling time, or sampling date that might explain the cha nges in sound production (Table 4). Because these differences in spotted seatrout courts hip sound occurred in the same grid over different dates, repeated gr id stations were considered independent data points. Table 2 Number of total grids and stations sample d with the percent of grids sampled in each region. Region Number of Grids Number of Stations Percent Grids Sampled A 25 100 50% B 25 100 53% C 30 120 45% D 30 120 41% E 30 120 47% F 30 120 43% G 20 80 46% Total 190 760 46%
22 Table 3 Number of grids repeated and percen tage of repeated grids displaying differences in spotted seatrout detections by region. Region Number of Repeated Grids Percent Repeated Grids wi th Differences in Spotted Seatrout Detections between Dates A 5 20% B 3 33% C 6 17% D 8 13% E 5 60% F 5 20% G 2 0% Total 34 24% Forty six percent of all grids were sampled during the Ma y through September sampling season (Figure 10). All sampled grids were examined for proximity of the four sampling stations to one another to account for potential overlapping in detected calls. Each grid was first scored for sampleable area (that with water depth > 1.5 m). Grids where the area was less than 15% were checked for distance between sampled stations. If this distance was less than 150 m, then those two stations were treated as one station. This decision was based on the assumption that those stations were close enough that the same group of fish could be detected in both locations. Two grids qualified as each having less than 15% of the area available for sampling with two stat ions within 150 m of one another. In each of these grids, the two proximal stations were considered one site, and two stations were thus removed from the universe reducing the number of total stations sampled to 758.
23 Table 4 Repeated grids displaying differences in spotted seatrout detections between sampling dates. Temperature (C), salinity (ppt), dissolved oxygen (mg/L), and time measurements for each grid are the average of the four sampling stations in each grid. Spotted seatrout detectio ns are the number of spotted seat rout heard in each station in the grid, with each station separated by a comma. Estimated number of spotted seatrout were categorized as either: 1=1-2 individuals, 2=3-5 individuals, 3=Small aggregation, 4=Large aggregation Region Grid # Date Grid Average Temp Grid Average Salinity Grid Average DO Grid Average Time Spotted Seatrout Detections A 65 8/5/04 30.1 19.4 8.7 20:33 2,2,4,4 A 65 8/27/04 30.6 17.3 9.7 22:02 0,0,0,1 B 51 5/5/04 24.9 23.9 7.6 21:13 0,0,0,0 B 51 8/6/04 30.0 20.1 6.7 21:01 0,2,1,0 B 51 8/30/04 30.8 18.8 5.1 20:50 3,3,3,3 C 297 6/1/04 30.0 29.9 7.1 20:39 2,0,3,0 C 297 9/7/04 28.3 26.2 8.1 20:23 1,3,4,3 D 269 6/24/04 32.4 28.5 7.4 20:50 0,0,1,4 D 269 7/19/04 27.9 23.6 5.1 23:19 0,0,0,0 E 274 8/15/04 29.1 29.4 8.5 20:26 3,4,3,2 E 274 9/12/04 29.2 29.1 5.4 23:43 0,0,1,0 E 295 7/22/04 29.9 30.9 8.4 21:42 3,2,2,2 E 295 9/12/04 29.0 23.3 6.52 23:09 0,1,0,1 E 353 6/30/04 31.5 33.0 7.8 21:44 0,0,4,0 E 353 7/22/04 30.2 32.1 9.4 23:04 0,0,2,0 F 391 7/23/04 29.9 31.8 6.9 21:04 3,3,2,2 F 391 9/13/04 28.6 27.8 5.0 22:45 1,0,0,0
24 Figure 10 2004 Tampa Bay hydrophone survey sampled grids. Sampled grids are indicated by the pi nk/coral color.
25 Geographic distribution of spot ted seatrout courtship sounds Spotted seatrout sounds (all numerical cate gories) were detected at approximately one-third of all stations (n=758) samp led (Figure 11) throughout Tampa Bay. The majority of the sounds (13% of all sampled stations) were made by 1-2 individual spotted seatrout (Table 5). The sound categories of 35 individuals, small aggregations, and large aggregations (regard less of distance from the hydrophone ) were each present at roughly 8% of all stations sampled. Although large aggregations were detected throughout most of Tampa Bay, they were not equally distributed amongst the seve n regions. They occurred most commonly in the lower bay and the eastern region of the middle bay (Figure 12). No large aggregations were detected in Hillsbor ough Bay. The regional differences were significant ( , n = 758, p<0.001) (Figure 13). Comp ared to the overall expected frequency (Ho=7.7%) of large aggregations thro ughout Tampa Bay, regions A, C and G had fewer aggregations than expected while re gions B, D, E, and F had more (Figure 14). However, only the differences in regions C ( , n = 758, p<0.01), E ( , n = 758, p<0.05), and G were significant ( , n = 758, p<0.01). Most aggregations (95%) occu rred in shoreline grids rather than offshore. These differences were statistically significant ( , n = 58, p<0.001). Although three stations with large aggregations occurr ed in open water grids, these stations had characteristics similar to shoreline grids. Depth did not exceed 3.3 m at any of these stations and two of these three stations had SAV substrata while the other station was a non-channel substratum.
26 Figure 11 2004 Tampa Bay hydrophone survey spotted seatrout detections.
27 Table 5 Spotted seatrout detections by category and distance. Spotted Seatrout Category Number of Stations Â“Directly onÂ–top ofÂ” Number of Stations Â“ClosebyÂ” Number of Stations Â“In the distanceÂ” Total Number of Stations Percent Stations Sampled (total = 758) 1-2 individuals 3 33 62 98 12.9% 3-5 individuals 0 28 35 63 8.3% Small Aggregation 0 35 24 59 7.8% Large Aggregation 0 39 19 58 7.6%
28 Figure 12 2004 Tampa Bay hydrophone survey spo tted seatrout la rge aggregation stations.
29 Figure 13 The number of stations within a region a spotted seatrout large aggregation was detected (black) and the total number of stations sampled within that region (gray). The percent of stations within a region that had large aggregation detections is listed above the bar. Region ABCDEFGNumber of Stations 0 20 40 60 80 100 120 140 Spotted seatrout aggregation stations Total stations sampled 6% 9% 0.01% 11.6% 13.3% 10% 0%Upper Bay Zone Middle Bay Zone Lower Bay Zone Hillsborough Bay Zone
30 Figure 14 Large aggregation percent by region compared to the expected percent (7.7%) of large aggregation detections. Region ABCDEFGLarge Aggregation Percentage 0 2 4 6 8 10 12 14 Expected % of Aggregation Detections (Ho=7.7%)Upper Bay Zone Middle Bay ZoneLower Bay ZoneHillsborough Bay Zone Large aggregations occurred most frequently over SAV substratum. The frequency of large aggregations differed significantly by substrate type ( , n=758, p<0.001). Although non-channel was the most frequently sampled substratum (n= 482), it had the lowest association (1.7% of stations sampled) with large aggregations (Table 6). Roughly one-quarter of the stations sampled over S AV had large aggregations nearby, the largest percentage of any substrata type (Table 6). SAV was significantly higher and nonchannel areas were significantly lower from the expected substrata frequency (SAV: ,
31 n =39, p<0.001, non-channel: , n =39, p<0.001). SAV was pres ent in all regions with the exception of Hillsborough Bay, but aggregation detections over SAV only occurred in regions D, E, and F. Large aggregations rare ly occurred (4.0% of stations sampled) in channels (Table 6). Structure was the least sa mpled substratum but associated with large aggregation sound more frequently than ch annel or non-channel (8.2% of stations sampled) with two stations at old range markers and two stations at bridges. Table 6 Substrata at all sampled stations and stations with large spotted seatrout aggregations categorized as Â“close-byÂ”. Pe rcent of each substratum used by Â“close-byÂ” large aggregations is indicated. Substrata Channel Non-ChannelStructure SAV All Sampled Stations 150 482 49 77 Large Aggregation Stations 6 8 4 21 % Used by Large Aggregations 4.0% 1.7% 8.2% 27.3%
32 Environmental variables associated w ith spotted seatrout courtship sounds Depth of large aggregation stat ions was significantly shallo wer than that of stations without large aggregations (Mann-Whitney Rank Sum test, n=758, p<0.001). Mean depth of stations containing large a ggregations categorized as Â“c lose-byÂ” was 2.8 m, ranging between 1.6-8.2 m (Figure 15). Figure 15 Mean depth (m) of stations without large aggregations versus large aggregation stations, +/one standard deviation. Numbers a bove and below error bars are the number sampled. 0 1 2 3 4 5 6 7 8 9 10 Stations without Large Aggregations Large Aggregation Stations RegionDepth (m)ABC DEFG 97 96 119 105 109 113 80 3 4 1 14 11 6Upper Bay Zone Middle Bay ZoneLower Bay ZoneHillsborough Bay Zone
33 The largest depth value associated with a la rge aggregation (8.2 m) was taken from under a bridge. All stations without large aggregat ions had an average depth of 4.2 m, ranging between 1.5-21.4 m. Mean depth of stations containing large aggregations by region ranged from 2.3 m (n=14) in region D to 3.4 m in regions B & C (n=5). In comparison, mean depth of stations without aggregations varied from 3.0 m (n=97) in region A to 5.3 m (n=109) in region E. As spotted seatrout pr esence was also analyzed according to grid type (shoreline versus open water) and gr id type was most likely related to depth, differences in mean grid type depth were examined. Significant differences between depths of shoreline and open water-categorized grids verified these categories were likely a function of depth (Mann-Whitn ey Rank Sum test, n=758, p<0.001). As stations containing large aggregations (Â“close-byÂ”) were found in shallower, shoreline areas of the bay, th e water temperature was signi ficantly warmer at these stations than at stations w ithout aggregations in the deeper, open water locations (MannWhitney Rank Sum test, n=758, p<0.001). Stations containing large aggregations had an average temperature of 30.3 C while all ot her stations averaged 29.3 C (Figure 16). Also temperature range was smaller at sta tions containing large aggregations (28.0-31.8 C) than at stations without aggregations (24.4-33.8 C). Regionally, mean temperature was relatively consistent for st ations without aggregations, ranging from 28.9 C (n= 96) in region B to 29.9 C (n=113) in region F.
34 Figure 16 Mean temperature (C) of stations wi thout large aggregations versus large aggregation stations, +/one standard deviation. Numbers a bove and below error bars are the number sampled. Region 0.56.512.518.524.530.536.542.5Temperature (C) 26 27 28 29 30 31 32 33 Stations without Large Aggregations Large Aggregation Stations ABC DEFGUpper Bay Zone Middle Bay ZoneLower Bay ZoneHillsborough Bay Zone97 96 119 105 109 113 80 3 4 1 14 11 6 Salinity differences between Â“close-byÂ” stations with and without large aggregations were marginally different (Mann-Whitney Rank Sum test, n=758, p=0.05). Mean salinity across all regions of stati ons containing large aggregations was 27.6 ppt and stations without aggregations was 26.2 ppt (Figure 17). The salinity range of stations containing large aggregations was 18.3-34.5 ppt while stations without aggregations had a larger range between 13.1-35.4 ppt. Salinit y averages at stations of both large aggregations and non-aggregations varied by regions with the two averages within one or two parts-per-thousands of one another at each region.
35 Figure 17 Mean salinity (ppt) of stations with out large aggregations versus large aggregation stations, +/one standard deviation. Numbers a bove and below error bars are the number sampled. 0.56.512.518.524.530.536.542.5 16 18 20 22 24 26 28 30 32 34 36 Stations without Large Aggregations Large Aggregation Stations Region Salinity (ppt)ABC DEFG 97 96 119 105 109 113 80 34 1 14 11 6Upper Bay Zone Middle Bay ZoneLower Bay ZoneHillsborough Bay Zone Dissolved oxygen (DO) was significantly gr eater at Â“close-byÂ” stations containing large aggregations than at stations without aggregations (Mann-Whitney Rank Sum test, n=758, p<0.001). DO values ranged between 5.3 -9.9 mg/L for stations with large aggregations and averaged 7.6 mg/L (Fi gure 18). Stations without aggregations experienced a much broader DO range of 0.2-12.61 mg/L, averaging 6.5 mg/L. The majority (82%) of stations w ith large aggregations were f ound at DO values greater than this non-aggregation mean (6.5 mg/L). Regionally, DO of stations with large
36 aggregations ranged from 6.9 mg/L (n=14) in region D to 8.3 mg/L in region A (n=3) and region E (n=11). Mean regional DO values of stations without a ggregations ranged from 5.6 mg/L in region G (n=90) to 7.2 mg/L in region E (n=109). Figure 18 Mean dissolved oxygen (mg/L) of stations without large aggregations versus large aggregation stations, +/one standard deviation. Numbers above and below error bars are the number sampled. Upper Bay Zone Middle Bay ZoneLower Bay ZoneHillsborough Bay Zone 2 3 4 5 6 7 8 9 10 11 12 Stations without Large Aggregations Large Aggregation Stations RegionDissolved Oxygen (mg/L)ABC DEFG 97 96 119 105 109 113 80 3 4 1 14 11 6
37 Comparison of spotted seatrout spawning locations with other sciaenids Although congeners, sand seatrout used di fferent areas to spawn and were more commonly detected within Tampa Bay than sp otted seatrout. Sand seatrout sounds were detected more frequently (53% of stations, n= 758) than spotted seatrout sounds (Table 7). Aggregation-level sound was also more common in sand seatrout (40% of all stations, n=758) than spotted seatrout (15% of all st ations). Sand seatrout aggregations were detected in all regions (Figure 19) whereas spotted seatrout aggregations were nearly absent in region G (n=1) and region C (n =5) (Figure 20). Although sand seatrout aggregations were more equally distributed geographically than spotted seatrout (Figure 21), regional differences in sand seat rout aggregations were significant ( , n =758, p<0.001). Table 7 Sand seatrout detections by category and distance. Sand Seatrout Category Number of Stations Â“Directly onÂ–top ofÂ” Number of Stations Â“ClosebyÂ” Number of Stations Â“In the distanceÂ” Total Number of Stations Percent Stations Sampled (total = 758) 1-2 individuals 0 27 36 63 8.3% 3-5 individuals 1 30 5 36 4.7% Aggregation 26 180 100 306 40.4%
38 Figure 19 2004 Tampa Bay hydrophone survey sand seatrout detections.
39 Figure 20 2004 Tampa Bay hydrophone survey spotte d seatrout and sand seatrout aggregations.
40 Figure 21 The number of stations within a regi on a sand seatrout aggregation was detected (black) and the total number of stations sampled within that region (gray). The percent of stations within a region that had large aggregati on detections is listed above the bar. Region ABCDEFGNumber of Stations 0 20 40 60 80 100 120 140 160 Sand seatrout Aggregation Stations Total Stations Sampled Upper Bay Zone Middle Bay Zone Lower Bay Zone Hillsborough Bay Zone19%50% 42.5% 56.3% 50% 26.9% 33.8% Most sand seatrout aggregations (70.3% n=306) occurred in open water grids ( , n = 758, p<0.001) while most spotted seatrout ag gregations occurred in shoreline grids. The frequency of sand seatrout aggregations differed significantly by substrate type ( , n=758, p<0.001). Non-channel was the most freq uently used substratum (34.6%, n=206) by sand seatrout aggregations while SAV was used the least (Table 8), the opposite of what was seen with spotted seatrout.
41 Table 8 Substrata at all sampled stations and stati ons with sand seatrout aggregations and percent of each substratum us ed by aggregation stations. Substrata Channel Non-Channel Structure SAV All Sampled Stations 150 482 49 77 Aggregation Stations 23 167 13 3 % Used by Aggregations 15.3% 34.6% 26.5% 3.9% Mean depth of sand seatrout aggregations (5.5 m, n=206) was nearly double that of spotted seatrout aggregati ons across all regions (Figure 22 ). Depths associated with sand seatrout aggregations (1.8-16.1 m) also exhibited a wider range than values associated with spotted seatrout aggregatio ns. Mean temperature (29.2 C), salinity (26.5 ppt), and dissolved oxygen (6.2 mg/L) for sand seatrout aggregations were lower than those associated with spotted seatro ut aggregation stations (Table 9).
42 Figure 22 Mean depth (m) of spotted seatrout and sand seatrout aggregation stations, +/one standard deviation. Numb ers above and below error ba rs are the number sampled. 0 2 4 6 8 10 12 Spotted seatrout aggregations Sand seatrout aggregations Depth (m)RegionABC DEFG 37 6 31 5 9 49 4 15 24 40 24 16 19 1Upper Bay Zone Middle Bay ZoneLower Bay ZoneHillsborough Bay Zone Table 9 Environmental variables at stations wh ere all three sciaenid species (spotted seatrout, sand seatrout, silver perch) were detected at a ggregation levels simultaneously. Species Depth (m) Temperature (C) Salinity (ppt) Dissolved Oxygen (mg/L) Spotted seatrout 2.8 30.1 28.0 7.4 Sand seatrout 5.5 29.2 26.5 6.2 Silver perch 3.9 28.5 27.9 6.6 All 3 species 4.2 30.1 29.3 6.8
43 Silver perch were also heard more frequently than spotted seatrout but less frequently than sand seatrout (Table 10). Silver perch c ourtship sounds were hear d at almost half of the stations (43.2 %, n= 758) and aggregati on level sound (24%) was the most frequently detected silver perch sound categ ory (Figure 23). In contrast to spotted seatrout and sand seatrout, silver perch aggregations were mu ch more evenly distributed geographically (Figure 24) without significant regional differences ( , n =181, p=0.069). All of the regions had between 15-33% of their sample d stations categorized as silver perch aggregation locations (Figure 25). Table 10 Silver perch detections by category and distance. Silver Perch Category Number of Stations Â“Directly onÂ–top ofÂ” Number of Stations Â“ClosebyÂ” Number of Stations Â“In the distanceÂ” Total Number of Stations Percent Stations Sampled (total = 758) 1-2 individuals 2 39 31 72 9.5% 3-5 individuals 3 49 22 74 9.8% Aggregation 25 104 52 181 23.9%
44 Figure 23 2004 Tampa Bay hydrophone survey silver perch detections.
45 Figure 24 2004 Tampa Bay hydrophone survey silver perch, spotted seatrout, and sand seatrout aggregation detections.
46 Figure 25 The number of stations within a region a silver perch aggregation was detected (black) and the total number of stations sampled within that region (gray). The percent of stations within a region that had large aggreg ation detections is listed above the bar. Region ABCDEFGNumber of Stations 0 20 40 60 80 100 120 140 160 Silver Perch Aggregation Stations Total Stations Sampled Upper Bay Zone Middle Bay Zone Lower Bay Zone Hillsborough Bay Zone20%29% 21.7% 32.8%17.5% 28.6% 15% Silver perch aggregation presence was sim ilar at stations from both shoreline and open water grids ( , n = 758, p=0.198). Silver perch ag gregations also did not differ significantly by substrate type ( , n =758, p=0.162), although they were most frequently detected over SAV (Table 11). Mean depth of silver perch aggregations (3.9 m, n=129) was between the mean depths of spotted s eatrout and sand seatrout, as was mean DO (Table 9). Mean values of temperature (28.5 C) at station with aggregations were less for silver perch than the other sciaenids, whereas mean salinities at stations with aggregations were simila r for all three species.
47 Table 11 Substrata at all sampled stations and sta tions with silver pe rch aggregations and percent of each substratum us ed by aggregation stations. Substrata Channel Non-Channel Structure SAV All Sampled Stations 150 482 49 77 Aggregation Stations 19 87 5 18 % Used by Aggregations 12.7% 18.0% 10.2% 23.4% Simultaneous detection of a spotted seatro ut aggregation with at least one other sciaenid aggregation occurred at nearly onequarter (n=29) of st ations with spotted seatrout aggregations (n=117). Spotted s eatrout aggregations were more commonly detected with sand seatrout aggregations (n =26) than with silver perch aggregations (n=13). At only ten stations (1.3 % of all sa mpled stations, n=758) were all three species simultaneously detected at aggregation levels When aggregations of all three species were heard at a station, they were always hear d in some combination of close-by or in the distance (Table 12). The stations where all three species were detected simultaneously occurred in five of the seven regions (regions B, D, E, F, and H) and the majority (n=6) occurred in shoreline-categor ized grids and over non-channel substratum (n=4). Mean salinity at these stations was greater than at aggregation stations for any one of the species, whereas average depth, temperature, and DO were intermediate (Table 9).
48 Table 12 Stations where all three sciaenid specie s (spotted seatrout, sa nd seatrout, silver perch) were detected at aggregation levels simultane ously. Aggregation distance categories are: 1:Â“directly on-top ofÂ”, 2: Â“clos e-byÂ”, or 3: Â“in the distanceÂ”. Date Region Substrata Depth (m) Spotted seatrout aggregation distance Sand seatrout aggregation distance Silver perch aggregation distance 5/26/04 B Structure 5.1 3 2 3 6/2/04 D SAV 1.9 2 3 2 6/2/04 D SAV 1.9 2 2 3 6/9/04 E Structure 2.4 2 3 2 6/9/04 E Channel 10.9 3 3 2 7/1/04 F Non-channel 3.1 2 2 3 7/19/04 D Non-channel 4.8 3 3 3 7/29/04 H Non-channel 2.0 2 3 2 8/6/04 B Structure 5.6 3 2 2 8/16/04 F Non-channel 3.8 2 2 2 Analysis of spotted seatrout courtship sounds There was not a significan t relationship between deci bel level (in the 200-300 Hz frequency range) and abundance category assign ed by ear. There was a large range in decibel level within each category an d much overlap between categories (R2=0.08) (Figure 26). The large aggreg ation category had the least sp read in sound le vel but still overlapped extensively with the other categorical ranges. Af ter reducing the data set to those recordings with only s potted seatrout calling, there sti ll remained a large amount of overlap (R2=0.20) (Figure 27). Similarly, even af ter correcting for distance by using only recordings of spotted seatrout closeby, (Figure 28) there was no clear relationship (R2=0.14).
49 Figure 26 Decibel level within 200300 Hz frequency range corresponding to number of spotted seatrout categorized by ear. All of the plotted r ecordings do not exclusively contain spotted seatrout calls and other sciaenid species may be present. Estimated number of spotted seatrout were categoriz ed as either: 1=1-2 individuals, 2=3-5 individuals, 3=Small aggreg ation, 4=Large aggregation Decibel Level 405060708090100110120C. nebulosus Number as Categorized by Ear 0 1 2 3 4 y = 0.0455x 2.7028 R = 0.0821
50 Figure 27 Decibel level within 200300 Hz frequency range corresponding to number of spotted seatrout categorized by ear. All of the plotted record ings are exclusively spotted seatrout calls. Estimated number of spotted seatrout were categoriz ed as either: 1=1-2 individuals, 2=3-5 individuals, 3=Sm all aggregation, 4=Large aggregation 405060708090100110120 0 1 2 3 4 C. nebulosus Number as Categorized by EarDecibel Levely = 0.0501x 1.5703 R = 0.1966
51 Figure 28 Decibel level within 200300 Hz frequency range corresponding to number of spotted seatrout categorized by ear. All of the plotted record ings are exclusively spotted seatrout calls detected close-by. Estimated nu mber of spotted seatrout were categorized as either: 1=1-2 individu als, 2=3-5 individuals, 3= Small aggregation, 4=Large aggregation 405060708090100110120 0 1 2 3 4 C. nebulosus Number as Categorized by EarDecibel Levely = 0.0527x 1.9116 R = 0.141
52 DISCUSSION Geographic distribution and environmental vari ables associated with spotted seatrout spawning Spotted seatrout spawning occurred in all zones (lower, middle and upper bay) except Hillsborough Bay, and over a wide range of salinities. Most aggregation-level sound occurred in the lower bay zone and eastern middle region; areas previously reported as having spawning activity (McMic hael and Peters, 1989; Lowerre-Barbieri, 2004). Although spawning did not occur as freque ntly in the upper bay zone, presence of large aggregations indicated that spotted seatrout we re not restricted to areas in proximity to the Gulf of Mexico. Mature spotted seatro ut have been reported to move to higher salinity areas in the summer months to spawn (Helser, et al., 1993). However, Tampa Bay spotted seatrout spawning locations were present across the latitudinal expanse of the bay and there was only a marginal differen ce in salinity between aggregation and nonaggregation stations. Although salinity aff ects spotted seatrout egg buoyancy and diameter as well as juvenile survival (Kucera et al., 2002; Holt and Holt, 2003), spotted seatrout are capable of spawning in a wide range of salinity. Optim al spawning salinities ranged from 15 ppt and 21 ppt in Louisian a (Saucier and Baltz, 1993) and spawning studies conducted in captivity generally maintained salinity between 25-30 ppt (Arnold et al., 1976; Brown-Peterson et al., 1988; Gray et al., 1991) although one study kept salinity within approximately 1 ppt of 35.4 ppt (Wis ner et al., 1996). The spread of spawning salinity values range from 7.0 ppt to 25.8 ppt in Louisiana (Saucier and Baltz, 1993) while Tampa Bay larvae were collected be tween 18 ppt and 32 ppt (McMichael and Peters, 1989). Even higher salinity values we re associated with la rval collections from
53 other estuaries with values of 36 ppt (P eebles and Tolley, 1988) and 48 ppt (Holt and Holt, 2003). The moderate range of spawni ng salinities in Tampa Bay (18.3-34.5 ppt) as well as the range in non-spawning areas (13.1-35 .4 ppt) were within th e reported range of spawning salinities for spotted seatrout, suggesting that in Tampa Bay salinity is not influential in determining spawning location provided extreme valu es are not involved. The percentage of locations with spaw ning aggregation differed regionally and was greater in those regions with more s horeline grids and SAV substrata. When compared to the percentage of aggregati ons throughout Tampa Bay (7.7%), the western region of the lower bay had significantly more aggregations and the western region of the middle bay had significantly less, while th eir counterparts did not differ from this expected number. While the number of aggreg ations in the two upper bay regions did not differ from the expected, Hillsborough Bay, sh aring similar latitude with the upper bay, had less than the expected with no larg e aggregation detections. The east/west discrepancies in the lower ba y are likely attributable to both available and sampled substrata. The western portion of the lower bay has more areas categorized by SAV than its eastern counterpart (or any other region) and it also ha d more stations sampled over SAV. Other variables, such as depth, temperature, salin ity, and dissolved oxygen were similar between the two regions. Discrepancie s in the middle bay could again be a result of reduced potential spawning habitat in the western region. Shorelin e grids held 95% of spawning locations and in all regions but th e western region of the middle bay, between 43-78% of grids were categorized as shorel ine and sampled according to this percent. Conversely, the middle bay western region only had 25% of grids categorized as shoreline. This region also had the fewest stations sampled over SAV (besides
54 Hillsborough Bay where no SAV was ever detect ed) whereas its eastern counterpart had three times the amount of stations sampled over SAV. Although spawning aggregations were dete cted in both regions of the upper bay zone, Hillsborough Bay, a zone of similar la titude, was devoid of spotted seatrout spawning. Larvae have been collected at the southern most end of Hillsborough Bay, although this area had a smaller amount coll ected than the middle and lower bay areas (McMichael and Peters, 1989). Hillsborough Bay traditionally has the poorest water quality and consistently experiences hypoxi a (Janicki, 2001; Jani cki, 2001). Reduced abundance of fish and crustaceans, poor flus hing rates, low dissolved oxygen values, and seagrass loss make Hillsborough Bay a poor nur sery habitat (Sykes and Finucane, 1966; Taylor, 1970; Lewis and Estevez, 1988) as well as spawning habitat. Although SAV coverage has increased from complete absen ce in 1982 to an estimat ed 192 acres in 1999, Hillsborough Bay still has the least amount of SAV coverage of any Tampa Bay region (Tomasko, 2002). It also had the lowest mean value of dissolved oxygen of all regions, likely a connection with relatively little SAV, the key spotted seatrout spawning substratum. SAV areas, especially when in proximity to channels or bottom contours, have consistently been reported as spotted seat rout spawning habitat (Mok and Gilmore, 1983; Holt, et al., 1985; Brown-Pete rson, et al., 1988; Crabtree and Adams, 1995). Spawning site selection has been attri buted to placing early life histor y stages in or near habitat types that will foster growth and survival (Peebles and Tolley, 1988) and early stage spotted seatrout eggs have been consistently collected over or near seagrass beds (Holt et al., 1985). SAV is also essential habitat for spotted seatrout larv ae (Holt and Holt, 2000)
55 and was the most important habitat variable associated with youngof-the-year spotted seatrout in Tampa Bay (McMichael and Pe ters, 1989; Nelson and Leffler, 2001). Spotted seatrout have been reported to spawn in a wide variety of habitats, besides those associated with SAV including deep moving wa ter between barrier islands (Saucier and Baltz, 1993), large bridges (Saucier and Baltz 1992), and barrier island beaches as well as on natural sand and shell reefs (Hein and Shepard, 1979). Othe r areas within the spotted seatrout range that have little or no SAV s upport spawning over available substrata such as soft bottom, oyster beds, and tidal marshes (Mahood, 1975; BrownPeterson and Warren, 2001; Lowerre-Barbieri et al., in review). In Tampa Bay, at approximately 75% of the stations with SAV substratum, spotted seatrout aggregations were not detected. Use of SAV as spawni ng habitat also varied regionally. Although SAV was present in all regions with th e exception of Hillsborough Bay, no aggregation detections over SAV occurred in the upper bay or the western region of the middle bay. Spawning aggregations were located over all su rveyed substrata types in Tampa Bay with structure as the most freque ntly used substratum following SAV. The four spawning aggregations associated with structure were split between bridges and old range markers. However, both the bridges and old range ma rkers were in the vi cinity of SAV. Spawning locations were primarily located in shallow, shoreline areas of Tampa Bay. Although the average depth varied regi onally, aggregations consistently used shallow areas regardless of the available depth. As most aggreg ations were detected over SAV, mean depth of aggregat ion stations was relatively shallow. The upper range of mean depth (3.4 m) associated with stations with aggregations occu rred in two regions (western upper and middle bay) where aggreg ations were not detected over SAV. The
56 deepest area used for spawning (8.2 m) was associated with an approximately 450 m-long bridge under the main span. Deeper areas have been implicated as sp awning habitat in othe r studies. Optimal spawning depth in a Louisiana acoustic study was reported between 4.0-8.0 m, with mean depth of aggregation sound occurring at 5.2 m (Saucier and Baltz, 1993). Spotted seatrout aggregations were acoustically detected in Indian River Lagoon, Florida over a deep channel area and in shallow SAV habitats (Mok and Gilmore, 1983). Similarly, gravid spotted seatrout females were collected in varying depths of water in Barataria Bay, Louisiana (Hein and Shepard, 1979). However, in Tampa Bay, the distinct differences between mean depths at aggregation and non-aggregation stations demonstrate the influence of this variable both independently and as an associ ate with substrata type. As temperature is also a function of depth, spaw ning locations were associated with warmer water temperatures than in areas devoid of spawning. Temperature is reported to affect s potted seatrout repr oductive output (BrownPeterson et al., 1988; Kupschus, 2004). Ini tiation of spawning has been linked to temperature, with spawning onset paralleling a 5 C increase to 23 C (Brown-Peterson et al., 1988). The lowest reported temperature at which spawning occurs (from Tampa Bay larvae back-calculations) is 20.4 C (McMichael and Peters 1989). Tampa Bay spotted seatrout began spawning at least one month later than usual in 2004 and this delay is likely due to cooler spring water temperat ures. Although the hydrophone survey started at the beginning of April, large aggregation s ounds were not detected until May when the water temperature rose significantly. The fi rst aggregation was detected at 29 C, and average water temperature of spawning st ations (30.3 C) throughout the season was
57 similar to the predicted optimum reproductive temperature (Kupschus, 2004; 29 C) and concurrent with other repor ted ranges of adult courtshi p calls and young-of-the year (Saucier and Baltz, 1992; 27.5-28.8 C; Sauc ier and Baltz, 1993; 24.5-33.5 C; Nelson and Leffler, 2001; 29.9-30.4 C). Spawning stations were co nsistently located in warmer areas of the bay. This relations hip is likely attributable to depth, as aggregations were habitually located in the shallo w, shoreline areas of Tampa Bay. Comparison of spawning locations between spotted seatrout and other sciaenids Aggregation sounds of spotted seatrout ra rely overlapped with aggregation sounds of other sciaenids. Silver perch were f ound across all regions, subs trate types, and grid types. Conversely, spotted seatrout and sand s eatrout used specific areas of the bay for spawning, indicating they may actively select spawning habitats. Similarly, in the Indian River Lagoon, Florida, silver perch aggreg ations were broadly distributed along the Intracoastal Waterway (ICW) whereas spotted seatrout aggregations were predominantly located within a spec ific southern section (Mok and Gilmore, 1983). While the primary silver perch spawning aggregation was locate d in the ICW, smaller aggregations were detected in shallower areas, some char acterized by SAV (Mok and Gilmore, 1983). Spotted seatrout spawning locat ions shifted temporally, w ith isolated aggregations occurring over shallow SAV early in the seas on, with a shift to deeper SAV areas as well as in the ICW later in the season (Mok and Gilmore, 1983). Differences in substrate and depth were the primary distincti ons between spotted seatrout and sand seatrout spawning habitats Although spotted seatro ut and sand seatrout are congeners and have courtship calls of similar frequency, their use of distinctly different habitats within the estuary app ears to segregate the two and minimize the
58 opportunity for cross-species fert ilization. Spotted seatrout an d sand seatrout are able to hybridize, and sand seatrout have been s hown to hybridize frequently with weakfish ( Cynoscion regalis ) along the east coast of Florida, but there have not been many spotted seatrout/sand seatrout hybrids detected in Tampa Bay (Mike Tringali, pers. comm.). Sand seatrout as well as silver perch also had a much higher percentage of aggregation detections than s potted seatrout. A number of factors could be responsible, including: (1) varying abundance by species; (2) speci es-specific variatio n in the level of sound associated with spawning; and/or (3) the interacti on between species-specific spawning diel periodicities and the sampli ng window. Acoustic interactions between silver perch and spotted seatro ut in the Indian River Lagoon indicated that differing diel periodicities may result from the two speci es sharing overlapping spawning locations (Mok and Gilmore, 1983). As peak acoustic activity of silver perch o ccurred later in the evening during the months when spotted seat rout were spawning, it was suggested that spotted seatrout were responsible for delayi ng the daily start time of the silver perch aggregation sounds. A similar pattern has been observed at Bunces Pass. The silver perch spawning season begins earlier th an the spotted seatrout seas on with the diel periodicity of the silver perch aggregation shifting to la ter in the evening once aggregation sounds of spotted seatrout have begun (Sarah Walters, pers. obs). If these three species have different diel periodicities a ssociated with aggregation calls, then the survey could potentially miss aggregation calls depending on the time certain habitats were sampled. Methodology review The mobile hydrophone survey is an ex cellent methodology for assessing the geographic distribution of c ourtship calls associated w ith spawning. However, a few
59 weaknesses are associated with this type of method and should be addressed. First, because of the nature of the survey, only one boat and one hydrophone were used to assess presence/absence of species-specific calls. Although distance of these calls were estimated, it is not possible to accurately a ssess the true distance of these fish without multiple hydrophones and sound propagation st udies. As each station has different substrata, depth, and acoustic interferences, additional te sting would have to be conducted at each station in or der to accurately determine the distance between the sound source and the hydrophone. However, more tr aditional techniques to assess spawning location can experience similar problems. Pl anktonic eggs and larvae can be quickly dispersed by tides, current, and wind from th eir original spawning site. Similarly, adult capture may occur just prior to spawning, but before the fish are on the spawning grounds. Another methodology issue involves the sampling window used during the study. The five hour window (starting at sunset roughly 20:00 EDT, and continuing until approximately 01:00 EDT) did not account fo r variability of aggregation-associated sound duration within the spawning season. Pr eliminary examination of the Bunces Pass spawning aggregation diel periodicity indicate d that although the average daily duration of aggregation-associated sound over the sp awning season was 5.9 hours, duration ranged from 3.0-12.3 hours (Walters et al., in review). Start and end times of aggregation sound varied as well, with the earliest start time beginning at 17:00 EDT and the latest start time occurring 1.5 hours after sunset at 22:04 EDT. End times ranged from 20:44 EDT to 05:30 EDT. Roughly a quarter of days did not ha ve aggregation sound beginning until after sunset. As sunset was the designated sampling start time for the hydrophone survey
60 and there appears to be diel variability associated with the start and duration of aggregation-associated sound, it is possibl e that hydrophone sampling occurred outside the window of peak sound production. Other studies have reported shorter spawning durations as well. Holt et al. (1985) f ound eggs only for a three hour period around sunset, estimating a spawning window between two hours pre-sunset and 2.5 post-sunset. Although the LARS at Bunces Pass assi sted in reporting sp awning aggregation diel trends, it was only one location used to represent the entire bay. Additionally, the spawning aggregation at Bunces Pass is one of the larges t aggregations in Tampa Bay and one of the only aggregations detected in a Gulf pass. As an anomaly, the spawning aggregation at Bunces Pass may behave di fferently than other estuarine spawning aggregations. Additional LARS at other spawning aggrega tion locations in Tampa Bay would help compare the trends at Bunces Pass to these sites as well as those sampled in the mobile hydrophone survey. These additional permanent monitors would provide further resolution to the diel and seasonal spawning periodicities within Tampa Bay. Multiple studies have verified that sound production is associated with spawning by coupling acoustic sampling with egg colle ction (Mok and Gilmore, 1983; Saucier and Baltz, 1992; Saucier and Baltz, 1993; Gilmore, 1994; Luczkovich et al., 1999), and adult collection (Crabtree and Adams, 1995; Luczkov ich et al., 1999; Lowerre-Barbieri, 2004). However, further research is necessary to dete rmine the level of sound that is consistently associated with gamete release. Spotted seatro ut were only considered to be spawning in large aggregations in this study in order to be conservative bu t spawning could be occurring in the smaller aggregations as well as with 3-5 individuals. Conversely, spawning may not be occurring throughout th e entire duration of aggregation sound.
61 Analysis of spotted seatrout courtship sounds Signal processing, using decibel level w ithin a given frequency range, did not match spotted seatrout abundance categories a ssigned by the human ear. Ultimately, it is not possible to pair decibel ranges with fish number categories as each individual category does not use an exclusive decibe l range. More comp lex signal processing strategies are needed to account for differen ces in sound levels that might be found in different situations. For example, one closeby spotted seatrout call could be louder than an aggregation located in the distance from the hydrophone. Aggregation density may also influence the decibel level. These signal pr ocessing issues must be solved before this type of analysis can be used to assess species, number, and distance.
62 LITERATURE CITED Arnold, C.R., J.L. Lasswell, W.H. Bailey, T.D. Williams and W.A. Fable, Jr. 1976. Methods and techniques for spawni ng and rearing spotted seatrout ( Cynoscion nebulosus ) in the laboratory. Proc. Annu. Conf. Southeast. Assoc. Fish Wildl. Agencies 30: 167-178. Brown-Peterson, N.J., P. Thomas and C.R. Arnold. 1988. Reproductive biology of the spotted seatrout, Cynoscion nebulosus in South Texas. Fishery Bulletin 86(2): 373-388. Brown-Peterson, N.J. and J.W. Warren. 2001. The reproductive biology of spotted seatrout, Cynoscion nebulosus along the Mississippi Gulf Coast. Gulf of Mexico Science 19: 61-73. Brown-Peterson, N. J., M.S. Peterson, D.L. Ni eland, M.D. Murphy, R.G. Taylor and J.R. Warren. 2002. Reproductive biology of female spotted seatrout, Cynoscion nebulosus in the Gulf of Mexico: differences among estuaries? Environmental Biology of Fishes 63: 405-415. Crabtree, R.E. and D.H. Adams. 1995. Spaw ning and fecundity of spotted seatrout, Cynoscion nebulosus in the Indian River Lagoon, Florida. Florida Marine Research Institute Technical Report 1-28. Fish, M.P. and W.H. Mowbray. 1970. Sounds of Western North Atlantic Fishes: A reference file of biological underwater sounds. The Johns Hopkins Press. Baltimore. 256 pp. Gilmore, G. Sound production and communication in the spotted seatrout. In: Bortone, S.A. Ed. 2003. Biology of the spotted seatrout. Marine Biology Series. CRC Press. Boca Raton, FL. 177-195. Gilmore, G. 1994. Environmental parameters as sociated with spawning, larval dispersal and early life history of the spotted seatrout, Cynoscion nebulosus (Cuvier). Interim Report, Harbor Branch Oceanographic Institution, Inc. Goodwin, C.R. Circulation of Tampa and Sa rasota Bays. 1989. In NOAA Estuary-of-the Month Seminar Series No. 11 Tampa and Sarasota Bays: Issues, Resources, Status, and Management. Proceedings of a Seminar held December 10, 1987 Washington DC. 49-64. 215 pp. Gray, J.D., T.L. King and R.L. Colura. 1991. E ffect of temperature and hypersalinity on hatching success of spotted seatrout e ggs. Progressive Fish Cult. 53: 81-84.
63 Hein, S. and J. Shepard. 1979. Spawning of spot ted seatrout in a Louisiana estuarine ecosystem. Proceedings of the Annual Confer ence of the Southeast Association of Fish and Wildlife Agencies 33: 451-465. Helser, T.E. and R.E. Condrey. 1993. Spotted seatrout distribution in four coastal Louisiana Estuaries. Transactions of the American Fisheries Society 122:99-111. Holt, J.G. and S.A. Holt. Effects of variable salinity on reproduction and early life stages of spotted seatrout. In: Bortone, S.A. Ed. 2003. Biology of the spotted seatrout. Marine Biology Series. CRC Press. Boca Raton, FL. 135-145. Holt, J.G. and S.A. Holt. 2000. Vertical distribu tion and the role of phys ical processes in the feeding dynamics of two larval sciaenids Sciaenops ocellatus and Cynoscion nebulosus Marine Ecology Progressive Series 193:181-190. Holt, J.G., S.A. Holt and C.R. Arnold. 1985. Diel periodicity of spawning in sciaenids. Marine Ecology Progressive Series 27:1-7. Janicki, A., R. Pribble and M. Winowitch. 2001. Examination of the spatial and temporal nature of hypoxia in Tampa Bay, Florida. Tampa Bay Estuary Program Technical Report #09-01 140pp. Janicki, A., R. Pribble, S. Janicki and M. Winowitch. 2001. An analysis of long-term trends in Tampa Bay water quality. Ta mpa Bay Estuary Program Technical Report #09-01 401pp. Kucera, C.J., C.K. Faulk and G. J.Holt. 2002. Th e effect of spawning salinity on eggs of spotted seatrout ( Cynoscion nebulosus Cuvier) from two bays with historically different salinity regimes. Journal of Experimental Marine Biology and Ecology 272:147-158. Kupschus, S. 2004. A temperatur e-dependant reproductive mode l for spotted seatrout (Cynoscion nebulosus) explaini ng spatio-temporal variati ons in reproduction and young-of-the-year recruitment in Florida es tuaries. ICES Journal of Marine Science 61:3-11. Lewis, R.R. and E.D. Estevez. 1988. The ecol ogy of Tampa Bay, Flor ida: An estuarine profile. U.S. Fish Wildl. Serv. Biol.Rep. 85 (7.18). 132 pp. Locascio, J.V. and Mann, D.A. 2005. Effects of Hurricane Charley on fish chorusing. Proc Roy. Soc. Lond. Biol Let. 1:362-365. Lowerre-Barbieri, S. 2004. Research on the re production of spotted s eatrout in Florida waters 1999-2003. The Florida Marine Res earch Institute Internal Report 20 pp.
64 Lowerre-Barbieri, S.K., L.R. Barbieri, M.K. Knowlton, C.F. Cotton, and J.R. Flanders. In review. Integrating passiv e acoustic and traditional harvesting methods to determine spotted seatrout spawning ac tivity and fecundity in Georgia waters. Transactions of the Amer ican Fisheries Society Luczkovich, J.J., H.J. Daniel, III and M.W. Sprague. 1999. Characterization of critical spawning habitats of weakfish, spotted s eatrout, and red drum in Pamlico Sound using hydrophone surveys. East Carolina University Final Report and Annual Performance Report 128 pp. Mahood, R.K. 1975. Spotted seatrout in coastal waters of Georgia. Proceedings of the Annual Conference of the Southeast Fish and Wild life agencies 29:195-207. McMichael, R.H. Jr. and K.M. Peters. 1989. Early life history of spotted seatrout, Cynoscion nebulosus (Pisces: Sciaenidae), in Tampa Bay, Florida. Estuaries 12: 98-110. Mok, H. and R.G. Gilmore. 1983. Analysis of so und production in estu arine aggregations of Pogonias cromis Bairdiella chrysoura and Cynoscion nebulosus (Sciaenidae). Bull. Inst. Zool., Academia Sinica 22 (2): 157-186. Nelson, G.A. and D. Leffler. 2001. Abundance, spatial distribution, and mortality of young-of-the-year spotted seatrout ( Cynoscion nebulosus ) along the gulf coast of Florida. Gulf of Mexico Science 30-42. Peebles, E.B and S.G. Tolley. 1988. Distribution, growth, and mortality of larval spotted seatrout, Cynoscion nebulosus : A comparison between two adjacent estuarine areas of southwest Florida. Bulle tin of Marine Science 42(3):397-410. Saucier, M.H. and D.M. Baltz. 1992. Hydrophone identification of spawning sites of spotted seatrout Cynoscion nebulosus (Osteichthys: Scianidae) near Charleston, South Carolina. Northeast Gulf Science 12(2): 141-145. Saucier, M.H. and D.M. Baltz. 1993. Spawni ng site selection by spotted seatrout, Cynoscion nebulosus and black drum, Pogonias cromis in Louisiana. Environmental Biology of Fishes 36: 257-272. Sykes, J.E. and J.H. Finucane. 1966. Occurrenc e in Tampa Bay, Florida, of immature species dominant in Gulf of Mexico comm ercial fisheries. Fish Bull. 65(2):369379. Tavolga, W.N. Sound production and detection. In: Hoar, W.S. and D. J. Randall Eds. 1969. Fish Physiology. Academic Press. New York. 135-205.
65 Taylor, J.L. 1970. Coastal development in Tamp a Bay, Florida. Marine Pollution Bulletin 1(10):153-156. Walters, S., S. Lowerre-Barbieri, J. Bickfo rd, L. Crabtree and D. Mann. In review. Preliminary results on seasonal and diel periodicity of a resident Cynoscion nebulosus spawning aggregation in Tampa Bay, Florida. 58th Proceedings of the Gulf and Caribbean Fisheries Institute. Wisner, B.B., K.L. Mertz and T.R. Ca po. 1996. Spawning frequency and egg production of captive spotted seatrout. The Pr ogressive Fish-Culturist 58:140-142.