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 22 Ka 4500
controlfield tag 007 cr-bnu---uuuuu
008 s2010 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0003311
Passive acoustic studies of estuarine fish populations of southwest florida
h [electronic resource] /
by James Locascio.
[Tampa, Fla] :
b University of South Florida,
Title from PDF of title page.
Document formatted into pages; contains X pages.
Dissertation (Ph.D.)--University of South Florida, 2010.
Includes bibliographical references.
Text (Electronic dissertation) in PDF format.
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
ABSTRACT: Recordings of fish sound production were made in Charlotte Harbor, Florida using Long Term Acoustic Recording Systems (LARS) programmed to record 10 seconds of sound every 10 minutes. Results demonstrated a strong circadian pattern in fish sound production that occurred within a few hours of dusk each evening. Sound production lasted on average 8.7 hrs each evening during the peak spawning season. LARS were deployed when Hurricane Charley crossed Charlotte Harbor in August, 2004. The hurricane did not inhibit nightly chorusing events of spawning fish. Rather, sound levels produced by spawning fish on the night of and 3 days after the hurricane were higher and lasted longer than any of the 9 days recorded prior to the hurricane. Acoustic time series data recorded at multiple sites in Charlotte Harbor during 2005 revealed changes in the spatial distribution of fish sound production in response to increased freshwater inflow and consequent decreased bottom dissolved oxygen concentrations in early June. Fish sound production decreased rapidly over several days at study sites in the northern portion of the harbor most immediately affected by changes in environmental conditions. Meanwhile, fish sound production increased at the study site furthest seaward where normoxic levels were sustained. By August levels of fresh water inflow decreased substantially, bottom dissolved oxygen levels increased and sound production resumed at sites previously affected by these conditions. Fish sound production began intermittently in February and ended in November. Peak levels were reached by mid-late April / early May and continued throughout the summer time. Seasonal patterns of sound production match the reported spawning periods of estuarine sciaenid species recorded. Black drum sound production was measured in the canal systems of Cape Coral and Punta Gorda, Florida during the 2004-2006 spawning seasons. The circadian pattern of sound production was similar to other sciaenids documented in Charlotte Harbor. Seasonal patterns of black drum sound production occurred during October through April and peaked in February. This seasonal period of sound production also matched patterns of black drum reproductive readiness and spawning reported in the literature for the Gulf of Mexico. A hydrophone array was used in the Cape Coral canal system to localize calling black drum and measure source levels and propagation of calls. Source level estimates averaged 165 dBRMS re: 1μPa SPL (SD=1.0) (n = 1,025). Call energy was concentrated in the fundamental frequency (94 Hz) and first two harmonics (188 Hz and 282 Hz). A square root model best described propagation of the fundamental frequency and first harmonic and a log 10 model best described the second harmonic. Based on the mean RMS source level, signal propagation, background levels, and hearing sensitivity, the communication range of black drum at the study site was estimated at between 33 and 108 meters and was limited by background levels, not auditory sensitivity. The timing and levels of sound production and egg production were compared in black drum. Eggs were collected hourly from 1800 0400 by surface plankton tows on two consecutive evenings while black drum sound production was continuously recorded. This sampling effort was conducted five separate times from January through April, 2006. Evidence of the time of spawning was indicated by the collection of blastodiscs (fertilized single cell eggs) or back calculated early cleavage stage eggs. Neither the timing nor the quantity of sound production was positively correlated with egg production on a nightly basis and the greatest densities of eggs were collected on evenings which had the lowest levels of sound production. This may have been due to differences in the fecundity of individual females spawning on the evenings when sampling was conducted.
Advisor: David A. Mann, Ph.D.
Fish sound production
x Marine Science
t USF Electronic Theses and Dissertations.
Passive Acoustic Studies of Estuarine Fish Populations of Southwest Florida by James Vincent Locascio A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy College of Marine Science University of South Florida Major Professor: David A. Mann, Ph.D. Aaron J. Adams, Ph.D. R. Grant Gilmore Jr., Ph.D. Ernst B. Peebles, Ph.D. Jose J. Torres, Ph.D. February 25, 2010 Keywords: fish sound production, fish hearing, fish spawning, sciaenidae, Charlotte Harbor Copyright 2010, James V. Locascio
Dedication This dissertation is dedicated to my family and to the memory of Dr. John R. Costigan.
i Acknowledgments There are many people to thank for thei r support and encouragement during the pursuit of this degree; foremost is my academic advisor, David Mann. Without his support this would not have been possible. I am grateful to him for his patience and the many dialogues that took place over the years during which so many ideas and intellectual breakthroughs occurred. I am al so very grateful to my committee members who were generous with their time a nd enthusiastic about helping me. My parents, siblings, and friends also pr ovided invaluable support of my efforts. Many times they listened patiently while I rambled through thoughts on some technical or computational difficulty I was facing. I wa s often able to redirect my thinking about the problem while describing it to them and although they may not have always understood the details they were more helpful than they know. The help I received from colleagues in the Mann and Peebles labs ha ve also been very much appreciated. Residents of Cape Coral and Punta Gorda, Florida provided invalu able support during the black drum research by allowing me the use of their docks and occasionally by feeding me. I am also grateful to Doug Myhre for r ecovering data off failed hard drives Â– twice! Financial support for this research wa s provided by a grant from the National Sea Grant College Program. Financial support was also provided through the College of Marine Science Student Endowed Fellowships which awarded me the Wachovia, C.W. Bill Young, and Knight Fellowships. Other f unding was provided by the Office of Naval Research for travel, the WomenÂ’s Internationa l Fishing Association, and the J.N. Â‘DingÂ’ Darling Education Foundation.
ii Table of Contents List of Tables ................................................................................................................ .......v List of Figures ............................................................................................................... .... vii Abstract ...................................................................................................................... ........ xi Introduction .................................................................................................................. .... xiv References .................................................................................................................... .... xvi Chapter One: Diel periodicity of fish s ound production in Charlo tte Harbor, FloridaÂ…Â…1 Introduction ..............................................................................................................1 Methods....................................................................................................................2 Results ...................................................................................................................... 6 Discussion ................................................................................................................7 References ........................................................................................................ ......11 Chapter Two: Effects of Hurrican e Charley on fish chorusing .........................................27 Introduction ............................................................................................................27 Methods..................................................................................................................28 Results ....................................................................................................................29 Discussion ..............................................................................................................31 References ..............................................................................................................31 Chapter Three: Seasonal Peri odicity of sound production by sc iaenids in Charlotte Harbor, Florida ...................................................................................... ....39
iii Introduction ............................................................................................................39 Methods..................................................................................................................40 Results ....................................................................................................................40 Discussion ..............................................................................................................41 References ..............................................................................................................43 Chapter Four: Timing of black drum ( Pogonias cromis ) sound production .....................49 Introduction ............................................................................................................49 Methods..................................................................................................................50 Results ....................................................................................................................53 Discussion ..............................................................................................................57 References ..............................................................................................................64 Chapter Five: Localization and source level estimates of black drum ( Pogonias cromis ) 83 Introduction ............................................................................................................83 Methods..................................................................................................................85 Results ....................................................................................................................92 Discussion ..............................................................................................................94 References ............................................................................................................105 Chapter Six: The quantitative and temporal relationship of sound production and egg production by black drum ( Pogonias cromis ) .........................................131 Introduction ..........................................................................................................131 Methods................................................................................................................132 Results ..................................................................................................................138 Discussion ............................................................................................................140
iv References ............................................................................................................144 Summary and Conclusions ..............................................................................................173 References ............................................................................................................180 About the Author ................................................................................................... End Page
v List of Tables Table 1.1. Summary stat istics of fish chorus events from Charlotte Harbor, Fl ..........16 Table 2.1. Daily parameter values of fish choruses prior to, on th e day of, and after the passage of Hurricane Charley ...............................................................34 Table 2.2. Polynomial models of fish chorus parameters ............................................35 Table 3.1 Cross Correlation of river flow and fish sound production ........................44 Table 4.1. Correlation results of tota l acoustic energy and maxi mum sound pressure level of black drum ................................................................................... .69 Table 4.2. Correlation, regression, and Mann-Whitney resu lts of black drum chorus parameters ............................................................................................. ......70 Table 4.3-A. Monthly means and standard deviations of chorus st art time, time of maximum sound pressure level and chorus end time .............................71 Table 4.3-B. Monthly means and standard deviations of chorus st art time, time of maximum sound pressure level and chorus end time .............................71 Table 4.3-C. Monthly means and standard deviations of chorus st art time, time of maximum sound pressure level and chorus end time .............................72 Table 4.4. Monthly means and standard deviations of maximum sound pressure level, dates of last c horus and calls produced by blac k drum at each study site ..73 Table 4.5. Correlation results of black drum chorus parameters from concurrently collected data at all study sites ....................................................................74
vi Table 5.1. Summary statistics of black drum source level estimates. ........................114 Table 5.2. Source level estimates of black drum and various marine mammals .......115 Table 6.1. Time series of developmental stages of black drum eggs ........................147 Table 6.2. Summary data of sp awning and sound production by black drum ...........159 Table 6.3. Fitted instantaneous morta lity rate equations and egg predator data ........160 Table 6.4. Distribution of black drum egg densities with depth .............................. 161 Table 6.5. Estimated black drum fe male spawning stock biomass and individuals .162
vii List of Figures Figure 1.1 Location of study s ite in Charlotte Harbor, Florida. ................................. 17 Figure 1.2-A. Time series of 300-400 Hz band raw data .............................................18 Figure 1.2-B. Time series of 300-400 Hz band raw data prior to 5-point smoothing .19 Figure 1.2-C. Time series of 300-400 Hz band data after 5-point smoothing .............20 Figure 1.3-A. Overlay plot of SPL data from consecutive 24 hr periods ....................21 Figure 1.3-B. Mean and standard devi ation of SPL data from consecutive 24 hr periods .......................................................................................... .........22 Figure 1.4. Delineation of chorus events in acoustic time series data .........................23 Figure 1.5. Start and end times and durations of chorus events ..................................24 Figure 1.6. Laborator y and field recordings of Cynoscion arenarius ..........................25 Figure 1.7. Species composition of c horuses and percent time calling per species .....26 Figure 2.1. Approximate track of Hu rricane Charley across Charlotte Harbor, Fl .....36 Figure 2.2. Composite spectrogram of acoustic energy associated with Hurricane Charley and fish choruses .........................................................................37 Figure 2.3. Spectrum level time series of acoustic energy associated with Hurricane Charley and fish choruses .........................................................................38 Figure 3.1. Study site locations of acoustic recordings in Charlotte Harbors ..............45 Figure 3.2-A Time series of fi sh sound production and river flow in 2005 ................46 Figure 3.2-B Two week means and sta ndard deviations of fi sh sound production and
viii river flow in 2005................................................................................... ..47 Figure 3.3 Regression of nightly ma ximum sound pressure levels and bottom dissolved oxygen concentrations in Charlotte Harbor .............................48 Figure 4.1. Black drum study site loca tions in Cape Coral and Punta Gorda, Fl. ......75 Figure 4.2. Diel periodici ty of black drum sound production ......................................76 Figure 4.3. Black drum chorus start time, end time and time of sunset (seasonal) .....77 Figure 4.4. Monthly means and standard de viations of chorus start and end times and times of maximum SPL ............................................................................78 Figure 4.5-A. Time series of black dr um sound production from all study sites and years ............................................................................................ ..........79 Figure 4.5-B. Monthly means and standard deviations of time series of black drum sound production from all stud y sites and field seasons .......................80 Figure 4.6. Daily maximum SPL with surface and bottom water temperature data ....81 Figure 4.7. Concurrent data of chor us start time from all study sites (2005 ................82 Figure 5.1. Study site location in Ca pe Coral, Florida of array recordings ...............116 Figure 5.2-A. Signal processing steps in call detection of black drum ......................117 Figure 5.2-B. Signal processing step s in peak detection of black drum calls ............118 Figure 5.3. Regression models of received levels vs. distance to source ..................119 Figure 5.4-A. Waveform of a black drum call ...........................................................120 Figure 5.4-B. Spectrogram of a black drum call ........................................................121 Figure 5.5. Frequency of maximum SPL of black drum calls at distance .................122 Figure 5.6. Transmission loss plot s of black drum calls at distance ..........................123 Figure 5.7-A. Acoustic backgrou nd level measurements in the study area ...............124
ix Figure 5.7-B. Signal-to-noise rati os of black drum calls in the study area ................125 Figure 5.8. Audiogram and s ource level estimates of black drum .............................126 Figure 5.9-A. Histogram of RMS rece ived and source levels of black drum calls ...127 Figure 5.9-B. Histogram of Peak rece ived and source levels of black drum calls ....128 Figure 5.10-A. Amplitude modulated r eceived levels of sequentially produced calls by a black drum ................................................................................ .129 Figure 5.10-B. Patterned movement of a black drum based on localized X-Y positions of consecutively produced calls ........................................................130 Figure 6.1. Study site lo cation in Cape Coral, Florida...............................................163 Figure 6.2. Seasonal time series of black drum sound production and egg summary of egg densities produced on nights samples were collected ......................164 Figure 6.3-A. Black drum sound producti on and timing of early stage eggs collected on 1/29/06 ....................................................................................... ....165 Figure 6.3-B. Black drum sound producti on and timing of early stage eggs collected on 2/14 and 2/15/06 ............................................................................16 5 Figure 6.3-C. Black drum sound producti on and timing of early stage eggs collected on 3/3 and 3/4/06 ................................................................................ 166 Figure 6.3-D. Back calculated spawning tim es and mean densities of early stage eggs collected on 3/3 and 3/4/06 .................................................................166 Figure 6.3-E. Black drum sound producti on and timing of early stage eggs collected on 3/20 and 3/21/06 ............................................................................16 7 Figure 6.3-F. Back calculated spawning tim es and mean densities of early stage eggs collected on 3/20 and 3/21/06 .............................................................167
x Figure 6.3-G. Back calculated spawning tim es and mean densities for early stage eggs collected on 4/6 and 4/7/06 .................................................................168 Figure 6.3-H. Back calculated spawning tim es and mean densities for early stage eggs collected on 4/6 and 4/7/06 .................................................................168 Figure 6.4-A. Regression of egg de nsity and maximum SPL including April data ..169 Figure 6.4-B. Regression of egg de nsity and maximum SPL excluding April data ..169 Figure 6.5. Regression of weighted aver age of time of spawning and temporal center of sound production ................................................................................17 0 Figure 6.6. Regression of weighted average of time of spawning and temperature ..170 Figure 6.7-A. Regression of nightly e gg production estimates a nd temperature including April data ............................................................................1 71 Figure 6.7-B. Regression of nightly egg production estimates and temperature excluding April data ............................................................................1 71 Figure 6.8. Regression of egg density and water depth .............................................172 Figure 6.9. Distribution of egg developmental stages and depth ...............................172
xi Passive Acoustics Studies of Estuarin e Populations of Southwest Florida James Vincent Locascio ABSTRACT Recordings of fish sound production we re made in Charlotte Harbor, Florida using Long Term Acoustic Recording Syst ems (LARS) programmed to record 10 seconds of sound every 10 minutes. Results de monstrated a strong circadian pattern in fish sound production that occurred within a few hours of dusk each evening. Sound production lasted on average 8.7 hrs each ev ening during the peak spawning season. LARS were deployed when Hurricane Charley crossed Charlotte Harbor in August, 2004. The hurricane did not inhibit nightly chorusing events of spawning fish. Rather, sound levels produced by spawning fish on the nigh t of and 3 days after the hurricane were higher and lasted longer than any of the 9 days recorded prior to the hurricane. Acoustic time series data recorded at multiple sites in Charlotte Harbor during 2005 revealed changes in the spatial distribut ion of fish sound production in response to increased freshwater inflow and conse quent decreased bottom dissolved oxygen concentrations in early June. Fish sound pr oduction decreased rapidl y over several days at study sites in the northern portion of the harbor most immediately affected by changes in environmental conditions. Meanwhile, fi sh sound production incr eased at the study site furthest seaward where normoxic levels were sustained. By August levels of fresh water inflow decreased substantially, bo ttom dissolved oxygen levels increased and
xii sound production resumed at sites previously affected by these conditions. Fish sound production began intermittently in February a nd ended in November. Peak levels were reached by mid-late April / early May a nd continued throughout the summer time. Seasonal patterns of sound production match the reported spawning periods of estuarine sciaenid species recorded. Black drum sound production was measured in the canal systems of Cape Coral and Punta Gorda, Florida during the 2004-2006 spawning seasons. The circadian pattern of sound production was similar to other sci aenids documented in Charlotte Harbor. Seasonal patterns of black drum sound produc tion occurred during October through April and peaked in February. This seasonal pe riod of sound production also matched patterns of black drum reproductive readiness and spaw ning reported in the lit erature for the Gulf of Mexico. A hydrophone array was used in the Cape Coral canal system to localize calling black drum and measure source levels and pr opagation of calls. Source level estimates averaged 165 dBRMS re: 1 Pa SPL (SD=1.0) (n = 1,025). Call energy was concentrated in the fundamental frequency (94 Hz) and fi rst two harmonics (188 Hz and 282 Hz). A square root model best desc ribed propagation of the funda mental frequency and first harmonic and a log 10 model best described the second harmonic. Based on the mean RMS source level, signal propagation, bac kground levels, and hearing sensitivity, the communication range of black drum at the study site was estimated at between 33 and 108 meters and was limited by background le vels, not auditory sensitivity. The timing and levels of sound producti on and egg production were compared in black drum. Eggs were collected hourly from 1800 Â– 0400 by surface plankton tows on
xiii two consecutive evenings while black drum s ound production was con tinuously recorded. This sampling effort was conducted five se parate times from Ja nuary through April, 2006. Evidence of the time of spawning was indicated by the collection of blastodiscs (fertilized single cell eggs) or back calculated early cleav age stage eggs. Neither the timing nor the quantity of sound producti on was positively correlated with egg production on a nightly basis and the greates t densities of eggs were collected on evenings which had the lowest levels of sound production. This may have been due to differences in the fecundity of individua l females spawning on the evenings when sampling was conducted.
xiv INTRODUCTION Sound production by fishes has probably been observed by humans for nearly as long as they have been pursuing fishes. Ea rly wooden hulled vessels would have acted as transducers providing fisherman with cues to locate sound produc ing aggregations. Among the earliest known scientific investigations of the subject are references made by Dufosse (1874) to AristotleÂ’s conclusions th at sound production was associated with a mechanism involving the swimbladder (Tower, 19 08). Many contributions were made in the latter part of the 19th century which focused mainly on the mechanisms of sound production (Dufosse, 1874; Sorensen, 1895; Tower, 1908). In the past several decades considerable effort has been invested to increase the knowledge base associated with the study of sound production and reception by fishes. This is due in part to increased recognition of the extent to which teleost fishes use sound and the practicality demonstrated by pa ssive acoustic methods for documenting reproductive behavior of soniferous fishes. Im portant advances in th e field have been the development of new recording technologies, th e identification of species-specific sounds, the behavioral context in which they are produ ced and the variety of anatomical features and physiological and neuronal processes re sponsible for sound production (Winn, 1964; Fish and Mowbray, 1970; Fine et al., 1977; Mok and Gilm ore, 1983; Fay and Simmons, 1999; Ladich and Popper, 2004). Considerable effort has also been dedicated to understanding the auditory sensitivity of fish es. Early experiments involved behaviorally conditioned responses to stimuli and in 1998 Kenyon et al. published the first auditory
xv brainstem response (ABR) results for a fish. Since the time of KenyonÂ’s ABR publication the technique, referred to now as auditory evoked potential (AEP) has been repeated on a wide variety of fishes to create audiograms. The research conducted for this dissertati on relied heavily on th e use of recording technology developed at the College of Marine Science. These recording systems were remotely deployed and capable of collecting lo ng term, high temporal resolution acoustic data. These data provided detailed acousti c time series of fish sound production on time scales not previously possible and were used to describe di el and seasonal patterns of sound production by estuarine sciaenids in Ch arlotte Harbor, Florida. Information collected by these recording syst ems was used in all six chapte rs of this dissertation. Other research focused on source level a nd signal propagation estimates of black drum calls recorded with a hydrophone array and pr ocessed with a loca lization algorithm. These data were combined with estimates of black drum auditory sensitivity and background sound pressure le vels at the study site to estimate the acoustic communication range of black drum. To my know ledge, this study is the first to relate complementary data of sound production and rece ption abilities for the same species with signal propagation and background sound pressure levels. To da te, source le vel estimates have been reported for only one other fish species, the toadfish ( Opsanus beta ). In the final chapter I report on the quant itative and temporal relationship between sound production and egg production by black drum. This study was conducted in a mostly enclosed canal basin of Cape Coral, Florida where it could be assumed that the eggs were collected from the same population of fish whose sounds were being recorded. Sampling took place on two consecutive ev enings, five times during the 2005-06
xvi spawning season. Few studies have quan titatively compared the output of sound production and egg production because of the di fficulties associated with confirming that the sampled population is isolated. The application of passive acoustics for monitoring reprodu ctive populations of sound producing fishes has been well establishe d by previous investigators. The results of these previous investigations have help ed to provide a framework for much of the research design used in this di ssertation. It is my intention that the resear ch conducted in this dissertation will further promote the us e of passive acoustic methods and help to establish direction for future research. REFERENCES Dufosse, A. (1874b). Â“Rescherches sur les bruts et les sons expressifs que font entendre les Poisons dÂ’ Europe et sur les or gans producteurs de ces phenomenes acoustiques ainsi que sur les appareils de lÂ’ audition de plusieurs de ces animauxÂ”. Annales des Sci. Naturelles ser 5, 19, 53 pp Fay, R. R., and Simmons, A. M. (1999). Â“The sense of hearing in fish and amphibians,Â” in Comparative hearing: fish and amphibians, edited by R. R. Fay and A. N. Popper (Springer-Verlag, New York), pp. 293-318. Fine, M. L., Winn, H. E., and Olla, B. L. (1977). Â“Communication in fishes,Â” in How Animals Communicate edited by T. A. Sebeok (Indiana University Press) pp. 472-518. Fish, M. P., and Mowbray, W. H. (1970). Â“S ounds of western North Atlantic fishes,Â” (John Hopkins University Press, Baltimore, Maryland).
xvii Kenyon, T. N., Ladich, F., and Yan, H. N. (1998). Â“A comparative study of hearing ability in fishes: the auditory brainste m response approach,Â” J. Comp. Physiol. A. 182, 307-318. Ladich, F., and A. N. Popper, (2004). Â“Paralle l evolution in fish hearing organs,Â” in Evolution of the Vertebrate Auditory System, edited by A. N. Popper and R. R. Fay (Springer-Verlag, New York), pp. 95-127. Mok, H. K. and Gilmore, R. G. (1983). Â“A nalysis of sound production in estuarine aggregations of Pogonias cromis Bairdiella chrysoura and Cynoscion nebulosus (Sciaenidae),Â” Bull. Inst. Zool., Academica Sinica. 22, 157-186. Sorensen, W. (1895). Â“Are the extrinsic muscle s of the air-bladder in som Siluroidae and the Â‘elastic springÂ’ apparatus of others subordinate to the vo luntary production of sounds? What is, according to our pr esent knowledge, the function of the Weberian ossicles,Â”? Journal of Anatomy and Physiology, 19:109-552. Taylor, RW. (1908). Â“The production of sound in the drumfishes, the sea-robin and the toadfish,Â” annals N.Y Acad. Sci. 18:149-180
1 CHAPTER ONE Diel Periodicity of Fish Sound Produc tion in Charlotte Harbor, Florida INTRODUCTION Males of many fish species, particularly members of the family Sciaenidae, produce species-specific sounds associated with courts hip and spawning (Fish and Mowbray, 1970; Tavolga, 1977; Sauc ier et al., 1992; Gilmore, 2003). Changes in levels of fish sound production have been correlated with spawning activity on daily and seasonal time scales (Mok and Gilmore, 1983; Luczkovich et al ., 1999). Many authors have reported maximal sound production to occur during the spawning season from the period of dusk to several hours after nightfall. Holt et al. (1985) documented maximal spawning of sciaenids occurred within th is same time of day by examining developmental stages of eggs collected in plankton samples. Because fish sound production serves as a useful proxy for spaw ning activity, hydrophone surveys provide a powerful, cost-effective and non-destruc tive method for documenting the time and location where spawning is taking place (Mok and Gilmore, 1983; Saucier and Baltz, 1993; Mann and Lobel, 1995; Luczkovich et al., 1999; Hood et al ., 1999; Locascio and Mann, 2005). Many investigations over the past several decades have demonstrated the usefulness of passive acoustics in the field of fisheries researc h. In particular, studies of soniferous spawning aggregations of fishes in the Indian River Lagoon, Florida, have
2 shown long-term site fidelity, with principa l spawning sites being used for more than 20 years (Mok and Gilmore, 1983; Gilmore, 1994; Gilmore, 2003). Promoting the use of passive acoustics as a research tool for c onservation and management, Luczkovich et al. (1999) delimited spawning areas of weakfish ( Cynoscion regalis ) in Pamlico Sound, North Carolina using hydrophone surveys and su ggested the use of passive acoustics for designating marine protected areas. Connaughton and Taylor (1995) demonstrated that physiological measurements of repr oductive readiness of weakfish, C. regalis peaked during the same time of year as sound production. While these and other studies have es tablished the use of passive acoustic techniques for locating indivi duals and aggregations of pa rticular species during their spawning stage (Rountree et al., 2002), the methods have all relied upon the use of mobile hydrophone surveys. Since the mobile hydrophone survey is labor intensive, it is difficult to sample the same sites over the course of the spawning season with a high degree of temporal resolution. Thus, it doe s not provide much information on how sound production varies over a range of time scales or how use of a spawning site may change over a season. The objective of this study was to record hi gh temporal resolution data (10 seconds every 10 minutes) on fish sound pr oduction in Charlotte Harbor, Florida over multiple consecutive days using a newly de veloped recording technology and to examine temporal patterns in s ound production of fishes. METHODS Charlotte Harbor, Florida is a 700-km2 coastal plain estuary receiving freshwater input from the Peace, Myakka, and Caloosahat chee Rivers, which together drain a basin
3 that exceeds 12,000 km2 (Hammet, 1990, Adams et al. 2004). It is the second largest open water estuary in the state, and has been designated an estuary of national significance by the Environmental Protection Agency and has five National Wildlife Refuges occurring within its watershed. The es tuary ranges from 0 Â– 7.5 m in depth with a mean of 2.0 m (USEPA, 1999). Tides in gr eater Charlotte Harbor are predominantly semi-diurnal (McNulty et al., 1972), with a mean diurnal range of 0.5 Â– 1.1 m (Wilzbach et al., 1999). The climate in Ch arlotte Harbor is subtropical, with infrequent freezes, and seasonal mean water temperatures range from 18 to 32o C with a salinity range from 0-36 ppt (Nelson, 1998; Adams et al., 2004). Long Term Acoustic Recording System (LARS) The LARS is a fully programmable ac oustic sampling and datalogging device capable of recording high temporal resolu tion data over extended time periods. It is composed of a Persistor CF2 computer, Oceanographic Embedded Systems 16S2 analog to-digital converter, and a cu stom signal conditioning circui t board developed at the University of South Florida College of Mari ne Science. In this study the LARS was programmed to record ten seconds of sound ev ery ten minutes with 16-bit resolution and a sampling rate of 2500 Hz with an anti-alia sing filter having a cut-off frequency of 1 kHz. Given the Nyquist criterion (Hartm ann, 1997), our effective frequency range for recording underwater sounds while avoiding aliasing (B radbury, 1998) of recorded signals was 0 1250Hz. Sound files were saved on a SanDisk 256 MB flash memory card. An HTI 96-min series hydrophone (sensitivity: -164 dBV/ Pa) with an underwater connector was mounted on the lid of an Ikelite 5810 underwater housing. The underwater (hydrophone) connector was wired to the signal conditioning board, which
4 also provided the power supply for the hydrophon e. Power was supplied with four D-cell batteries connected in series. We deployed th e LARS in the northern portion of Charlotte Harbor (Fig 1.1.) from May 7 through June 10, 2003. The water depth at the study site was approximately 3.5 m and the bottom was composed of muddy sand. Using SCUBA, the underwater housing was secured to a land anchor, installed in the sediment, with galvanized steel chain. When deployed, the underwater housing remained positively buoyant approximately 0.5 m above the bottom. The LARS was retrieved using SCUBA and data were downloaded in th e field to a laptop computer. Data Analysis Data were processed using Qlogger, a custom MATLAB (v6.5) program. Each ten second file was analyzed with a 2500-point Fast Fourier Transform (FFT) to generate a power spectrum. Average spectrum levels were calculated for each 100 Hz wide band and examined for relative concentration of ac oustic energy. The greatest concentration of acoustic energy was found in the 300-400 Hz band and all time series analyses were done on these data. The time series of average sound pressure levels (SPL) for each 10 second recording period within th e 300-400 Hz band were smoothed using a 5-point moving average to reduce variability in the data from sounds produced by boat traffic. Spectrographic analyses were used to verify the origin of sound production (i.e. boat traffic or fishes) and to provide positiv e species identification where possible by descriptions of known species calls reported by previous au thors. The frequency of occurrence of species calls was examin ed between 1650 Â– 0440 hours by scoring the presence or absence of each specie s in two hour bins for all nights.
5 Mean daytime background SPL was calculated from the period 0500 Â– 1700 hours EST using the smoothed data in the 300-400 Hz band. A threshold 3 standard deviations above mean daytime background levels was es tablished to mark the start and end of nightly chorus events and calculate their duration. Occasionally the SPL crossed above the threshold for short periods outside of ni ghtly chorus event times. These occurrences were associated with boat traffic and were not included in the values calculated for chorus duration. We calculated the daily mean maximum SPL and mean time of occurrence and the correla tion between daily maximum SPL and chorus duration. Captive Recordings To validate their presence, we recorded the call of a captive sand seatrout ( Cynoscion arenarius ) (SL = 189 mm) to compare to fish calls recorded in the field. Captive recordings were made using an HT I 96-min series hydrophone (sensitivity -164 dBV/Pa) and Nomad Jukebox recorder with 16 -bit resolution and 48 kHz sampling rate. This file was later converted to 16-bit, 2500 Hz sampling rate using Cool Edit software to generate spectrographs and oscillographs. The Nomad recorder was calibrated with a 1 Volt input signal prior to recordings ma de in the lab. Hydrophone placement was approximately 5 cm away from the fish duri ng lab recordings. Field recordings were subsampled by selecting six calls randomly (3 each from preand post-chorus time periods) from each of twelve randomly selected nights of field recordings (n=72). Call duration (time from beginning of the first pulse to the beginning of the last pulse of the call) and pulse period (call duration divided by one less than the total number of pulses within the call) were measured from the time domain signal, and the calls were compared spectrographically.
6 RESULTS A plot of the 300-400 Hz band raw time se ries data shows nightly peaks in SPL along with smaller shorter duration peaks occurring at various times throughout each day (Fig. 1.2-A). By listening to these signals while review ing their details spectrographically, daytime peaks were eas ily distinguishable as boat traffic while nighttime peaks (louder and longer in duration) were attributable to calling fishes. Smoothing the data with a 5-point moving average substantially reduced the daily variability in SPL while preserving the overall trend associated with fish sound production (Fig. 1.2-B and 1.2-C). A well defined diel pattern in fish so und production was recorded throughout the study period. Sharp increases in sound pre ssure levels due to fish calling occurred regularly each day during the late afternoon / early evening hours (Fig. 1.3-A and 1.3-B). The threshold for establishing start and end times of chorus events was set to 85.4 dB re 1 Pa, three standard deviations (5.4 dB) above mean daytime background levels (80 dB re 1 Pa) (Fig. 1.4). The mean start time of choruses was 1726 hours EST, mean end time was 02:10h EST, and mean duration of chorus events was 8.7 hr per night (Fig. 1.5 and Table 1.1). Mean maximum daily SPL was 117 dB and occurred at the mean time of 2236 hours EST. A weak negative correlation existed between daily maximum SPL and chorus duration (r = -0.24, p = 0.18, df = 32). Call duration and pulse period estimated fr om waveform and spectral analyses of the labrecorded C. arenarius were 174 ms and 43.5 ms, respectively (Fig. 1.6-A and 1.6-B). For this same individua l the received call level wa s calculated at 130 dB re 1 Pa. The mean call duration and pulse peri od of 72 field-recorded calls were 234.8 ms
7 (SD=93 ms) and 45.8 ms (SD= 5.5 ms), respectiv ely. The number of pulses per call from field recordings ranged from 2 to 12 (mode=6 ). A comparison between the lab recording and a selected field recording with the same number of pulses (four) is shown in figures 1.6-C and 1.6-D. Both lab and field record ed calls had a dominant frequency around 350 Hz. The sound recorded in the laboratory had an initial peak in SPL that was not as prominent in the field recordings. Sand seatrout calls occurre d every night during all tim e periods and dominated sound production during chorus even ts. We were able to confirm the presence of silver perch ( Bairdiella chrysoura ), oyster toadfish ( Opsanus beta ) and spotted seatrout ( Cynoscion nebulosus ) by comparing spectrographs of our field recordings to those of sound-truthed captive and in situ recordings by Fish and Mowbry (1970) and Mok and Gilmore (1983). Spotted seatrout and toadfish calls occurred less frequently than sand seatrout calls, but still occurre d during all time periods examin ed (Fig. 1.7). Silver perch calls were only recorded during the latter por tion of the evening (t ypically after midnight) as the chorus faded (Fig. 1.7). Six other t ypes of sounds were infrequently recorded during evening hours. All of these were low frequency, pulsed s ounds and are believed to have been produced by fishes whose sounds have not yet been documented. DISCUSSION Field data collected in this study reveal ed diel patterns in fish sound production with a level of resolution not previously possible. Chorus events lasted for many hours each night and were not highly variable in maximum recorded SPL, or start and end times. Chorus duration, which ranged from 6 to 13 hours, was more variable than other
8 parameters (CV=17.8%) and may have been due to some change(s) in environmental conditions or simply a factor of spatial distribution of fishes when calling began or ended for the evening. Longer term time series data will allow us to examine the potential of lunar and other environmental influences on spawning activity. Maximum levels of drumming activity in the evening, generally from 1800 to 2200 hours, have been noted in the field for a number of sciaenid species, including Cynoscion xanthus (Fish and Cummings, 1972), C. nebulosus and Pogonias cromis (Mok and Gilmore, 1983; Saucier and Baltz 1993; Connaughton and Taylor, 1995) Bairdiella chrysoura (Mok and Gilmore, 1983; Connaughton a nd Taylor, 1995), and the Japanese species, Nibea albiflora and Argyrosomus argentatus (Takemura et al., 1978; Connaughton and Taylor, 1995). Many of these authors also reported the sound level to drop substantially by early morning (~ 0000-0300 hours). Reproductive effort and outcome are ma ximized by bringing together large numbers of males and females at relativel y precise times especially for broadcast spawners (Johannes, 1978; Holt et al., 1985). The onset of darkness provides a general cue throughout a population along with the si gnaling of sound production to effectively form spawning aggregations during the eveni ng hours. Night time spawning also confers the advantage of dispersing planktonic eggs during darkness so egg densities are lower during the following daylight period which ma y reduce vulnerability to visual predators (Holt et al., 1985). The study period corresponds to peak s easonal spawning times for all species recorded and so it is not surprising that the sound level was consistently high and chorus events lasted many hours each night. Longer term acoustic time series data could be
9 compared with histological and ichthyopla nkton data to examine how different methods compare for delimiting the spawning season. Exploring the quantitative relationship between sound production and egg production over daily and seasonal time scales is needed to estimate how these variables are correlated over differe nt time scales and determine if there is always egg pr oduction where there is sound production. This study was conducted in a deeper, cha nneled area of Charlo tte Harbor which may explain why sand seatrout was the most prominent species recorded, as this is their documented habitat (Hoese and Moore, 1998). Sand seatrout was the only species that reached chorus levels, which are characterized by high temporal overlap of calls and high sound pressure levels. All othe r species were only heard as individuals or small groups and occurred less frequently during chorusing. It is possible that ca lls of other species were masked by sand seatrout calls, but it may also be possible that other species avoided calling during this time period. Silver perch cal ls were only recorded near the end of and after chorusing, which may support the idea that species adjust thei r calling schedules relative to others. Mok and Gilmore (1983) al so found that silver perch calling occurred later in the evening and suggested that dela yed silver perch calling was due to loud group calling by spotted seatrout earlier in the eveni ng. This suggests that louder calling may provide a competitive advantage during courtship and spawning. The number of calling indivi duals contributi ng to the sound level of nightly chorusing events is impossible to assess from our recordings made with a single hydrophone. However, the low variability within the data suggests that the number of calling individuals is consistent within the active space of th e hydrophone on a nightly basis, or that the sound level reaches an asym ptote at which point a large increase in the
10 number of calling individuals results in a small increase in sound level. It may be possible to use a hydrophone array to estimat e of the number of individuals producing sound. The pulse period of field recorded calls closely matched that of the captive C. arenarius recording and showed littl e variation (coefficient of variation 12.1%) despite a range of 2-12 pulses per call. The conservative nature of this parameter indicates that it could be used for species-specific recogn ition among fish. The importance of pulse periods for species recognition has been doc umented for damselfishes (Spanier, 1979; Mann et al., 1997) and anurans (Loftus-Hillis and Littlejohn, 1971; Ma nn et al., 1997). Mann et al. (1997) documented low variation (CV=2.8%) in the pulse period of striped cusk-eel calls which ranged from 1-27 pulses and suggested that other cusk eel species would have different pulse periods. Sand seatrout produce a Â‘purringÂ’ sound very similar to the call of the weakfish ( Cynoscion regalis ), a closely related species oc curring on the Atla ntic coast. Luczkovich et al., (1999) characte rized the Â‘purrÂ’ of the weakfish from field recordings as a series of pulses having a mean repetiti on rate of 15.4 pulses/s at a mean dominant frequency of 360 Hz (n=7). Connaughton et al., (2000) described weakfish calls as having 2-15 pulses per call at 20 pulses /s. Because sounds produced by fishes are species-specific and because there has b een a long standing debate on whether arenarius and regalis are two separate species or simply re gional variations of the same species a detailed comparison of their cal ls would be interesting. In coastal and estuarine systems such as Charlotte Harbor, FL research is often directed at understanding how ecosystem function may be affected by anthropogenic
11 influences. Typically this type of influen ce is described as hab itat loss or degradation due to changing land use patterns (e.g. devel opment) or hydrologic alterations (dams or water withdrawal) (Hobbie, 2000). Ecosystem components, such as seagrasses, oysters, or fish assemblages are studied over time a nd used as indicators of potential changes occurring within the ecosystem. Often thes e data require labor intensive and costly methods and do not necessarily yield resu lts that are immediately accessible or informative of ecosystem function in response to stressors. Passive acoustics represent a cost-effective complementary approach to studying ecosystem function by documenting the important life history stage of reproduction of soniferous fishes with great precision on a wide range of temporal and spatial scales. Future research should focus on collecting concurrent environm ental data to help explain variability in fish sound production in relation to natura l and anthropogenic influences. REFERENCES Adams, A. J., Locascio, J.V., and Robbins, B.D. (2004). Â“Microhabitat use by a postsettlement estuarine fish: evidence from relative abundance and predation among habitats,Â” J. Exp. Mar. Biol. Ecol. 299:17-32. Bradbury, J.W., and Vehrencamp, S.L. (1998). Â“P rincipals of animal communication,Â” Sinauer Associates, Inc. Sunderland, Massachusetts. Connaughton, M. A., and Taylor, M.H. (1995) Â“Seasonal and daily cycles in sound production associated with spawning in the weakfish, Cynoscion regalis, Â” Environ. Biol. Fishes 42:233-240. Connaughton, M. A., Taylor, M.H., and Fine, M. L. (2000). Â“Effects of fish size and
12 temperature on weakfish disturbance calls : implications for the mechanism of sound generation,Â” J. of Exp. Biol. 203:1503-1512. Fish, J. F., and Cummings, W. C. (1972). Â“A 50-dB increase in sustained ambient noise from fish. ( Cynoscion xanthulus ),Â” J. Acoust. Soc. Am. 52:1266-1270. Fish, M.P., and Mowbry, W.H. (1970). Â“Sounds of Western North Atla ntic Fishes,Â” The Johns Hopkins Press, Baltimore, Maryland. Gilmore, R. G., Jr. (1994). Â“Environmental pa rameters associated wi th spawning, larval dispersal and early life hist ory of the spo tted seatrout, Cynoscion nebulosus (Cuvier),Â” Report to Marine Research Institute, Florida Department of Environmental Protection, St. Petersburg, Florida. Gilmore, R. G., Jr. (2003). Â“Sound produc tion and communication in the spotted seatrout,Â” Pages 177-195. in S. Bortone, editor. Biology of the spotted seatrout. CRC Press Boca Raton, Florida. Hammet, K. M. (1990). Â“Land us e, water use, streamflow characteristics, and waterquality characteristics of the Charlotte Harbor inflow area,Â” Florida U.S. Geological Survey Water Supply Paper 2359-A. Hartmann, W. M. (1997). Â“Signals, sound, and sens ation,Â” American institute of physics. Woodbury, New York. Hobbie, J. E. (2000). Â“Estuarine science: the key to progress in Coastal Ecological Research,Â” Pages 1-4 in J. E. Hobbie, editor. Estuarine science a synthetic approach to research and practi ce. Island Press Washington, D.C. Hoese, D. H. and Moore, R.H. (1998). Â“Fishes of the Gulf of Mexico ; Texas, Louisiana, and adjacent waters,Â” Texas A&M Univ ersity Press, College Station, Texas.
13 Holt, J. G., Holt, S.A. and Arnold, C.H. (1985). Â“Diel periodicity of spawning in sciaenids,Â” Mar. Ecol. Prog. Ser. 27:1-7. Hood, P., Crabtree, R. and Murphy, M. (1999). Â“Preliminary investigations of red drum spawning habitat in Tampa Bay, Florida, Â” Florida Marine Research Institute. American Fisheries Society Conference. Johannes, R. E. (1978). Â“Reproductive strategies of coastal marine fishes in the tropics,Â” Environ. Biol. of Fishes 3:65-84. Locascio, J. V. and Mann, D.A. (2005). Â“E ffects of Hurricane Charley on fish chorusing,Â” Roy. Soc. Biol Lett. (1) 3:362:365. Loftus-Hillis, J. J. and Littlejohn, M.J. (1971) Â“Pulse repetition rate as the basis for mating call discrimination by two sympatric species of Hyla, Â” Copeia 1971: 154156. Luczkovich, J. J., Sprague, M.W., Johnson, S.E. and Pullinger, R.C. (1999). Â“Delimiting spawning areas of weakfish Cynoscion regalis (family sciaenidae) in Palmico Sound, North Carolina using passive hydr oacoustic surveys.Â” Bioacoustics 10:143-160. Mann, D. A. and Lobel, P.S. (1995). Â“Passive acoustic detection of sounds produced by the Damselfish, Dascyllus albisella (Pomacentridae),Â” Bioacoustics 6:199-213. Mann, D. A., Bowers-Altman, J. and Rountr ee, R.A. (1997). Â“Sounds produced by the striped cusk-eel Ophidion marginatum (Ophidiidae) during courtship and spawning,Â” Copeia 3:610-612. Mok, H. K. and Gilmore, R.G. Jr. (1983). Â“Ana lysis of sound production in estuarine fish Aggregations of Pogonias cromis Bairdiella chrysoura and Cynoscion nebulosus
14 (Sciaenidae),Â” Bulletin of the Zoological Institute of Academ ia Sinica 22:157186. McNulty, J. K., Lindall, W.N. and Sykes, J. R. (1972). Â“Cooperative Gulf of Mexico estuarine inventory and study, Florid a: Phase 1: Area Description,Â” NOAA Technical Report, National Marine Fisheries Service Circulation 378. Nelson, G. A. (1998). Â“Abundance, growth, a nd mortality of young-of -the-year pinfish, Lagodon rhomboids, in three estuaries along th e gulf coast of Florida,Â” Fish. Bull. 96:315-328. Rountree, R. C., Goudey, C., Hawkins, T., Luczkovich, J.J., and Mann, D. A. (2002). Â“Proceedings of Listening to fish: An international workshop on the applications of passive acoustic applications in marine fisheries,Â” Massachuse tts Institute of Technology Sea Grant. Cambridge, Massachusetts. Saucier, M. H. and Baltz, D.M. (1992). Â“Spa wning site selection by spotted seatrout, Cynoscion nebulosus and black drum, Pogonias cromis in Louisiana,Â” Environ. mental Biol. Fish 36:257-272. Saucier, M. H., Baltz, D.M. and Roumillat, W.A. (1992). Â“Hydrophone identification of spawning sites of spotted seatrout Cynoscion nebulosus (Osteichthys: Sciaenidae) near Charleston, South Carolina,Â” Northeast Gulf Science 12:141-145. Spanier, E. (1979). Â“Aspects of species recognition by sound in four species of damselfishes, genus Eupomacentrus (Pisces: Pomacentridae),Â” Z. Tierpsychol. 51:301-316. Takemura, A, Takita, T. and Mizue. K. ( 1978). Â“Studies on the unde rwater sound, VII: Underwater calls of the Japanese marine dr um fishes (Sciaenidae),Â” B. Jpn. Soc.
15 Sci. Fish. 44:121-125. USEPA (United States Environmental Protec tion Agency). (1999). Â“E cological condition of estuaries in the Gulf of Mexi co,Â” EPA 620-R-98-004. Environmental Protection Agency, Office of Research and Development, National Health and Environmental Effects Laboratory, Gulf Ecology Division. Gulf Breeze, Florida. Wilzbach, M. A., Cummins, K.W., Rojas, L. M., Rudershausen, P.J., and Locascio, J.V. (1999). Â“Establishing baseline seagrass para meters in a small estuarine bay,Â” Pages 125-135 in S. Bortone, editor. Seagrasse s: monitoring, ecology, physiology, and management. CRC Pre ss, Boca Raton, Florida.
16 Table 1.1. Summary statistics of nightly chorus parameters recorded in Charlotte Harbor, Florida from May 7 through June 10, 2003 (hrs is time of day; h, m, s are the amount of hours, minutes, and seconds). Parameter n Mean SD CV % Maximum SPL (dB) 34 117 2.36 1.98 Time of maximum SPL 34 2236 hrs 0.7 h 2.95 Chorus start time 34 1726 hrs 1.0 h 6.46 Chorus end time 34 0210 hrs 1.0 h 6.46 Chorus duration 34 8.7 h 1.5 h 17.8
17 Figure 1.1. Study site location in upper Charlotte Harbor, Florid a where acoustic recordings of fish sound production were made from May 7 through June 10, 2003.
18 5.7.03 5.12.03 5.17.03 5.22.03 5.27.03 6.1.03 6.5.03 6.10.03 70 80 90 100 110 120 DateSPL dB (re: 1uPa/sqrt(Hz))Figure 1.2-A. Time series of 300-400 Hz band raw acoustic data recorded in Charlotte Harbor, Florida from May 7, 2003 through June 10, 2003. Ten seconds of acoustic data were recorded every ten minutes.
19 5.22.03 5.23.03 5.24.03 5.25.03 5.26.03 70 80 90 100 110 120 DateSPL dB (re: 1uPa/sqrt(Hz)) Figure 1.2-B. Selected dates of raw acoustic tim e series data shown in Fig 1.2-A prior to smoothing with a 5-point moving average.
20 5.22.03 5.23.03 5.24.03 5.25.03 5.26.03 72 82 92 102 112 122 DateSPL dB (re: 1uPa/sqrt(Hz)) Figure 1.2-C. Shown here are the acoustic tim e series data presented in Figure 1.2-B. after smoothing with a 5 point moving average.
21 14:00 16:00 18:00 20:00 22:00 00:00 02:00 04:00 06:00 08:00 10:00 12:00 14:00 70 80 90 100 110 120 Hours (EST)SPL dB (re: 1uPa) Figure 1.3-A. Shown are five-point smoothe d acoustic time series data of consecutive 24-hour periods overlaid on the same figure. Data were recorded during May 7, through June 10, 2003 in Charlotte Harbor, Florida.
22 14:00 16:00 18:00 20:00 22:00 00:00 02:00 04:00 06:00 08:00 10:00 12:00 14:00 70 80 90 100 110 120 Hours (EST)SPL dB (re: 1uPa/sqrt(Hz)) mean SPL + St. Dev. St. Dev. Figure 1.3-B. Mean and standard deviation of 5-point smoothed acoustic time series data featured in figure 1.3-A are shown here.
23 5.27.03 5.28.03 5.29.03 5.30.03 5.31.03 6.1.03 70 80 90 100 110 120 DateSPL dB (re: 1 uPa/sqrt(Hz)) Figure 1.4. Delineation of nigh tly chorus events using a threshold established as 3 standard deviations above daytime sound pr essure levels recorded during 0500 Â– 1700 hrs. The threshold is indicated by the arrow on the data of 5.27.03.
24 5.7.03 5.10.03 5.15.03 5.20.03 5.25.03 5.30.03 6.4.03 6.9.03 0 200 400 600 800 1000 1200 1400 1600 1800 2000 Datetime of day (hrs) 6 7 8 9 10 11 12 13 14 chorus duration (hrs) chorus duration chorus start chorus end Figure 1.5. Start and end times of nightly chorus events (left axis) and duration of chorus events (right axis) for each evening, May 7 through June 10, 2003. Sunset was at 19:05 EST on May 7 and 19:23 EST June 9, 2003
25 0 0.05 0.1 0.15 0.2 0.25 -0.05 0 0.05 Amplitude 0 0.05 0.1 0.15 0.2 0.25 -0.02 -0.01 0 0.01 0.02 0.03 Time (ms)Frequency Frequency Frequency Frequency 0 0.05 0.1 0.15 0.2 0.25 0 500 1000 Time (ms)Frequency 0 0.05 0.1 0.15 0.2 0.25 0 500 1000 Figure 6A Figure 6B Figure 6C Figure 6D Figures 1.6-A, B, C, D. Lab (A &B) and field recordings of a Cynoscion arenarius call. A. Time waveform of lab recorded C. arenarius call. B. Spectrogr am of lab recorded C. arenarius call. C. Time waveform of field recorded C. arenarius D. Spectrogram of field C. arenarius
26 1650 1840 1850 2040 2050 2240 2250 0040 0050 0240 0250 0440 0 10 20 30 40 50 60 70 80 90 100 Time of DayPercentage of Nights Calling C.a. O.b. C.n. B.c. Figure 1.7. Percentage of nights each species was documented producing sound during the time periods on the x axis. C.a.= Cynoscion arenarius ; O.b.= Opsanus beta ; C.n.= Cynoscion nebulosus ; B.c.= Bairdiella chrysoura.
27 CHAPTER TWO Effects of Hurricane Charley on Fish Chorusing INTRODUCTION Hurricanes are acute, catastrophic events that have the capacity to severely alter natural ecosystems (Mallin et al. 1999). Ec ological impact studies following hurricane events have traditionally included an assessment of habitat alteration, consequent changes in community structure and recovery rates in areas where pre-hurricane data were available for comparison. Although insightful, such a posteriori investigations do not report on the acute responses of biota during, or on temporal scales more immediate to the time of a hurricane strike. Indeed, it is an exceedingly rare occasion for a direct hurricane strike to occur at the precis e time and location of an ongoing field study (Spiller et al. 1998). However on 13 August 2004 this is exactly what occurred in Charlotte Harbor, Florida. The passage of Hurricane Charley directly through Charlotte Harbor provided a unique opportunity to doc ument the acoustic energy associated with the storm and to investigate whether such a major catastrophic event altered the calling behavior of soniferous fish. Spawning is an important part of the lif e history of fishes, and spawning output can influence subsequent recruitment of juvenile fishes. Males of many fishes, particularly members of the family Sci aenidae, produce speciesspecific courtship
28 sounds, which can be identified with hydrophone surveys to determine when and where spawning is taking place (Mok & Gilmore 1983; Saucier et al. 1992; Saucier & Baltz 1993; Mann & Lobel 1995; Luczkovich et al. 1999; Zelick et al. 1999; Gilmore 2003). Many marine fishes are broad cast spawners and spawn at dusk and during the night, which may reduce the risk from visual predat ors on planktonic eggs and adults (Hobson et al. 1978; Robertson 1983; Holt et al. 1985). METHODS To investigate diel patterns of fi sh sound production, a long-term acoustic recording system (LARS) was deployed in Charlotte Harbor, Florida from 4Â–17 August 2004. The LARS was custom built by the Univers ity of South Florida College of Marine Science and consisted of a Persistor Micr ocomputer (CF2), an Oceanographic Embedded Systems 16-bit A/D board (AD16S2) and a custom built signal conditioning board that allowed signal calibration and use of anti-aliasing filters. The LARS was held in an underwater housing and connected to an ex ternal hydrophone (sensi tivity: K164 dBV re 1 mPa; High-Tech, Inc. HTI: 96 min). The underwater housing was attached by chain to a tie-down anchor and remained positively buoya nt some 50 cm above the bottom of the harbor at a 3.5m depth. Acoustic data were sampled at 3333 Hz for 10 s every 10 min and recorded to onboard Compact Flash memory. These data were processed using QLOGGER, a custom MATLAB (v. 6.5) program (Mathworks, Inc.). Each 10 s file was transformed with a 3333 point fast Fourier transform (FFT) to generate a power spectrum. Data were averaged in 100 Hz bins to analyze the timi ng associated with chorusing. To analyse the
29 fish sound production, the timeseries of av erage spectrum level sound between 500 and 600 Hz was smoothed using a 5-point moving av erage to reduce variability from sounds produced by boat traffic. To quantitatively ch aracterize chorusing events, four parameters were established from the smoothed time-seri es, including daily chor us start times, end times, duration and maximum sound pressure level (SPL). The chorus start time was defined as the time when the SPL exceeded 90 dB. The chorus end time was defined as the time when the SPL decreased to less than 90 dB after the chorus had started. The 90 dB SPL threshold used to define chorus start and end times was calculated as three standard deviations, (12 dB) above mean (78 dB SPL) daytime (05:00Â–17:00 h) SPL. Although conservative, this method provides an objective way to establish parameter values. Chorus duration was the chorus end time minus the chorus start time. To investigate the potential influence of Hu rricane Charley on the cy clic sound producing behaviour of fish we fit the time-series of each parameter to a fourth-order polynomial equation. Data points from the day of the hur ricane were excluded from the models to allow for a comparison of pred icted versus observed values. RESULTS Charley, a strong category 4 hurricane, produced wind speed in excess of 226 kph (140 mph) as it crossed Charlotte Harbor on a north-northeasterly track at an approximate speed of 35.5 kph (22 mph) on the afternoon of 13 August 2004. Charley was the strongest hurricane to hit th e United States since Hurricane Andrew in 1992. The eye of the storm crossed into the harbour over th e barrier islands of North Captiva and Cayo Costa south of Boca Grande Pass at approximately 16:00 h eastern daylight saving
30 time (EDT) and departed through the Peace Rive r corridor at approximately 17:00 h EDT (Pasch et al. 2004; Fig. 2.1). By 14:00 h EDT on 13 August, low-fre quency acoustic signals generated by Charley were received by the LARS. At this time, the outer bands were circulating over the harbour, but the ey e of the storm was in the Gulf of Mexico 80 km to the southwest. The loudest received signals (raw, unsm oothed data; 118 dB, 0Â–100 Hz frequency bin) associated with the storm were recorded at approximately 16:00 h EDT when the hurricaneÂ’s eye, approximately 8 km in diam eter, was entering the harbour and the inner bands were directly over the study site (Elect ronic Appendix, audio S2). Signals received from the hurricane were absent or minimal by 17:30 EDT, at which time the first discernable fish calls were recorded. Acoustic energy related to the hurricane was mainly below 400 Hz and concentrated from 0Â–100 Hz, whereas most of the energy associated with the fish chorus comprised mainly of sand seatrout (Cynosci on arenarius) was between 500 and 600 Hz (Fig. 2.2). An analysis of the 500Â–600 Hz band showed there was pronounced diel periodicity in fish sound pr oduction (Fig. 2.3). Nightly chor using events commenced on average at 19:32h EDT and lasted nearly 7 h. Four parameters were used to characterize chorusing events including chorus start tim e, chorus end time, chorus duration and maximum SPL. With the exception of chorus end time, all models provided a fit sufficient for predicting data points on the day of the hurricane within 95% confidence. Observed values were in close agreement w ith predicted values and were within the standard deviation of the mean for each parameter (Tables 2.1 and 2.2).
31 DISCUSSION The ability of the model to predict results well within the inherent variability of each parameter tested indicates the hurricane had no immediate deleterious effect on fish sound production. In fact, data recorded from the three night s after the hurr icane showed increased calling by fish with the highest ma ximum sound levels and start times up to 2.5 h earlier than previous nights. The night of the hurricane was transitional from previous days to these high levels of sound production. Th is could be attributable to the influence of the hurricane on the distribution of fishes or the result of a longer-term cycle in chorusing behavior that may have taken pl ace without the occurrence of the hurricane. While this study found no immedi ate negative impact on populations of chorusing fishes, it is possible that a delayed onset of lo wered oxygen concentrations resulting from increased freshwater inflow associated with th e hurricane could impair fish chorusing and spawning activity in affected areas of Charlotte Harbor REFERENCES Gilmore Jr, R. G. (2003). Â“Sound production and communication in the sp otted seatrout,Â” In Biology of the spotted seatrout (ed. S. A. Bortone), pp. 77Â–195. Boca Raton, Florida: CRC Press. Hobson, J., Godbout, R. and Arnold, C. R. (1978). Trophic relations hips among fishes and plankton at Eniwetok Atoll, Marshall Islands. Fish. Bull. 76, 133Â–153. Holt, J. G., Holt, S. A., and Arnold, C. R. (1985). Â“Diel periodicity of spawning in sciaenids,Â” Mar. Ecol. Prog. Ser. 27, 1Â–7. Luczkovich, J. J., Sprague, M. W., Johnson, S. E. and Pullinger, R. C. (1999).
32 Â“Delimiting spawning areas of weakfish C ynoscion regalis (family Sciaenidae) in Pamlico Sound, North Carolina, using passive hydroacoustic surveys,Â” Bioacoustics 10, 143Â–160. Mallin, M. A. (1999). Â“Hurricane effects on wate r quality and benthos in the Cape Fear watershed: natural and anthropoge nic impacts,Â” Ecol. Appl. 9, 350Â–362. Mann, D. A. and Lobel, P. S. (1995). Â“Passive acoustic detection of sounds produced by the damselfish, Dascyllus albisella (Pomacentridae),Â” Bioacoustics 6, 199Â–213. Mok, H. K. and Gilmore Jr., R. G. (1983). Â“Analysis of sound production in estuarine fish aggregations of Pogonias cromis, Baridiella chrysoura, and Cynoscion nebulosus (Sciaenidae),Â” Bull. Inst. Zool. Acad. Sinica 22, 157Â–186. Pasch, J. P., Brown, D. P. and Blake, E. S. (2004). Â“Tropical cyclone report Hurricane Charley 9Â–14 August 2004,Â” National Hu rricane Center, NOAA, FL, USA. (http//www.nhc.noaa.gov/2004cha rley.shtml?#tcreports) Saucier, M. H. and Baltz, D. M. (1992). Â“Spa wning site selection by spotted seatrout, Cynoscion nebulosus, and black drum, Pogoni as cromis, in Louisiana,Â” Environ. Biol.Fish. 36, 257Â–272. Saucier, M. H., Baltz, D. M. and Roumillat, W. A. (1992). Â“Hydrophone identification of spawning sites of spotted seatrout Cynos cion nebulosus (Osteichthys: Sciaenidae) near Charleston, South Carolina,Â” Northeast Gulf Sci. 12, 141Â–145. Spiller, D. A., Losos, J. B., and Schoener, T. W. (1998). Â“Natural restoration of the species area relation for a lizard after a hurricane,Â” Science 281, 695Â–697. Robertson, D. R. (1983). Â“On the spawning be havior and spawning cycles of eight surgeonfishes (Acanthuridae) from the I ndo-Pacific,Â” Environ. Biol. Fish. 9, 193Â–
33 223. Zelick, R., Mann, D. A. and Popper, A. N. (1999) Â“In Acoustic commu nication in fishes and frogs comparative hearing: fish and amphibians,Â” (ed. A. N. Popper & R. R. Fay), pp. 363Â–411. New York: Springer.
34 Table 2.1. Parameter values of fish chorus events recorded during ea ch evening of the study. Date 2004 Chorus Start Time EDT Chorus End Time EDT Chorus Duration (hrs) Daily Max SPL dB 8/4 19:54 02:24 6.5 111.8 8/5 19:24 03:44 8.3 114.0 8/6 20:04 02:04 6.0 109.1 8/7 20:34 02:54 6.3 102.5 8/8 20:54 02:04 5.2 102.8 8/9 20:14 01:04 5.5 102.4 8/10 21:04 02:34 5.5 107.8 8/11 20:44 02:14 5.5 115.1 8/12 20:14 02:24 6.2 116.3 8/13 18:34 02:24 7.8 115.7 8/14 17:24 03:34 9.2 120.7 8/15 17:04 02:54 8.8 124.3 8/16 17:44 03:14 8.5 120.0 Mean 19:32 02:38 6.9 112.5 S. Dev. 1 h. 23 min. 36 min 1.44 7.2 SPL dB = sound pressure level in deci bels (SPL relative 1 Pascal / Hz0.05)
35 Table 2.2. Polynomial models of fish chorus event parameter valu es fit to time series data excl uding the night of the hurrican e. Predicted chorus parameter values were calculated for the night of the hurricane from each model. Parameter Fitted Polynomials: R2 Mean S. Dev. Predicted Observed Chorus End y = -8E-05x4 + 0.0021x3 0.0179x2 + 0.0497x + 0.037 0.49 02:39 0:37 1:03 1:24 Chorus Start y = 0.1595x4 4.3448x3 + 36.002x2 96.718x + 854.25 0.93 19:37 1:24 17:42 17:34 Chorus Duration y = -0.0049x4 + 0.1328x3 1.103x2 + 2.9326x + 4.8645 0.88 6h 48min 1h 29min 7.7 7.8 Max SPL y = -0.0128x4 + 0.27 99x3 1.4011x2 0.7256x + 115.67 0.91 112.2 7.5 120.2 115.7
36 Figure 2.1. Map of Charlotte Ha rbor, Florida showing the st udy site location (star) and the approximate path of Hurricane Charley as it crossed the harbor during the afternoon of August 13, 2004. P e a c e R i v e rM y a k k a R i v e rGulf Of Mexico Charlotte Harbor Path of CharleyStudy SitePunta Gorda 1600 EDT 1615 EDT 1630 EDT 1645 EDT 1700 EDT
37 Figure 2.2. Composite time series of power spectra calculated w ith a 512-point Fast Fourier Transformation of each 10 second acous tic data file recorded from 8/12/2004 to 8/142004. Color bar indicates spectr um level sound (dB re 1Pa/Hz0.5). Fish choruses are shown on each of the three days and sounds associated with the hurricane are shown on 8/13/2004. Frequency (Hz) 13:54 19:54 02:54 7:54 13:54 19:54 02:54 7:54 13:54 19:54 02:54 1600 1500 1400 1300 1200 1100 1000 900 800 700 600 500 400 300 200 100 20 40 60 80 100 120 140 Fish Chorus Fish Chorus Fish Chorus Hurricane 8/12/2004 8/13/2004 8/14/2004
38 Figure 2.3. Spectrum level time series of 5-point smoothed data of 100-Hz wide frequency bins containing the greatest ac oustic energy associated with fishes (500-600 Hz bin) and Hurricane Charley (0-100 Hz bin) The dashed line at 90 SPL dB indicates the chorus threshold. Tick marks on the x-ax is correspond to midnight on each date. 8-4 8-5 8-6 8-7 8-8 8-9 8-10 8-11 8-12 8-13 8-14 8-15 8-16 60 70 80 90 100 110 120 130 Date (2004)SPL (dB re 1 uPa/sqrt(Hz)) 0 100Hz 500 600Hz
39 CHAPTER THREE Seasonal Periodicity of Sound Production in Sciaenids of Charlo tte Harbor, Florida INTRODUCTION Estuarine environments are highly pr oductive and experience a broad range of conditions due to variation in freshwater inflow, nutrient leve ls, and land use changes that may occur within the watershed. Biotas th at live and reproduce in these systems, including several species of sciaenids, are useful for unde rstanding ecosystem function. Because sciaenids produce sound in associ ation with courtship and spawning, field recorders can be used to document their re productive activity and can show how habitat use changes over different time periods and in response to environmental conditions. Charlotte Harbor is the second largest ope n water estuary in Florida and receives fresh water from the Peace and Myakka rivers at the northern end and the Caloosahatchee at the southern end. Portions of the estu ary become hypoxic after a substantial volume of freshwater has been received and stratification of the water column occurs. This effect naturally creates a change in the spatial di stribution of aquatic biota, and may cause mortality if the onset of these conditions is more rapid than the organismÂ’s ability to relocate to suitable habitat. The redistribu tion of biota also occu rs in response to the change in the location of th e zooplankton maximum subsequent to increased freshwater inflow (Peebles, 2002).
40 In this study, long term ac oustic recording systems (LAR S) were used at various locations in Charlotte Harbor during 2005 to document patterns in sound production by fishes. The objectives were to understand the general patterns of sound production on a seasonal basis and changes in levels of sound production in areas affected by hypoxia due to increased freshwater inflow. METHODS LARS were programmed and deployed and acoustic data were processed according to the methods in Locascio and Mann, 2008 (Chapter One). Study sites were located in the northern, west-cen tral, and southern areas of Ch arlotte Harbor north of Pine Island, Florida (Fig. 3.1). Water quality da ta including temperature, salinity and dissolved oxygen (mg/L) were recorded at 0.5m depth intervals when LARS were periodically retrieved to downl oad data. These data were collected approximately every one to three months. Maximum sound pressu re levels were regr essed against bottom dissolved oxygen concentrations from data collected in April through October. Maximum sound pressure levels were cross-corre lated with river flow data gauged at the USGS Shell Creek near Punta Gord a station from June 1 Â– June 10. RESULTS Sound production by Sciaenid fishes including sand seatrout ( Cynoscion arenarius ), spotted seatrout ( Cynoscion nebulosus ), and silver perch ( Bairdiella chrysoura ) was recorded from February throug h November, 2005 in Charlotte Harbor, Florida. The majority of sound producti on was by sand seatrout. Sound production
41 occurred slightly earlier in th e season at the northern sites but ended at approximately the same time of season throughout the harbor. A change in the spatial di stribution of sound production occurred in association with a rapid, high volume increase of freshwat er inflow to the ha rbor during the first week of June which persisted at elevated le vels throughout the first half of August (Figs. 3.2-A and 3.2-B). Stratification of the water column and decreased bottom oxygen concentrations (< 2.0 mg/L) were recorded at all sites on June 17 and August 22 except for CH 6. By October 9 bottom oxygen concen trations were above 2.0 mg/L at all sites along with resumed fish sound production. Bo ttom concentrations of dissolved oxygen and maximum daily sound pressure levels were positively correlated during periods of April through October. Lin ear regression explained 61% of the variability in maximum sound pressure levels as a function of dissolv ed oxygen concentrations for study sites in the northern part of the harbor (not in cluding sites 5 and 6) (Fig. 3.3). Cross correlations of daily maximum s ound pressure levels and river flow over the ten days, 6-01 Â– 6-10 were negative and st rongest at 0 lag for sites CH 1, CH 7, and CH 3 and at a -1 lag for CH 5. Cross correlatio ns were positive at 0 lag for CH 6 and at a -1 day lag for CH 4. Site CH 4 was nega tively correlated with river flow when the duration of days used in the cross correla tion increased from 10 to 20 (6-01 Â– 6-20) (Table 3.1). DISCUSSION The seasonal duration of sciaenid so und production recorded in this study approximates the seasonal spawning durat ion for these species (Johnson, 1978). If
42 acoustic sampling was only conducted in the nort hern portions of th e harbor, a bi-modal pattern in sound production would have seemed apparent. However, when taking into account the data from the southern site (CH 6) it becomes clear that high levels of sound production occurred from late March through la te October in the Ch arlotte Harbor. The pattern in sound production documented in this study demonstrates a typical response by estuarine fishes to changes in en vironmental conditions associated with an increase in freshwater inflow to the estu arine system. Low or depleted bottom oxygen concentrations combined with a shift seaward of the zooplankton maximum would explain the lowered levels of sound production in the northern harbor with concurrent increase in sound production at the southern st udy site and a return of sound production in the northern harbor when bottom waters became re-oxygenated. The response of fish sound production to increased flows appeared to be almost immediate. At most of the northern sites maximum sound pressure levels began to decline within days of flow increase and within approximately two w eeks minimum sound pressure levels were reached. Sound production at CH 4 continued at higher levels for longer, following the increase in freshwater inflow, than the ot her northern sites and even the second site furthest south (CH 5). This is why the cros s-correlation for this site remained positive until the time series used was increased to a 20 day period. Reasons for this may be associated with specific hydrodynamic conditions that exit in this area which preserved this site for longer. This study demonstrated that acoustic monito ring of fish represents another useful method for studying estuarine ecosystem function. Much consideration has been given to understanding the response of the estuary to managed freshw ater inflow. One of the
43 major difficulties in assessing this is in the ability to make meani ngful measurements of processes on more immediate time scales. Remote water quality monitoring stations represent a major investment for studying estu arine quality. Hydrophones could be easily accommodated within the sensor array on thes e platforms and would provide biological data as a complement to an otherwis e mainly abiotic suite of information. REFERENCES Johnson GD (1978). Â“In: Johnson GD (ed) Developm ent of fishes of the mid-Atlantic bight: an atlas of egg, larval, and juve nile stages volume IV Carangidae through Ephipidae,Â” US Fish and Wildlife Se rvice Department of the Interior, Washington D.C. p 235-236 Peebles, E. B. (2002). Â“Temporal resolution of biological and physical influences on bay anchovy Anchoa mitchilli egg abundance near a river-plu me frontal zone,Â” Mar. Ecol. Prog. Ser. 237:257-269
44 Table 3.1. Cross correlation re sults of river flow and ma ximum sound pressure levels (SPL) of fish choruses using the ten and tw enty day periods: 6/1 Â– 6/10 (columns 2&3) and 6/1 Â– 6/20 (columns 4&5). Maximum SPL were negatively correlated at all sites except CH 6, the southern-most from the river mouth. Lags of 0 days at sites CH 1, 2, and 3 demonstrate the rapid influence of fr eshwater inflow on sound production at these northern sites over the 10 day period. La gs of 1 and 2 days at CH 4 and CH 5 demonstrate delays in the effect of freshw ater inflow on sound production at these sites over the 10 day period. Cross correlations performed over the 20 day period show the same trends among the sites (negative/positive) but lag periods cha nge as a result of changes in volume of fresh water inflow. Ri ver flow data were collected by the United States Geological Survey (USGS) gauge at the Shell Creek near Punta Gorda station (02298202). 6/1 6/10 6/1 6/10 6/1 6/20 6/1 6/20 Site lag r lag r CH 1 0 -0.78 -4 -0.54 CH 2 0 -0.73 0 -0.66 CH 3 0 -0.74 -6 -0.54 CH 4 -1 0.74 -6 -0.66 CH 5 -2 -0.74 0 -0.48 CH 6 0 0.66 0 0.45
45 Figure 3.1. Study site locations in Charlotte Harbor where long term acoustic recording systems were deployed during 2005. Ten sec onds of sound were recorded every ten minutes at all study sites. Charlotte Harbor
46 70 100 130 70 100 130 70 100 130 Sound Pressure Level dB (re: 1uPa) 70 100 130 CH 2 CH 3 CH 4 CH 6 70 100 130 70 100 130 CH 1 CH 5 Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan 0 2,500 5000 Flow (cfs) Shell Creek Figure 3.2-A. Acoustic time series data of fish sound production recorded in Charlo tte Harbor, Florida 2005. Fish sound produ ction began in early February and increased throughout the season until approximately the beginning of November. A substantial drop in sound pressure levels (SPL) occurred in ea rly June at all study sites except for CH6, the furthest south from the river mouth, which increased during this same time. The drop in SPL is associated w ith increased freshwater inflow and a consequent decrease in b ottom water dissolved oxygen concentrations. A s eaward shift in the zooplankton maximum woul d likely also have occurred following th e increased freshwater inflow to the harbor caus ing a spatial redistribut ion of estuarine biota.
47 70 100 130 70 100 130 70 100 130 70 100 130 Sound Pressure Level dB (re: 1uPa) 70 100 130 70 100 130 2/1 3/1 3/29 4/26 5/24 6/21 7/19 8/16 9/13 10/11 11/8 12/6 1/3 0 2500 5000 Flow (cfs) CH 1 CH 2 CH 3 CH 4 CH 5 CH 6 Shell Creek Figure 3.2-B. Calculated two week mean and standard deviations of acoustic time series of data of fish sound production shown in Figuure 3.1-A.
48 0 1 2 3 4 5 6 7 70 80 90 100 110 120 130 Dissolved Oxygen (mg/L)Maximum Sound Pressure Level (dB re: 1uPa) y = 4.98*x + 82.23 r2 = 0.61 Figure 3.3 Regression of maximum nightly sound pressure level and bottom dissolved oxygen concentrations data during April through October from the study sites located in northern Charlo tte Harbor (CH 1, 2, 3, 4, and 5). These sites and time periods were selected because they are located within the range of influence of freshwater infl ow and during seasonal periods of fish sound production.
49 CHAPTER FOUR Timing of Black Drum Sound Production INTRODUCTION Knowledge of the timing and location of spawning provides fundamentally important information for the management of fish species; and this knowledge can be acquired for soniferous species using hydrophone surveys (Mok and Gilmore, 1983; Luczkovich et al., 1999). Traditional met hods for acquiring this knowledge rely upon the collection and examination of fish for spawning condition via gonad histology and gonado-somatic indices and also by back-calcu lating the time of spawning from the ages of eggs or larvae at the time of collec tion (Peters and McMich ael, 1990; Nieland and Wilson, 1993; Fitzhugh et al., 1993). In some cas es direct observations of spawning have been made at fish aggregation sites usi ng SCUBA, remote cameras, and submersible vehicles, (Domeier and Colin, 1997; Erisma n and Konotchick, 2009). While effective, these methods are labor intensive and costly and are not practical for providing high resolution quality data over an entire spawning season or at multiple sites within a season. Many fishes produce sounds associat ed with reproductive behavior and the use of hydrophone recordings to document patterns in acoustic behavior has been conducted for many years (Breder, 1968; Luczkovich et al., 1999; Gilmore, 2003; Ma nn et al. 2008). Relatively recent advances in technology ha ve made low-cost submersible acoustic
50 recording systems available for this purpose. These recording systems, along with signal processing software, now represent the most practical method available to collect longterm high resolution acoustic da ta on spawning behavior of soniferous fishes, many of which include commercially and recreationa lly managed species. Such data can be collected synoptically over wide spatial scales and in remote environments that may not be accessible by other methods. When acousti c data are combined with environmental data collected on the same time scales a grea t deal can be learned about the ecology of sound production and spawni ng site selection. The black drum is a large, long-liv ed sciaenid that ranges from the Bay of Fundy to Argentina (Sutter et al. 1986; Hoese and M oore, 1998). In the Gulf of Mexico black drum spawn in bays and estu arine habitats from late fa ll through early spring (Murphy and Taylor, 1989; Peters and McMichael, 1990) and produces high intensity sounds associated with courtship and spawning (Mok and Gilmore, 1983; Saucier and Baltz, 1993) that may exceed 170 dB re: 1Pa (Locasci o and Mann, in press). In this study our objectives were to record high temporal resolution acoustic da ta and describe patterns of black drum sound production during the spaw ning season and to compare them to previously collected traditional reproductive data that were used to document the spawning season of black drum. METHODS Long-Term Acoustic Recording Syst ems (LARS) were deployed in residential estuarine canals at one site in Punta Gorda and at three sites in Cape Coral, Florida (Fig. 4.1) to document patterns of sound producti on by black drum during their spawning
51 season. One LARS was deployed at the P unta Gorda site from March 22 Â– May 3, 2004 and December 12, 2004 Â– May 4, 2005. At Cape Coral sites 1 and 3 (CC 1 and CC 3) LARS were deployed from February 12 Â– April 6, 2005 and at Cape Coral site 2 (CC 2) from February 12 Â– May 6, 2005 and October 21, 2005 Â– June 7, 2006. Surface and bottom water temperature data were record ed at CC 2 during the October, 2005 Â– June, 2006 deployment using Hobo temperature data loggers (model UA-002-08; Onset Computers) programmed to record data at ten minute interval s. The surface temperature data logger was attached to a buoy and su spended 0.5m below the surface. The bottom temperature data logger was attached to th e LARS, and recorded temperature data at 0.5m above the bottom. During this deploym ent the LARS stopped recording after the first week and was reprogrammed and rede ployed on December 3, 2005. With this exception, all LARS functioned according to schedule. Two LARS models were used for re cordings: a Persistor CF2 computer (sample rate 2,634 Hz) and a Toshiba Pocket PC model E755 (sample rate 11,025 Hz). The Persistor-based LARS was used for all r ecordings except the Oc tober 21, 2005 Â– June 7, 2006 deployment at CC 2, where the Pocket PC was used. High Tech Inc. 96-min series hydrophones were used on all LARS (sensitivity -164dB re: 1V /1Pa and flat frequency response of 2 Hz Â– 37 kHz). The sensitivity of each recorder was calibrated by recording a 0.1 Vpeak sinusoidal voltage. LARS were anchored and remained positively buoyant 0.5 m above the bottom. Water depth at all sites was appr oximately 7 m and the bottom was a soft composite of mud, sand, and clay. Each 10 second file was analyzed with a fast Fourier transform (FFT) to generate a power spectrum from which the band sound pressure level in 100 Hz wide bins was
52 calculated. The SPL was greatest in the 100 Â– 200 Hz band and a five point moving average of data in this frequency range was used for all time series analyses. Mean daytime SPL was calculated from 0700 Â– 1500 hour s for each site and season separately. A chorusing threshold was defined as 2 standa rd deviations above mean daytime SPL and used to delineate the start and end times of nightly black drum chorus events. Chorus duration, total acoustic energy (TAE), and ma ximum SPL were then calculated for each time series using a custom MATLAB script. To be considered a chorus event the SPL were required to exceed the threshold for a minimum of five consecutive points. This controlled for the infrequent cases where SPL briefly exceeded thresholds during the daytime due to vessel noise, weather, or occasional daytime calls made by black drum. TAE (dB re 1 Pas) was calculated by integr ating the acoustic ener gy (after converting SPL to Pa) over the duration of time SPL exceeded the threshold (e.g. summing the area under the curve). This is si milar to the calculation of sound exposure level (ASEL) (ANSI S1.1-1994). Correlations were performed between TAE and maximum SPL and between chorus duration and maximum SPL on time series from each site and season. The purpose was to evaluate the use of the maxi mum SPL to represent the total black drum sound production that occurred on a nightly basi s. Linear regressi ons were performed between chorus start time and sunset and co rrelations were performed between chorus start and end times, between chorus start time and time of maximum SPL, and between chorus start time and maximum SPL for each time series. Data were examined for normality based on standardized kurtosis and skewness. If data were non-normal a Spearman correlation was performed instead of the Pearson correlation. The ascending
53 and descending slopes of nightly chorus even ts from each time series were calculated from the chorus start to the time when the sound level first reached 6 dB below maximum SPL and from the time the sound level decreased 6dB below maximum SPL to the chorus end. Differences between ascending and des cending slopes were tested using the MannWhitney non-parametric test. Correlati ons were performed between sites using concurrent data of chorus start time, chorus end time, chorus duration, and time of maximum SPL recorded during February 14 Â– Ap ril 6, 2005 at all sites. Alpha values for these correlations were adjusted using se quential Bonferroni tests to correct for experiment-wise error (Sokol and Rolf 1995) Cross correlations were performed between nightly maximum SPL, including ni ghts where calling occurred but chorus thresholds were not exceeded, and th e corresponding surface and bottom water temperature data for the 2005-2006 CC 2 time se ries. Fourier analysis was used to examine patterns of lunar peri odicity in black drum sound production in the Punta Gorda time series recorded during December, 2004 Â– May, 2005 and the CC 2 time series recorded during October, 2005 Â– June, 2006. RESULTS: TAE and maximum SPL, and duratio n and maximum SPL were positively and significantly correlated for all sites and seasons (Table 4.1) The association between TAE and maximum SPL was high (r = 0.95 to 0.9 9), and thus maximum SPL was used to quantitatively represent black drum sound produ ction on a nightly basis. The only other soniferous fish species we found in a qualita tive review of the acoustic recordings was the gulf toadfish ( Opsanus beta ). The fundamental frequency of gulf toadfish is
54 approximately 280Hz, (Fine and Thorson, 2002) and so their calls di d not contribute to SPL calculated in the 100-200 Hz fr equency band for black drum. Black drum demonstrated a di el pattern of sound production. Calling was occasionally recorded during the mid mo rning and through the early afternoon but increased dramatically during the late after noon/early evening and chor us durations lasted up to 12 hours during peak season (Fig. 4.2). Regressions of chorus start time and time of sunset resulted in higher r2 values for the shorter time series which began during February, 2005 (CC 1, CC 2, CC 3) and March, 2004 (PG) (r2 = 0.39 to 0.54), than for the two longer time series that covered th e majority of the season, CC 2 2005-2006 (r2 = 0.04) and PG 2004-2005 (r2 = 0.02). Data used for all correlations were normally distributed except for data of chorus start time from CC 3 which was slightly non-normal and so Spearman correlations were performed with these data. Chorus start and end times were negatively, and in most cases signif icantly, correlated for all time series which indicated that later start times generally m eant earlier end times and conversely earlier start times were associated with later end times. Correlations of chorus start time and maximum SPL were strongly negative and signif icant for all time series, indicating that earlier chorus start times were associated with a higher maximum SPL. However, chorus start time and time of maximum SPL were posit ively and relatively weakly correlated. Results of these correlations are detailed in Table 4.2 and data are graphically represented in Fig. 4.3. Monthly means of chorus start a nd end times were more variable than the monthly mean time of maximum SPL however th e variability about the means for each of these parameters were similar among months (Table 4.3, Fig. 4.4). The ascending slopes measured from the chorus start to the tim e corresponding to a sound level 6 dB below the
55 maximum SPL were significantly greater than the descending slopes for all time series (Table 4.2). A distinct seasonal pattern was ev ident in black drum sound production recorded at each site during each season (Fig 4.5-A). Black drum calls were recorded as early as the third week of October, 2005 at CC 2, which was the earliest deployment of a LARS. Threshold levels were first exceeded for 12 hour durations at this site in mid-late December, 2005 on four different nights. Be ginning with the first week of January, 2006 threshold levels were exceeded each night co ntinually through the first week of April, 2006 and calls at sub-threshold levels were last recorded on April 10, 2006. Black drum calls were recorded on the first day (D ec. 13) of the PG 2004-2005 deployment. Threshold levels were first exceeded duri ng the first week of January, 2005 and last exceeded during the third week of April, 2005. Black drum calls were last recorded during this deployment on April 27, 2005. For both the CC 2 2005-2006 and PG 20042005 time series monthly mean maximum SP L were greatest during January through March and peaked in February, but the January and February values were nearly identical at CC 2 and February and March values were mo re similar than the Ja nuary value at PG. The time series which began in Februar y, 2005 (CC sites) and March, 2004 (PG site) produced maximum SPL patterns which confor med to those established by the longer time series at their respective sites (Table 4.4, Fig. 4.5-A and 4.5-B). The last black drum calls were recorded on April 5, 2005 at each of the CC sites and on May 1, 2004 at the PG site. The CC 2 2005-2006 time series bega n and ended somewhat abruptly based on supra-threshold levels. Data recorded at each of the CC s ites during 2005 demonstrated a similar pattern at the end the season. In c ontrast to the CC sites, maximum SPL during
56 both years at the PG site increased and decrea sed more gradually at the start and end of each season and sound production continued for 2 Â– 3 weeks longer. Maximum SPL at PG were also generally lower and more variable than at the CC sites. Time series data of maximum SP L and corresponding temperature data for CC 2 are shown in Fig. 4.6. The grea test coefficient produced by the cross correlation of maximum SPL and bottom temperature was 0.81 at 0 days lag. The correlation coefficient produced by surface temperature and maximum SPL at 0 days lag was -0.14 and the greatest coefficient was -0.4 at 22 days lag. Surface temperatures ranged from approximately 17.5 to 26o C during the period we recorded black drum sound production (12/4/05 4/10/06), but fluctuated within a range of about 18 to 22o C during 12/4/05 2/20/06 (x = 20.3, stdv = 1.8, n = 78) and 22.5 to 26o C (x = 23.9, stdv = 1.2, n = 48) during 2/21 Â– 4/10/06. Surface temperatures du ring these two periods were significantly different (t = -14.8, p = < 0.01) Cross correlations between maximum SPL and surface temperature for each of these time periods pr oduced maximum correlation coefficients of -0.33 at 2 days lag for the period 12/4/05-2/ 20/06 and -0.69 at 1 day lag for the period 2/21 Â– 4/10. Bottom temperatures ra nged from approximately 17 to 24o C over the entire time series and were less variable than su rface temperatures. The seasonal peak in maximum SPL occurred when surface and bottom temperatures were between 18 and 22o C during early January th rough late February. Concurrent data of chorus start, chorus end, and chorus duration were positively correlated between all sites except PG and CC 1 (Table 4.5). A stronger association however existed among the Cape Coral sites for each of these variable s and in particular for chorus start time (Fig. 4.7). Correlations of time of maximu m SPL between sites
57 demonstrated weak, insignificant associations and were either slightly positive or negative. The FFT results of maximum SPL time series did not indicate black drum sound production occurred on a lunar cy cle at either CC 2 or PG. DISCUSSION: The black drum spawning season has been defined within th e Gulf of Mexico by various authors through histol ogical examination of oocyte development, gonado-somatic indices, and the collection of eggs, larvae, and juven iles (Murphy and Taylor, 1989; Peters and McMichael, 1990; Nieland a nd Wilson, 1993; and Fitzhugh et al. 1993). Results of these studies are in general agr eement and place the spawning season from late fall through early spring, includ ing some temporal variabilit y due latitude, with peak spawning during February and March. Seas onal patterns of black drum sound production recorded in this study at different sites and during different years are consistent with the spawning season defined in the literature. These results demonstrate that passive acoustics can be as effective as traditi onal methods have been at documenting the seasonal reproductive period of black drum. Figure 4.5-B features gonado-somatic index data of black drum, reprinted from a study by Fitzhugh, et al. (1993), to illustrate the relationship between reproductive condition and sound production during the spawning season. While time series of black drum sound production at differe nt sites and years conformed to the same general seasonal patter n, clear similarities and differences existed between them. Sound production at PG varied by only 1 day for the date of the last chorus and 4 days for the date of the last call recorded between the 2004 and 2005
58 seasons. The dates of the la st recorded chorus and call were identical among CC sites during 2005, and differed from the CC 2 2006 time series by only 2 and 5 days respectively. Sound production last ed 2-3 weeks longer at the PG site than the CC sites. We do not have data to explain the similar ities between years at the same sites or the differences between the PG and CC sites, howev er the simplest explanation could be that water temperature differences were responsible for these patterns. The PG and CC sites are only 40 km apart and so the influence of latitude alone may not be sufficient for creating enough difference in water temperatur e, but local effects including exposure to sun, wind, and influence of adjacent wate r bodies may contribute to cause the temperature differences responsible for the later end to seasonal calling in PG. Differences were also evident in the lower and more variable maximum SPL recorded at PG and CC 3 and this may be in formative of the distri bution of calling fish within the active space of the hydrophones. B ecause black drum source levels were not reported to be highly variable among indivi duals (Locascio and Mann, in press) the patterns of maximum SPL at these sites are likely not due to lower intensity calls but rather from fishes at greater and more vari able distances from the hydrophone. Both sites were located within smaller, narrower areas of the canal systems compared to CC 1 and CC 2 and may have accommodated fewer fish es pecially if (male) black drum establish territories requiring some amount of space be tween individuals. Another interesting pattern was apparent in the corr elation of concurren tly collected data from all sites. The higher correlations among the CC sites for parame ters of chorus timing demonstrates that acoustic signaling by black drum probably o ccurs in the context of a communication network, where the calling behavi or initiated by some indivi duals elicits responses by
59 others and propagates throughout the population. We were not able to confirm this pattern in the Punta Gorda canal system because of having only a single study site. Two previous studies used hydropho ne recordings to i nvestigate black drum spawning behavior. Saucier and Baltz (1993) conducted mobile hydrophone surveys in coastal southeast Louisiana and recorded black drum during January through April, in 15.0o C -24.0o C water temperatures, with peak sound production during March and April. The highest SPL were recorded in 20.8o C ( 1.01) and 18.9o C ( 1.43) water temperatures for presumed large and mode rately sized black drum aggregations, respectively. Mok and Gilmore (1983) al so conducted mobile hydrophone surveys and recorded black drum during the winter and ea rly spring in Indian River Lagoon, Florida. They reported maximal sound produc tion during January in 18.0 to 20.0o C water temperatures and no sound production occurred below 15.0o C. Although water temperature did not r each the apparent 15.0o C lower limit for sound production during our study, the temperature range over which bl ack drum were record ed (bottom: 17 24o C, surface: 17.5 26o C) and the range associated with highest levels of sound production (18 22o C) were consistent with previous studies. The black drum is a demersal species which could account for the higher correlation between SPL and bottom water temp eratures. The higher correlation between surface water temperatures at a 1 day lag and sound production during the latter half of the season could indicate black drum were hi gher in the water column or possibly that this was a response to increas ing photoperiod, which would be positively correlated with temperature. The range of water temper atures associated with black drum sound production has also been reported for spaw ning. Peters and McMichael (1990) back
60 calculated larval black drum birthdates from collections made in Tampa Bay, Fl. and estimated water temperatures were 16-20o C during the early part of the spawning season and 21-24o C during the peak season. Holt et al. (1988) collected black drum eggs in water temperatures of 18 to 25o C in the Gulf of Mexico ne ar Port Aransas, Texas. Within the black drumÂ’s U.S. geographic range spawning has been documented to occur later in the year at more northern latitudes (Murphy and Taylor, 1989), but apparently within the same water temperature range. In Chesapeake Bay for example, black drum spawn from late April through June (Wells and Jones, 2002) when water temperatures are within approximately the same range repor ted for the Gulf of Mexico during the spawning season (Maryland Department of Natural Resources, 2009). Johnson (1978) estimated that black drum spawning at the mouth of Chesapeake Bay probably occurred in water temperatures of 15 21 o C. Peters and McMichae l (1990) also noted peak spawning occurred around new and full moons a nd suggested this was due to increased tidal amplitude. We found no lunar period icity associated with black drum sound production, but since the precise relationship between the tim ing of sound production and spawning has not been explained for black dr um it is possible that the timing of these different behaviors vary. Studies have been conducted in which associations with moon phase and fish sound production were reporte d (Breder, 1968; Gilmore, 2003; Mann et al,. 2008) and Aalbers (2008) found increased calling rates were associated with spawning, which occurred throughou t the lunar cycle but more so at the time of the new moon to four days after. Establishing chorus start and end times at 2 standard deviations above mean daytime levels is a conservative approach fo r measuring the timing of chorus parameters
61 because the earliest and latest parts of the chorus are ignored. Ne vertheless, sufficient variability in chorus duration measurements resulted, and strong a ssociations between TAE and maximum SPL were ev ident in correlations for a ll time series. The high correlation of these two parameters qualifie d maximum SPL to quantitatively represent nightly black drum sound production. While TAE would appear to be a better choice because it is a more comprehensive measur e, maximum SPL has the advantage of not depending on what threshold level is chose n, whereas TAE tends to increase at lower threshold levels. The correlation between ma ximum SPL and chorus duration was not as strong as maximum SPL and TAE because th e threshold points from which chorus duration was measured did not account for vari ability in duration of signal amplitude at levels above the threshold as did the TA E calculation. Still, correlation results of maximum SPL and duration were relatively stro ng for all time series and ranged from 0.59 to 0.93. In a previous study, we found a weak negative relationship between maximum SPL and chorus duration of sand seatrout, Cynoscion arenarius (r = -0.24) (Locascio and Mann, 2008). The relatively low co rrelation in this case was because data were collected during the peak of the spaw ning season when signal amplitude and chorus duration were consistently high and variability was low compared to the range that exists when data are recorded across an entire spawning season as in the present study. The diel pattern of black drum sound production recorded in this study is consistent with general descriptions found in the literature for many fishes which produce sound associated with courtship and spawning. Calling typically star ts to increase from low daily background levels within an hour or two prior to sunset, increases rapidly, and reaches maximal levels within a few hours af ter sunset (Breder, 1968; Takemura et al.
62 1978; Mok and Gilmore, 1983; Connaughton and Taylor, 1995; Luczkovich et al. 1999; Gilmore, 2003; Aalbers, 2008). In the prev ious hydrophone studies by Saucier and Baltz (1993) and Mok and Gilmore (1983) black dr um calling was noted as early as 1300 and 1400 hr, respectively and the majority of sound production occurred between 1800-2200 hr. The earliest chorus start time we reco rded occurred at 1510 hr on Feb. 19, 2005 at CC 2 but individual calls were occasi onally recorded throughout the day. Earlier chorus start times occurred during the peak spawni ng season and generally corresponded to later end times and more of ten to higher maximum SPL (Table 4.2). This pattern, along with later chorus start times at the be ginning and end of the season, was responsible for the low r2 with time of sunset for the two longer time series. Connaughton and Taylor (1995) discovered a si milar pattern of hi gh intensity calling earlier in the day which lasted later into the evening during the p eak spawning season of weakfish. Various physiological indicators of reproductive readiness in weakfish, including increased plasma androgen levels and hypertrophy of sonic muscle in males, were evident during the seasonal period of maximal sound production and spawning. This pattern has also been documented for spotted seatrout (Brown-Peterson, 2003) and toadfish (Fine and Pennymaker, 1986). It is li kely that similar conditions would exist in black drum, contributing to the patterns we documented of earlier and increased sound production during their peak reproductive period. For all time series the rate of in crease in SPL along the ascending slope of the chorus event was significantly higher than the rate of decrease along the descending slope, and monthly means of each were not highly variable over seasonal periods. Similarly, monthly mean values for time of maximum SPL were not highly variable over
63 the course of the spawning season despite the relatively high variability in mean chorus start and end times on this time frame. Rate s of changing SPL during fish chorus events are not available in the literat ure, although in cases where suffi cient diel time series data have been collected authors have observ ed a relatively rapid onset of calling and substantial increase in SPL over moderate ly short time periods (Breder, 1968; Connaughton and Taylor 1995; Locascio a nd Mann, 2008; Mann et al. 2008). One mechanism for this may be for one individua lÂ’s calls to elicit responses from other individuals resulting in rapi dly increased SPL as calling activity spreads throughout the network of fish. This has been proposed to serve as a general means of aggregating individuals for spawning while at the same time creating the opportunity among (male) individuals to compete acoustically for a ch ance at reproduction. There is currently more evidence to suggest that calling rate s of individuals are not highly variable (Connaughton and Taylor, 1995; Locascio and Mann, in press) and so increased SPL and calling rates of a group are more likely due to more individuals calling as opposed to individuals calling more (Connaughton and Ta ylor 1995), although few definitive data exist for fish on this subject and more are ne eded. The more gradual changes associated with the descending slope of th e chorus event may therefore be due to fewer individuals calling as this part of the evening progresses. This may simply be a consequence of sonic muscle fatigue but it may also be related to fluctuating androgen levels, as demonstrated on a seasonal time frame for weakfish. To our knowledge no data has been published yet that demonstrates androgen levels in fish fluctuate on a daily cycle however it would seem to be possible especially during th e spawning season. Rubow and Bass (2009) recently discovered diel and seasonal differe nces in the function of the hormonally
64 modulated vocal motor system of the mids hipman. Bass and Zakon (2005) demonstrated that injections of 11-ketotestosterone increas ed call duration in toadfish and midshipman for up to 2hr and concluded that Â‘11KT modul ation of the vocal pattern generator may contribute to increased calling and the tran sition from a non-calling to a calling stateÂ’ (Bass and Zakon, 2005). It would then seem possible that decreased calling could be associated with decreases of 11KT on a daily basis. In this study it was demonstrated that the timing of black drum sound production strongly corresponds with the seasonal spaw ning period described in the literature. Long-term acoustic recording systems can ther efore be used to complement traditional methods for defining the spawning season which are far more costly, labor intensive, and destructive. Inferences about habitat quality can also be made from these acoustic data since spawning site selection should place early life history stages in habitats beneficial for growth and survival (Peebles and To lley, 1988). Therefore, in addition to documenting the timing and location of repr oductive behavior of soniferous fishes, acoustic surveys can be a convenient and useful way to evaluate pa tterns of habitat use and in this context environmental data s hould be recorded alo ng with acoustic data. REFERENCES Aalbers, S.A. (2008). Â“Seasonal, diel, and lu nar spawning periodici ties and associated sound production of white seabass ( Atractoscion nobilis ),Â” Fish Bull 106:143-151 American National Standard Institute (1994) In: Acoustical Terminology. Standards Secretariat Acoustical Society of Amer ica 120 Wall Street, 32nd Floor New York, New York 100053993, p 9
65 Bass, A.H., and Zakon, H.H. (2005). Â“Sonic and electric fish: At the crossroads of neuroethology and behavioral neur oendocrinology,Â” Horm Behav 48:360-372 Breder, C.M. (1968). Â“Seasonal and diurnal occu rrences of fish sounds in a small Florida Bay,Â” Bull. Am. Mus. Nat. Hist. 138:327-378 Brown-Peterson, N. (2003). Â“The reproducti ve biology of spotted seatrout. In: Bortone SA (ed) The Biology of Seatrout, CRC Press, Boca Raton, FL., p 99-133 Connaughton, M.A. and Taylor, M.H. (1995). Â“Seasonal and daily cycles in sound production associated with spawning in the weakfish, Cynoscion regalis,Â” Environ. Biol. Fishes 42:233-240. Domeier, M.L. and Colin, P.L. (1997). Â“Tr opical reef fish spaw ning aggregations: defined and reviewed,Â” Bull Mar Sci 60(3): 698-726 Erisman, B.E. and Konotchick, T.H. (2009). Â“ Observations of spawning in the Leather Bass, Dermatolepis dermatolepis (Teleostei: Epinephelidae), at Cocos Island, Costa Rica,Â” Environ Biol Fish 85:15-20 Fine, M.L. and Pennymaker, K.R. (1986). Â“Hor monal basis for sexual dimorphism of the soundproducing apparatus of the oy ster toadfish. Exp Neurol 92:289-298 Fine, M.L. and Thorson, R.F. (2002). Â“Cre puscular changes in emission rate and parameters of the boatwhistle adve rtisement call of the gulf toadfish, Opsanus beta ,Â” Environ Biol Fish 63:321-331 Fitzhugh, G.R., Thompson, B.A., and Snider, T. G. III (1993). Â“Ovarian development, fecundity, and spawning frequency of black drum Pogonias cromis in Louisiana,Â” Fish Bull 91:244-253
66 Gilmore, R.G. Jr. (2003). Â“Sound production and communication in the spotted seatrout,Â” In: Bortone S (ed) Biology of the spo tted seatrout. CRC Press, Boca Raton, FL, p 177Â–195 Hoese HD, Moore RH (1998) In: Fishes of the Gulf of Mexico. Texas A&M University Press, College Station, Texas pp 240-241 Holt S.A., Holt G.J., and Young-Abel, L. (1988). Â“A procedure for identifying sciaenid eggs. Contrib. Mar. Sci. Suppl. 30:99Â–108 Johnson, G.D. (1978). Â“ Pogonias cromis black drum,Â” In: Johnson GD (ed) Development of fishes of the mid-Atlan tic bight: an atlas of egg, larval, and juvenile stages volume IV Carangidae th rough Ephipidae. US Fish and Wildlife Service Department of the In terior, Washington D.C. p 235-236 Locascio, J.V. and Mann, D.A. (2008) Â“Diel pe riodicity of fish sound production in Charlotte Harbor, Florida,Â” Trans. Am. Fish. Soc. 137:606Â–615 Locascio J.V., Mann D.A. (in review) Â“Localiz ation and source level estimates of black drum ( Pogonias cromis ) calls,Â” J. Acoust. Soc. Am. Luczkovich, J.J., Sprague, M.W., Johnson, S.E., Pullinger, R.C. (1999). Â“Delimiting spawning areas of weakfish Cynoscion rega lis (family Sciaenidae) in Pamlico Sound, North Carolina, using passive hydr oacoustic surveys,Â” Bioacoustics 10:143Â–160. Mann, D.A., Locascio, J.V., Coleman, F.C., Koenig, C.K. (2008). Â“ Goliath grouper Epinephelus itajara sound production and movement patterns on aggregation sites,Â” Endang. Species. Res. 7:229-236 Maryland Department of Natural Resources (2009) Accessed 8 December www.eyesonthebay.net
67 Mok, H.K., Gilmore, R.G. (1983). Â“Analy sis of sound producti on in estuarine aggregations of Pogonias cromis Bairdiella chrysoura and Cynoscion nebulosus, Â” (Sciaenidae). Bull Inst Zool Academica Sinica 22:157-186 Murphy, M.D., and Taylor, R.G. (1989). Â“Rep roduction and growth of black drum, Pogonias cromis in northeast Florida,Â” Northeast Gulf Sci 10:127-137 Nieland, D.L., and Wilson, C.A. (1993). Â“R eproductive biology and annual variation of reproductive variables of black drum in the northern Gulf of Me xico,Â” Trans. Am. Fish. Soc. 122(3):318-327 Peebles, E.B., and Tolley, S.G. (1988). Â“Distr ibution, growth and mo rtality of larval spotted seatrout, Cynoscion nebulosus : A comparison between two adjacent estuarine areas of southwest Florida,Â” Bull Mar Sci 42:397-410 Peters, K.M., and McMichael, R.H. (1990). Â“E arly life history of the black drum Pogonias cromis (Pisces: Sciaenidae) in Tampa Bay, Florida,Â” Northeast Gulf Sci 11:39-58 Rubow, T.K., and Bass, A.H. (2009). Â“Reproduc tive and diurnal rhyt hms regulate vocal motor plasticity in a teleos t fish,Â” J Exp Biol 212:3252-3262 Sutter, F.C., Waller, R.S., and Mcllwain, T.D. (1986). Â“Species profiles: life histories and environmental requirements of coastal fisheries and invertebrates (Gulf of Mexico)-black drum,Â” U.S. Fish Wild l Servo BioI Rep 82(11.51, U.S. Army Corps of Eng. TR EL82-410 p. Saucier, M.H., and Baltz, D.M. (1993). Â“Spa wning site selection by spotted seatrout, Cynoscion nebulosus, and black drum, Pogoni as cromis, in Louisiana,Â” Environ Biol Fish 36:257Â–272
68 Sokol, R.R. and Rolf, J.F. (1995). Biometry (third edition). WH Freeman, New York Takemura, A., Takita, T., and Mizue, K. ( 1978). Â“Studies on the underwater sound VII: underwater calls of the Japanese marine dr um fishes (Sciaenidae),Â” Bull Jpn Soc Sci Fish 44:121Â–125 Wells, B.K., and Jones, C.M. (2002). Â“Yield -per-recruit analysis for black drum, Pogonias cromis along the east coast of the Un ited States and management strategies for Chesapeake Bay,Â” Fish Bull 99:328-337
69 Table 4.1. Correlation results of total acous tic energy (TAE) and maximum sound pressure level (Max SPL) and chorus duration a nd Max SPL for acoustic time series data reco rded at each study site. Mean background sound pressure levels and chorus thresholds are calculated from band level measurements of 100-200 Hz. mean and stdv TAE, TAE, chorus duration,chorus duration, of background chorus thres hold Max SPL Max SPLMax SPL Max SPL Site Dates dB SPL (re: 1 Pa) dB SPL (re: 1 Pa)d.f. r p r p PG 3/22/04 5/3/04 85.0 (4.1) 93.2 25 0.97 < 0.001 0.76 < 0.001 PG 12/12/04 5/4/05 85.7 (4.3) 94.3 82 0.98 < 0.001 0.79 < 0.001 CC 1 2/12/05 4/6/05 91.6 (3.2) 98.0 42 0.99 < 0.001 0.93 < 0.001 CC 2 2/12/05 5/6/05 90.7 (3.8) 98.3 1020.99 < 0.001 0.75 < 0.001 CC 2 12/3/05 6/7/06 90.0 (3.5) 97.0 42 0.99 < 0.001 0.85 < 0.001 CC 3 2/12/05 4/6/05 93.1 (3.5) 100.1 39 0.95 < 0.001 0.59 < 0.001
70 Table 4.2. Correlation results of chorus para meters, regression results of chorus star t time and time of sunset, and compariso ns of ascending and descending slopes of chorus amp litude using the Mann-Whitney test for acousti c time series data recorded at each study site. Asterisks denote a Spearman correlation was performed in stead of a Pearson correlation due to non-normality of data. chorus start time, chorus start time, chorus start time chorus start time, ascending slope descending slope Ma nntime of sunset chorus end time time of max SPL max SPL to -6db max SPL from -6dB max SPL Whitney Site Dates d.f. r^2 P r p r P r p mean stdv mean stdv p PG 3/22/04 5/3/04 25 0.41 <0.001 -0.65 0.001 0. 28 0.12 -0.67 < 0.001 2.1 0.8 1.5 0.85 0.003 PG 12/12/04 5/4/05 82 0.02 0.33 -0.56 < 0.001 0.19 0.1 -0.74 < 0.001 2.4 1.44 1.7 1.1 0.004 CC 1 2/12/05 4/6/05 42 0.54 <0.001 -0.34 0.03 0.41 0.006 -0.78 < 0.001 3.1 1.32 1.7 0.62 <0.001 CC 2 2/12/05 5/6/05 102 0.43 <0.001 -0.28 0.07 0. 38 0.01 -0.78 < 0.001 2.9 0.95 2.3 0.87 0.006 CC 2 12/3/05 6/7/06 42 0.04 0.06 -0.61 < 0.001 0. 02 0.81 -0.85 < 0.001 3 1.42 1.8 0.9 <0.001 CC 3 2/12/05 4/6/05 39 *0.39 <0.001 *-0.27 0.1 *0.31 0.06 *-0.56 < 0.001 1.8 0.85 1.2 0.88 <0.001
71 Tables 4.3-A, 4.3-B, 4.3-C. Monthly means a nd standard deviations of chorus start tim e (4-A), time of maximum SPL (4-B), and chorus end time (4-C) for acoustic time series data recorded at each study site 4.3-A Chorus Start Time December January February March April ALL Site Dates mean stdv mean stdv mean stdv mean stdv mean stdv mean stdv PG 3/22/04 5/3/04 19:30 0:2620:200:2720:02 0:36 PG 12/12/04 5/4/05 21:130:3419: 441:2819:43 0:5621:120:3620:12 1:15 CC 1 2/12/05 4/6/05 17:400:5018:32 0:5420:450:0718:19 1:02 CC 2 2/12/05 5/6/05 17:590:3718:23 0:4120:381:1218:26 1:00 CC 2 12/3/05 6/7/06 20:201:2917:391:04 17:411:0618:54 0:4220:391:3018:23 1:20 CC 3 2/12/05 4/6/05 18:190:3118:59 0:55 18:48 0:56 4.3-B Time of maximum SPL December January February March April ALL Site Dates mean stdv mean stdv mean stdv mean stdv mean stdv mean stdv PG 3/22/04 5/3/04 21:40 0:5021:300:3921:34 0:42 PG 12/12/04 5/4/05 22:591:1222: 430:5722:19 0:4722:310:4422:35 0:55 CC 1 2/12/05 4/6/05 21:001:0321:13 0:5822:551:1721:11 1:04 CC 2 2/12/05 5/6/05 21:200:5621:32 0:4621:400:3421:29 0:49 CC 2 12/3/05 6/7/06 21:430:5022:281:12 21:590:5621:40 0:3621:501:0122:01 0:59 CC 3 2/12/05 4/6/05 21:500:5721:16 1:18 21:50 1:08
72 4.3-C Chorus End Time December January February March April ALL Site Dates mean stdv mean stdv mean stdv mean stdv mean stdv mean stdv PG 3/22/04 5/3/04 23:56 0:3522:450:3823:10 0:50 PG 12/12/04 5/4/05 23:500:511: 03 0:520:54 0:5923:270:440:30 1:05 CC 1 2/12/05 4/6/05 1:02 0:410:38 1:0223:540:340:39 1:06 CC 2 2/12/05 5/6/05 1:00 0:341:35 0:4823:030:230:41 0:52 CC 2 12/3/05 6/7/06 23:561:143:29 1:33 2:24 1:001:06 0:4223:201:061:54 1:49 CC 3 2/12/05 4/6/05 0:30 0:350:09 0:45 0:14 0:46
73 Table 4.4. Monthly means and standard devi ations of maximum sound pre ssure band level measurements of 100 Â– 200 Hz and dates of the last black drum chorus event and last recorded black drum calls (sub-thres hold levels) at each study site and season. December January February Ma rch April May Last Last Site Dates mean stdv meanstdvmeanstdv meanstdv meanstdvmeanstdvChorus Call PG 3/22/04 5/3/04 111.012.6 99.1 6.5 91.0 0.0 4/21/20045/1/2004 PG 12/12/04 5/4/05 88.4 3.4 101.911.6116.78.9 113.47.8 98.3 7.4 90.1 0.4 4/22/20054/27/2005 CC 1 2/12/05 4/6/05 126.94.2 116.812.3 105.25.5 4/4/2005 4/5/2005 CC 2 2/12/05 5/6/05 128.53.1 118.613.7 96.6 6.4 96.9 0.7 4/4/2005 4/5/2005 CC 2 12/3/05 6/7/06 96.6 7.0 129.48.6 130.74.5 123.84.8 98.3 6.0 96.6 1.7 4/6/2006 4/10/2006 CC 3 2/12/05 4/6/05 119.06.5 111.57.7 101.02.0 4/4/2005 4/5/2005
74 Table 4.5. Correlation results of chorus para meters from concurrently recorded acous tic time series data at all study sites. The Cape Coral sites were more highly corre lated with one another than with the Punta Gorda site. This indicates that the Cape Coral population of black drum may function on a ne twork level of acoustic comm unication where individuals function as nodes. Asteri sks denote alpha values adjusted for experiment-wise error. Sites Chorus Start Time Chorus End Time Chorus Duration Time of max SPL r p d.f.r p d.f.r p d.f.r p d.f. PG, CC 1 -0.06 0.71 40 0.31 0.05 39 0.15 0.35 40 0.210.21 39 PG, CC 2 0.37 *0.02 40 0.29 0.07 40 0.41 0.01 40 -0.080.64 40 PG, CC 3 0.16 0.34 38 0.32 *0.05 38 0.26 0.11 38 0.120.48 38 CC 1, CC 2 0.72 <0.001 42 0.57 <0.001 42 0.79 <0.001 42 0.190.24 42 CC 1, CC 3 0.72 <0.001 39 0.38 0.02 39 0.52 <0.001 39 -0.020.91 39 CC 2, CC 3 0.77 <0.001 39 0.61 <0.001 39 0.69 <0.001 39 -0.100.53 39
75 Figure 4.1. Study site locations within the estuarine canal systems of Punta Gorda and Cape Coral were acoustic r ecordings of black drum sound production were made. In Cape Coral, green is the location of site 1, re d is the location of s ite 2, and yellow is the location of site 3. Blue repr esents the location of the sing le study site in Punta Gorda, Florida.
76 1000 1000 1200 1400 1600 1800 2000 2200 0000 0200 0400 0600 0800 70 80 90 100 110 120 130 140 Time of DaySound Pressure Level dB (re: 1uPa) Figure 4.2. Diel periodicity of black dr um sound production demonstrated by a 24hoverlay plot of data recorded from 2/15/05 Â– 3/16/05 at CC 2. Sound production increases dramatically during the late after noon/early evening and is sustained well above daytime background levels for many hours during peak spawning season. The grey portion of the diel color bar at the top of the figure bar repres ents the range of sunset and sunrise over range of dates th ese data were recorded.
77 2/13/05 2/23/05 3/5/05 3/15/05 3/25/05 4/4/05 1500 2000 0100 0600 Time of Day 1/6/05 1/25/05 2/14/05 3/6/05 3/26/05 4/15/05 1500 2000 0100 0600 Time of Day 12/11/05 12/30/05 1/19/06 2/8/06 2/28/06 3/20/06 4/9/06 1500 2000 0100 0600 DateTime of Day CC 2 2005-2006 PG 2005 CC 3 2005 Figure 4.3. Times of chorus start (empty circle), chorus end (fille d circle), and sunset (line) represented by select time ser ies data. Chorus start times were negativ ely correlated with chorus end times and so ear lier chorus start times generally corresponded to later chorus end times and longer chorus duration. Regressions of chorus start time and suns et demonstrated weak associations betwee n these parameters over longer time series and moderately strong and positive associati ons over the shorter time series which beg an in the middle or later part of the spawning season.
78 16:00 20:00 00:00 04:00 Time of Day CC2 2005-06 CC2 2005 CC1 2005 CC3 2005 Dec Jan Feb Mar Apr 18:00 20:00 22:00 00:00 02:00 MonthTime of Day PG 2004-05 PG 2004 Figure 4.4. Monthly means and standard deviat ions of chorus start and end times and time of max sound pressure levels (SPL) fo r all time series. Chorus start times are indicat ed by the bottom series of lines, chorus end times by the top series of lines, and time of max SPL by the middle series of lines in each figure. Mean chorus st art and end times were more vari able than the mean time of max SPL.
79 75 100 130 75 100 130 75 100 130 75 100 130 Sound Pressure Level dB (re: 1uPa) 75 100 130 12/4 1/1 2/1 3/1 4/1 5/1 6/1 75 100 130 DatePGI 2004 CC 3 2005 CC 1 2005 CC 2 2005 PGI 2004-2005 CC 2 2005-2006 Figure 4.5-A. Acoustic time series data from all study site locations and years. Ten seconds of acoustic da ta were recorded e very ten minutes. Sound pressure levels were calcula ted as the band level of 100 Â– 200 Hz. Incr eased sound pressure levels during the l ate winter and early spring are consis tent with patterns of reprodu ctive readiness of black drum during their spawning season in th e Gulf of Mexico. Patterns in sound production are similar between years for each study site.
80 Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul 0 2 4 6 8 10 12 Mean GSI (+/1 sd) 80 100 120 140 Mean Max SPL (+/1 sd) 80 100 120 140 Mean Max SPL (+/1 sd) CC2 2005-2006 CC2 2005 CC3 2005 CC1 2005 PG 2004-2005 PG 2004 Female 1986-1987 Male 1986-1987 Figure 4.5-B: Monthly means and standard de viations of sound pressure level data fr om all sites and years (top two figures) al ong with gonado-somatic index (GSI) data reprinted from Fitzhugh et al., 1993 (bottom figure). Patter ns in the timing and levels o f black drum sound production are in general agreement w ith patterns in GSI data collected from coastal waters of Louisiana. Peak GSI data from coastal Louisiana occur sli ghtly later in the year than peak levels of s ound production recorded in southwest Florida due to the influence of latitude and temperatur e on spawning and sound production, which o ccur earlier at lower latitudes.
81 12/4 12/23 1/12 2/1 2/21 3/13 4/2 4/21 80 100 120 140 Sound Pressure Level dB (re: 1uPa) 12/4 12/23 1/12 2/1 2/21 3/13 4/2 4/21 16 18 20 22 24 26 2005 2006Temperature C max SPL CC2 2005-06 Bottom water temp. Surface water temp. Figure 4.6. Nightly maximum sound pressure level data from the CC 2 2005-2006 time seri es (top figure) and bottom and surface water temperature data (bottom figure). Ten seconds of acoustic data were recorded every ten minutes during the period of 12/4 /05 Â– 4/10/06. Sound production was most highly corre lated with bottom water temperatures on a seasonal basis (r = -0.81), and was greatest when surface and bottom temperatures were approximately 18 to 22o C.
82 1600 1800 2000 2200 0000 Time of Day 1600 1800 2000 2200 0000 Time of Day 1600 1800 2000 2200 0000 Time of Day 2/13 2/23 3/5 3/15 3/25 4/4 1600 1800 2000 2200 0000 Date 2005Time of DayCC 2 2005 CC 3 2005 CC 1 2005 PGI 2005 Figure 4.7. Concurrently recorded data of chorus start time from all study site loca tions are shown. Correlations were greate st among the Cape Coral sites for chorus timing parameters (see Table 4.5) which may indicate the population of black drum in Cape Coral function on a network level of communication where information is transmitted th roughout the population via individuals acting as nodes.
83 CHAPTER FIVE Localization and Source Level Estimates of Black Drum ( Pogonias cromis ) Calls INTRODUCTION The knowledge base associated with th e study of sound producing fishes has expanded greatly over the past several decad es. This is due in part to increased recognition of the extent to which teleos t fishes use sound and the practicality demonstrated by passive acoustic methods for documenting reproductive behavior of soniferous fishes. Increased development of recording technologies and data processing algorithms have lead to the production of l ong term, high temporal resolution time series data of fish sound production on daily and seasonal time scales (Locascio and Mann, 2008; Mann et al., 2008; Mann and Grothues, 200 9). Important advances in the field have been the identification of species-speci fic sounds, the behavioral context in which they are produced and the vari ety of anatomical features and physiological and neuronal processes responsible for sound production (Winn, 1964; Fish and Mowbray, 1970; Fine et al., 1977; Mok and Gilmore, 1983; Fay and Simmons, 1999; Ladich and Popper, 2004). Considerable effort has also been dedicated to understanding the auditory sensitivity of fishes. Early experiments i nvolved behaviorally c onditioned responses to stimuli and in 1998 Kenyon et al. published th e first auditory brainstem response (ABR) results for a fish. Since the time of KenyonÂ’ s ABR publication the te chnique, referred to
84 now as auditory evoked potential (AEP) has been repeated on a wide variety of fishes to create audiograms. Despite the extensive research on hearing and sound production, few studies have investigated the parameters necessary for li nking these aspects in a way that allows communication ranges to be estimated. Ther e have been studies on the propagation of damselfish sounds (Mann and Lobel, 1997; Ma ruska et al., 2007), oyster toadfish (Fine and Lenhardt, 1983), midshipman (Bass and Clark, 2003) two species of freshwater gobies (Lugli and Fine, 2003), and one study on the source levels of oyster toadfish (Barimo and Fine, 1998). There have been no reports of propagation of sciaenid sounds (Mann et al., 2008) and only one estimate of a potential range of source levels from a single silver perch call (Sprague and Luczkovich, 2004). Source level estimates along with habitat specific signa l transmission loss, background sound levels and hearing sensitivit y data are required for estimating the communication range of soniferous fishes. Sour ce level data are also required in passive acoustic studies where estimates of the number of vocalizing individuals within the active space of a hydrophone are desired. Estimates of source levels along with signal transmission loss can be made using an ar ray with a minimum of 3 hydrophones. Signal arrival time differences between each hydrophone are used to localize the position of the source. The resulting distances betw een the source and hydrophones are regressed against the received levels at each hydrophone and this provides an estimate of signal transmission loss. By definition a source le vel is the sound pressure level (SPL) at 1 meter away from the source. Therefore, the transmission loss estimate (slope of regression) is used together with the sourceÂ’s location to ba ck-calculate the SPL at the
85 required 1 meter distance from the source. Th is method also has the potential to track the movements of vocalizing fishes and there by gain important behavioral information associated with the sound production. Thes e localization technique s have been used a great deal with marine mammals to both tr ack them and estimate source levels (Janik, 2000; Clark and Ellison, 2000; Hayes et al., 2000; Gedamke et al., 2001; Mellinger, 2005; Miller, 2006). Canal side residents of Cape Cora l, Florida have long complained of loud, low frequency Â‘boomingÂ’ sounds occurring in thei r homes after dark during winter months. Our suggestion to the city that a species of fish may be the source was received with mixed feelings. Upon conducting an acoustic su rvey of the canal systems we were able to confirm that indeed a population of voca lizing black drum inhabited the canals and was responsible for the sonic disturbance experienced by some community members. Black drum are a large, demersal fish in the Sciaenid family that are well-known for producing sounds. More importantly, we discove red an excellent field site for studying black drum sound production and associ ated behavior in great detail. The main goals of this study were to measure source levels and propagation of black drum calls and to devel op an algorithm for automatic de tection of calls and specific peaks within calls for localizing the positi on of calling fish. We also measured the hearing of a black drum to estimate potent ial acoustic communication range based on source level, propagation loss, back ground levels, and hearing thresholds. METHODS A. Field Recordings
86 Recordings of black drum were ma de from a dock in a sea-walled estuarine canal system of Cape Coral, Florida (2633 '3.00"N 8159'18.50"W)(Fig. 5.1) on March 13, 2005 using a four-element linear array of High Tech Inc. 96-min-series hydrophones (sensitivity: -164 dB re: 1V/1Pa and flat frequenc y response of 2 Hz-37 kHz ) connected to a calibrated Alesis adat HDXR 24 multi-track hard disk recorder which sampled at 44.1 kHz and 24-bit resolution. H ydrophones were spaced relative to each other at 0, 8, 24, and 56 meters. A fifth, laterally offset hydrophone was used only for resolving the left-right ambigu ity of the array. Hydrophone cables were anchored to the bottom with lead weights and hydrophone elements were buoyed 0.5 meters above the bottom with a small foam collar. Water depth at the site was seven meters and the bottom was a composite of soft clayey-silt. Acoustic data were recorded continuously from 1900 Â– 2100 hours. Temperature, salinit y, and depth data were recorded with a Eureka Manta multi-probe and used to cal culate the speed of sound in the water (Mackenzie, 1981). The water column wa s well mixed (i.e. not stratified) and atmospheric conditions were clear and calm during recordings. B. Automatic Signal Detection and Localization In order to estimate source levels (SL) the distances betw een the source and the hydrophones must be reliably known. In this study such distance estimates were made by localizing the sourceÂ’s position with a MATLAB routine called Ishmael 1.0 (Mellinger, 2001). This program uses a hype rbolic function based on differences in signal arrival times at each hydrophone and a l east-squared-error fit to estimate the X-Y position of the source. Required input fo r the localization algorithm includes the
87 hydrophone positions in Cartesian coordinates, the speed of sound in the water, and the signal arrival time differences between hydropho nes. To determine signal arrival time di fferences, a detection algorithm was created using MATLAB v7.5.0 to automatically find the same peak within an individual call on the four different recording tracks and then calculate the signal arrival time differences between them. An illustrated summary of the steps used in this process are shown in figures 5.2-A and 5.2-B. This program worked by first detecting individual calls and then using the results as reference points, dete cted a target peak with in the first part of individual calls. To detect calls, time series of acoustic data were first low-pass filtered using a ten-point moving average and the resu lting time series was rectified (absolute values calculated) and then enveloped using a 4410 point moving average. This signal processing has the effect of making the target of interest more easily detectable because the variability in amplitude over time is diminished leaving a broadened, more generalized peak (e.g. individua l calls) available for detection. An amplitude threshold was used to determine the index where the signal amplitude exceeded this level. The resulting index values from the call detection process were used as the basis for peak detection within individual calls, b ecause it was typically the largest peak. The first negative peak of each call was target ed. The search for the target peak was conducted on the unfiltered time series data and based on an amplitude threshold setting and a range of index values occurring prior to and after the index va lue of the reference peak from the call detection process. Th e MATLAB function, Â‘findpeaksÂ’ was used to identify and return the index value of the firs t negative peak within the search range that exceeded the amplitude threshold. This approach required some fine tuning of the search
88 criteria (i.e. amplitude and range of point s searched) because th e call detection is performed on filtered data and peak detection is performed on unfiltered data. Also the signal-to-noise ratio may change or the received level (RL) or calling rate may change over the course of the acoustic recordings. It was most effective to detect the peaks on the track with the highest RL for target signal first and then use these results to search for the corresponding peak on other tracks. B ecause the inter-hydrophone distances, speed of sound, and sample rate were known we c ould limit the search for the corresponding peak on the other tracks. Fo r example, if two hydrophones ar e eight meters apart, the speed of sound is 1520 m/s, and the sample rate is 44.1 kHz, then the number of search points to find the target on the second track must be limited to 232 (i.e. 8/1520*44.1=232). From these detected peaks, signal arrival time differences between hydrophones were calculated and used in the localization algorithm to estimate the location of calling fish. The detection program also calculate d the RL of calls for each track as peak and RMS dB re: 1Pa SPL from the unfiltered time series data, using the equation: 20*log10(signal) + 164 + 20, where: Â‘signalÂ’ is RMS or peak of the call, +164 is the hydrophone calibration (dBV/Pa), and +20 is the Alesis calibration from a 1V input signal. The RMS calculations were limite d to the first 15,000 points of each call (340 ms) to avoid the possibility of including en ergy from a temporally overlapping call of another fish. Call duration was measured fo r 50 calls that did not contain energy from the overlapping call of another fish. The RMS level of these 50 calls calculated from the first 15,000 points was compared to the RMS leve l calculated from th e entire call.
89 The uncertainty associated with the estimated X-Y location of the source is reported by Ishmael as mean error in milliseconds, which we converted to meters by multiplying the value by the calculated sp eed of sound in the water (1520 m/s) and dividing by 1000. If the localization erro r for a call was greater than 1.5 m the localization was reattempted using a cu stom MATLAB program,Â’HotWav2Â’ which incorporates the localization algorithm used in Ishmael. Hotwav2 allo ws the user to view multi-track waveforms, manually select the targ et peak on different tracks and instantly view and save the localization re sults. If the localization er ror was still greater than 1.5m the results were not used and the X-Y positi on was linearly interpol ated from the prior and following X-Y positions of localized cal ls. The total distance swam and mean swimming speed were calculated by summing th e distances between successive localized X-Y positions of an individual calling fish recorded over a 59 minute period. For this same individual we calculated the calling period from the results of the call detection algorithm. A total of 81 calls from 5 other individual fish were localized using the semiautomatic HotWav2 routine. C. Source Levels Received levels (RL) were regressed against distance using the log10(distance) model which is commonly referred to as th e cylindrical spreading loss model when used for shallow water propagation. A variety of alternative models were also tested and a square root(distance) model was determined to provide the best overall fit. This fitted regression equation of the square root model was used to estimate SLs as the y-intercept at 1 meter. A separate SL estimate was calculated from each of the four pairs of RL and distance data per call. The mean of these f our estimates from each call was used as the
90 single SL estimate of the call. SLs were not estimated for calls in which the X-Y position was linearly interpolated. SLs were estimated from calls of 6 different individual fish. D. Call Energy and Propagation The relative concentr ation of acoustic en ergy within individual black drum calls was analyzed with a 15,000-point Fast Fourie r Transform (FFT) to create power spectra with 3 Hz wide bins. From this we dete rmined the frequency bin with the greatest concentration of acoustic energy and the fundamental frequency ( n ) along with associated harmonics. For each frequency, n to 3 n the SPL was regressed against distance to determine the model which provided the best fit to the data. The slope of the best fitting model was used as the transmission loss (TL) es timate. Data used for this analysis were calls produced by a single individual (n=944) at many different locations and therefore distances from the hydrophones. FFT results from a sub-sample of these calls (n=50) were plotted based on four different mean distance groups of 2.4 (SD=0.47), 8.9 (SD=0.85), 15.5 (SD=0.97), and 47.4 (SD=0.96) me ters. Background levels and signalto-noise ratios (SNR) were estimated from these same 50 calls. Background levels were calculated as the dBRMS re: 1 Pa SPL of the 50 ms immediately preceding the signal (e.g. fish call). The SNR was calculated as RMS SPL signal Â– RM S SPL background. A theoretical cutoff frequency for the study area was calculated using the absolute cutoff frequency equation for a non-rigid lower bounda ry (Urick, 1983, Rogers and Cox, 1988). The sound speed value of 1549m/s for clayey-s ilt was used in the equation (Hamilton and Bachman, 1982). E. Auditory Evoked Potentials
91 A single black drum (FL= 42 cm), of undetermined gender was collected with a seine net and tested for auditory sensitivity using the auditory evoked potential method (AEP). The fish was kept in an aquarium overnight and tested the following day. The experiment was performed inside an a udiology booth in a cylindrical steel tank (height=120 cm, diameter=50 cm, and wall thickness=1.5 cm) which was closed at one end, oriented upright and filled with 30ppt oxyge nated seawater. During the experiment the fish was secured in a soft mesh harness, which restricted movement but allowed for normal respiration, and was suspended from a la boratory stand so that the top of the head was 10 cm below the water surface. Three subdermal stainless steel needle electrodes (Rochester Electro-Medical) were used fo r the experiment. The signal recording electrode was inserted approximately 1.0 cm in to the medulla region of the head, adjacent to the midline. A reference electrode wa s inserted the same depth into the dorsal musculature and a ground electr ode was placed directly in th e water within 10 cm of the fish. Acoustic stimuli and AEP waveform recordings were conducted following the methods of Egner and Mann (2005) but w ith the following modifications: acoustic stimuli were presented as phase alternating 50 ms pulsed tones at 75, 150, 300, 450, 600, and 900 Hz and 1000 signal presentations were averaged to measure the evoked response at each level and frequency using 5 dB attenuatio n steps. Auditory sensitivity thresholds for each frequency were defined as the presen ce of a peak at twice the frequency of the acoustic stimulus and a minimum of 3 dB above background levels.
92 RESULTS The call detection algorithm successf ully identified 962 sequential calls produced by an individual fish over a 59 minute period. Of this total, automatic peak detection on all four tracks and localization of the sour ce to within 1.5 m of error was successful for 805 calls. Reattempted localization using HotWav2 was successful for 139 of the remaining 157 calls. The X-Y positions of the remaining 18 calls were linearly interpolated. The square root model provided a bett er overall fit to the data compared to other models including the conventional cylindrical spreading loss model (Fig. 5.3), which is often the default choice for estimating TL in shallow water in the absence of empirical data. The square root model resulted in a slightly higher r-squared value and was centered better on the data, especially for ca lls close to a hydrophone so it was able to provide more accurate SL estimates. This is evident when comparing the RL and location of the closest call to a single hydrophone to the SL s estimated for this call by each model. For example, the highest RL was 166 dBRMS re: 1Pa SPL and was located 0.95 m from the nearest hydrophone. The square r oot model for all calls predicted an SL of 168 dBRMS re: 1Pa SPL and the cylindrical model predicted an SL of 174 dBRMS re: 1Pa SPL. A total of 1,025 SL estimates were calculated from calls pro duced by six different individual fish over a 59 minute period. SL estimates ranged from 159 to 174 dBRMS re: 1Pa SPL and 171 to 185 dBpeak re: 1Pa SPL and averaged 165 dBRMS re: 1Pa SPL (SD=1.1) and 175 dBpeak re: 1Pa SPL (SD=0.7). Variability in SL estimates between individuals was low (Table 5.1). Mean call duration was 600 ms (n = 50, SD = 22). RLs
93 based on the first 15,000 sample points of a call (or 340ms) were on average 2 dB higher than RLs based on the entire call (600ms) a nd variability was quite low (n=50, SD = 0.02). Black drum produced tonal calls having multiple harmonics. Throughout the course of the call, durations of positive and negative peaks increased while cycle periods decreased and an additional drop in amplitude was evident at about 300 ms (Fig. 5.4-A). The majority of energy was concentrated in the fundamental frequency (94 Hz) and first two harmonics at 188 Hz and 282 Hz (Fig. 5.4-B) The greatest concentration of acoustic energy was most commonly found in the 94 Hz frequency bin, although this became somewhat more variable as distance from the source to the hydrophones increased (Fig. 5.5). Energy levels of higher order harmonics were much lower regardless of distance from source to receiver. The square root model provided the be st fit for RL vs. distance for the entire call, fundamental frequency and first harmonic. A logarithmic model provided the best fit to the second harmonic (Fig. 5.6-A and 5.6-B). Background levels ranged from the mid 120s to mid 140s dBRMS re: 1Pa SPL at each of the four distance groups (Fig. 5.7-A) and most of the background noise can be attr ibuted to other calling black drum. SNRs also ranged approximately 20 dB for each distance group and there was considerable overlap in SNR between each group (Fig. 5.7-B). The maximum and minimum SNR were 39 dB at 0.95 m and -1.7 dB at 47 m. The theoretical cut-off frequency equation predicted that frequencies below 282 Hz s hould not propagate, however frequencies well below (e.g. 94 Hz) this estimated cu toff did propagate.
94 From the 962 calls produced by an individual fish over 59 minutes we calculated a mean calling period of 3.6 seconds (SD=0.48) This calculation included a 28 second pause in calling after 39 minutes, 7 seconds of steady calling. Th e total distance swam over the 59 minute period was estimated at 1,035 m and occurred in an area approximately 20 m x 40 m. Swimming beha vior included roaming throughout this area and occasionally performing recurrent loopi ng and backtracking patterns over distances of about 5 Â– 10 m (Fig. 5.10-A and 5.10-B). The mean swimming speed was 0.3 m/s (SD=0.15) or about 0.5 body lengths per second based on size at maturity estimates of approximately 0.6 m fork length for males (Fitz hugh et al., 1993; Niel and et al., 1993). Results of the AEP experiment indicated that of the frequencies tested the fish was most sensitive to those below 450 Hz and the grea test auditory sensitivity occurred at 300 Hz with a 94 dBRMS re: 1Pa SPL threshold (Fig. 5.8). DISCUSSION Although SLs have been reported for several species of marine mammals, data on SLs of fishes are very rare. In fact, we are aware of only one ot her published report on fish SLs in which calls of oyster toadfish were recorded at one meter away from artificial nest sites. SL estimates aver aged 126 (SD=2.7) and 123(SD=4.5) dB RMS re: 1Pa SPL for toadfish boatwhistle and agonistic gr unts, respectively (Barimo and Fine, 1998). Sprague and Luczkovich (2004) estimated a possible SL range of 128-135 dBRMS re: 1Pa SPL for a single silver perch call record ed from an ROV. The mean SL estimate of black drum calls in our study was 165 dBRMS re: 1Pa SPL (SD =1.0) which has approximately 8000x greater intensity than estimates for toadfish and 1000-5000x greater
95 intensity than that of the silver perch call. Table 5.2 shows how SL estimates of black drum compare to those of various marine mammals reported in the literature. Vocalization types presented in this table are used for soci al interaction as opposed to echolocation clicks commonly used for interr ogating the environment and/or foraging, and which may typically exceed 200 dBpeak-peak re: 1Pa SPL (Rasmussen, 2002; Madsen and Wahlberg, 2007). In the study by Barimo and Fine, reco rdings were made at precisely 1 meter away from five artificial nest si tes placed on the bottom which were inhabited by vocalizing male toadfish. The study was conducted in a natural, yet experimentally well controlled environment. Their results demonstrated that the variability in call parameters, including SL estimates of individuals, likely reflect ed real differences in the levels of sound produced as opposed to differences introdu ced through experimental error or via propagation effects. Black drum are demers al, however we cannot be certain at what depth the fish were when they called because we used a one dimensional horizontal linear array. The depth of the sound source and receiv er can greatly influen ce the variability of call parameters over distance in shallow wa ter (Forrest et al., 1993 Mann and Lobel, 1997). Although we cannot be certain that the va riability in estimates of black drum SLs relates as well to actual differences in the levels produced, the degree of variability is similar to that of the toadfish levels record ed by Barimo and Fine. This comparison is a bit complicated because toadfish measurements were made directly at 1 meter and the black drum measurements were back-calcu lated to 1 m using the fitted regression equation. Also, while our SL estimates were from six different indi viduals, 92% of calls analyzed were produced by a single fish.
96 Propagation of sound in shallow water is a notoriously complex phenomenon. Transmission loss (TL) is fr equency dependent and influe nced by a wide range of variables from the condition of the sea surf ace to the physical/chemical properties of the water to the composition of the bottom and the va riety of structural objects that may exist in between. The cylindrical spreading loss model can provide a rough estimate of TL in shallow water where multipath effects are exp ected but site-specific information about factors affecting propaga tion are unknown (Urick, 1983; Lurton, 2002). Use of this model is well represented in the literatur e and it is common to find results that lie somewhere between the cylindrical (3 dB loss per distance doubling) and spherical spreading loss models (6 dB loss per dist ance doubling) (e.g. Ma nn and Lobel, 1997). When fit to our data (of the entire call) this model estimated a transmission loss of 5.6 dB per distance doubling. This model did not, howev er, provide the best fit to the data and tended to overestimate SLs by about 6 dB co mpared to the square root model. Signal attenuation on average was great est in the first few meters for frequencies of 94, 188, and 282 Hz although only the 282 Hz band behaved more in accordance with the expected log based cylindrical/spherical sp reading loss models. Given that the cut off frequency was calculated to be 282 Hz this result is not surprising. What is surprising is the propagation of the 94 Hz and 188 Hz bands which are well below the estimated cut off frequency. This may be due to the fact that the acoustic impedance of the water, which has a sound speed of 1520 m/s, and that of the bottom sediment, which has a sound speed estimated at 1549 m/s are similar e nough so that the phys ical bottom is not effectively the acoustic bottom. Fine and Le nhardt (1983) explained similar results this way in their study on propagation of toadfish calls made over a soft bottom. The cut-off
97 frequency is dependent on depth and bottom type (i.e. rigid vs. soft). A rigid bottom will produce a stricter cut-off, as evidenced in studies of goby sound production where signals attenuated as much as 30 dB over a 5 to 50cm distance in rock lined shallow water streams (Lugli and Fine, 2003 and 2007). A so fter bottom may allow more energy to propagate below the theoretical cut-off as demo nstrated by our data (F ig. 5.5) and by Fine and Lenhardt (1983). We are not certain how th ick the layer of soft bottom sediment at our study site is, but most certainly a layer of limestone exists belo w it at some point and this rigid layer may serve as the acoustic botto m. In this scenario it is likely that a velocity gradient would exist as a result of compaction of sedimentary materials with depth which would also result in the refraction of sound b ack up towards and into the overlying water which is the source medium Under these conditions low frequency sounds have small reflective losses at low gr azing angles (Urick, 1983). If fish were close to the bottom when they were calling, and black drum are demersal, then grazing angles would have been low. The linear array configuration we used was adequate for our purposes of localizing calling fish and estimating SLs and TL. Localization of the source is optimal when the distance between hydrophones is on the same order of magnitude as the distance from the source to the array (Moehl et al., 2003; Madsen and Wahlberg, 2007). The maximum inter-hydrophone distance of our array was 56m and the range of estimated distances from the source to the ar ray was approximately 1 to 50m. We were able to record a large number of black drum calls over a relatively short time period (2 hours) and the detection algorithm performed r easonably well on these data. A key to the success of the algorithm was in its ability to find at least one refere nce point (the first
98 negative peak) per call that could be used to then locate the same poi nt on other tracks. The search for the target peak was based on signal amplitude and a limited time range. This detection strategy worked in most cas es because, as with toadfish (Fine and Thorson, 2008), black drum generally avoided or minimized overlap of their calls. When it failed it was usually because the call was somehow distorted on one of the tracks and the relevant peak could not be identified. This was due to either propagation error in some form or partial call overl ap. In some instances a ca ll on one track was entirely masked on the most distant track by the ca ll of another fish close to the hydrophone associated with this most distant track. Inci dents of fully overlapped calls such as this were rare in our data and only seemed to occur when fish were calling near the opposite ends of the array which would have separa ted them by a distance of perhaps 50m or more. Partial call overlap involved only the end portion of one call, (approximately the last 10-20%) where there was minimal energy, and the beginning of another. This was sometimes problematic because we used the first negative peak of the call in the localization algorithm and this is also why we limited the RMS calculations to the first 15,000 sample points of each call. The semiautomatic algorithm (HotWav2) was useful for overcoming problems associat ed with call overla p (< 10% of the time) and for calls that were far enough away from all hydrophones so that amplitudes of RLs were too low. Avoidance of call overlap or call al ternation has been docum ented for anurans and is believed to serve as a means of male-mal e spacing and territoria l maintenance (Zelick, et al. 1999; Grafe, 2005). Density dependent eff ects have also been associated with this form of calling behavior among anurans. Fo r example, as the density of calling males increases, the distances between individua ls decreases and calling males at greater
99 distances become ignored thereby reducing th eir active space. (Zelick et al., 1999). It may be possible that similar de nsity dependent effects also o ccur in the calling behavior of black drum which would e xplain the incidents of fully overlapped or masked calls. This temporal patterning of calls by black dr um is in contrast to the highly overlapping calls of the sand seatrout ( Cynoscion arenarius ) in which it becomes virtually impossible to identify individuals during chorusing even ts (Locascio and Mann, 2008) recorded with a single hydrophone. Ninety two percent of localized calls were produced by an individual fish at regular intervals while it swam broadside a nd close to the array during our recording. Our interpretation that this was a single animal, rather than multiple animals at close proximity to one another and moving synchr onously, is based on the regular calling rate, short and consistent distances between loca lized positions and the generally directed pattern of movement that was evident when localized positions of successive calls were linked together. A similar interpretation wa s made by Gedamke et al. (2001) for minke whales and Hayes et al. ( 2000) for blue whales from localizations of consecutive vocalizations. Madsen and Wahlberg (2007) al so point out that lo calizing consecutive vocalizations is very useful for validating re sults and improving precision. We believe in the separate identity of the five other indivi dual fish for which we estimated source levels because of the consistency in their locations over short time periods. The more than 1000 SL estimates made in this study demonstrates that given a large enough sample size in an acoustically act ive area, the highest RLs can be expected to approximate or overlap th e range of SLs (Fig. 5.9-A and 5.9-B). The array data also demonstrate that it is possibl e to observe behavior associ ated with calling by mapping the
100 directed movements of fish. Our fortuitously tracked individual swam over 1 kilometer within a relatively small area at a slow, steady rate while calling at regular intervals. During this time the fish would occasionall y swim in a looping pattern several times in one general area then move on to a nearby location and repeat a similar pattern or appear to be pacing back and forth around a central location or object (contac t the authors for an animation of these data) (Fig. 5.10-A and 5.10B). This behavior is reminiscent of lekking but we cannot confirm this because we have no data on the territorial distribution of other males due to the cons traints of our array. Howeve r, an additional ten or so individuals appeared to be involved in the gr oup of calling fish that occurred within the active space of the hydrophones, and if a larger more geometrically complex array had been used it would likely have been possi ble to resolve the locations and movement patterns of these other fish as we ll. Also of interest is the brief interruption in calling that occurred after a 39 minute period when the individual fish paused for 28 seconds and then resumed calling at its previous rate. This is noteworthy because Connaughton and Taylor (1996) described conti nuous calling by male weakfish ( C. regalis ) prior to spawning but a temporary cessation in calli ng during female pursuit and spawning. After gamete release the male often return ed to the bottom and began calling again. Cessation of calling just prior to gamete rele ase has also been documented for red drum (Guest and Laswell, 1978; Ramcharitar et al., 2006) and Atlantic croaker (Connaughton et al., 2003; Ramcharitar et al., 2006). There are few data on calling rates of individual fish or estimates of energetic costs associated with calling by fishes (Mit chell et al., 2008). In contrast to many terrestrial species Amorim et al. (2002) showed that oxygen consumption requirements
101 for calling by the oyster toadfish ( O. tau ) were negligible. Mi tchell et al. (2008) concluded that calling by this species was limited by glycogen de pletion resulting in sonic muscle fatigue and estimated that 300 ms long calls produced at 200 Hz and a rate of 15 times/minute would only be sustainable for about 5 minutes. We documented that an individual black drum was capable of producing a 600 ms call at 94 Hz, on average, every 3.6 seconds (~16 times/minute) for at l east 59 minutes. The nearly twelve fold difference between these species in the amount of time over which calls may be produced might not be too surprising if the black drumÂ’s sound production mechanism was similar to the more efficient single twit ch mechanism of the weakfish ( C. regalis ), a fellow sciaenid (Connaughton et al., 1997; Sprague, 200 0). However, it instead appears to be more similar to that of the oyster toadfish which Fine et al. (2001) described as an inefficient damped system requiring multiple rapid muscle contractions to produce its prolonged boatwhistle call. Evidence for this interpretation includes the apparent slowing of contraction and relaxation times of the s onic muscle along with a decrease in signal amplitude reflecting muscle fatigue (Fig. 5.4-A) This interpretation is also supported by the drop in frequency of the fundamental a nd harmonic bands towards the end of the call (Fig. 5.4-B) (Fine et al., 2001). The midshipman ( P. notatus ) is a unique example of a high endurance sound producer capable of maki ng a continuous Â‘humÂ’ in the 80-100 Hz range for up to one hour (Ibarra, 1983). Dense banks of mitochondria on the sonic muscle fiber and low contracti on rates are attributed to this speciesÂ’ calling ability (Bass and Marchaterre, 1989). The greater calling en durance of the black drum can, in part, be accounted for by a lower sonic muscle contrac tion rate, but a closer look at the muscle fiber, which is intrinsic to the swimbl adder, and swimbladde r morphology would be
102 useful for understanding how calling is sustai ned over long periods and why the fish is able to produce such a high intensity sound. An early investigation of the characteristics of the black drum swimbladder and sonic mu scle first published in 1904 (Harmer, et al., 1922) reported that the Â‘air-b ladder and its muscles const itute the most powerful soundproducing organ yet found in any fish. Unfo rtunately, no further details were given. Mok and Gilmore (1983) originally described three different black drum call types; the staccato, loud drum a nd short grunt. We observed only the loud drum call type in our recordings but when handled in cap tivity the staccato, a disturbance call, was produced. Our description of fundamental frequency and multiple harmonic call structure is consistent with that of Mok a nd Gilmore. However they described the loud drum call as being composed of two distin ct parts, a shorter lower amplitude Â‘boonÂ’ sound followed by a longer higher amplitude Â‘b oundÂ’ sound. An alternative explanation is that this is instead the result of partial ca ll overlap of two fish at different locations. In the single black drum specimen we tested, auditory sensitivity was highest for frequencies below 450 Hz and peaked at 300 Hz. Ramcharitar and Popper (2004) also showed that black drum auditory sensitivity was highest for frequencies below 500 Hz with the greatest average peak sensitivity (although not signifi cantly so) at 300 Hz. Ramcharitar et al. (2006) showed similar auditory sensitivity for two other sciaenids, spot ( L. xanthurus ) and weakfish ( C. regalis ) and Horodysky et al. ( 2008) demonstrated this same trend for six other sc iaenid species, including C. regalis Despite variations in experimental design, which is not unusual fo r AEP experiments, trends in frequency sensitivity were strongly consistent among th ese studies. Mean threshold levels for frequencies of greatest sensitivity generally ra nged from 90 to just over 100 dB re: 1 Pa
103 SPL with the exception of Ramcharitar and PopperÂ’s black drum data which ranged, between 85 and 90 dBRMS re: 1Pa SPL. These differences are not necessarily large considering that 5 dB resolution per frequency was used in each of the studies and that standard deviations of mean threshold values appeared to be of a pproximately this same amount. Despite having a sample size of just one fish our AEP results do seem reasonable in the context of these other studies. There are many examples of fishes with highly correlated relationships between frequencies of highest sound production and frequencies of best hearing and other examples where there is a poor correlation (Ladich, 2000; Popper and Schilt, 2008; Maruska, 2007). For sciaenids, the data de monstrate that there is a correlati on between the frequency of sound production and best he aring (Horodysky et al., 2008). We found the majority of energy in black drum calls concentrated below 300 Hz and greatest at 94 Hz; auditory sensitivity was best below 450 Hz and peaked at 300 Hz. While there is not a direct match in peak frequencies of hear ing and sound production for black drum there is substantial overlap in these ranges (Fig. 5.8). A more detailed i nvestigation of the coevolution of the acoustic communication sy stem of sciaenids is warranted. Our data on SLs, TL, background leve ls and auditory sensitivity of black drum allowed us to estimate the poten tial communication range of this species at our study site. The SPL vs. distance regression was used to estimate the distance at which the SPL would match the upper (146 dBRMS) and lower limits (127 dBRMS) of background sound levels. Based on these parameters we esti mated a range of 33 Â– 108 meters over which one fish could theoretically be detected by another. Under these conditions background levels and not auditory sensitivity limite d the communication range of calling fish.
104 Background levels were dominated by calls produced by other fi sh, so in effect the signal produced by one fish created the background level for the following fish in the calling sequence. Background levels and therefore co mmunication distances fluctuate mainly as a function of temporal call ove rlap and level, which is a function of distance. The consistent range of background levels across all distances reflects the consistency in timing and location of calls produced by diffe rent individual fish. Ramcharitar and Popper (2004) showed a decrease in auditory sensitivity of black drum in the 300-600 Hz range when exposed to broad-band white no ise at 136 dB re: 1Pa. Pure tones are typically used in AEP studies, however fish do not produce pure t ones and in nature broad-band background levels are not realistica lly flat. In addition to standard AEP methods including masking paradigms, behavior ally conditioned responses to recordings of a species own call masked by ambient backgr ound noise recorded in the field would be useful approach for understandi ng auditory sensitivity in black drum and other species. Fine and Lenhardt (1983) pointed out that there was no evidence proving fish actually do communicate over long di stances (>3 m) and to our knowledge this is still the case. The high intensity SLs and minimization of call overlap of bl ack drum make them an excellent candidate for testing this. Grafe (1996) successfully influenced call alternation in African painte d reed frogs by performing playback experiments. If playback experiments could be used to a ffect temporal calling patterns in a group of localized black drum then communication di stances could be estimated, at least for calling males. This is the first study to report both source level estimates and auditory sensitivity for a sciaenid, a family renowned for its s ound producing abilities. The array data also
105 provided new behavioral information associat ed with calling furthe r demonstrating the value and relative simplicity of array record ings. Direct measurements of site specific sound propagation and transmission loss made from array data are critical to the accuracy of SL estimates, (Madsen and Wahlberg, 2007) and can increase knowledge on the nature of sound propagation in complex, shallow wa ter environments. In this study, for example, we found that propagation of black dr um sounds violated the theoretical cut off frequency estimate and RLs of calls at various distances were not best described by the commonly used log10 model. Using the blac k drum as a model, the Cape Coral study area offers exceptional prospects for disc overing more extensive details about the acoustic behavior of individual fish and of groups in the context of communication networks. Future research efforts conducte d at this study site could be designed to provide information for comparative analyses with theory from the better studied acoustic communication behavior of anurans, bi rds and insects. REFERENCES Amorim, M. C. P., McCracken, M. L., and Fi ne, M. L. (2002). Â“Metabolic costs of sound production in the oyster toadfish, Opsanus tau ,Â” Can. J. Zool. 80, 830-838. Au, W.W.L., Pack, A. A., Lammers, M. O., He rman, L.M., Deakos, M.H., and Andrews, K. (2006). Â“Acoustic properties of humpback whale songs,Â” J. Acoust. Soc. Am. 120, 1103-1110. Barimo, J. F., and Fine, M. L. (1998). Â“Rel ationship of swim-bla dder shape to the directionality pattern of underwater sound in the oyster toadfish,Â” Can. J. Zool. 76, 134-143.
106 Bass, A. H., and Clark, C. W. (2003). Â“The physical acoustics of underwater sound communication,Â” in Handbook of Auditory Research edited by A. M. Simmons, A. N. Popper, R. Fay (Springe r-Verlag, New York), pp. 15-64. Bass, A. H., and Marchaterre, M. A. (1989). Â“Sound-generating (sonic) motor system in a teleost fish ( Porichthys notatus ): sexual polymorphism in the ultrastructure of myofibrils,Â” J. Comp. Neurol. 286, 241-253. Clark, C. W., Ellison, W. T., (2000). Â“Calibratio n and comparison of the acoustic location methods used during the spring migration of the bowhead whale, Balaenamysticetus off Pt. Barrow, Alaska 1984-1993,Â” J. Acoust. Soc. Am. 107(6), 3509-3517. Connaughton, M. A., and Taylor, M. H. ( 1996). Â“Drumming, courts hip, and spawning behavior in captive weakfish, C ynoscion regalis,Â” Copeia. 195-199. Connaughton, M. A., Fine, M. L. and Taylor, M. H. (1997). Â“The effects of seasonal hypertrophy and atrophy on fiber morphol ogy, metabolic substrate concentration and sound characteristics of the weakfi sh sonic muscle,Â” J. Exp. Biol. 200, 2449 2457. Connaughton, M. A., Lunn, M. L., Fine, M. L., Taylor, M. H. (2003). Â“Characterization of sounds and their use in two sciaenid spec ies: weakfish and Atlantic croaker,Â” in Listening to fish: passive acoustic ap plications in marine fisheries, edited by R. Rountree, C. Goudey, and T. Hawkins. Massachusetts Inst itute of Technology, Sea Grant College Program, Techni cal Report MITSG 03-2, Cambridge, Massachusetts. Egner, S. A., and Mann, D. A. (2005). Â“ Auditory sensitivity of sergeant major damselfish
107 Abudefduf saxatilis from post-settlement juvenile to adult,Â” Mar. Ecol. Prog. Ser. 285, 213-222. Fay, R. R., and Simmons, A. M. (1999). Â“The se nse of hearing in fish and amphibians,Â” in Comparative hearing: fish and amphibians, edited by R. R. Fay and A. N. Popper (Springer-Verlag, New York), pp. 293-318. Fine, M. L., Winn, H. E., Olla, B. L. (1977). Â“Communication in fishes,Â” in How animals communicate edited by T. A. Sebeok (Indi ana University Press) pp. 472-518. Fine, M. L., and Lenhardt, M. L. (1983). Â“S hallow-water propagati on of the toadfish mating call,Â” Comp. Biochem. Physiol. 76A, 225-231. Fine, M. L., Malloy, K. L., King, C. B., Mitc hell, S. L., and Cameron, T. M. (2001). Â“Movement and sound generation by the toad fish swimbladder,Â” J. Comp.Physiol. A. 187, 371-379. Fine, M. L. and Thorson, R. F. (2008). Â“Use of passive acoustics for a ssessing behavioral interactions in individual toad fish,Â” Trans. Am. Fish. Soc. 137, 627-637. Fish, M. P., and Mowbray, W. H. (1970). Â“S ounds of western North Atlantic fishes,Â” (John Hopkins University Press, Baltimore, Maryland). Fitzhugh, G. R., Thompson, B. A., and Snider, T. G. III (1993). Â“Ovarian development, fecundity, and spawning frequency of black drum Pogonias cromis in Louisiana,Â” Fish Bull. 91, 244-253. Forrest, T. G., Miller, G. L., and Zagar, J. R. (1993). Â“Sound Propagation in shallow water: implications for acoustic communication by aquatic animals,Â” BioAcoustics 4, 259-270. Gedamke, J., Costa, D. P., and Dunstan, A. (2001). Â“Localization and visual verification
108 of a complex minke whale vocalizati on,Â” J. Acoust. Soc. Am. 109, 3038-3047. Grafe, T. U. (2005). Â“Anuran chorus es as communication networks,Â” in Animal Communication Networks edited by P. McGregor (Cam bridge University Press, Cambridge, UK), pp. 277-299. Grafe, T. U. (1996). Â“The function of call alte rnation in the African reed frog (Hyperolius marmoratus): precise call timing prev ents auditory masking,Â” Behav. Ecol. Sociobiol. 38, 149Â–158. Guest, W. C., and Laswell, J. L. (1978) Â“A note on courtship behavior and sound production of red drum,Â” Copeia. 337-338. Hamilton, E. L., and Bachman, R. T. (1982). Â“S ound velocity and rela ted properties of marine sediments,Â” J. Acoust. Soc. Am. 72(6), 1891-1904. Harmer, Sir, S. F., Herdman, W. A., Bridge T. W. and Boulenger, G. A. (1922). Â“ Fishes, Ascidians etc. The Cambridge Natural History Vol VII ,Â” Macmillan and Co. Ltd. London. Pp 358-362. Hayes, S. A., Mellinger, D. K., Croll, D. A., Costa, D. P., and Borsani, J. F. (2000). Â“An inexpensive passive acoustic system for recording and localizing wild animal sounds,Â” J. Acoust. Soc. Am. 107, 3552-3555. Horodysky, A. Z., Brill, R. W., Fine, M. L., Musick, J. A., and Latour, R. J. (2008). Â“Acoustic pressure and particle motion thresh olds in six sciaenid fishes,Â” J. Exp. Biol. 211, 1504-1511. Ibarra, R. M., Penny, L. T., Ebeling, A. W., Va n Dykhuizen, G., and Cailliet, G. (1983). Â“The mating call of the plainfin midshi pman fish, Porichthys notatusÂ” in Predators and Prey in Fishes, edited by D. L. Noakes, D. G. Linquist, G. S.
109 Helfman, and J. A. Ward (Dr. W. Junk Publishers, The Hague), pp. 205-212. Janik, V. M. (2000). Â“Source levels and the estimated active space of bottlenose dolphin ( Tursiops truncatus ) whistles in the Moray Firth, Scotland,Â” J. Comp. Physiol. A. 186, 673-680. Kenyon, T. N., Ladich, F., and Yan, H. N. (1998). Â“A comparative study of hearing ability in fishes: the auditory brainstem response approach,Â” J. Comp. Physiol. A. 182, 307-318. Ladich, F. (2000). Â“Acoustic communication and th e evolution of hearing in fishes,Â” Phil. Trans. R. Soc. Lond. B. 355, 1285-1288. Ladich, F., and A. N. Popper, (2004). Â“Paralle l evolution in fish hearing organs,Â” in Evolution of the Vertebrate Auditory System, edited by A. N. Popper and R. R. Fay (Springer-Verlag, New York), pp. 95-127. Locascio, J. V., and Mann, D. A., (2008). Â“Die l periodicity of fish sound production in Charlotte Harbor, Florida,Â” T. Am. Fish. Soc. 137, 606-615. Lugli, M. and Fine, M. L. (2003). Â“Acousti c communication in two freshwater gobies: Ambient noise and short-range propagation in shallow streams,Â” J. Acoust. Soc. Am. 114, 512-521. Lugli, M. and Fine, M. L. (2007). Â“Stream ambient noise, spectrum and propagation of sounds in the goby Padogobius martensii: sound pressure and particle velocity,Â” J. Acoust. Soc. Amer. 122, 2881-2892. Lurton, X. (2002). Â“Underwater aco ustic wave propagation,Â” in An Introduction to Underwater Acoustics; Prin cipals and Applications, edited by Philippe Blondel (Praxis Publishing Ltd., Chichester, UK) pp. 11-60. Mackenzie, K. V. (1981). Â“Nine-term equation for sound speed in the oceans,Â” J. Acoust.
110 Soc. Am. 70(3), 807-812. Madsen, P. T., Wahlberg, M. (2007). Â“Reco rding and quantification of ultrasonic echolocation clicks from free -ranging toothed whales,Â” D eep-Sea Research Part I. 54, 1421-1444. Mann, D. A. and Lobel, P. S., (1997). Â“Pr opagation of damselfish (Pomacentridae) courtship sounds,Â” J. Acoust. Soc. Am. 101, 3783-3791. Mann, D. A., Locascio, J. V., Coleman, F. C., and Koenig, C. C. (2009). Â“ Goliath grouper Epinephelus itajara sound production and movement patterns on aggregation sites,Â” Endang. Species Res. 7(3), 229-236. Mann, D. A., and Grothues T. M. (2009). Â“Short term upwelling events modulate fish sound production at a mid-Atlantic O cean observatory,Â” Mar. Ecol. Prog. Ser. 375, 65-71. Maruska, K. P., Boyle, K. S., Dewan, L. R ., Tricas, T. C. (2007). Â“Sound production and spectral hearing sensitivity in th e Hawaiian sergeant damselfish, Abudefduf abdominali ,Â” J. Exp. Biol. 210, 3990-4004. McDonald, M. A., Hildebrand, J. A., Wiggins, S.M., Thiele, D., Glasgow, D., and Moore, S. E. (2005). Â“Sei whale sounds recorded in the Antarctic,Â” J. Acoust. Soc. Am. 118(6), 3941-3945. Mellinger, D. K. (2001). Â“Ishmael 1.0 UserÂ’s Guid e. ISHMAEL: Integrated System for Holistic Multi channel Acoustic Exploration and Localization,Â” Na tional Oceanic and Atmospheric Admini stration Technical Memo randum OAR PMEL no. 120 34 pp. Mellinger, D. K. (2005). Â“Automatic detection of regularly repeating vocalizations,Â” J.
111 Acoust. Soc. Am. 118(3, pt2), 1940. Miller, P. J. O. (2006). Â“Diversity in sound pr essure levels and estimated active space of resident killer whale vocalizations,Â” J. Comp. Physiol. A. 192, 449-459. Mitchell, S., Poland, J., and Fine, M. L. (2008) Â“Does muscle fatigue limit advertisement calling in the oyster toadfish Opsanus tau ?,Â” Anim. Behav. 76(3), 1011-1016. Moehl, B., Wahlberg, M., Madsen, P. T., Heerfordt, A., and Lund, A. (2003). Â“The monopulsed nature of sperm whale clicks,Â” J. Acoust. Soc. Am. 114, 1143-1154. Mok, H. K. and Gilmore, R. G. (1983). Â“A nalysis of sound production in estuarine aggregations of Pogonias cromis Bairdiella chrysoura and Cynoscion nebulosus (Sciaenidae),Â” Bull. Inst. Zool., Academica Sinica. 22, 157-186. Nieland, D.L., and Wilson, C. A. (1993). Â“Rep roductive biology and an nual variation of reproductive variables of black drum in the northern Gulf of Mexico,Â” Trans. Am. Fish. Soc. 122(3), 318-327. Phillips, R., Niezrecki, C., and Beusse, D. (2004). Â“Determination of West Indian manatee vocalization levels and rate,Â” J. Acoust. Soc. Am. 115(2), 2486. Popper, A. N. and Schilt, C. R. (2008). Â“Hearin g and acoustic behavior: basic and applied considerations,Â” in Fish Bioacoustics edited by J. Webb, A. N. Popper, and R. Fay (Springer, New York), pp. 17-48. Ramcharitar, J. U., and Popper, A. N. (2004). Â“Masked auditory thres holds in sciaenid fishes: A comparative study,Â” J. Acoust. Soc. Am. 116, 1687-1691. Ramcharitar, J. U., Higgs, D. M., and Popper, A. N. (2006). Â“Audition in sciaenid fishes with different swim bladder-inner ear c onfigurations,Â” J. Acoust. Soc. Am. 119, 439-443.
112 Ramcharitar, J. U., Gannon, D. P., and Popper, A. N., (2006). Â“Bioacoustics of fishes of the family sciaenidae (croakers and dr ums),Â” Trans. Am. Fish. Soc. 135, 14091431. Rasmussen, M. H., Miller, L. A., and Au, W. W. L. (2002). Â“Source levels of clicks from free-ranging white beaked dolphins ( Lagenorhynchus albirostris Gray 1846) recorded in Icelandic wate rs,Â” J. Acoust. Soc. Am. 111, 1122-1125. Rasmussen, M.H., Lammers, M., Beedholm, K., Miller, L. A. (2006). Â“Source levels and harmonic content of whistles in whit e-beaked dolphins (Lagenorhynchus albirostris),Â” J.Acoust. Soc. Am. 120(1), 510-517. Sirovic, A. (2006). Â“Blue and fin whale acousti cs and ecology off Antarctic Peninsula,Â” Diss. Abst. Int. Pt. B Sci. & Eng. 67(4), 1862. Rogers, P. H., and Cox, M. (1988). Â“Underwat er sound as a biological stimulus,Â” in Sensory Biology of Aquatic Animals, edited by J. Atema, R. Fay, A. N. Popper, and W. N. Tavolga (Springe r-Verlag, Berlin), pp. 131-149. Sprague, M. W. (2000). Â“The single sonic mu scle twitch model for the sound-production mechanism of the weakfish, Cynoscion regalis ,Â” J. Acoust. Soc. Am 108(5), 2430-2437. Sprague, M. W., Luczkovich, J. J. (2004). Â“Meas urement of an individual silver perch Bairdiella chrysoura sound pressure level in a field re cording,Â” J. Acoust. Soc. Am. 116, 3186-3191. Teloni, V., Zimmer, W.M.X., and Tyack, P.L. (2005). Â“Sperm whale trumpet sounds,Â” BioAcoustics 15(2), 163-174. Urick, R. J. (1983). Â“Propagation of sound in the sea: transmission loss, I and II,Â” in Principals of Underwater Sound, 3rd edition, edited by Diane Heiberg and Janet
113 Davis (Peninsula Publishing, Los Altos, CA) pp. 99-200. Wang, K., Wang, D., Akamatsu, T., Fujita, K., Sh iraki, R. (2006). Â“Estimated detection distance of a baiji's (Chinese river dolphi n, Lipotes vexillifer) whistles using a passive acoustic survey method,Â” J. Acoust. Soc. Am. 120(3), 1361-1365. Winn, H. E. (1964). Â“The biological significance of fish sounds,Â” in Marine BioAcoustics edited by W. N. Tavolga (Pergamon, New York), pp. 213-231. Zelick, R., Mann, D. A., and Popper, A. N., (1999). Â“Acoustic communication in fishes and frogs,Â” in Comparative Hearing: Fish and Amphibians edited by R. Fay and A. N. Popper (Springer Verlag, New York), pp. 363-411.
114 Table 5.1. Mean RMS and Peak source level es timates of six different individual fish and all fish combined. SL is mean source le vel (SPL dB re: 1Pa), STDV is standard deviation, and CV is the coefficient of variation expressed as a percentage. Fish 1 Fish 2Fish 3Fish 4Fish 5Fish 6 All Fish Mean RMS SL 164.9 167.7170.6161.1161.9161.9 164.7 STDV 1.09 1.131.470.820.950.41 0.98 CV % 0.66 0.680.860.510.590.25 0.59 Mean Peak SL 175.3 178.3180.9171.5173.7172.3 175.3 STDV 0.70 1.151.360.790.900.47 0.90 CV % 0.40 0.640.750.460.520.27 0.51 N 944 362 8377 1025
115 Table 5.2. Source level estimates of black drum and of various marine mammals reported in the literature. All source level estimates are reported as (dB RMS SPL re: 1 Pa). Source Level Author Species Mean (+/SD) Range Vocalization Type Phillips et al. west indian manatee ( Trichechus manatus ) 112.0 vocalization Miller killer whale ( Orcinus orca) 140.2 (4.1) whistle Wang et al. Chinese river dolphin ( Lipotes vexillifer ) 143.0 (5.8) whistle Miller killer whale ( Orcinus orca) 146.6 (6.6) variable Miller killer whale ( Orcinus orca) 152.6 (5.9) stereotyped McDonald et al. sei whale ( Balaenoptera borealis ) 156.0 (3.6) tonal call Rasmussen et al. white-beaked dolphin ( Lagenorhynchus albirostris ) 118 -167 whistle Gedamke, et al. minke whale ( Balaenoptera acutorostrata ) 150 -165 stereotyped Janik bottlenose dolphin ( Tursiops truncatus) 158.0 (0.6) --169 whistle Au et al. humpback whale ( Megaptera novaeangliae ) 151 -173 song Locascio and Mann black drum ( Pogonias cromis ) 165.0 (1.0) 159 -174 tonal call Teloni, et al. sperm whale ( Physeter macrocephalus) 172.0 (n=1) trumpet Sirovic blue whale ( Balaenoptera musculus ) 189.0 (3.0) call Sirovic fin whale (Balaenoptera physalus) 189.0 (4.0) call
116 Figure 5.1. Study site location wi thin the estuarine canal syst em of Cape Coral, Florida where array recordings were made. Th e study site was in a sea-walled basin, approximately 240 x 265m in surface area and 7m deep with a soft bottom.
117 0 1 2 3 4 5 6 -0.4 -0.2 0 0.2 0.4 Amplitude (voltag 0 1 2 3 4 5 6 0 0.2 0.4 Amplitude (voltage) 0 1 2 3 4 5 6 0 0.5 0.1 Time (sec)Amplitude (voltage) 1. Low Pass Filter2. Rectify 3. Envelope Figure 5.2-A. Sequence of signal processing steps used to iden tify individual calls: acoustic da ta are first low-pass filtered with a 10point moving average (top figure) then rec tified (e.g. absolute values calculated, mi ddle figure) and finally enveloped with a 4,410 point moving average (bottom figure). The da shed line indicates the detec tion threshold in signal am plitude used to identify individual calls.
118 0 10 20 30 40 50 60 -0.5 0 0 5 Amplitude ( V 0 10 20 30 40 50 60 -0.2 0 0.2 Amplitude (V) 0 10 20 30 40 50 60 -0.1 0 0.1 Amplitude (V) 0 10 20 30 40 50 60 -0.05 0 0.05 Time ( msec ) Amplitude (V) Track 01 Track 02 Track 04 Track 03 Figure 5.2-B. Peak detection within a si ngle call on the four different simultaneou sly recorded tracks of the hydrophone array Time of arrival differences are calculated for th e same peak between different tracks, in dicated by arrows, to use for localizing th e position of the calling fish.
119 0102030405060 Distance (m) 130 140 150 160 170dB RMS re:1uPa SPLy = 174.87-18.61*Log10(distance) r ^2 = 0.90 dashed line y = 168.74-3.85*SqRt(distance) r ^2 = 0.93 solid line Figure 5.3. Square root and Log10 model fits of received levels regressed against distance from source location to each hydrophone The square root model provided a better over-all fit to the data and was used to estimate source levels of black drum calls.
120 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 Time (sec)Amplitude 5.4-A. Time (sec)Frequency Hz 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 100 200 300 400 500 600 700 800 900 1000 1100 1200 5.4-B. Figures 5.4-A and 5.4-B: Waveform (A) and spectrogram (B) of a single black drum call. Black drum produce high intensity single pulsed tonal calls with multiple harmonics.
121 0 10 20 30 40 50 60 94 188 282 Distance (m)Frequency Hz Figure 5.5. The frequency containi ng the greatest concentration of acoustic ener gy within black drum calls recorded at various distances are plotted. The maximum energy is mainly concentrated in the fundamental frequency ( 94Hz) at all distances but this varies more with increasing distance from the source possibly due to propagation eff ects or variability in call production.
122 0 5 10 15 20 25 30 35 40 45 50 -35 -30 -25 -20 -15 -10 -5 0 Distance ( m ) Transmission Loss (dB) 94Hz (SQRT) 188Hz (SQRT) 282Hz (LOG10) (C) Cylindrical (S) Spherical S 94Hz 282Hz 188Hz C Figure 5.6-A. Plot of slopes (e.g. transmissi on loss) from the regression equations of received levels vs. distances for the f undamental frequency (94 Hz) and first two harmonics (188 & 282 Hz) of black drum calls. Slopes of the cylindrical (e.g. Log10) and spher ical (e.g. Log20) spreading loss models are plotted for reference.
123 0 94 188 282 376 470 564 658 752 846 940 1034 1128 60 80 100 120 140 160 Frequency HzdB RMS re:1uPa SPL 2.4m 8.9m 15.5m 47.4m Figure 5.6-B. Results of Fast Fourier Tran sformation (FFT) of the mean sound pressure level (dB re: 1Pa) of a subsample of 50 black drum calls recorded at four different mean distance groups. The majority of energy is concentrated in the fundamental frequency (94Hz) and the first two harmoni cs (188 & 282 Hz). Transmission loss data of the fundamental frequency and first harmonic are similar and data of both frequencies were best fit with a square root model. The 282 Hz harmonic was best fit by a Log model (see Figure 5.6-A).
124 0 5 10 15 20 25 30 35 40 45 50 125 130 135 140 145 150 Distance (m)Background Sound Pressure Levels dB RMS re: 1uPa Figure 5.7-A. Background sound pressure levels (root mean square dB re: 1Pa) calculated from the 50 millisecond period preced ing each call at four separate distance groups from the location of the source. Background levels are within a similar range at al l distance groups. The source of most background noi se is the calls of other black drum.
125 0 5 10 15 20 25 30 35 40 45 50 -5 0 5 10 15 20 25 30 35 40 Distance (m)Signal to Noise Ratio dB RMS re: 1uPa SPL Figure 5.7-B. Signal-to-Noise Ratios (SNR) for each of the four distance groups shown in Fig. 5.7-A based on data of black dru m calls and background levels. The SNR is calculated by subtraction of the background level from th e signal level (both are expr essed as the root mean square sound pressure level dB re: 1Pa). The SNR ranges over approximately 20 dB for each distance group and decreases with increasing distance from the source.
126 0 75 150 300 450 600 900 80 90 100 110 120 130 140 150 160 170 Frequency HzAmplitude dB RMS re: 1uPa SPL black drum call AEP Figure 5.8. Audiogram of a single black drum measured by auditory evoked potential (AEP ) along with mean source level estimate s of the fundamental frequency and associated harmonics of black dr um calls. The horizontal dashed lines represent the upper and lower limits of background sound pressure leve ls recorded at the study site. The freque ncy of greatest aud itory sensitivity (3 00 Hz) and that of call intensity (94 Hz) are not di rectly matched however considerable overlap exists in the range of frequencies whe re auditory sensitivity and call intensity are greatest.
127 140 145 150 155 160 165 17 0 0 10 20 30 40 50 60 dB RMS re:1uPa SPLNumber of Calls RMS Source Levels RMS Received Levels Figure 5.9-A. Frequency distribution of re ceived sound pressure levels and source sound pressure levels of black drum calls (r oot mean square dB re: 1 Pa). The highest received le vels overlap the range of source level estimates.
128 145 150 155 160 165 170 175 180 0 10 20 30 40 50 60 70 dB peak re:1uPa SPLNumber of Calls Peak dB Source Levels Peak dB Received Levels Figure 5.9-B. Frequency distribut ion of received sound pressure levels and source sound pressure levels of black drum calls (p eak dB re: 1 Pa). The highest received levels ove rlap the range of sour ce level estimates.
129 0 20 40 60 80 100 120 140 160 180 160 165 170 175 180 Call Numbe r dB RMS re: 1uPa SPL demonstrated in fig. 10B Figure 5.10-A. Waveform of received levels of consecutively produced black drum ca lls which demonstrate cyclical movement patterns of the calling fish. The amplitudes of the received leve ls rise and fall as the fish moves closer to and further from the hydrophone. An example of this patterned movement is de monstrated in Figure 10-B us ing Cartesian coordinates.
130 -12 -12 -11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 6 7 8 9 10 11 Distance (m)Distance (m) Loop 1 Loop 2 Loop 3 Figure 5.10-B. Localized X-Y positions of consecutively produced calls shown in Figure 5.10-A wh ich demonstrate recurrent loop ing patterned movement by the fish. Three sepa rate consecutive loops are plotted with st art and end points of each designated by a n Â‘SÂ’ and an Â‘EÂ’. S1 S2 S3 E3 E2 E1
131 CHAPTER SIX The Quantitative and Temporal Relations hip of Egg Production and Sound Production by Black Drum ( Pogonias cromis ) INTRODUCTION Acoustic time series recorded ove r seasonal periods have demonstrated that patterns in fish sound production approximate patterns in spawni ng condition over these periods (Mok and Gilmore, 1983; Connaughton and Taylor, 1995; Locascio and Mann, in review). Associations between sound pr oduction and spawning have been made for sciaenids by simultaneously collecting eggs via plankton surveys and making acoustic recordings of sound production via hydrophone surveys (Mok and Gilmore, 1983; Saucier and Baltz, 1993; Luczkovich et al., 1999 ). These studies have documented the coincidence of sound production and spawning on a qualitative basis. The difficulties in measuring the relationship between these vari ables on a quantitative basis are related to the uncertainties in knowing that the fish whose sounds are be ing recorded are from the same population responsible for producing the eggs which are collected. Single hydrophone recordings are omni-directional a nd cannot resolve the precise location of calling fish and, in most cases, the numbers of calling fish within the active space of the hydrophone due to partial or complete overlap of calls. Hydrophone arrays must be used to localize the position of calling fish. Establishing the sp atial origin of egg production is
132 also difficult and would requi re information on how the e ggÂ’s trajectory, as a passive particle, would behave in the prevailing hydrodynamic conditions. An alternative solution to local izing the origin of sound production and egg production within an open-field environment is to address th e problem within a confined natural environment. Such an environment was found in a residential estuarine canal system of Cape Coral, Florida inhabite d by a sound-producing population of spawning black drum. The study area was a moderately small, mostly enclosed basin which allowed us to be reasonably sure that we were sampling eggs of the same population of fish whose sounds were being recorded. Th e primary objectives of this study were to quantitatively relate the timing and levels of sound production and egg production by black drum during different periods of their spawning s eason. Sampling was conducted on two consecutive evenings so that instantaneous egg mortality rates could be calculated for the cohort produced on the first evening of sampling. METHODS Cape Coral is a coastal city in southwest Florida with an extensive sea-walled canal system (~200+ miles) that is connect ed to Charlotte Har bor. This study was conducted in a dead-end 240 x 265m basin of the canal system (Fig. 6.1). An acoustic time series of black drum sound production wa s recorded at the Cape Coral study site from December Â– April, 2006 using a longterm acoustic recording system (LARS model: Toshiba Pocket PC E755) programme d to record 10 seconds of sound every 10 minutes (sample rate: 11,025 Hz). High Tech Inc. 96-min series hydrophones (sensitivity: -164dB re: 1V/1Pa and flat frequency res ponse of 2 Hz Â– 37 kHz) were
133 used with all recording systems in this st udy. The LARS was deployed in approximately 7m of water where it remained stationary and positively buoyant 0.5m above the bottom. Recordings from this system provided a hi gh resolution acoustic time series of black drum sound production over the majority of the spawning season, in cluding the peak. Surface and bottom water temperature data were recorded for the duration of the study using Hobo temperature data loggers (m odel UA-002-08; Onset Computers) programmed to record data at ten minute intervals. The su rface temperature data logger was attached to a surface buoy and suspende d 0.5m below it. The bottom temperature data logger was attached to the LARS. To investigate the temporal and quantitative relationships between egg production and sound production surface plankton tows were conducted hourly from 1800 Â– 0400 hrs to collect eggs on two consecutive ni ghts while acoustic recordings were made simultaneously at five locations in the canal basin using LARS programmed to record 10 seconds of sound / 10 minutes (sample rate 11, 025 Hz) (Fig. 6.1). Sampling took place on the evenings through early mornings of Jan. 29-31, Feb. 14-16, March 3-5, March 2022, and April 6-8, 2006. The schedule of plankt on tows was based on prior knowledge of the timing of nightly black drum chorus events and was intended to cover the duration of a chorus event completely. Plankton samp les were collected using a 333 micron mesh net with rectangular frame (0.25m high x 0.5m wide) and calibrated flow meter (General Oceanics, model 2030R) towed at the surface by a motorized canoe at approximately 2 knots along course covering the entire st udy area (Fig. 6.1). Each tow required approximately 20 minutes to complete a nd began 20 minutes prior to each hour. Collection times used in analyses were ther efore designated as the temporal midpoint of
134 the tow (e.g. 10 min prior to ea ch hour). Flow meter data was recorded at the beginning and end of each tow to calculate the volum e of water sampled in cubic meters (m3). Tide data was recorded before each tow by measuring the change in water column height relative to a fixed mark on a dock piling. Plankton samples were preserved in 10% formalin immediately after each tow and tr ansferred to 50% isopropyl after 48 hrs. A LARS was deployed near each corner of the basi n and a fifth in the center to calculate the mean sound pressure level of black drum vo calizations throughout the study area. LARS were deployed in the afternoon prior to th e first plankton tow and retrieved sometime after the last plankton tow on the second night of sampling. Failure of the trolling motor used on the canoe ended plankton sampling on Ja n 30 after the fourth plankton tow of the second evening (2040-2100 hrs) had been comp leted. A series of plankton tows were conducted at the study site on March 14, 2007 to examine the depth distribution of black drum eggs. Three 10-minute tows (~2 knots) were made using a 0.5 m diameter, 333 m mesh net between 2000 Â– 0048 hrs at each of the surface (0 Â– 0.25m), 1m, 2m, and 3m depths and a single tow was made at 4m. Samp les from these collections were preserved as before. Acoustic and Plankton Data Processing / Analysis All acoustic data were analyzed with a fa st Fourier transform (FFT) to generate a power spectrum from which the band level s ound pressure level (BSPL re: 1Pa) in 100 Hz wide bins was calculated. The SPL was greatest in the 100 Â– 200 Hz band and a five point moving average of the BSPL in this frequency range was used for all time series analyses.
135 The developmental stages and densit ies of black drum eggs were evaluated from each plankton sample collected. Approximately 10-20% (by volume) of each sample was sub sampled by inverting the contents of the entire sample 5 times in a 1.0 L graduated cylinder and decanting a subsample into a sm aller graduated cylinder and measuring its volume. The total number of black drum eggs in the subsample was counted and used to estimate the total number of eggs in the entire sample. The total egg density (eggs/m3) of each sample was calculated by dividing the estimat ed total number of eggs in the entire sample by the volume of water filtered (m3) during the plankton tow. The developmental stage of the first 100 eggs (or all eggs if less than 100) in each subsample were characterized under a dissecting light micros cope according to the following stages: 1. Blastodisc; 2. 2 cells, 3. 4 cells; 4. 8 cel ls; 5. 16 cells; 6. 32 cells; 7. 64 cells; 8. 10. Morula (early, mid, late); 1113. Blastula (early, mid, late); 14-16. Gastrula (early, mid, late); 17-19. Early Embryo ( early, mid, late); 20-22. Tail Bud (early, mid, late); 23-25. Tail Free (early, mid, late); 26-28. Late Embr yo (early, mid, late); 29. Hatching. In this developmental scheme, Â‘blastodiscÂ’ refers to a single fertilized undifferentiated cell. It was then possible to estimate the density of individual stages (or relative ages) as a percentage of the total egg density of each collection. Time of spawning was interpreted by the presence of blastodiscs and by back-c alculation of the 2-cel l through blastula stages for samples collected in March and Ap ril. Because egg developmental rates are mainly a function of water temperature, and to a lesser degree egg diameter, (Pauly and Pullin, 1988) developmental rates of red drum eggs ( Sciaenops ocellatus ) (Holt et al., 1985; Reese, R. pers. comm.) incubated in th e same water temperatures as black drum egg collections were applied to back-calcula te the time spawned of black drum eggs.
136 Back-calculated spawning times were grouped into 1hr bins since sampling was limited to this time resolution and densities were co rrected for using a calculated mortality rate, and average densities were reported for each 1 hour bin. Egg diameter of both species is similar; black drum are approximately 0.9 Â– 1.1 mm and red drum are approximately 0.8 Â– 1.0 mm, but may vary due to salinity and pa rental condition (Johnson, 1978). Densities of egg predator taxa were estimated fro m samples collected at 2200, 0000, 0200 hrs each evening. Estimates of nightly egg production (eggs/m3) were calculated by averaging egg densities from all collections made after sp awning ended for the evening. These data were regressed against nightly maximum sound pressure levels to examine the quantitative relationship between levels of sound production and egg production. Maximum sound pressure level is highly corr elated with measures of total acoustic energy and can be used to quantitatively re present nightly black drum sound production (Locascio and Mann, in review). The wei ghted average of the time of spawning was based on the egg densities and collection times of blastodiscs and the back-calculated spawning times of later stage eggs. The wei ghted average of the time of spawning was regressed against the tempor al center of sound production (e.g. non-weighted, midpoint) to examine the temporal relationship betw een sound production and spawning. Chorus start times and spawning time were also co mpared over the course of the season. The weighted average of the time of spawning a nd nightly egg production estimates were also regressed against temperature data.
137 Instantaneous daily mortality rates were estimated from e gg densities and relative ages of the cohort spawned on the first even ing of sampling and were calculated from samples collected during both evenings. The mortality equation was used: Nt = N0e-zt where N0 is the initial number of eggs at time 0, Nt is the number of eggs remaining at time t ( > t0) and Z is the instantaneous daily mortality rate coefficient. Relative ages of eggs were calculated as the difference in the weighted average of the time of collection of successive developmental stages from the wei ghted average of the time of collection of blastodiscs. Averages were weighted by the number of eggs of each developmental stage in the sample. Egg densities for each relative age were calculated as the percentage of the total egg density of the sample. If the same developmental stage occurred in multiple plankton samples for the same evening the average density for that age was used in the mortality model. Only egg density and age data from samples co llected after spawning ended for the evening were used in the mortality equation. Egg densities of samples collected fro m various depths were fit with a logarithmic model. The fitted equation was used to in terpolate egg densities at each midpoint between sampled depths. This allowed es timates of egg densities and the relative percentage of eggs at each depth from the surface to 4 m. The total number of eggs within the 0.25 m surface laye r of the study area was estimated by multiplying nightly egg production (eggs/m3) by the volume of the surface layer. The total number of eggs in the water column (to 4 m) was estimated by dividing the total number of eggs in the surface layer by the percentage of eggs in this layer. The relationship between depth and developmental stage was also examined. Female spawning stock biomass was estimated
138 from the total number of eggs in the water column in conjunction with batch fecundity estimates for black drum re ported in the literature. RESULTS Black drum sound production and egg production occurred on all evenings samples were collected however neither th e timing nor quantity of sound production was well correlated with egg production on a night ly basis (Fig. 6.2). The April samples which contained the highest densities of e ggs collected during the study had the lowest levels of sound production. The toadfish ( Opsanus beta ), a nest-building species, which produces sounds at a fundamental frequency of about 282 Hz was the only other fish species recorded during the study. Two cohorts were apparent in sa mples collected each night; one spawned on the previous night and the second spawned on the current night of sampling. In February, when water temperatures were approximately 17.6 Â– 18.3o C, egg development slowed and eggs from a third cohort from 2 nights prior were collected. Blastodiscs were collected on every evening except for the Januar y samples. Back calculation to the time of spawning for early cleavage stages colle cted in February was not possible because developmental rate data were not available fo r these water temperatures. However, this was also not necessary because blastodiscs we re only collected in the 2250 hr sample on both evenings and older eggs of this cohor t were not collected before or younger eggs after this sample time. This isolated th e 2250, 1 hr sample bin as the time spawning occurred on both February evenings. A simila r pattern was evident in the January data but the earliest developmental stage found in samples was 32-cell. The duration of
139 spawning on all other sampling nights during th e season ranged from 3 to 5 hours. The tide was ebbing during spawning times on all nights except for a very low amplitude flood tide within the first two hours of spawning in April. Spawning occurred earlier in the day as the season progressed and in the la te March and April samples it was apparent from the collection of advanced early cleavag e stage eggs that spawning occurred prior to the time of the first plankton sample. To th e contrary, chorus start time became gradually later and chorus duration shorte r as the spawning season waned. The numbers of eggs of each developmental stage for all plankton samp les collected are detailed in Table 6.1. The timing of sound production and egg producti on (blastodisc through blastula stages) and sound production and back calculated times early cleavage stages would have been spawned are illustrated in figures 6.3-AÂ– 6.3-H. The regression of egg producti on against maximum sound pressure levels demonstrated a negative a relationship; 90% of the variability in egg production was explained by maximum sound pres sure level data (Fig. 6.4-A). Greatly increased levels of egg production coupled with low levels of sound production in the April data strongly leveraged this result. Recal culation of the regression w ith the April data removed however, still demonstrated that a negative re lationship existed; 42% of the variability was explained (Fig. 6.4-B). The regression of the weighted aver age of the time of spawning and the temporal cente r of sound production explained 27% of the variability in the time of spawning (Fig. 6.5). The relati onship between the weighted average of the time of spawning and temperature was strongl y negative and 94% of the variability in time of spawning was explained by temperatur e (Fig. 6.6). Temperature explained 52% of the variability in egg production (Fig. 6.7-A) and 54% when the April data were
140 removed from the regression (Fig. 6.7-B). Data of spawning times and timing of sound production for each date are summarized in Table 6.2. Estimates of instantaneous egg mortality rates increas ed throughout the season, ranging from a low of 0.62 d-1 in January to a high of 2.72 d-1 in April (Table 6.3). Egg predators quantified in the pl ankton samples included Ctenoph ora sp. and Hydrozoan sp.; Chaetognaths, unconfirmed as a predator of e ggs, were also quantified. Ctenophores did not appear in the samples until April and Hydr ozoans were only sparsely represented in the samples prior to April, Chaetognaths were present in all plankton samples and were positively correlated with mortality rates (r = 0.57, p = 0.11). Combined Ctenophores and Hydrozoans were also positively correla ted with mortality rates (r = 0.66, p = 0.22). Copepods, mainly A. tonsa appeared highly abundant in all samples but were not quantified. Egg densities were logarithmically di stributed with depth (Fig. 6.8). Egg developmental stage was relatively evenly di stributed with depth (Fig. 6.9). The 0.25 m depth surface layer contained 33.9% of eggs in the 4 m water column (Table 6.4). Based on mean batch fecundity and eviscerated body we ight values of black drum sampled in the northern Gulf of Mexico (Fitzhugh et al. 1993, Niela nd et al. 1993) and Argentina (Macchi et al., 2002) estimates of spawning stock biomass ranged from 1.6 to 1491.3 kg and numbers of individual fish ranged from 0.2 to 28.3 during the study period (Table 6.5). DISCUSSION The results of this study demonstr ate that patterns in sound production were not useful for predicting patterns in egg producti on by black drum on a daily scale but do
141 provide accurate characteri zation of spawning periodicity on a seasonal basis. The former result may seem unexpected given th e strong agreement betw een peak seasonal patterns of sound production and peak states of reproductive readine ss in sciaenids (Mok and Gilmore, 1983; Connaughton and Taylor, 1995; Locascio and Mann, in review). The lack of a strong positive a ssociation between sound produc tion and egg production, both in timing and levels, shows that courtship behavior (e.g. fish cal ling) and spawning are not always relative on a daily scale. Sound production has been reported to occur before, during, and after spawning in sciaenid s (Mok and Gilmore, 1983; Connaughton and Taylor, 1995) and we recorded extended pe riods of sound production prior to spawning earlier in the season and well after spawning occurre d later in the season. The lack of a positive association between egg production and sound production may also possibly be explained by the tendency for long-lived, se rial spawners to have highly variable reproductive output (Fitzhugh et al., 1993). Bl ack drum may live 50 Â– 60 years (Murphy and Taylor, 1989) and may spawn approximate ly every 4 days (Fitzhugh et al., 1993). Black drum are also highly fecund, and older, larger fish typically produce larger batch sizes of generally higher egg quality (Macch i et al., 2002). Estimated batch fecundity was reported by Nieland to range from 0.51 to 2.42 million ova and was positively correlated with fish size. Hi gher egg densities in April were likely the result of larger more fecund females spawning in the study ar ea. Estimates of spawning stock biomass and numbers of females based on data of Fitzhugh and Nieland ar e unrealistically low prior to April. This indicates that smaller or simply less fecund fish were in the study area prior to April and larger more fecund females or a greater number of smaller, less fecund females entered and spawned in the study area in April. The lower batch size
142 estimates reported by Macchi et al. (2002) provide the only reas onable estimates of spawning stock biomass for data prior to Apri l also indicating the possible transition to larger, more fecund females spawning in the study area in April. The timing and duration of daily sp awning in this study was generally consistent with reports of a circad ian spawning pattern documented for sciaenids occurring approximately -2 to +4 hours around dusk (Ho lt, et al., 1985; Aalber s, 2008). Spawning occurred as late as 4+ hours after sunset in February, based on bl astodisc collection and as early as the 15:50 hr bin (15:10 Â– 16:00 hr) for late March and April based on backcalculated spawning times. Most of the vari ability in the timing of spawning could be explained by water temperature data (r2 = 0.94). As the season progressed and temperatures increased (along with photoperiod) spawning oc curred earlier in the day. Aalbers (2008) documented a similar trend in spawning by white seabass ( A. nobilis ). Temperature and photoperiod are re garded as the principal determinants of reproductive development. Discerning the effect s of each is difficult but in general, it is expected that photoperiod has a greater influe nce in temperate climates and temperature in tropical climates, with possible interac tions between them and local environmental cues in effect as well (Pankhurst and Port er, 2003). So where does the black drum, a temperate sciaenid which ranges from the north east region of the U.S. to Argentina fit into this scheme? There is evidence to i ndicate temperature is the more influential environmental mediator of spawning. The ra nge of water temperature in which black drum spawn (15 Â– 25o C) is consistent over a wide lat itudinal range in the U.S. (south Florida to Chesapeake Bay) while photoperiod ov er this range is more variable (in other words, black drum spawn in the same range of water temperatur es, but not photoperiod
143 over a wide latitudinal range). Other ev idence comes from two previous hydrophone surveys of black drum sound production (Mok and Gilmore, 1983; Saucier and Baltz, 1993) which both documented that sound production did not occur when water temperatures dropped below 15o C but resumed once water temperature rose above this limit. Also, in this study spawning occurred earlier in the evening in January when water temperature was 20.5o C than in cooler February water temperatures (17.6 Â– 18.3o C), despite the slightly longer photoperiod in February. Mortality rates increased throughout the study period concomitantly with predator densities. Cowan et al., (1992) estimated mo rtality rates of black drum eggs as 0.02 Â— 0.03 d-1 in mesocosm studies using hydromedus a and ctenophore predators. Mortality rates in this study were much higher than those reported by Cowan et al., but direct comparisons are not possible due to differences in scale. Although chaetognaths were the most abundant taxa quantified, th ey are only anecdotally consider ed to be egg predators. A simple experiment could be c onducted to test this. Instead, it is likely that the apparent high densities of copepods (not quantified) instead would have s upported the chaetognath population. Likewise, copepod naupl ii would also have supporte d the diet of black drum larvae. The gelatinous hydrozoans and ct enophores appeared towa rd the end of the spawning season likely in conjunction with the spring warming and increase in production. These taxa are vor acious consumers of fish eggs and larvae and their increased abundance would explain increased morta lity rates, such as w ith the April data. It may also have been possible that chaet ognaths and other taxa unaccounted for also contributed to egg mortality.
144 The estimates of developmental times of red drum eggs were consistent with the developmental rates qualitativel y observed in the plankton coll ection data, but this study would have benefitted from making direct measurements of time to developmental stage at ambient water temperatures. It would also have been informative to estimate the population density of adults in the study area as well. Although most of the eggs were concentrated in the surface layer, many eggs of various developmental stages also existed at other depths. Including oblique tows along with surface tows (concurrently if possible) would improve the sampling design. This study also demonstrated that modified natural systems can be highly productive areas for fish es and are logistically well suited for studying them. REFERENCES Aalbers, S.A. (2008). Â“Seasonal, diel, and lu nar spawning periodici ties and associated sound production of white seabass ( Atractoscion nobilis ),Â” Fish Bull 106:143-151 Connaughton, M.A., and Taylor, M.H. (1995). Â“Seasonal and daily cycles in sound production associated with spawning in the weakfish, Cynoscion regalis,Â” Environ. Biol. Fish. 42:233 Â–240 Cowan, J.H., Birdsong, R.S., Houde, E.D., Priest J.S., Sharp, W.C., Mateja, G.B. (1992). Â“Enclosure experiments on surv ival and growth of black drum eggs and larvae in lower Chesapeake Bay,Â” Estuaries 15(3):392-402 Fitzhugh, G.R., Thompson, B.A., Snider, T.G. III (1993). Â“Ovarian development, fecundity, and spawning frequency of black drum Pogonias cromis in Louisiana,Â” Fish Bull 91:244-253
145 Holt, .J.G., Holt, S.A., and Arnold, C.R. (1985). Â“Diel periodicity of spawning in Sciaenids,Â” Mar. Ecol. Prog. Ser. 27:1-7 Johnson, G.D. (1978). Â“ Pogonias cromis black drum. In: Johnson GD (ed) Development of fishes of the mid-Atlantic bight: an atlas of egg, larval, and juvenile stages volume IV Carangidae through Ephipid ae,Â” US Fish and Wildlife Service Department of the Interior, Washington D.C. p 235-236 Locascio, J.V. and Mann, D.A. (in review). Â“D iel and seasonal timing of black drum ( Pogonias cromis ) sound production,Â” Luczkovich, J.J., Sprague, M.W., Johnson, S.E., Pullinger, R.C. (1999). Â“Delimiting spawning areas of weakfish Cynoscion rega lis (family Sciaenidae) in Pamlico Sound, North Carolina, using passive hydroacoustic surveys,Â” Bioacoustics 10:143Â–160. Macchi, G.J., Acha, E.M., and Lasta, C. A. (2002). Â“Reproduction of black drum (Pogonias cromis) in the R o de la Plata estuary, Arge ntina,Â” Fish. Res.59: 83-92 Mok, H.K., and Gilmore, R.G. Jr. (1983). Â“A nalysis of sound production in estuarine aggregations of Pogonias cromis Bairdiella chrysoura and Cynoscion nebulosus (Sciaenidae). Bull Inst Zool Academica Sinica 22:157-186 Murphy, M.D., and Taylor, R.G. (1989). Â“Rep roduction and growth of black drum, Pogonias cromis in northeast Florida,Â” Northeast Gulf Sci 10:127-137 Nieland, D.L., and Wilson, C.A. (1993). Â“Rep roductive biology and annual variation of reproductive variables of black drum in the northern Gulf of Me xico,Â” Trans. Am.
146 Fish. Soc. 122(3):318-327 Pauly, D., and Pullin, R.S.V. (1988). Â“Hatching ti me in spherical, pelagic, marine fish eggs in response to temperature and e gg size,Â” Environ. Biol Fish. 22(4) 261-271 Panhurst, N.W., and Porter, M.J.R. (2003). Â“Cold and dark or warm and light: variations on the theme of environmental control of reproduction,Â” Fish Pshyiol. Biochem. 28: 385-389 Saucier, M.H., and Baltz, D.M. (1993). Â“Spa wning site selection by spotted seatrout, Cynoscion nebulosus and black drum, Pogonias cromis in Louisiana,Â” Environ Biol Fish 36:257Â–27
147 Table 6.1. Number of black drum eggs in each developm ental stage from 100 randomly selected eggs from each plankton sample collected. Developmental stages are listed numeric ally, 1 Â– 28 and nominally with abbreviations which are explained in the methods of chapter 6. In many samples two cohorts were present; those spawned on the night samples were collected and those spawned on the previous night. February samples contained three separate cohorts due to slowed developmental rates associated with low water temperatures. January 29 30, 2006 1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 Tow date time Bldsk EC 2 EC 4 EC 8 EC 16 EC 32 EC 64 EM M M L M E B M B L B E G 1 1/29/2006 1740-1800 2 1/29/2006 1840-1900 3 1/29/2006 1940-2000 4 1/29/2006 2040-2100 5 1/29/2006 2140-2200 6 1/29/2006 2240-2300 61 7 1/29/2006 2340-0000 6 5 8 1/30/2006 0040-0100 5 8 9 1/30/2006 0140-0200 6 1 10 1/30/2006 0240-0300 8
148 0 11 1/30/2006 0340-0400 5 4 3 1 1 1/30/2006 1740-1800 2 1/30/2006 1840-1900 3 1/30/2006 1940-2000 4 1/30/2006 2040-2100 continuedÂ….. 15 16 17 18 19 20 21 22 2 3 2 4 2 5 2 6 2 7 2 8 MG LG E EE M EE L EE E TB TB E TF M T F L T F E L E M L E L L E H 1 1/29/2006 1740-1800 100 2 1/29/2006 1840-1900 100 3 1/29/2006 1940-2000 100 4 1/29/2006 2040-2100 100 5 1/29/2006 2140-2200 93 7 6 1/29/2006 2240-2300 26 13 7 1/29/2006 2340-0000 3 3 2 8 1/30/2006 0040-0100 4 2 9 1/30/2006 0140-0200 5 3 4 10 1/30/2006 0240-0300 2 0 11 1/30/2006 0340-0400 1 5
149 1 1/30/2006 1740-1800 100 2 1/30/2006 1840-1900 100 3 1/30/2006 1940-2000 100 4 1/30/2006 2040-2100 78 22 February 14 16, 2006 1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 Tow date time Bldsk EC 2 EC 4 EC 8 EC 16 EC 32 EC 64 EM M M L M E B M B L B E G 1 2/14/2006 1740-1800 2 2/14/2006 1840-1900 3 2/14/2006 1940-2000 4 2/14/2006 2040-2100 5 2/14/2006 2140-2200 6 2/14/2006 2240-2300 26 7 2/14/2006 2340-0000 60 8 2/15/2006 0040-0100 1 65 9 9 2/15/2006 0140-0200 1 86 10 2/15/2006 0240-0300 84 11 2/15/2006 0340-0400 8 5 1 2/15/2006 1740-1800 2 2/15/2006 1840-1900 3 2/15/2006 1940-2000 4 2/15/2006 2040-2100 5 2/15/2006 2140-2200
150 6 2/15/2006 2240-2300 1 7 2/15/2006 2340-0000 57 8 2/16/2006 0040-0100 33 9 2/16/2006 0140-0200 31411 10 2/16/2006 0240-0300 11211 11 2/16/2006 0340-0400 1 1 1 6 continuedÂ… 15 16 17 18 19 20 21 22 2 3 2 4 2 5 2 6 2 7 2 8 MG LG E EE M EE L EE E TB TB E TF M T F L T F E L E M L E L L E H 1 2/14/2006 1740-1800 23 7 7 2 2/14/2006 1840-1900 27 7 3 3 2/14/2006 1940-2000 24 7 6 4 2/14/2006 2040-2100 22 7 8 5 2/14/2006 2140-2200 42 5 8 6 2/14/2006 2240-2300 35 3 9 7 2/14/2006 2340-0000 319 8 2/15/2006 0040-0100 178 9 2/15/2006 0140-0200 121 10 2/15/2006 0240-0300 142 11 2/15/2006 0340-0400 15
151 1 2/15/2006 1740-1800 89 1 1 2 2/15/2006 1840-1900 964 3 2/15/2006 1940-2000 3265 3 4 2/15/2006 2040-2100 20 755 5 2/15/2006 2140-2200 100 6 2/15/2006 2240-2300 963 7 2/15/2006 2340-0000 871 8 2/16/2006 0040-0100 82 9 2/16/2006 0140-0200 72 10 2/16/2006 0240-0300 67 11 2/16/2006 0340-0400 73 March 3 -5, 2006 1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 Tow date time Bldsk EC 2 EC 4 EC 8 EC 16 EC 32 EC 64 EM M M L M E B M B L B E G 1 3/3/2006 1740-1800 6 2 3/3/2006 1840-1900 3 3/3/2006 1940-2000 13 4 6 4 5 4 3/3/2006 2040-2100 37 15 2 5 3/3/2006 2140-2200 8 8 6 3/3/2006 2240-2300 9 1 1 4 7 3 7 3/3/2006 2340-0000 16 10 7 6 3 1
152 8 3/4/2006 0040-0100 9 4 8 4 2 9 3/4/2006 0140-0200 5 4 4 6 10 3/4/2006 0240-0300 4 2 5 7 1 11 3/4/2006 0340-0400 1 0 0 1 3/4/2006 1740-1800 9 25 4 2 3/4/2006 1840-1900 19 7 4 8 19 9 3 3/4/2006 1940-2000 18 49 20 6 4 2 4 3/4/2006 2040-2100 1 1 14 18 4 3 2 2 5 3/4/2006 2140-2200 6 0 3 4 6 6 3/4/2006 2240-2300 7 3/4/2006 2340-0000 8 3/5/2006 0040-0100 4 4 5 6 9 3/5/2006 0140-0200 2 1 7 9 10 3/5/2006 0240-0300 1 9 7 0 1 1 11 3/5/2006 0340-0400 continuedÂ… 15 16 17 18 19 20 21 22 2 3 2 4 2 5 2 6 2 7 2 8 MG LG E EE M EE L EE E TB TB E TF M T L T E L E M L L L E H
153 F F E 1 3/3/2006 1740-1800 19 18 2 3/3/2006 1840-1900 15 41 3 3/3/2006 1940-2000 21 44 3 4 3/3/2006 2040-2100 5 41 5 3/3/2006 2140-2200 1 2 6 3/3/2006 2240-2300 3 7 3/3/2006 2340-0000 1 2 8 3/4/2006 0040-0100 1 9 3/4/2006 0140-0200 10 3/4/2006 0240-0300 11 3/4/2006 0340-0400 1 3/4/2006 1740-1800 2 3/4/2006 1840-1900 26 10 3 3/4/2006 1940-2000 17 17 4 3/4/2006 2040-2100 1 5 3/4/2006 2140-2200 1 6 3/4/2006 2240-2300 7 3/4/2006 2340-0000 8 3/5/2006 0040-0100 9 3/5/2006 0140-0200 10 3/5/2006 0240-0300 11 3/5/2006 0340-0400 March 20 -22, 2006 1 2 3 4 5 6 7 8 9 11111
154 0 1 2 3 4 Tow date time Bldsk EC 2 EC 4 EC 8 EC 16 EC 32 EC 64 EM M M L M E B M B L B E G 1 3/20/2006 1740-1800 8 1 2 4 2 3/20/2006 1840-1900 1 4 2 9 4 16 1 1 3 3/20/2006 1940-2000 1 1 34 2 0 6 13 4 3/20/2006 2040-2100 1 5 2 4 52 5 3/20/2006 2140-2200 71 8 48 6 3/20/2006 2240-2300 2 2 6 0 3 6 7 3/20/2006 2340-0000 122 9 05 8 3/21/2006 0040-0100 7 3 8 5 4 9 3/21/2006 0140-0200 7 1 2 6 1 10 3/21/2006 0240-0300 38 2 2 11 3/21/2006 0340-0400 1 3/21/2006 1740-1800 2 2 3/21/2006 1840-1900 1116 178206 2 3 3 3/21/2006 1940-2000 511 2 1 4 5 4 1 0 4 3/21/2006 2040-2100 1 1 6 8 3 5 3/21/2006 2140-2200 9 82
155 6 3/21/2006 2240-2300 4 0 5 46 7 3/21/2006 2340-0000 9 0 1 0 8 3/22/2006 0040-0100 3 0 9 3/22/2006 0140-0200 10 3/22/2006 0240-0300 11 3/22/2006 0340-0400 continuedÂ… 15 16 17 18 19 20 21 22 2 3 2 4 2 5 2 6 2 7 2 8 MG LG E EE M EE L EE E TB TB E TF M T F L T F E L E M L E L L E H 1 3/20/2006 1740-1800 8 5 2 3/20/2006 1840-1900 6 2 3 3/20/2006 1940-2000 7 4 3/20/2006 2040-2100 5 3/20/2006 2140-2200 6 3/20/2006 2240-2300 7 3/20/2006 2340-0000 8 3/21/2006 0040-0100 1 9 3/21/2006 0140-0200 20 10 3/21/2006 0240-0300 607 11 3/21/2006 0340-0400 263143
156 1 3/21/2006 1740-1800 6 2 3/21/2006 1840-1900 44 3 3/21/2006 1940-2000 4 4 3/21/2006 2040-2100 5 3/21/2006 2140-2200 6 3/21/2006 2240-2300 7 3/21/2006 2340-0000 8 3/22/2006 0040-0100 655 9 3/22/2006 0140-0200 298 10 3/22/2006 0240-0300 6337 11 3/22/2006 0340-0400 266014 April 6 -8, 2006 1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 Tow date time Bldsk EC 2 EC 4 EC 8 EC 16 EC 32 EC 64 EM M M L M E B M B L B E G 1 4/6/2006 1740-1800 6524 2 4/6/2006 1840-1900 42513 65 4 5 1 6 3 4/6/2006 1940-2000 1146 3316 1 7 1 9 3 0 4 4/6/2006 2040-2100 5 2 2 7 3 5 4/6/2006 2140-2200 9 73 6 4/6/2006 2240-2300 8 6 1 4 7 4/6/2006 2340-0000 8 0 2 0
157 8 4/7/2006 0040-0100 6 5 2 8 9 4/7/2006 0140-0200 10 4/7/2006 0240-0300 11 4/7/2006 0340-0400 1 4/7/2006 1740-1800 13 6 11 2 4/7/2006 1840-1900 1201115 9810 3 3 3 4/7/2006 1940-2000 1111 5617 6 26 4 4/7/2006 2040-2100 5 4 9 4 6 5 4/7/2006 2140-2200 6 9 22 6 4/7/2006 2240-2300 4 2 5 8 7 4/7/2006 2340-0000 4 8 5 2 8 4/8/2006 0040-0100 5 3 8 9 4/8/2006 0140-0200 10 4/8/2006 0240-0300 11 4/8/2006 0340-0400 continuedÂ… 15 16 17 18 19 20 21 22 2 3 2 4 2 5 2 6 2 7 2 8 MG LG E EE M EE L EE E TB TB E TF M T F L T F E L E M L E L L E H
158 1 4/6/2006 1740-1800 2 6 3 2 4/6/2006 1840-1900 4 3 4/6/2006 1940-2000 4 4/6/2006 2040-2100 5 4/6/2006 2140-2200 6 4/6/2006 2240-2300 7 4/6/2006 2340-0000 8 4/7/2006 0040-0100 7 9 4/7/2006 0140-0200 982 10 4/7/2006 0240-0300 8020 11 4/7/2006 0340-0400 30646 1 4/7/2006 1740-1800 21 3 1 2 4/7/2006 1840-1900 11 3 4/7/2006 1940-2000 4 4/7/2006 2040-2100 5 4/7/2006 2140-2200 6 4/7/2006 2240-2300 7 4/7/2006 2340-0000 8 4/8/2006 0040-0100 57 9 4/8/2006 0140-0200 3367 10 4/8/2006 0240-0300 107812 11 4/8/2006 0340-0400 74 26
159 Table 6.2. Summary data of egg production and spawning, sound production, and water temperature data for all sampling events. Egg Production Wt. Avg. Blastodisc Sunset Start End Duration Date (m3) Time of Spawning Collect Time Time Spawning Spawning Spawning 1/29/06 71.0 18:10 1/30/06 2/14/06 16.9 22:50 22:50 18:20 22:50 22:50 1h 2/15/06 4.8 22:50 22:50 18:20 22:50 22:50 1 h 3/3/06 21.2 20:00 17:50, 19:50 18:30 17:50 21:50 4 h 3/4/06 24.9 17:50 17:50, 18:50 18:30 16:20 19:20 3 h 3/20/06 44.2 18:00 17:50, 18:50 18:40 17:00 22:10 5 h 10 m 3/21/06 93.5 17:20 17:50, 18:50 18:40 15:50 19:00 3 h 10 m 4/6/06 2889.2 17:40 19:50 18:48 15:50 19:20 4 4/7/07 943.7 17:40 18:50 18:48 15:50 19:30 3 h 40 m Table 6.2. Continued. Spawning Maximum Temporal Center of Start End Duration Date Temp. oC SPL Sound Production Chorus Chorus Chorus 1/29/06 20.5 127.2 21:20 17:00 1:40 8h 40m 1/30/06 128.4 16:50 2/14/06 17.6 125.9 22:35 18:30 2:40 8h 10m 2/15/06 18.3 131.9 21:40 17:30 1:50 8h 20m 3/3/06 23.4 122.7 21:20 18:20 0:20 6h 3/4/06 23.6 122.9 21:30 18:20 0:40 6h 20m 3/20/06 25.1 117.0 21:40 19:10 0:10 5h 3/21/06 25.7 113.7 21:45 19:10 0:20 5h 10m 4/6/06 25.1 89.5 NA NA NA NA 4/7/07 25.6 89.9 NA NA NA NA
160 Table 6.3. Black drum egg mortality estimates generated from the standard mortality e quation. Data of egg density and age of a single cohort collected hourly ove r two consecutive nights (1800 Â– 0400 hr) were used in the mortality models. Time series of egg development used in the models began with the blastodisc stage (single cell) for all dates except 1/29 Â– 1/30 and generally end ed with the late embryo stage. Th e Â‘ZÂ’ term is the instantaneous daily mortality rate from the fitted equation. Densities of egg predators collected during sampling ev ents are also presented. Date Fitted Mortality Equation Z r2 Ctenophora (m3) Hydrozoan (m3) *Chaetognatha (m3) 1/29 1/30 y = 66.1789*(EXP(-0.61518*x)) -0.62 0.35 0.3 2/14 2/15 y = 17.5472*(EXP(-0.91141*x)) -0.91 0.44 0.02 0.98, 0.68 3/3 3/4 y = 8.99884*(EXP(-1.65792*x)) -1.66 0.62 3.10, 17.2 3/20 3/21 y = 19.4855*(EXP(-1.78385*x)) -1.78 0.7 0.31 1.47, 1.40 4/6 4/7 y = 1940.51*(EXP(-2.17401*x)) -2.17 0.53 0.34, 0.72 0.30, 1.39 6.44, 17.94 *Chaetognaths are anecdotally reported to be egg predators
161 Table 6.4. Distribution of egg densities with depth. Egg co llections were made and densiti es calculated for each depth listed in the column entitled, Â‘measured mean egg densitiesÂ’. Measured egg densities were then used to model egg densities, with a logarithmic func tion (Fig. 6.9), at all depths including depth intervals not sampled. More than 50% of the eggs were estimated to occur in the upper 0.5m of the w ater column. MeasuredModeled Depth (m) Mean Egg Densities (m3)Egg Densities (m3)% at Depth 0.125 448.8439.733.9 0.5 269.820.8 1 149.6184.814.2 1.5 135.110.4 2 122.599.97.7 2.5 72.55.6 3 55.450.23.9 3.5 31.32.4 4 9.514.91.1
162 Table 6.5. Estimates of egg production (m3) in the 0.25m depth sampled area, in the entire study area to 0.25 m depth, and in the entire study area to 4.0 m depth ar e presented. Estimates of female spawning stoc k biomass and numbers of individual females present in the study area on evenings when egg collections were made are based on ba tch fecundity reported by Macchi et al. (92,886.0); Fitzhugh et al. (262295.1); a nd Nieland et al. (496790.0). Fecundity estimat es were reported as eggs/kilogram of w et, eviscerated body weight. Egg Production (m3) Egg Production (m3)Egg Production (m3) sampled area entire study area entire study area SSB Kg SSB Kg SSB Kg females (n)females (n) Date 0.25m depth 0.25m depth 4.0m de pth *Macchi *Fitzhugh*Nie land*Fitzhugh *Nieland 1/29/06 71.04 1,153,726.623,406,205.9036.7 13.06.92.10.8 1/30/06 2/14/06 16.89 274,278.93809,767.688.7 126.96.36.199.2 2/15/06 4.82 78,311.95231,204.362.5 0.90.50.10.1 3/3/06 21.17 343,784.071,014,971.2410.9 3.92.00.60.2 3/4/06 24.88 403,975.431,192,677.2612.8 188.8.131.52.3 3/20/06 44.24 718,475.342,121,191.3722.8 184.108.40.206.5 3/21/06 93.54 1,519,074.274,484,840.3948.3 17.19.02.81.0 4/6/06 2889.19 46,918,964.78138,521,251.411491.3 528.1278.886.631.5 4/7/07 943.70 15,325,178.3145,245,305.13487.1 172.591.128.310.3
163 Figure 6.1. The canal basin in Cape Coral, Florida where black drum sound production and egg production studies were conducted is shown. Black drum eggs were collected during hourly surface plankton tows from 1800 Â– 0400 hr. along the course indicated by the light blue lines while acoustic recordings were made for 10 seconds every 10 minutes at each location indicated by the yellow filled circles.
164 12/4 12/18 1/1 1/15 1/29 2/12 2/26 3/12 3/26 4/9 70 80 90 100 110 120 130 140 Date 2006SPL dB (re: 1uPa) Figure 6.2. Seasonal time series of black drum sound production recorded in Cape Coral, Florida during 2005 Â– 2006. The two consecutive evenings when egg production and sound production data were collected are represented in red and egg density estima tes (m3) for each of the two evenings are listed below them. N/A, 71.0 16.9, 4.8 21.2, 24.5 44.2, 93.5 28 89.2, 943.7
165 Figure 6.3-A. Sound pressure le vels (right yaxis) and the timing and densities of early cleavage (EC) black drum eggs ( left yaxis) collected during January, 2006. The earliest developmental stage collected on this evening was 32-cells at 22:50 hr. Figure 6.3-B. Sound pressure levels (right yaxis) and the timing a nd densities of early cleavage (EC) black drum eggs (left yaxis) collected duri ng February, 2006. Blastodiscs (single cell) were coll ected on both evenings at 22:50 hr.
166 Figure 6.3-C. Sound pressure levels (right yaxis) and the timing a nd densities of early cleavage (EC) black drum eggs (left yaxis) collected during March 3 and 4, 2006. Blastodiscs (single cell) were collected on both evenings. Figure 6.3-D. Back-calculated times and mean densities of blastodiscs (single cell) based on collection of early cleavag e eggs during March 3 and 4, 2006. Resolution of backcalculated spawning times is limited to 1 hour Back-calculated densities of blastodiscs are corrected for by applying the results of the mortality equation to the age of early cleavage eggs at th e time of collection. 3 3 4 3 3 4
167 Figure 6.3-E. Sound pressure levels and the timing and densities of early cleavage (EC) black drum eggs collected during March 20 and 21, 2006. Blastodiscs (single cell) were collected on both evenings. Figure 6.3-F Back-calculated times and mean de nsities of blastodisc s (single cell) based on collection of early cleavag e eggs during March 20 and 21, 2006. Resolution of backcalculated spawning times is limited to 1 hour Back-calculated densities of blastodiscs are corrected for by applying the results of the mortality equation to the age of early cleavage eggs at th e time of collection. 3 20 21 3 20 21
168 Figure 6.3-G. Sound pressure levels and the timing and densities of early cleavage (EC) black drum eggs collected during April 6 and 7, 2006. Blastodiscs (single cell) were collected on both evenings. Figure 6.3-F Back-calculated times and mean de nsities of blastodisc s (single cell) based on collection of early cleavag e eggs during April 6 and 7, 2006. Resolution of backcalculated spawning times is limited to 1 hour Back-calculated densities of blastodiscs are corrected for by applying the results of the mortality equation to the age of early cleavage eggs at th e time of collection. 4 6 7 4 6 7
169 Figure 6.4-A. Linear regression of egg density (e.g. production) against maximum nightly sound pressure levels including April data. Egg density data are log transformed. Figure 6.4-B. Linear regression of egg density (e.g. production) against maximum nightly sound pressure leve ls excluding April data.
170 Figure 6.5. Linear regression of the weighted average of th e time of spawning against the temporal center of sound production. The we ighted average of the time of spawning is based on times blastodiscs were collected a nd the back-calculated times early cleavage and blastula stages would have been blasto discs and the densitie s at these times. Figure 6.6. Linear regression of the weighted average of the time of spawning against water temperature excluding the January data.
171 Figure 6.7-A. Regression of black drum egg production against water temperature including April data. Egg dens ity data are log transformed. Figure 6.7-B. Linear regression of black dr um egg production against water temperature excluding April data.
172 Figure 6.8. Logarithmic regression of egg dens ity against depth. Th e fitted equation was used to interpolate egg densities at interval s between sampled depths to estimate total egg densities throughout the 4m water column. Figure 6.9. Linear regression of egg developmental stage ag ainst depth. Data of egg developmental stages found at depth are qualita tive only (e.g. densities of these stages are not used).
173 SUMMARY and CONCLUSIONS Sound production associated with the reprodu ctive behavior of estuarine sciaenids was the research focus of this dissertation. The majority of data were collected in Charlotte Harbor, Florida and adjacent canal systems of Cape Coral and Punta Gorda using programmable remotely deployed acousti c recorders which provided high temporal resolution data of fish sound production over exte nded periods. These data were used to examine aspects of the temporal and spatial periodicities of fish sound production on diel and seasonal time scales and in response to changes in environmental conditions. In additional studies, a hydrophone array was used to localize the posit ion of calling black drum and estimate their source levels and potential communication ranges within the canal system of Cape Coral. The tem poral and quantitative relationship of egg production and sound production of black drum wa s also investigated at this same study site location in Cape Coral. In Charlotte Harbor proper, four species were mainly represented in the acoustic recordings of fish sound production; these were the sand seatrout ( Cynoscion arenarius ), the spotted seatrout ( Cynoscion nebulosus ), the silver perch ( Bairdiella chrysoura ) and the oyster toadfish ( Opsanus beta ) was also commonly record ed. Other unidentifiable sounds of biological origin were also recorded. Nightly chor uses were usually comprised of calls produced by these most popular spec ies but were dominated by sand seatrout at all study sites, most of which were located in deeper, un-vege tated areas greater than 2 m
174 depth. The timing of calls during the even ing choruses varied somewhat by species (chapter one). Sound production in Charlotte Harbor, on a diel scale, was typical of the circadian patterns reported previously for these and ma ny other soniferous fish species in which sharp increases in sound production occur in th e late afternoon or early evening and may be sustained for several hours before they dec line somewhat less rapidly than they began. This same general pattern was also document ed for black drum in the canals of Cape Coral and Punta Gorda. This circadia n pattern of sound production has been hypothesized to function as a form of rallying ca ll to bring individuals together in time and space for reproductive activity. Calling at night is believed to confer an advantage for signal transmission because background sound pressure levels are generally lower due to calmer atmospheric conditions. Spawning at night, under the cove r of darkness, is supposed to confer an advantage for eggs ag ainst visual predation. No new evidence was collected during this research to interpret these existing understandings of the diel pattern of nighttime calling and spawning in fi sh sound production much differently from the existing knowledge, but specific experiment s could be designed to test this more rigorously. Seasonal patterns of fish sound production in Charlotte Harbor and the adjacent canal systems agreed well with the documented spawning period of each species recorded. In Charlotte Harbor proper, sound production began around the beginning of February and increased gradually to peak valu es by mid April or early May. Peak values were sustained throughout the summer months then waned in the fall, typically around early November. For each of the major co ntributing species, the acoustic records
175 provided an accurate representation of their seasonal spawning period. In Cape Coral and Punta Gorda black drum sound production was recorded as early as October. Sound pressure levels increased shar ply in January and were sust ained at high levels through March and ceased during April. This patte rn of sound production is in agreement with the documented spawning season of this specie s in the Gulf of Mexico. These acoustic time series data all demonstrate the useful ness of passive acous tic recordings for monitoring periods of reproductive activity of soniferous fish es remotely and with high resolution. Also demonstrated is the usefulness of these methods for studying the reproductive ecology of sound producing fishes as models of estuarine ecosystem function in response to environmental stressors Changes in the spatial distribution of fish sound production in Charlotte Harbor were associated with increased volumes of freshwater inflow and conse quent decreases in bottom c oncentrations of dissolved oxygen (chapter three). An early passive acoustic study was conducte d by Breder (1968) from a dock in Lemon Bay, Florida during 1961 65. Lem on Bay is a narrow, shallow water body located between the barrier islands and mainland near E nglewood, Florida some 20 km west of Charlotte Harbor proper. The major ity of sounds recorded during this study were produced by the catfish, Ariopsis felis (formerly, Galeichthys ) and the toadfish Opsanus beta Recordings were made over diel and s easonal time frames and patterns of sound reported for each species were generally simila r to those that have been recorded since the time of his study; evening choruses of the catfish and day/ night sound production by the toadfish occurred during spring through fall which is the spawning period. A third commonly recorded but unidentif ied fish call, was described by Breder as the Â‘repeaterÂ’,
176 which consisted of up to twelve repetitious so ft tapping sounds. This sound was recorded mainly during the winter months and was abse nt during the summer. This is an unusual period for fish sound production to occur in this region with the exception of the black drum which the Â‘repeaterÂ’ clearly is not, based on the acoustic description. Breder speculated the source of the Â‘repeate rÂ’ may have been the red drum ( Sciaenops ocellatus ), however the red drum spawning season in this region occurs during July Â– October. Although the Lemon Bay habitat was suitable for spotted seatrout ( Cynoscion nebulosus ) and specimens were collected in fish traps at the study site, Breder did not report sound production by it in his study. C. arenarius the dominant sound producer recorded in Charlotte Harbor was also not r ecorded in the Breder study. This is likely explained by the difference in habitat types between the deeper Ch arlotte Harbor study sites and the shallower Lemon Bay study site and consequent limitations of the signal transmission in Lemon Bay. The percolator sound type described for A. felis in the Breder study was not conspicuous in the reco rdings made in Charlotte Harbor, however the species was very likely pr esent at times near the study sites and may have been recorded but this was not found in the reco rdings which were examined closely for species composition. Breder found no real influence of tide on sound production, a possible slight lunar influence, and that temperature was th e major variable influencing sound production on a seasonal basis. These c onclusions are also consistent with data collected from Charlotte Harbor. While results of passive acoustic st udies of fish sound production have demonstrated that it is an effective tool for estimating the seasonal extent of spawning activity by many sound producing fishes a co mparison between daily levels of sound
177 production and egg production by black drum did not reveal a positive relationship. This may have been due to the great variability in batch size and fecundity of similarly sized black drum females or females of different sizes and fecundities may have entered the study area on different days when eggs were collected. This result may underlie the importance of interpreting measurements of sound production in the context of male behavior (courtship) and not as a direct measurement of secondary productivity (e.g. spawning), at least in the case of black drum. Gilmore (1994 and 2007), however, found a positive correlation between egg/larval abundance and sound pressure levels/call occurrence by silver perch ( Bairdiella chrysoura ), spotted seatrout ( Cynoscion nebulosus ), and sand seatrout ( Cynoscion arenarius ). The contrasti ng results of these studies may be due to differences associated with variability of f ecundity and batch size among these species, the sex ratio in the sp awning aggregations where data were collected, and possibly in th e experimental methods used in each study. Further examination of the relationship between e gg and sound production le vels is warranted. Hydrophone array recordings we re used to estimate sour ce levels of black drum calls. When source level estimates were combined with measurements of auditory sensitivity, signal transmission loss, and backgr ound levels at the study site estimates of the potential communication range of black drum were possible. An additional fortuitous result of this study was the ability to track the movement of an individual fish by localizing its position when it called, which occurred every 3. 5 seconds on average over one hour. A larger more complex hydrophone arra y would have been able to resolve this information for other calling individuals in th e study area too. Direct visual observations of fish behavior associated with sound pr oduction are not often po ssible therefore much
178 could be learned with this technique about the spacing and interactions of individuals within a spawning aggregation and the role of sound production in sexual selection of black drum. Acoustic and visual signaling during courtship and spawning by terrestrial animals and fishes is generally performe d by the male. While mating systems and parental care vary within and between thes e classes of animals se xual selection theory would predict that similarities would exist in their courtship behavi or. In the case of terrestrial animals, birds fo r example, sound production may be used alone as in the burrowing petrels (Bretagnolle, 1990) or in co njunction with visual displays and/or exaggerated male features as in the sa ge grouse (Vehrencamp et al., 1989) during courtship or associated male -male agonistic behavior. Pa rallels to this have been observed in fishes, particularly in reef fish es where visual observations of courtship and territorial behavior have been possible. For example, the domino damselfish (Mann and Lobel, 1998) performs a Â‘signal jumpÂ’ in which Â‘popsÂ’ and Â‘chirpsÂ’ are produced in association with rapid upward swimming followed by a sharp dive (e.g. the Â‘jumpÂ’) usually near a nest site. While more obser vations have been made on the behaviors of terrestrial animals than fishes (for obvious reasons), some of the information learned from them can be applied to fishes and al so used to further test hypotheses of fish behavior, including sexual sel ection. In this regard, a pa rticular aspect of social modulation of reproduction that w ould be especially interesting to investigate is the role of male sound quality in female choice by bl ack drum and how the call qualities are interpreted by the female, if in fact they are at all.
179 The usefulness of passive acoustics for studying the timing and location of reproductive behavior in s ound producing fishes has long been known. However, the methods have not yet become incorporated in to the regular protocol of scientists who study the reproductive biology and ecology of many of the sound producing species, including the drums and groupers. Advancem ents in recording technology and data analysis software, combined with lower equipm ent costs will increase the availability and use of passive acoustic methods in the near future. The recording instrumentation available today can record high temporal re solution data for periods in excess of one year. The advantage of remotely deployed in strumentation which co llects high temporal resolution data is that while trends in normal activity on different time scales are recorded, rare events are also recorded and their effects on normal activity are evident. This was the case for the data recorded on hurricane Charley, which provided new information about fish behavior during a hurricane. The ability to collect enormous acoustic tim e series data sets on wide spatial and temporal scales means that many sounds of unknown origin will be recorded. It is relatively straightforward for the trained obser ver to categorize the or igin of a recorded sound as biological or not and to a slightly lesser extent whether it was produced by a fish, marine mammal, or invert ebrate, although exceptions to th ese generalities will likely be discovered. Future efforts should fo cus on positively identifying the origin of unknown sounds. This can be done via direct in situ observations if conditions of visibility allow or by capture and isolati on of the sound producer, which may also be done in situ under the right circumstance.
180 From a resource assessment and manage ment perspective, passive acoustics represent a powerful tool, as evidenced by the data in chapter three. To realize this potential environmental data must also be collected in conjunction with sound production data whenever possible to help explain the vari ability in acoustic time series data of fish sound production. Many water qua lity monitoring stations now exist in coastal and near shore environments and a hydrophon e could easily be incorporat ed into the sensor array for this purpose. Estuarine systems where information is needed on the impact of regulated freshwater inflow on ecosystem f unction are especially relevant to these methods. REFERENCES Breder CM Jr. (1968) Seasonal a nd diurnal occurrences of fish sounds in a small Florida Bay. Bull. Am. Mus. Nat. 138(6):327-377 Bretagnolle V (1990) Behavioural affinities of the Blue petrel Halobaena caerulea. Ibis 132:102-105 Gilmore RG Jr. (1994) Environmental parame ters associated with spawning, larval dispersal and early life histor y of the spo tted seatrout, Cynoscion nebulosus (Cuvier). Final Program Rev., Contract N o. LCD 000. Mar. Res. Inst., Fla. Dept. Environ. Protection, St. Petersburg, Fla. Gilmore RG Jr. (2007) Fish monitoring fi nal report 2006 work for the RECOVER monitoring and assessment plan: Northern Estuaries. Final Report to Florida Oceanographic Society, Stuart, Flor ida, & SFWMD, W. Palm Beach. Mann DA, Lobel PS (1998) Acoustic be havior of the damselfish ( Dascyllus albisella ):
181 behavioral and geographic variation. Environ. Biol. Fishes. 51(4):421-428 Vehrencamp, SL, Bradbury JW, and Gibson RM ( 1989) The energetic cost of display in male sage grouse. Anim. Behav. 38(5):885-896
182 ABOUT THE AUTHOR James V. Locascio received a bachelo rÂ’s degree in marine biology from Texas A&M University at Galveston. He gained a va riety of lab and field experience thereafter ranging from molecular biology and genetics to fisheries research in Alaska and later traveled extensively through Central America. Prior to enrolling at the College of Marine Science (CMS) he worked at a small marine la b and also as a profe ssional guide in the J.N. Â‘DingÂ’ Darling National Wildlif e Refuge on Sanibel Island, Florida. His graduate research on sound production of fishes has been published in peer reviewed journals and featured on the Discove ry, Science, and Hist ory channels, in New York Times Science and on the Colbert Report. The experience has been entertaining as well as educational. As a graduate student at the CMS, James was awarded the J.N. Ding Darling Education Foundation Scholarship, International Wo menÂ’s Fishing Scholarship (twice), Wachovia Endowed Fellowship (C MS), C.W. Bill Young Fellowship (CMS), and the Knight Fellowship (CMS). James looks forward to life after graduate school, sharing his knowledge with others and suppor ting those who desire to learn and achieve in the way he has been supported by so many throughout his education.