Can Passive Acoustics be Used to Estimate the Length of Atlantic Croaker (Micropogonias undulatus) within the Pamlico Estuary?
Loading...
Date
2010
Access
Authors
Krahforst, Cecilia S.
Journal Title
Journal ISSN
Volume Title
Publisher
East Carolina University
Abstract
The Atlantic States Marine Fisheries Council (AFMFC) attributes the decline in the Atlantic croaker (Micropogonias undulatus) stock to over-fishing practices. The majority of collected Atlantic croaker are small, young of year (YOY) individuals, which may lead to changes in age and size-class structure within the population. North Carolina and other southeastern states receive economic benefits from the fishing practices of Atlantic croaker and other important Sciaenid fishes. The goal of this research is to determine if Atlantic croaker populations can be monitored using passive acoustics by ascertaining if sound-production of a population will provide the listener with information such as length, maturity, or sex. An instrumented tripod (ITPod) was deployed to estimate the physical parameters (i.e. currents, water quality, waves, and turbidity) of the water and a passive acoustic recording system recorded environmental sounds (<10 kHz) every 10 s at 15 min intervals at two sites in the Pamlico Sound from June to November 2008. Once a month at each site, Atlantic croaker were collected using a juvenile otter trawl and a gillnet, in addition an echosounder unit was simultaneously deployed to determine the size-selectivity of the nets. Laboratory results of Atlantic croaker sound production revealed that the fundamental frequency is inversely correlated to the length of the fish. Based on these captive fish recordings, a linear regression analysis revealed that total length (TL, mm) was inversely related to fundamental frequency (F[subscript]0, Hz), where F[subscript]0 = 1073.95 - 3.12 (TL), (R[superscript]2 = 0.84). Sex was not a significant factor influencing fundamental frequency for developing Atlantic croaker. Analysis of the length, mass, and area of the swimbladder, sonic muscles, and gonads revealed that all of these internal structures affected the fundamental frequency, but all were also highly correlated with the length of Atlantic croaker. Therefore, the data indicate that the fundamental frequency (Hz) in field recordings could be used to predict length (mm) with the same data using the following empirical relationship:TL = 305.323 - 0.270 (F[subscript]0). This linear regression equation collected from Atlantic croaker in captive conditions was used to predict the average length of Atlantic croaker from the passive recordings at the two field sites. These predictions were compared to the average lengths of Atlantic croaker collected in the trawl and gillnet, as well as the average length of the fish calculated from an active acoustic echosounder. When comparing the mean predicted length estimates from the passive acoustic device to the mean length of Atlantic croaker collected in the trawls, there was a significant difference for all months (ANOVA, p<0.05) except June (ANOVA, p=0.37). An analysis comparing the lengths of all fishes collected in the trawl, gillnet, and active acoustics showed the selectivity of the nets. There was a significant difference in mean fish length among gear-types (ANOVA, p<0.05) for each month. Because each net was size selective when considered alone, I assumed they would not be size selective when the lengths of fishes collected in the gillnet and trawl net were combined. There was no significant difference between the fish community length estimates of the nets and that of the active acoustic echosounder (ANOVA, p>0.05) for all months except June (ANOVA, p=0.01). Next, a comparison was made between the Atlantic croaker the predicted mean length from the passive acoustic recordings, the mean lengths from the trawls, and the mean length predicted from the active acoustic echosounder surveys. There was no significant difference between the mean Atlantic croaker total lengths from the trawl and all of the fish in the active acoustics (ANOVA, p>0.05), except for June (ANOVA, p=0.001). Therefore, I conclude that the passive acoustic estimates of Atlantic croaker lengths (predicted from fundamental frequency) obtained in the laboratory did not correctly predict the observed length-structure of the Atlantic croaker population. This result may be due to the cutoff frequencies of the sites; in shallow water, the dominant frequency of an acoustic signal can change as it propagates. If the wavelength of the low-frequency component of a signal exceeds the water depth (the cutoff frequency), than these frequencies are essentially filtered from the recorded sound. Another factor influencing cutoff frequency is that of the resonance of trapped bubbles under the surface of the sediment. These bubbles can cause their own echoes when the sound reverberates off of the surface of the bubble. The bubble reverberations will add higher frequency components to the original sound. Together these phenomena would have increased the dominant frequency at the hydrophone, leading to an underestimate of length-structure for an Atlantic croaker population (the higher dominant frequencies would predict smaller fish observed). My research shows that Atlantic croaker length is inversely related to the fundamental frequency of the "croak" call, but the acoustic environment and the effects of the cut-off frequency need to be better understood prior to using this method in the field to estimate Atlantic croaker lengths.