2pAB – Sound production of the critically endangered totoaba: applying underwater sound detection to fish species conservation

Goldie Phillips – gphillips@sci-brid.com
Sci-Brid International Consulting, LLC
16192 Coastal Hwy
Lewes, DE 19958

Gerald D’Spain – gdspain@ucsd.edu
Catalina López-Sagástegui – catalina@ucr.edu
Octavio Aburto-Oropeza – maburto@ucsd.edu
Dennis Rimington – drimington@ucsd.edu
Dave Price – dvprice@ucsd.edu
Scripps Institution of Oceanography,
University of California, San Diego
9500 Gilman Drive
San Diego, CA 92093

Miguel Angel Cisneros-Mata – macisne@yahoo.com
Daniel Guevara – danyguevara47@hotmail.com
Instituto Nacional de Pesca y Acuacultura (INAPESCA) Mexico
Del Carmen, Coyoacán
04100 Mexico City, CDMX, Mexico

Popular version of paper 2pAB
Presented Tuesday afternoon, December 3rd, 2019
178th ASA Meeting, San Diego, CA

The totoaba (Figure 1), the largest fish of the croaker family, faces a severe illegal fishing threat due largely to the high value of its swim bladder (or buche; Figure 2) in Asian markets. While several conservation measures have been implemented in the Gulf of California (GoC) to protect this endemic species, the totoaba’s current population status remains unknown. Passive acoustic monitoring (PAM) – the use of underwater microphones (called hydrophones) to detect, monitor, and localize sounds produced by soniferous species – offers a powerful means of addressing this problem.

Croaker fishes are well known for their ability to produce sound. Their characteristic “croaking” sound is produced by the vibration of their swim bladder membrane caused by the rapid contraction and expansion of nearby sonic muscles. As sound propagates very efficiently underwater, croaks and other sounds produced by species like the totoaba can be readily detected and recorded by specialized PAM systems.

However, as little is known about the characteristics of totoaba sounds, it is necessary to first gain an understanding of the acoustic behavior of this species before PAM can be applied to the GoC totoaba population. Here we present the first step in a multinational effort to implement such a system.


Figure 1. Totoaba housed at CREMES


Figure 2. Totoaba swim bladder.

We conducted a passive acoustic experiment at the aquaculture center, El Centro Reproductor de Especies Marinas (CREMES), located in Kino Bay, Mexico, between April 29 and May 4, 2019. We collected video and acoustic recordings from totoaba of multiple age classes, both in isolation and within group settings. These recordings were collectively used to characterize the sounds of the totoaba.

We found that in addition to croaks (Video 1) captive totoaba produce 4 other call types, ranging from short duration (<0.03s), low-frequency (<1kHz) narrowband pulses, classified here as “knocks” (Video 2), to longer duration, broadband clicks with significant energy over 10kHz. There is also indication that one of the remaining call types may function as an alarm or distress call. Furthermore, call rates and dominant call type were found to be dependent on age.

Video 1. Visual representation (spectrogram) of a croak produced by totoaba at CREMES. Time (in minutes and seconds) is shown on the x-axis with frequency (in kHz) displayed on the y-axis. Sounds with the greatest amplitude are indicated by warmer colors.

Video 2. Visual representation (spectrogram) of a series of “knocks” produced by totoaba at CREMES.

As PAM systems typically produce large amounts of data that can make manual detections by a human analyst extremely time-consuming, we also used several of the totoaba call types to develop and evaluate multiple automated pre-processing/detector algorithms for a future PAM system in the GoC. Collectively, results are intended to form the basis of a totoaba population assessment that spans multiple spatial and temporal scales.

1pAB4 – Combining underwater photography and passive acoustics to monitor fish

Camille Pagniello – cpagniel@ucsd.edu
Gerald D’Spain – gdspain@ucsd.edu
Jules Jaffe – jjaffe@ucsd.edu
Ed Parnell – eparnell@ucsd.edu

Scripps Institution of Oceanography, University of California San Diego
La Jolla, CA 92093-0205, USA

Jack Butler – Jack.Butler@myfwc.com
2796 Overseas Hwy, Suite 119
Marathon, FL 33050

Ana Širović – asirovic@tamug.edu
Texas A&M University Galveston
P.O. Box 1675
Galveston, TX 77550

Popular version of paper 1pAB4 “Searching for the FishOASIS: Using passive acoustics and optical imaging to identify a chorusing species of fish”
Presented Monday afternoon, November 5, 2018
176th ASA Meeting, Victoria, Canada

Although over 120 marine protected areas (MPAs) have been established along the coast of southern California, it has been difficult to monitor their ability to quantify their effectiveness via the presence of target animals. Traditional monitoring methods, such as diver surveys, allow species to be identified, but are laborious and expensive, and heavily rely on good weather and a talented pool of scientific divers. Additionally, the diver’s presence is known to alter animal presence and behavior. As one alternative to aid and perhaps, in the long run, replace the divers, we explored the use of long-term, continuous, passive acoustic recorders to listen to the animals’ vocalizations.

Many marine animals produce sound. In shallow coastal waters, fish are often a dominant contributor. Aristotle was the first to note the “voice” of fish, yet only sporadic reports on fish sounds appeared over the next few millennia. Many of the over 30,000 species of fish that exist today are believed to produce sound; however, the acoustic behavior has been determined in less than 5% of these biologically and commercially important animals.

Towards the goal of both listening to the fish and identifying which species are vocalizing, we developed a Fish Optical and Acoustic Sensor Identification System (FishOASIS) (Figure 1). This portable, low-cost instrument couple’s a multi-element passive acoustic array with multiple cameras, thus allowing us to determine which fish are making which sound for a variety of species. In addition to detecting sporadic events such as fish spawning aggregations, this instrument also provides the ability to track individual fish within aggregations.


Figure 1. A diver deploying FishOASIS in the kelp forest off La Jolla, CA.

Choruses (i.e., the simultaneous vocalization of animals) are often associated with fish spawning aggregations and, in our work, FishOASIS was successful in recording a low-frequency fish chorus in the kelp forest off La Jolla, CA (Figure 2).

Figure 2. Long-term spectral average (LTSA) of low-frequency fish chorus of unknown species on June 8, 2017 at 17:30:00. Color represents spectrum level, with red indicating highest pressure level.

The chorus starts half an hour before sunset and lasts about 3-4 hours almost every day from May to September. While individuals within the aggregation are dispersed over a large area (approx. 0.07 km2), the chorus’ spatial extent is fairly fixed over time. Species that could be producing this chorus include kelp bass (Paralabrax clathratus) and halfmoons (Medialuna californiensis) (Figure 3).

Figure 3. A halfmoon (Medialuna californiensis) in the kelp forest off La Jolla, CA.

FishOASIS has also been used to identify the sounds of barred sand bass (Paralabrax nebulifer), a popular species among recreational fishermen in the Southern California Bight (Figure 4).

Figure 4. Barred sand bass (Paralabrax nebulifer) call.

This article demonstrates that combining multiple cameras with multi-element passive acoustic arrays is a cost-effective method for monitoring sound-producing fish activities, diversity and biomass. This approach is minimally invasive and offers greater spatial and temporal coverage at significantly lower cost than traditional methods. As such, FishOASIS is a promising tool to collect the information required for the implementation of passive acoustics to monitor MPAs.