Acoustic Abundance Estimates Research Articles

Acoustic estimates of sperm whale abundance in the Mediterranean Sea as part of the ACCOBAMS Survey Initiative

Acoustic surveys for sperm whales (Physeter macrocephalus) were conducted in the Mediterranean Sea in summer 2018 as part of the vessel-based component of the ACCOBAMS Survey Initiative (ASI). Equal-spaced zigzag transects provided uniform coverage of key sperm whale habitats and were surveyed using a towed hydrophone array deployed from a research vessel at speeds of 5-8 knots. A total of 14,039 km of tracklines were surveyed in the western basin, Hellenic Trench and Libyan waters, with an acoustic coverage of 10% realised for sperm whales. During these surveys, 254 individual sperm whales were detected on the trackline, with an additional 66 individuals off-track. Sperm whales were only seen ten times on-track, with an additional 16 off-track sightings. Estimates of slant range to echolocating whales were used to derive density estimates through both design- and model-based distance sampling methodologies. An acoustic availability of 0.912 (sd = 0.036) was derived from via published models. When correcting for availability bias, a design-based abundance estimates of 2,673 individuals (95% CI 1,739-4,105; CV = 0.21) was derived for the surveyed blocks, which incorporated most known sperm whale habitat in the Mediterranean Sea. The equivalent model-based estimate was 2,825 whales (2,053-3,888; CV = 0.16). Over 97% of detected whales were in the western basin, with highest densities in the Algerian and Liguro-Provencal Basins between Algeria and Spain/France. In the eastern basin, detections were sparse and concentrated along the Hellenic Trench. A density surface modelling (DSM) exercise identified location and benthic aspect as being the most instructive covariates for predicting whale abundance. Distance sampling results were used in a power analysis to quantify the survey effort required to identify population trends. In the most extreme scenario modelled (10% per annum decline with decennial surveys), the population could have dropped by 90% before the decline was identified with high statistical power. Increasing the regularity of surveys would allow population trends to be detected more expediently. Mediterranean sperm whales are listed as Endangered on the IUCN’s Red List and the need for urgent conservation measures to reduce injury and mortality remains paramount for this unique sub-population.

Ex situ target strength of yellow croaker (Larimichthys polyactis) in a seawater tank

We conducted a TS study of yellow croaker (Larimichthys polyactis), a major commercial fish in Asia that has not been studied before. Ours is the first study that examines the relationship between target strength (TS) and fish length using 38, 70, and 120 kHz split-beam echosounders. We measured TS in five size groups of yellow croaker (Larimichthys polyactis) based on length (8.2–26.5 cm total length) in a seawater tank and calculated the Kirchhoff-ray mode (KRM) model. When the slope of the least-squares regression line was forced to 20 in the TS equation, the resulting values for the constant term (b20) at each frequency (38, 70, and 120 kHz) using the caged method were –67.69, –67.52, and −68.83 dB, respectively. The swimming angle (mean ± S.D.) in the net cage was − 4.2 ± 21.9°. The b20 values for the KRM model were –65.01, –64.74, and −66.27 dB at 38, 70, and 120 kHz, respectively. In both methods, the KRM model’s b20 values were relatively higher than those of the caged method by 2.68, 2.78, and 2.56 dB at 38, 70, and 120 kHz, respectively. Such TS ex situ measurements can be applied to improve the yellow croaker’s acoustic abundance estimates.

Sperm whale acoustic abundance and dive behaviour in the western North Atlantic

Sperm whales are an ideal species to study using passive acoustic technology because they spend the majority of their time underwater and produce echolocation clicks almost continuously while foraging. Passive acoustic line transect data collected between June and August 2016 were used to estimate a depth-corrected acoustic abundance and study the dive behaviour of foraging sperm whales in the western North Atlantic Ocean. 2D localizations (n = 699) were truncated at a slant range of 6500 m and combined with the multipath arrivals of surface reflected echoes to calculate 3D localizations (n = 274). Distance sampling using depth-corrected perpendicular distances resulted in a 10.5% change in the acoustic abundance estimate (2199 whales, CV = 14.6%) compared to uncorrected slant ranges (1969 whales, CV = 14.1%), and a detection function that was a better fit for the data. Sperm whales exhibited multiple foraging strategies, with bottom phases occurring at depths of 400–800, 800–1200, or > 1200 m, accounting for an average 39.2, 49.5, or 44.9% of the total recorded dive time, respectively. These results suggest that estimating 3D localizations using acoustic line transect data improves acoustic abundance estimation and can be used to answer population level questions about foraging ecology.

Fisheries acoustics in Norway—Wide band data, autonomous platforms and deep learning

This presentation gives an overview of recent achievements in fisheries acoustics associated with the Center for Research Based Innovation in Marine Acoustic Abundance Estimation and Backscatter Classification (CRIMAC), Norway. We present a data processing pipeline from raw acoustic data to survey indices of abundance for fisheries assessments as a test bench. This includes the raw data as well as manually worked up labels for acoustic target classification (ATC). We also collect data from dedicated CRIMAC surveys and organize broadband data across a range of species, using different echo sounder setting, to find optimal settings for various standard surveys. We have trained deep neural networks on labelled acoustic data for ATC and tested this in the pipeline. Semi-supervised methods are also developed, requiring only 10% of the training data compared to the supervised models. IMR has a developed a strategy for phasing in unmanned platforms in our surveys. The process is stepwise, where we first augment existing surveys to maintain time series integrity; this approach has been tested on our sand eel and sprat surveys. We also embed our deep learning models in containers to facilitate simple deployment. This allows adaptive survey strategies, which is a step towards fully autonomous acoustic surveys.

Исследование акустической силы цели японской скумбрии в северо- западной части Тихого океана in situ

Significant concentrations of Chub mackerel (Scomber japonicas) were found in 2014 in the area of the Southern Kuril Islands (Northwestern Pacific Ocean (NWPO)). Commercial fishing for Chub mackerel was resumed in the NWPO in 2016. In this connection, exploratory work and acoustic-trawl surveys are carried out to identify distribution and abundance of Chub mackerel in the NWPO. Acoustic abundance estimates of the stock size of Chub mackerel is important for fishing. However, the reflectivity of Chub mackerel is not well understood. Collections of in situ target strength (TS) signals and corresponding fish size measurements for Chub mackerel were obtained as a result of research in the NWPO (2015–2017). The calibrated scientific echo-sounder “EK‑60 Simrad” with two split-beam echo sounders of respectively 38 and 120 kHz was used to collect sonar data. Measurement of the TS of Chub mackerel was conducted at the trawl layer, which corresponds to the trawl opening per the trawling area. The echogram selection process and the filtering process of single targets of Chub mackerel are realize in the “SALTSE” software package (“TINRO”). The dependence of mean target strength (TS, dB) on mean fish length (L, cm) for was estimated as 20 log (L) — 72,8 dB at 38 kHz and as 20 log (L) — 73,6 dB at 120 kHz. The obtained values backscattering strength (Sv) of Chub mackerel in the NWPO are compared with the values Sv of other studies of the reflective properties of Scombridae.

Size- and density-dependent overwinter mortality of anchovy juveniles in the Bay of Biscay

For small pelagic fish, recruitment success is mainly determined by the degree of survival over the early stages of life. In the case of European anchovy in the Bay of Biscay, major mortality events take place during spring and summer, and thus, the juvenile population in autumn is considered as an indicator of the next year recruitment. However, overwinter mortality could be an additional process regulating anchovy recruitment success in this area. Anchovy juveniles winter mortality was estimated over eleven years, between 2003 and 2013, using acoustic estimates of juvenile abundance (in autumn) and age-1 adults (the following spring) from two independent surveys. The results show that mortality is highly variable among years, with values ranging between 0.004 and 0.016 d−1. Size- and density-dependent mortality were also evaluated during the analysis. Observed size distributions of juveniles in autumn were projected to the following spring assuming different size specific growth rate distributions in the population. Thereafter, the projected size frequency distributions were compared with the observed ones. The results indicate that mortality is higher for small individuals in most of the years (7 of 11). The mortality rate is positively correlated with juvenile abundance in autumn but not with mean juvenile length. The results suggest that although a density-dependent effect could be one of the main factors regulating the survival of juveniles during winter, size-dependent mortality could also impact recruitment. Understanding the factors that trigger size-dependent mortality during winter could contribute to the management of this fishery.

Song variability and change in the Australian pygmy blue whale song—Implications for passive acoustic abundance estimation

Blue whales were heavily exploited throughout the early 20th century with many populations hunted to near extinction. Today, several known sub populations of blue whale exist, separated by geographic range and the acoustic signals they produce. The eastern Indian Ocean subpopulation of pygmy blue whales is easily identified by the production of a characteristic Australian song type. Long term collection of passive acoustic data from the Perth Canyon, Western Australia and Portland, Victoria, have revealed unprecedented levels of variability in the vocal behaviour of Australian pygmy blue whales. These levels of variability have significant implications for any attempts to assess the abundance of this population with acoustic techniques. Further, these findings may indicate higher levels of cognitive capability than previously expected in blue whales. The mechanisms behind variability in song production are unclear, though there is research to suggest that changes in vocal behaviour may be culturally driven or caused by changes in the ambient noise and environmental conditions. These findings highlight shortfalls in the current methods for measuring acoustic abundance, as well as indicate that higher than expected levels of variability may exist in the vocal behaviour of other pygmy blue whale subpopulations.

Acoustic identification of marine species using a feature library

Sonars and echosounders are widely used for remote sensing of life in the marine environment. There is an ongoing need to make the acoustic identification of marine species more correct and objective and thereby reduce the uncertainty of acoustic abundance estimates. In our work, data from multi-frequency echosounders working simultaneously with nearly identical and overlapping acoustic beams are processed stepwise in a modular sequence to improve data, detect schools and categorize acoustic targets by means of the Large Scale Survey System software (LSSS). Categorization is based on the use of an acoustic feature library whose main components are the relative frequency responses. The results of the categorization are translated into acoustic abundance of species. The method is tested on acoustic data from the Barents Sea, the Norwegian Sea and the North Sea, where the target species were capelin (Mallotus villosus L.), Atlantic mackerel (Scomber scombrus L.) and sandeel (Ammodytes marinus L.), respectively. Manual categorization showed a high conformity with automatic categorization for all surveys, especially for schools.

Marine mammal passive acoustics applied to the monitoring of long-term trends in beaked whale abundance and to the derivation of a behavioral risk function for exposure to mid-frequency active sonar

Knowledge of Cuvier’s (Ziphius cavirostris) and Blainville’s (Mesoplodon densirostris) beaked whales’ distinct dive and vocal behavior has allowed for the development of multiple methods of passive acoustic abundance and density estimation (Marques et al., 2009, Moretti et al., 2010). These methods are being applied to multiple years of data to estimate long-term trends in abundance for Blainville’s beaked whales at the Atlantic Undersea Test and Evaluation Center (AUTEC) in the Bahamas and at the Pacific Missile Range Facility (PMRF) in Hawaii, and for Cuvier’s beaked whales at the Southern California Offshore Range (SCORE). These passive acoustic beaked whale dive data were combined with sonar and Range ship-track data to derive a behavioral risk function for Blainville’s beaked whales at AUTEC, and are being extended to Cuvier’s beaked whales at SCORE and Blainville’s beaked whales at PMRF. The behavioral risk function maps the probability of beaked whale dive disruption as a function of the receive le...

Characterizing diversity and variation in fish choruses in Darwin Harbour

AbstractThe diversity, intensity, and periodicity of fish sounds can provide a wealth of information on spatial and temporal distribution of soniferous fish and, on occasion, which environmental factors these choruses are driven by. Such information can help predict species presence and understand their movement patterns in the long term. At three sites in Darwin Harbour, Australia, sea-noise loggers on the harbour floor recorded ambient noise over a 2-year period. Many fish calls and nine different chorus types were detected over 50 Hz to 3 kHz. Source species were speculated for four of the choruses and source levels, a precursor to passive acoustic abundance estimates, were identified for two of these. Other calls displayed similarities to choruses detected elsewhere in Australia. All choruses displayed diel cycles with semi-lunar patterns present for three of the chorus types. Time of sunset and temperature were also significantly related to the presence of the most predominant chorus and while not statistically significant, height of high tide and salinity also appeared related. A lack of frequency and temporal partitioning in calling across the choruses in hours of darkness (after sunset) illustrates the complexity of monitoring communities of different vocal species. The study has outlined some of the patterns biological sounds exhibit, which has significant implications for sampling strategies when using soundscapes for temporal and spatial predictive modelling.

Measuring the distribution, abundance, and biovolume of zooplankton in an oligotrophic freshwater lake with a 710 kHz scientific echosounder

AbstractAcoustic surveys of the distribution and abundance of freshwater zooplankton were conducted in Lake Giles, an oligotrophic freshwater lake. Volume backscatter data from a 710 kHz scientific echosounder were converted to high‐resolution spatial and temporal numerical density estimates of small zooplankton. Vertical net tows of a 153 μm mesh closing bongo net at multiple depth intervals provided both identification of the types and sizes (0.5–1.5 mm length) of crustacean zooplankton present in the lake as well as an independent measurement of zooplankton numerical density. Net and acoustic estimates of zooplankton abundance, biovolume, and distribution were very similar. The improved resolution of the high‐frequency acoustic sampling provides insight into several aspects of freshwater zooplankton ecology including: separation of migrating and non‐migrating zooplankton, high resolution measurements of in situ zooplankton biovolume, calculation of in situ vertical velocities of migrating zooplankton, and fine‐scale (sub‐meter) horizontal and vertical zooplankton distribution during daytime, nighttime, and vertical migration events. These methods allow for more detailed and accurate estimates of zooplankton distribution than traditional net sampling methods can provide, including determining the total abundance of organisms within a specific habitat. They also provide higher resolution data in both space and time of smaller zooplankton taxa than have been measured previously in freshwater ecosystems.

Industry acoustic surveys of spawning southern blue whiting on the Bounty Platform, New Zealand

Annual acoustic surveys of spawning southern blue whiting (Micromesistius australis) on the Bounty Platform, southeast of New Zealand, have been carried out using industry vessels since 2004. In most years, surveys were carried out from a single vessel, while in 2009, acoustic data were collected from three vessels. The survey approach in all years was the same—vessels with calibrated Simrad ES60/ES70 echosounders and hull-mounted 38kHz transducers conducted aggregation-based surveys using an adaptive design. Surveys attempted to cover all areas of high southern blue whiting density. In most years there were multiple snapshots of the same aggregation. The resulting biomass was used as a relative estimate of spawning southern blue whiting abundance. There was a very large (seven-fold) increase in estimated biomass of southern blue whiting at the Bounty Platform from 2006 to 2007, which was due to the recruitment of one very strong year class (2002) into the spawning population. The estimated biomass from 2008 was also high, but biomass declined by a factor of four in 2009. The observed decline in acoustic estimates between 2008 and 2009 was too great to be explained solely by fishing and average levels of natural mortality. The very large changes in estimated abundance between years, and also between snapshots within a year, are related mainly to changes in survey temporal and spatial coverage, and illustrate an important limitation on interpretation of aggregation-based acoustic abundance estimates. In each snapshot an unknown proportion of the spawning aggregation is surveyed, and almost certainly not the entire spawning stock. Survey coverage depended on both the amount of survey time available (which is often limited by commercial constraints) and the behaviour of the fish (e.g., the extent and density of the aggregation, and the timing of spawning). It is therefore difficult to incorporate the resulting series of abundance estimates into a formal stock assessment model as a time series. Despite this, industry acoustic surveys of the Bounty Platform have led directly to management decisions and changes in catch limits.

Hydroacoustic studies on the commercially important squid Sthenoteuthis oualaniensis in the South China Sea

Knowledge of target strength (TS) is essential for accurate assessment of fish abundance using acoustic methods. The in situ TS (dB) of purpleback flying squid Sthenoteuthis oualaniensis in the South China Sea was measured with 70 and 120kHz split-beam echosounders. Biological samples were collected using a light falling-net with a valid working depth of 50m. Acoustic data sets in which S. oualaniensis contributed more than 90% of the total catch by number were selected to analyse in situ TS and determine relationships between TS and squid mantle length (ML). In situ TS increased with squid density when density was greater than a cut-off point, indicating a bias from accepting multiple targets at high squid density. At 70kHz the TS to ML of squid was: TS70kHz=19.13logML−77.84 (n=9, R2=0.8782, ML: 7.5–25cm); at 120kHz the TS to ML of squid was: TS120kHz=26.68logML−88.00 (n=9, R2=0.8521, ML: 7.5–25cm). This study provides new data for improving the acoustic abundance estimation of squid S. oualaniensis, and may be a useful reference for TS estimation of other squid species.

A joint stock assessment for the anchovy stock of the northern and central Adriatic Sea: comparison of two catch-at-age models

Anchovy (Engraulis encrasicolus, L.) is one of the most important commercial species of the northern and central Adriatic Sea, as well as one of the most productive fisheries in the whole Mediterranean. In the Adriatic Sea the stock of anchovy is shared between Italy, Croatia and Slovenia. A joint stock assessment was carried out using catch data from all the fleets for the time interval 1975-2009. Analyses were performed using estimates of natural mortality at age obtained by means of two different methods and two population dynamics methods based on the analysis of catch-at-age data: Laurec-Sheperd virtual population analysis (VPA) and integrated catch-at-age (ICA), both tuned to acoustic estimates of abundance. Gislason’s estimates for natural mortality appeared to be more realistic and were thus preferred for short-lived species. The general trend of biomass and fishing mortality is similar for the two models, highlighting the major collapse of the stock in 1987. Nevertheless, ICA has enough flexibility to combine all the data available without adding too much complexity in comparison with a VPA approach and seems to perform better in terms of the spawning stock biomass/recruitment relationship and diagnostics (i.e. the retrospective pattern). For the stock status, the exploitation rate from ICA is higher than the suggested threshold of 0.4 proposed by Patterson for small pelagic species.

In situ target strength measurements of the scyphomedusa Chrysaora melanaster

Large medusae can dominate the pelagic biomass in marine environments, but their contribution to acoustic scattering remains poorly characterized. Here we report on in situ target strength (TS) measurements of Chrysaora melanaster, a dominant species of scyphomedusa in the North Pacific and Arctic, made with 38 and 120kHz split-beam echosounders in the Chukchi Sea. The average TS of a ∼25cm bell diameter C. melanaster is estimated to be −60.1dB re 1m2 at 38kHz and −60.5dB re 1m2 at 120kHz. There was no evident relationship between mean jellyfish size and TS, which is likely attributable to a combination of a small range in average jellyfish size at the sampling sites and geographic variability in TS corresponding to factors other than size and/or errors in determining the size of the jellyfish. The TS of C. melanaster is similar to that of a small swimbladdered fish, but on a mass-specific basis, the TS of these jellyfish is substantially lower. These measurements will aid in acoustic estimates of jellyfish abundance and in partitioning the relative contribution of jellyfish to acoustic backscatter in mixed species assemblages.

Measurement of single-fish target strength in the South China Sea

We measured the target strength (TS) of three commercial fish species: whitespotted spinefoot (Siganus canaliculatus), black porgy (Acanthopagrus schlegelii), and creek red bream (Lutjanus argentimaculatus), in the South China Sea. The TS of caged or tethered fish (n=76 total) was measured using a Simrad EY60 portable scientific echosounder at 120 kHz. We evaluated the relationship between TS and total length (TL, cm) for the three species. This is the first attempt to use split-beam acoustics to measure single-fish TS in the South China Sea by Chinese researchers. Our results will improve the accuracy and precision of acoustic abundance estimates of commercially important species and further the development of underwater acoustic survey techniques in fisheries in the South China Sea.

Efficient trawl avoidance by mesopelagic fishes causes large underestimation of their biomass

Mesopelagic fishes occur in all the world's oceans, but their abundance and consequently their ecological significance remains uncertain. The current global estimate based on net sampling prior to 1980 suggests a global abundance of one gigatonne (10 9 t) wet weight. Here we report novel evidence of efficient avoidance of such sampling by the most common myctophid fish in the Northern Atlantic, i.e. Benthosema glaciale. We reason that similar avoid- ance of nets may explain consistently higher acoustic abundance estimates of mesopelagic fish from dif - ferent parts of the world's oceans. It appears that meso pelagic fish abundance may be underestimated by one order of magnitude, suggesting that the role of mesopelagic fish in the oceans might need to be revised.

A revised target strength–length estimate for blue whiting (Micromesistius poutassou): implications for biomass estimates

Abstract Pedersen, G., Godø, O. R., Ona, E., and Macaulay, G. J. 2011. A revised target strength–length estimate for blue whiting (Micromesistius poutassou): implications for biomass estimates. – ICES Journal of Marine Science, 68: 2222–2228. Acoustic abundance estimates of blue whiting have generally been higher than estimates based on catch data. One explanation has been that the relationship between acoustic target strength (TS) and length is too low and hence overestimates the number of fish. Measurements of TS were conducted during surveys of blue whiting in March/April 2003–2007 to the west of the British Isles from several different measurement platforms, and also during August 2005 in the Norwegian Sea. Results from these experiments confirm the view that the existing TS–length relationship is too low. A new TS–length relationship is proposed that is ∼5 dB higher. Blue whiting TS is considerably higher than observed and modelled for a similar species, southern blue whiting (Micromesistius australis).

Toward Estimating the Variance in Acoustic Surveys Based on Sampling Design

This paper develops a sampling method to estimate the integral of a function of the area with a strategy to cover the area with parallel lines of observation. This sampling strategy is special in that lines very close to each other are selected much more seldom than under a uniformly random design for the positions of the parallel lines. It is also special in that the positions of some of the lines are deterministic. Two different variance estimators are derived and investigated by sampling different man made signal functions. They show different properties in that the estimator that estimate the biggest variance gives an error interval that, in some situations, may be more than ten times the error interval computed from the other estimator. It is also obvious that the second estimator underestimates the variance. The author has not succeeded to derive an expression for the expectation of this estimator. This work is motivated towards finding the variance of acoustic abundance estimates.

Diel interactions between sprat and mackerel in a marine lough and their effects upon acoustic measurements of fish abundance

We have investigated how diel changes in fish behaviour may affect acoustic estimates of fish abundance in a marine system. Studies in Lough Hyne (Ireland), a marine lake with only a narrow and shallow connection to the sea, showed the presence of large numbers of sprat ( Sprattus sprattus), mackerel ( Scomber scombrus), and zooplankton. Fish species were identified from regular rod and line fishing. Zooplankton were identified from net hauls. Acoustic measures of zooplankton were low with little diel variation. The sprat formed dense schools during the day and dispersed into the water column at night. The acoustic estimate of sprat abundance at night, obtained by means of a standard hydroacoustic technique and protocol, was more than double the estimated biomass during the day. We have considered whether the lower fish estimate during daytime resulted from acoustic shadowing due to aggregation of the fish into dense schools. However, no decrease in echo energy was evident from the top to the bottom of the schools, and there was no reduction in the seabed echo beneath the schools. Acoustic shadowing was therefore not responsible for the diel differences in the estimates of abundance. Instead, we suggest that the target strength of individual sprat diminished during the day as a result of attacks from predatory mackerel. We observed echoes from gas released by the sprat as they gathered into dense plumes close to the seabed. Compression of the gas remaining within the swim bladder as the fish were moving deeper would also reduce swim bladder volume. Finally, negative buoyancy due to reduced swim bladder volume may in addition have forced the fish to change tilt angle to compensate for sinking. All these effects will reduce the target strengths of individual fish and lead to underestimation of fish abundance based on daytime surveys.