Examples of deployments for active acoustic studies (not to scale): (1) a multibeam echosounder on an ROV; (2) a single-beam echosounder on a mooring; (3) a single-beam echosounder on a USV; (4) an ROV with a single-beam and multibeam echosounder; (5) two echosounders, up- and down-facing, on a glider; (6) a research vessel with single- and multibeam echosounders; (7) a stationary bottom platform with a single-beam echosounder.

 

By Dr. Haley Viehman

Aquatic research and monitoring questions often revolve around what’s happening in the water column, such as fish quantity and movements, zooplankton distribution, or vegetation canopy height. Active acoustic instruments, a.k.a. echosounders, are well-suited to addressing these questions. Active acoustics has a long history in fisheries stock assessment, and more recently its applications have expanded into other sectors, including aquaculture, ocean renewable energy, hydropower and more.

Echosounders emit pulses of sound of a specific frequency band into the water, and then listen for the sound to return as it gets scattered by objects in the water (“backscatter”). This is similar to echolocation by dolphins and whales. One iteration of sending out a sound pulse and listening for its return is called a “ping.” How long it takes for the transmitted sound to return indicates how far away an object is, and the properties of the returned sound (such as pulse amplitude and shape) can reveal properties of the object itself. This sending and receiving of sound differentiates active acoustics from passive acoustics, which involves only receiving sound by use of a hydrophone.

Sound travels rapidly in water (1,450 to 1,500 m/s), so echosounders can sample a large volume of water at once with high spatiotemporal resolution, and they can be deployed for long periods of time. Compared to cameras, echosounders require no light, are relatively unaffected by turbidity, and can sample much larger volumes of water. More invasive physical sampling methods, such as trawls, hook and line, or baited traps, cannot compare to the spatiotemporal coverage and resolution of active acoustic systems.

These versatile instruments can be deployed on virtually any platform, moving or stationary. Echosounders have traditionally been deployed in the hulls of vessels, but are now used on USVs, AUVs, gliders, moorings, pilings, dams, and sometimes even airborne drones to fit any survey design and monitoring or research question.

 

Stationary seafloor platforms can be deployed with upward-facing echosounders and other instruments for long periods of time, such as this one for the Fundy Ocean Research Center for Energy. (Credit: FORCE)

 

Common applications of active acoustics include fisheries stock assessment, for example, in large-scale surveys to enumerate freshwater, estuarine, or marine fish stocks; on stationary platforms in rivers to quantify migrations of diadromous species (e.g., salmon); and on commercial fishing vessels to improve species detection and reduce bycatch. In aquaculture, active acoustics is used in and around fish pens and macroalgae farms to gather information on organism size, health, and responses to environmental factors; detect harmful jellyfish; inspect nets for damage and biofouling; and monitor the effects of aquaculture on the surrounding environment. In ocean renewable energy, active acoustics help to monitor fish and other aquatic organisms at offshore wind, wave, and tidal power sites, where traditional survey methods (e.g., trawls) may not be safely deployed or provide the necessary resolution and coverage. For hydropower applications, echosounders monitor fish abundance and distribution in rivers in relation to dams, e.g., pre- and post-dam modification or removal, and to assess the effectiveness of fishways under light and dark conditions. Power plant intakes can benefit from active acoustics to detect approaching jellyfish swarms. At marine protected areas, echosounders are used to map and monitor fish densities in and around the sites and assess their effects. In habitat mapping, active acoustics can detect underwater vegetation, map canopy height above the bottom, and detect anthropogenic effects. Underwater gas seeps can also be located and mapped via active acoustics.

Types of Echosounders

Echosounders come in many different forms, each with different strengths and applications. All include a transducer and transceiver, which can be separate units or combined. To generate a ping, the transceiver sends an electrical signal to the transducer, which translates that voltage to an acoustic pulse emitted into the water. The transducer receives any backscattered acoustic energy from the emitted pulse and converts it back to voltage, which is digitized by the transceiver. A computer controls the echosounder in real time or pre-configures its deployment plan (if autonomous), records data files, and processes the data. Echosounders can be grouped into two broad categories: single-beam and multibeam. Choosing the correct hardware depends on the questions being asked and the capabilities of each system.

 

Data from two types of echosounders. Left: an echogram from a single-beam echosounder, showing various scatterers from the surface to the seafloor over time. Right: one ping from an imaging sonar, showing a school of fish and their “shadows” near the seafloor.

 

Single-Beam Echosounders. These emit sound in one “beam” extending outward from the transducer face. Recreational fish finders are in this category, though the echosounders commonly used for monitoring or research (often called “scientific echosounders”) are usually highly engineered systems that, when calibrated, produce measurements that can be compared across surveys and sites. The sound spreads as it travels, resulting in a roughly conical acoustic beam that can extend for hundreds of meters depending on the environment and acoustic frequency. Data from single-beam echosounders are visualized in an “echogram,” which displays backscatter in two dimensions (depth or range from transducer versus time), with backscatter strength mapped to a color scale.

Single-beam echosounders provide information on the distance (range) to objects in the water column, and when calibrated, backscatter values can be transformed into biomass estimates. Some single-beam echosounders, known as “split-beam” echosounders, comprise multiple sectors forming the beam. Split-beam echosounders can locate individual objects within the beam in 3D, which allows more accurate size estimates, as well as tracking of individual objects as they pass through the beam.

Like all sampling tools, single-beam echosounders have limitations. One is species identification. Although the characteristics of backscattered sound can help differentiate object types, identifying an organism in an echogram to species level typically requires additional supporting information. This may come from direct or indirect species and size observations obtained from physical samples, visual techniques (e.g., stereo cameras), or genetic methods (e.g., environmental DNA). Single-beam echosounders are additionally not well suited for sampling water directly adjacent to boundaries, such as the water’s surface, the seafloor or man-made structures, because biological backscatter may be masked by the strong signal from these objects. Though suitable for sampling pelagic organisms, other tools may be needed for observing demersal species, such as imaging sonars. Recent advances in wide-band data collection and processing can improve resolution of single-beam echosounders near boundaries and may provide more insights for species identification.

 

Single-beam echosounder data processing can include detecting and omitting unwanted backscatter (e.g., from the bottom and air entrained at the surface), smoothing the data, detecting fish schools, and removing all non-school backscatter so only the schooling species can be quantified. The images here are examples of these steps carried out in Echoview.

 

Multibeam Echosounders. These send and receive sound in multiple beams angled outward from the echosounder to sample a fan-shaped “swath” of water. These include echosounders typically used for bathymetric surveys, which sample a huge volume of water and, hence, can also be useful for assessing water column features (e.g., biological aggregations, seaweeds or bubble plumes from gas seeps). Omnidirectional sonars sample a long-ranging, 360° swath of water by sending and receiving sound in a “donut” around themselves.

In contrast, imaging sonars, or “acoustic cameras,” are multibeam systems that use very high-frequency sound (usually in megahertz) and a fast ping rate to produce video-like data of a shorter-range swath, similar to medical ultrasounds. These systems produce highly intuitive data with fine detail, making them excellent tools for observing individual organism size and behavior, inspecting habitat or structures, and identifying species.

Active Acoustic Data Analysis

Choosing the right hardware and deployment for a given research or monitoring task is essential, but it is equally important to consider how the resulting data will be converted into usable information. Active acoustic data sets are incredibly variable, but processing can be broken down into six general steps.

The first is data exploration. This includes loading the data; organizing the various data streams (such as data from multiple acoustic frequencies, collected concurrently); and visualizing and inspecting the data for potential issues or interesting features.

The second is data calibration, which ensures the active acoustic measurements are quantitatively accurate and mapped correctly in space and time. This includes applying correct environmental parameters (for example, sound speed) and calculating any necessary offsets to the backscatter values. Calibration offsets are typically derived by analyzing backscatter data from a known target (usually a tungsten-carbide or copper sphere of specific dimensions), and the calibration data are ideally collected in situ before and after survey data collection.

The third step is data cleaning, including removing unwanted components of the data set, such as backscatter from non-targeted objects (e.g., the seafloor or entrained air near the surface). Other unwanted components include acoustic or electrical noise; for example, due to other acoustic instruments operating in the vicinity or to an unclean power source. Once identified, unwanted or contaminated data points can be removed entirely or replaced with representative values (e.g., a local average). The presence of noise degrades a data set and any subsequent results, regardless of the cleaning methods used. Sources of electrical and acoustic noise should therefore be identified and rectified before collecting survey data.

The fourth step is target detection and tracking: identifying and delineating the scatterers of interest in the data set. This may include detecting backscatter likely to come from individual fish (when fish are separated enough) and tracking those fish across pings, or delineating schools of fish when they are packed closely together.

 

Imaging sonar data processing often includes removing the background, smoothing the data and detecting fish, as shown here. Echoview was used to detect and track fish over time, providing fish size and movement metrics for analysis.

 

The fifth step is target classification: the identification of scatterers that have been detected in the data. This is one of the more challenging aspects of active acoustics. Classification might include categorizing backscatter according to taxonomic group, body size, bottom type, etc. Approaches to target classification will vary with survey goals and instrumentation. For example, if data are collected concurrently at multiple frequencies, groups of scatterers may be classified by comparing backscatter from one frequency to the next. The size, shape, behavior, and/or location of detected schools or individual targets may additionally be useful for classifying certain species of fish or other organisms.

The sixth step is target characterization: calculating metrics for the targets of interest in calibrated, cleaned and classified data. This can be, for example, average backscatter (converted to biomass) from each grid cell of the echogram, or swimming behavior metrics from individually detected and tracked fish (e.g., swimming speed, direction and tortuosity).

Processing Software

Data processing requirements should be considered and procured alongside hardware, including storage, computing power, and analysis software. Active acoustic data sets can be quite large, and more data or more complex processing steps require a more powerful computer. Most echosounders are supplied with software for data replay that may also provide basic processing capabilities, and this can be sufficient for some applications. However, specialized software is often required when processing goals require more comprehensive features, or flexibility to cover a broader range of technologies, deployment types, and data quality.

Echoview

Echoview is a global industry standard for water column echosounder data processing, enabling easy visualization and interaction with more than 75 different active acoustic file formats from 18 different echosounder manufacturers. The software includes a large selection of built-in data processing and automation tools to produce quantitative results from echosounder data.

Echoview has been employed across the full spectrum of active acoustic applications. For example, Echoview was used to automate the processing of several years of active acoustic data collected at the Fundy Ocean Research Centre for Energy (FORCE) in Nova Scotia, Canada, for multiple research objectives. At such sites, fast currents pull air bubbles deep into the water column, requiring their backscatter to be identified and removed to study the backscatter from fish accurately and understand their interactions with tidal energy devices. Echoview’s power and flexibility was invaluable for handling data collected in this challenging environment.

Learning Active Acoustics

Effectively using active acoustics for research or monitoring requires an understanding of the physics of underwater sound, the biology of the organisms of interest, and instrument operation. For example, the best echosounder type and frequencies to use will depend on the sound scattering properties of the sampled organisms and the environment they inhabit. Data collection settings, such as transducer orientation, pulse duration and ping rate, must be chosen carefully because they affect what can be learned from the resulting data set.

Once data are collected, software solutions such as Echoview can make working with the files relatively easy, but the steps needed for data cleaning, identifying scatterers of interest, and interpreting their backscatter must be guided by prior knowledge on how sound interacts with the species present. Active acoustics can present a steep learning curve as introductory literature in this space is generally sparse (though improving).

At Echoview Software, we deliver training courses in active acoustics theory and data processing, provide consulting and client services, and offer advice on any questions you might have about active acoustics and its applications. You can learn more by reaching out to us at: info@echoview.com.

Dr. Haley Viehman is a hydroacoustics specialist at Echoview Software in Tasmania, Australia.