Abstract
Motivation
Scientific echosounders are the standard tool in fisheries science for investigating the abundance, distribution, behavior, and ecology of fish, and have been used for monitoring around tidal energy devices. Echosounders have been deployed on the sea floor in a stationary upward-facing orientation for monitoring around gravity-based tidal energy devices but have also been deployed at the sea surface in a downward-facing orientation, either for mobile surveys or on ships at anchor. The advent of floating tidal energy platforms provides an opportunity to deploy echosounders at the sea surface in a long-term, stationary, downwardfacing orientation for monitoring. However, the strong currents that make tidal channels attractive for energy production are often dominated by turbulent hydrodynamic features and associated artefacts that can vary over the course of tidal cycles and may hinder the use of echosounders and other active acoustic technologies. Understanding the extent to which turbulent hydrodynamic features impact the use of a bottom-mounted or surface-deployed echosounder is important for designing effective monitoring systems.
In partnership with the Pathway Program, Sustainable Marine Energy (Canada) Ltd. and the Fundy Ocean Research Center for Energy, undertook a series of studies to understand whether deployment location impacted the efficacy of echosounder technology for monitoring by assessing the relative performance of surface-deployed instruments and a bottom-mounted echosounder. The bottom-mounted echosounder was the Simrad Wideband Autonomous Transceiver (WBAT: from the Simrad EK80 suite of echosounders) mounted on the Fundy Advanced Sensory Technology (FAST) autonomous underwater platform and deployed in Grand Passage, NS in the vicinity of the Sustainable Marine floating tidal energy platform (i.e., PLAT-I). The surface-deployed instruments were deployed via pole mount attached to the leading edge of the starboard pontoon of the PLAT-I platform and included a Sculpin HDC-SubC optical video camera, a Gemini 720is multibeam imaging sonar, and a downward-facing Simrad Wideband Transceiver (WBT: from the Simrad EK80 suite of echosounders).
The primary goal was to collect data to compare target detections for identifying the best placement of echosounders for monitoring in the vicinity of the PLAT-I deployed in a high flow environment. Thus, the objectives of the three studies were to i) investigate the near-surface target detection capabilities of a bottom-mounted, upward-facing, echosounder using a surfacedeployed optical video camera (Study 2A), ii) address this same objective over a greater detection range using a multibeam imaging sonar (Study 2B), and iii) investigate the relative performance of a bottom-mounted upward-facing, and a surface-deployed downward-facing echosounder for target detection (Study 2C). This final goal was addressed by identifying the extent of target detection interference due to air entrained in the water by surface waves and turbulence. The PLAT-I rotors were parked during data collection for all three studies. Because of safety concerns for personnel, the PLAT-I and instruments, the FAST platform was deployed at locations in the vicinity of the PLAT-I, but at distances that precluded the possibility of 3 overlapping sampling volumes between the instruments mounted on the PLAT-I and the FAST platform.
Summary of Findings
No fish images were captured in the 170 hours of optical video data examined for Study 2A and Study 2B. This result was unexpected given that during Study 2A, within the water depths interrogated by the optical camera, the upward-facing echosounder recorded signals consistent with the presence of fish in 55% of the 8.3 hours of echosounder data not obfuscated by entrained air. The images captured by the video camera were of sufficiently high-resolution that the absence of fish likely reflected a lack of fish passing within the camera’s field-of-view. Although the optical video data could not be used to cross-reference targets detected by the echosounder, the echosounder data collected during Study 2A contributed to our understanding of the importance of localized hydrodynamic regimes on the ability to collect useable data.
For Study 2B, the image resolution of the imaging sonar was insufficient to identify targets beyond two ambiguous categories: “single fish/debris” and “turbulence/fish/school of fish”. A third category denoted instances when the PLAT-I mooring chain was within the imaging sonar field-of-view. During this study, in nominally 100% of the echosounder observation time periods, signals that could be interpreted as fish were detected at depths that coincided with the depth range interrogated by the imaging sonar (i.e., the top 11 m of the water column). However, the imaging sonar identified potential detections for only 22% of the observation periods. The source of this discrepancy likely stems from non-overlapping sample volumes due to FAST platform deployment location for this study. Although the lack of overlapping sample volumes precluded definitive cross-referencing between the two instruments, optical cameras and imaging sonars have been shown to be valuable monitoring tools elsewhere and have value for monitoring in Grand Passage. As with Study 2A, the echosounder data collected during Study 2B contributed to our understanding of the importance of localized hydrodynamics for the collection of useable data.
Analyses of signal interference due to entrained air (Study 2C) suggested a strong difference in the hydrodynamic regimes at the deployment locations of the PLAT-I and the FAST platform with consequences for the proportion of useable data at each site. Signal interference manifested in several ways: i) the presence (PLAT-I site) or absence (FAST site) of a pronounced tide-phase asymmetry in the proportion of data excluded from analyses (due to the persistence and depth penetration of entrained air), ii) the presence (PLAT-I site: flood) or absence (PLAT-I: ebb, FAST site: flood and ebb) of a pronounced negative relationship between flow-speed and the proportion of useable data, and iii) the presence (FAST site) or absence (PLAT-I site) of a reduction in the proportion of useable observation periods with increasingly restrictive minimum acceptable proportions of useable water column.
The pronounced tide-phase asymmetry at the PLAT-I site appears to be a consequence of its deployment location downstream from Peter’s Island on the flood tide (upstream on the ebb tide). Consequently, the PLAT-I was deployed within the turbulent field generated by the interaction of the flood-tide with Peter’s Island and its associated bathymetry. The turbulence 4 and associated entrained air had significant consequences for the collection of useable data on the flood tide, reducing the useable proportion of 10-minute observation periods to ≤ 30% on the flood tide; the proportion of useable observation periods on the ebb tide ranged from 85- 95%. Going forward, this pronounced asymmetry indicates that useable data collected on the flood tide will be minimal at the current PLAT-I deployment location and will have important consequences for understanding the risk to fish during the flood tide phase.
Although comparative echosounder data were not available for Study 2A or Study 2B, analyses of the proportion of “useable” data collected at the FAST deployment sites for all three studies support the hypothesis that bathymetry associated with Peter’s Island creates a pronounced difference in the hydrodynamic regimes associated with the flood and ebb tides at the PLAT-I site and nearby locales (Study 2A and 2B). During Study 2C, the FAST platform was deployed just outside the direct downstream flow associated with Peter’s Island and the echosounder data showed little tide-phase asymmetry but did reveal a deterioration in the proportion of useable observation periods on both the flood and ebb tide when increasing the minimum proportion of the water column deemed as “useable”. While the lack of a pronounced asymmetry for the FAST site in Study 2C indicated that more data was useable on the flood tide relative to the echosounder placed at the PLAT-I, it should be noted that on the ebb tide, the PLAT-I site had more “useable” data as highlighted above.
Analysis of the useable proportion of individual observations (pings) within 1-m depth bins in the depths-of-interest for the PLAT-I (i.e., 1-8 m depth) revealed the same pattern found in the data analyzed across the entire water column for the presence (FAST site: Study 2A and 2B; PLAT-I site: Study 2C) or absence (FAST site: Study 2C) of tide-phase asymmetry.
Given that the echosounders used here were from the same Simrad EK80 suite of echosounders and deployed with identical data collection parameters, the ability to detect or define the boundary of entrained air was not affected by which echosounder was used. Nor was it affected by deployment at the sea surface in a downward-facing orientation, or on the sea floor in an upward-facing orientation. The sea surface vs. sea floor positioning of the echosounders does, however, have important data collection and analytical consequences that must be considered.
The ensonification beam emitted by a transducer is cone-shaped with the apex at the transducer and the diameter of the cone increasing with distance. It follows then that the region closest to the transducer will result in a highly restricted sampling volume and leave substantial proportions of water unsampled. Therefore, to maximize the volume of sampled water, the transducer should be placed furthest from the region of interest.
As with the considerations of the hydrodynamic regime when selecting a deployment location, one must also consider the consequences in the vertical dimension. If the entrained air so common to tidal energy sites is between the transducer and the target of interest, the acoustic beam will be scattered before encountering the target and will compromise the collection of quantitative data for estimating density and abundance.
Major Take-Aways
These studies demonstrate the importance of the influence of hydrodynamics on the ability to collect useable quantitative data required for analyses and reporting on fish abundance and distribution. Deployment location of an echosounder can have a profound impact on the ability to monitor throughout the tidal cycle, such that the hydrodynamic regime will influence when and where you can observe.
To obtain quantifiable and comparable data on targets of interest, the echosounder must be deployed such that the acoustic beam encounters the targets before encountering entrained air.
Additionally, to maximize the sampling volume, and thereby the likelihood of a fish encountering the acoustic beam, the echosounder should be deployed furthest from the region of interest.
Topics for Continued Research
Because of the fast-flowing and turbulent waters found in tidal streams, the in situ sampling (e.g., via net tows/trawl surveys) of acoustic targets (fish) commonly done in open waters is not possible in tidal streams. Therefore, ground-truthing the identity and size of acoustic targets observed in tidal streams remains an unresolved and ongoing topic of research in the hydroacoustics community.
The presence of entrained air makes monitoring fish in tidal channels particularly problematic. Currently, there is no proven strategy to observe fish with sufficient field-of-view or resolution within a highly turbulent environment. Without the ability to observe fish throughout the entire tidal cycle, quantifying risks of turbines to fish will remain difficult. Therefore, to understand potential risks, it is important to continue to focus on developing methods to detect and observe fish, or establish means by which to infer fish presence and behavior, within turbulent tidal channels.