Background
Nortek Netherlands specializes in Vessel-Mounted Acoustic Doppler Current Profilers (VM-ADCPs). While vessel-mounted systems provide acoustic access to remote locations, they also introduce challenges. The speed of the vessel must be removed from measured currents by a source resilient to dropouts and free of motion. Further, many Nortek users suspend their VM-ADCP from a side-mounted pole, making it vulnerable to mechanical vibrations, flow turbulence, and bubble layers. In addition, the instrument may be exposed to a host of onboard noise and interference sources. Artifacts encountered in low frequency VM-ADCP output motivated the deployment of a Nortek VM Ocean 100 kHz system, which has four beams for currents and bottom track and a fifth beam echosounder. Because suitable depths and currents are not accessible from the Netherlands, we planned our test in the Strait of Gibraltar (SoG) region.
Flow in the SoG consists of an eastward, buoyant, approximately 125 meter (m) thick Atlantic Ocean surface layer and an underlying westward, saline-rich, Mediterranean Sea layer (Sverdrup et al., 1942). While Atlantic Water is nutrient-rich, which is ideal for acoustic scattering, the Mediterranean Sea is clear due to its oligotrophic properties. Evaporation in the Mediterranean Sea exceeds precipitation and river inflow (Garrett et al. 1993 and Gilman and Garrett, 1994, both cited in Gomis et al., 2006), creating a water deficit that drives the two-layer, estuarine circulation. Currents in the SoG are rapid, bathymetry is complex, and mixing is rigorous (Sverdrup et al., 1942). Barotropic forcing mechanisms include the Levantar winds (easterly), Poniente winds (westerly), atmospheric pressure gradients, and tidal forcing. Westward tidal currents through the SoG reduce Atlantic inflow, while eastward tidal currents enhance it (Sverdrup et al., 1942). The build up and release of tidal energy at Camarinal Sill into an eastward train of internal solitary waves (solitons) is well studied (Bolado-Penagos et al., 2023). Solitons have been observed in the Alboran Sea up to 200 km eastward of Camarinal Sill (Pistek and La Violette, 1999; Apel and Worcester, 2000, both cited by Bolado-Penagos et al., 2023) and up to 50 hours later (Bolado-Penagos et al., 2023). Solitons are also reflected westward from Camarinal Sill (Global Ocean Associates, 2004), and northwestward internal solitary waves likely reflected off of the Moroccan slope are identified with in situ and satellite imagery (Roustan et al., 2024). Heightened vertical mixing from breaking of internal waves at Camarinal Sill results in upwelling of nutrients in the Atlantic layer (Bolado-Penagos et al. 2023 and references therein). In summary, a scatter-rich surface layer, tidally driven current speeds, depths exceeding 800 m, and water column structure constitute an ideal sampling location.
Methods
Nortek used a Nortek VM Ocean 100 kHz system with accompanying Nortek VM software. Details on the setup, motion correction, signal quality and instrument performance are laid out in the sections below.
Setup
A side-mounted pole deployment of a Nortek VM Ocean 100 kHz was planned with the goals of sampling currents, monitoring instrument performance, validating motion correction algorithms, and testing a flow noise assembly designed to allow streamlines to flow over the transducers without being disrupted by its sharp edges. On 8 Sept – 12 Sept. 2025, the VM-ADCP was deployed with an Advanced Navigation GNSS. The GNSS has dual heading and NTRIP corrections from a nearby base station. The GNSS’s internal AHRS outputs angular velocity, which is used in motion correction solutions. A horizontal position accuracy of 0.01 m and heading accuracy of 0.2 ° are listed in technical specifications. Time stamps were synced to the μs level, and a sample rate of 10 Hz was used.
Refer to Figure 1. The AS ONE is a 22.85 m long vessel fabricated from naval steel with two propulsion engines (339 kW) located on either side of the vessel about 7 meters (m) from the rear. Two ANSI316 mounting brackets were welded to the starboard (right) side of the vessel about 8 to 10 m from propulsion engines and away from freshwater tanks. The mounting consists of a 5.67 m-long steel pole with flanges to accommodate instrument positions. Mounting depths of 2.10 and 2.85 meters were possible. The pole was positioned into the brackets with a crane. Once in place, two ropes were extended from the instrument diagonally upward and secured to the side of the vessel. The setup was observed to remain quite stable during sailing at the maximum speed of 9 knots.
The AS ONE was docked in the Port of Algeciras, allowing access to both the Bay of Algeciras (Bay) and the Mediterranean Sea. The Bay provided a relatively sheltered environment and depths up to about 250 m. The waters beyond provided strong currents and depths greater than 850 m. Two-way navigation, no anchor rules in SoG, and a decision to avoid Moroccan territorial waters shaped sailing routes. Track timing incorporated soliton travel times from Camarinal Sill presented in Bolado-Penagos et al., 2023 in conjunction with spring high tide to increase the chance of intercepting internal waves. With this we posed the question:
Can a VM Ocean (100 kHz) be used to capture oceanographic features in the SoG region?
Motion Correction
Measurements from an ADCP deployed from a moving vessel contain both currents and vessel speed. The vessel speed must be subtracted to obtain the current. This motion correction step uses speed over ground from the ADCP’s bottom track. The bottom track ping returns a strong signal from the sea bed which contains only the vessel speed since the current here is zero. Bottom track data gaps are filled with speed over ground from the GNSS. The GNSS is typically mounted to a pole and subject to swaying from heading, pitch, and roll. These oscillations enter the corrected current and manifest as vertical ripples (Figure 2).
The VM-ADCP and the GNSS are not co-located, and GNSS antenna swing needs to be accounted for when correcting for vessel motion using GNSS velocities. These adjustments require knowledge of the three dimensional distance vector between the instruments, the lever arm. This unknown vector has ship coordinates of XYZ, where x is the long axis of the vessel, y is the transverse axis, and z is up and down. Nortek Netherlands has implemented an algorithm to correct for antenna swing using angular velocity output from the GNSS’s internal AHRS. Correcting the GNSS speed over ground this way mitigate ripples in GNSS-corrected currents. A time series plot of bottom track-corrected currents (BT), GNSS corrected-currents (GNSS), and lever arm-corrected currents (LA) illustrates the impact of the new algorithm.
Signal quality
Many steps are taken to limit the VM-ADCP’s exposure to noise and interference including short cables, instrument grounding, elimination of generators, and direct battery connections. Still, distortions are sometimes seen in low frequency VM-ADCP currents, beam amplitude, beam correlation, and, in some cases, echograms. These may appear as continuous stochastic stripes (variable width and spacing) of heightened amplitude, as shown in Figure 3. This represents an extreme case of acoustic noise from a Nortek VM Ocean (55 kHz) suspended from the side of a vessel in Oslo Fjord over a 200 m deep bed.
In addition, electromagnetic interference (EMI) may manifest as periodic geometric patterns including V’s, diagonal stripes, or pin stripes.
Sources of acoustic noise encountered in VM-ADCP measurements include propeller cavitation and bubble propagation and associated broadband acoustic noise; signal reflections off of the hull, brackets, or structures; bubble layer under the hull; and other acoustic instrumentation including depth sounders. Common sources of EMI include engine alternators, generators, cables, ground loops. The propulsion system and associated anthropogenic sound are well described by Tanttari and Hynninen, 2022. Acoustic, hydrodynamic, and machinery noise are interconnected by structural vibrations of the hull; the cause may be impossible to distinguish (Tanttari and Hynninen, 2022).
While Nortek’s clean installation measures provide protection against many sources, the VM-ADCP remains vulnerable to acoustic noise from the propeller, cavitation from various sources, EMI from the engine, flow across the transducers, and turbulent flow over sharp transducer edges. To test the plausibility of the last source, a plastic flow noise assembly was generated with a 3D printer and mounted to the VM-ADCP
(Figure 4).
While the intent was to compare raw bottom track amplitude from tracks sailed back-to-back over the same location with and without the flow noise assembly, obtaining such track pairs proved infeasible. Instead, three tracks were sailed without the flow noise assembly and the remaining tracks were sailed with it. Analysis of time series plots of speed over ground, noise floor (NF) estimates, and beam amplitude revealed major contributing factors to vertical distortions, regardless of the flow noise assembly. A time series plot for one track is presented. Bottom track and profiling range of the three tracks sailed without the flow noise assembly is addressed under Results, Instrument Performance.
Instrument Performance
Eighteen tracks were sailed over five days, six of which include depths over 300 m. Valid bottom track (VBT), the percentage of pings with bottom lock for a given track, was computed for all tracks. The profiling range of the six deep tracks was quantified using the amplitude threshold of the estimated NF plus 2 dB. The estimated NF is the minimum amplitude of the last 30 cells for each beam for a given track. These were averaged for an estimated NF for each track. Water column structure yielded micro pockets of low correlation that were not removed, however, a definitive margin between high (> 50%) and low (< 50%) correlation tended to coincide with or lie just beneath the amplitude threshold. This delineation likely represents the boundary between an upper layer of scatterer-rich Atlantic Water combined with suspended particulate matter and a lower layer of Mediterranean Outflow Water. The correlation threshold of 50% was fully implemented to omit regions impacted by electromagnetic interference (EMI) where applicable, as was the case with the Sea Track described under Results. The echosounder NF was estimated with the uncorrected return signal (Rx) intensity of the last 30 cells and range was quantified with that threshold plus 2 dB.
Results
Two tracks are presented (Figure 5), both sailed with the flow assembly installed. The impacts of the motion correction algorithm are analyzed with the Bay Track (104075_20250909T131852UTC), and signal quality is analyzed with the Sea Track (104075_20250911T082712UTC).
Motion Correction
The 7 km-long Bay Track begins with a segment in the Sea and spans depths of 100 to 400 m. Pitch and roll span as much as 7 and 10 degrees, respectively. Maximum angular velocity rates are ωx = 14.5 deg./s, ωy = 4.9 deg./s, ωz = 2.7 deg./s, where x, y, and z are ship coordinates. This track has a sample rate of 2 sec and valid bottom track of 96.7%, achieved as deep as 500 m.
The top panels of Figure 6 display current magnitude and pitch, roll, and heading for the Bay Track. Bottom track corrected-current includes unaltered GNSS-corrections where the bottom track dropped out. This represents the currents output before the lever arm algorithm was developed. The GNSS-corrections are performed with unaltered GNSS speed over ground. The LA-correction applies the LA-corrected GNSS speed over ground which minimizes the vertical rippling, leaving a smooth appearance.
Signal quality
The 30 km-long Sea Track is located at the Mediterranean exit of the SoG. Depths range from 400 to 800+ m deep. Pitch and roll span up to 5 and 12 deg., respectively. Maximum angular velocity rates of ωx =7.7 deg./s, ωy = 7.9 deg./s, ωz = 2.8 deg./s.
Figure 7 shows several parameters for the Sea Track. Speed over ground spans 0.2 m/s to 3.6 m/s and consists of relatively steady segments, increasing segments, and multiple abrupt increases. The NF spans about 30 to 50 dB for all beams and is lower for fore beams (1 and 4) than for aft beams (2 and 3). Beam amplitude and the 120 kHz echogram display distortions consistent with EMI: thin repeated pin stripes and diagonal stripes. Acoustic noise may also be present in this track.
The red solid rectangle highlights a period of heightened, increasing speed over ground (up to and greater than 2.5 m/s, marked with dashed line), a raised NF, and distortions consistent with EMI, especially in aft beams. As speed increases, distortion levels and NFs increase.
The black dashed rectangle highlights stretches of moderate and steady speed over ground less than 2.5 m/s that coincide with a low NF in all four beams and distortion-free amplitude.
Black arrows highlight an example of an abrupt increase in speed over ground that is likely an engine rev. This coincides with amplitude distortions and jumps in NF to 34 dB for Beam 1 (fore), 45 dB for Beam 2 (aft), 43 dB for Beam 3 (aft), and 37 dB for Beam 4 (fore).
Similar observations were confirmed in other tracks, including in the echograms of those with the echosounder enabled. Distortions consistent with EMI and acoustic noise coincide with elevated noise floors, while distortion-free data only occurs with a low NF in all beams. Lower sail speeds (< 2.5 m/s) coincide with low NF in all beams but does not guarantee it. In addition, engine specific tests were made (recall that the VM-ADCP was on the starboard side of the vessel): when the starboard engine was off and the port engine was revved, low NF in all beams and distortion-free output prevailed. In a test where the starboard engine was revved and the port engine was off, the NF was raised.
Two of the three tracks sailed without the flow noise assembly contain some periods of sail speed < 2.5 m/s, but the NF is not low in all beams. This condition only occurs when the engine appears to be idle. The absence of periods of low sail speed and low NF in all beams without the flow noise assembly leaves this question unanswered: Does the flow noise assembly provide some shielding of the VM-ADCP from engine/propeller noise?
Instrument Performance
Of the 18 tracks sailed, 15 had valid bottom track (VBT) over 88%. Beds of two tracks with lower VBT were largely out of range (depths > 800 m). For one of these two tracks, VBT was 5% with a maximum bottom lock depth of 824 m. For the other, VBT was 52% with a maximum bottom lock depth of 860 m (the Sea Track). The third, which had a VBT of 27%, had the second highest estimated NF (44 dB) of the 18 tracks and high speed over ground (2.5+ m/s for more than half of the track and a maximum speed of 4 m/s). Figure 6 shows bottom track dropouts for the Bay Track with grey vertical lines in the last panel. These occur with heightened vessel motion. No dropouts are seen with minimal pitch and roll, despite speed over ground as high as 4.2 m/s. VBT for this track was VBT 96.7%.
Two of the three tracks sailed without the flow noise assembly had high VBT (92% and 97%). These had stretches of elevated speed over ground (2.5+ m/s) and roll spanning a few to just over five degrees and very moderate pitch. The third track sailed without the flow noise assembly is described above with a VBT of 27%, elevated NF and speed over ground, roll spanning a few to ~ten degrees, and very moderate pitch.
The maximum profiling range for the six deeper tracks (> 300 m) spanned 240 to 320 m. The highest range is observed in the Sea Track, which had the lowest NF of 33.0 dB along with lower NF segments (29 dB) coinciding with slow sail speed (< 2.5 m/s). The echosounder was enabled for four of the six deeper tracks and proved to be sensitive to vessel EMI and acoustic noise, producing a maximum range of 150 to 260 meters.
Figure 8 shows additional output for the Sea Track, which crossed into shallower depths (~ 400 m) at the hairpin turn near the middle of the track. An eastward layer of Atlantic Water several tens to 300 m thick and an underlying layer of westward, Mediterranean Outflow Water are acoustically resolved. Current speeds approach 2.0 m/s in the upper layer, span about 0.2 to 1.0 m/s in the lower layer and are < 0.2 m/s in the thin layer in between. The echogram and Beam 1 amplitude show oscillatory, high intensity structure containing plumes that extend beyond the upper layer to depths of about 200 to 320 m. This structure appears to have extended the profiling range.
Conclusion
The updated motion correction algorithm produces smooth, continuous currents. The 10 Hz sample rate reduces the time elapsed between vessel speed corrections and velocity pings. Further, the spreading of the four VM-ADCP beams over deep, sloped beds distorts bottom track-derived vessel speed, while the GNSS gives a single value. Improved current output is demonstrated in a track with a combination of relatively deep (480 m) and shallow (85 m) water, sloped beds (~ 5 to 10 m/m), substantial vessel motion, and slight vessel motion combined with vessel speed as high as 4.2 m/s. For a track with multiple turns, the lever arm predicted by the algorithm is also confirmed. A sample rate of 10 Hz is now the GNSS default rate for Nortek systems, allowing for robust monitoring and removal of vessel motion.
This test demonstrated that the flow noise assembly does not prevent artifacts in output. Tracks sailed without the flow noise assembly have similar bottom track range, profiling range, and rates of bottom lock to those sailed with the assembly, to the extent that this can be ascertained with tracks sailed in the Bay. On the other hand, the only periods of low NF and distortion-free output in all beams in tracks without the flow noise assembly occurred with an apparently idle engine. To investigate whether the flow noise assembly provides a level of shielding, further tests are necessary.
Analysis of time series plots yielded several conclusions:
- Low NF in all beams coincides with distortion-free output.
- Heightened NF coincides with varying levels of distortions.
- Aft beam NF is consistently greater than fore beam NF, even during periods of low NF in all beams.
- Distortions are more pronounced in aft beams than fore beams.
- A relatively steady engine and moderate (up to 2.5 m/s) speed over ground generally (but not always) coincide with a low NF in all beams and distortion-free output.
- Higher speed over ground (2.5 + m/s) raises the NF and distorts output.
- An idle engine coincides with decreasing speed over ground, low NF in all beams, and distortion-free output.
- Engine revs (both engines) manifest as steep inclines in speed over ground and jumps in the NF.
- Revs in the starboard engine (same side as the VM-ADCP) with the port engine off coincided with a raised NF, especially in aft beams.
- Revs in the port engine with the starboard engine off coincided with a low NF in all beams; the distance between the port engine and/or the keel of the vessel shield the VM-ADCP.
Despite these challenges, the VM Ocean (100 kHz) bottom track performed well. Valid bottom track rates were generally high in this data set, and bottom lock was achieved as deep as 860 m. VBT was lower in tracks where the bed was out of range or where there was substantial pitch and roll.
Nortek recommends hull mounting of low frequency instruments. We conduct side-mounted tests because many users employ this convenient setup. As a result, profiling and echosounder ranges were inhibited by onboard factors such as acoustic noise and EMI. The values presented here represent performance during suboptimal conditions. We expect greater range on vessels with hull mounting and that afford greater distance between propulsion systems and the VM-ADCP.
The VM Ocean (100 kHz) captured the dynamics of the SoG region. The Sea Track features two-layered tidal exchange and oscillations consistent with internal waves. Daugharty et al., 2026 demonstrate that while that asymmetry characteristic of solitons was not observed in oscillatory structure, an internal wave amplitude of 70 m in this track is roughly consistent with observations described by Morozov and Velarde, 2019. Water column structure in beam amplitudes and echograms exhibit plumes of suspended particulate matter that were likely upwelled by vertical mixing introduced by energetic tides and internal waves at Camarinal Sill. This particle-rich condition extended the profiling range beyond the upper layer of scatterer-rich Atlantic Water.
Future steps include a deployment with the full suite of VM Ocean (100 kHz) echosounder frequencies enabled. A test with the VM Ocean (100 kHz) in listening mode with and without the flow noise assembly would indicate whether a dampening of background noise levels was measurable. A Nortek VM Ocean 55 kHz deployment would allow an assessment of capabilities and test motion correction algorithm performance with longer pings and three beams. A setup more aligned with recommendations would provide a true test of performance.
References
- Bolado-Penagos, M., Sala, I., Jesús Gomiz-Pascual, J., González, C.J., Izquierdo, A., Álvarez, Ó., Vázquez, Á., Bruno, M., Van Haren, H., 2023. Analysis of internal soliton signals and their eastward propagation in the Alboran Sea: exploring the effect of subinertial forcing and fortnightly variability. Prog. Oceanogr. 217, 103077. https://doi.org/10.1016/j.pocean.2023.103077
- Daugharty, M., Kamminga, S., Valk, J., van Dorp, R., López, O., 2026. Acoustic observations of stratified Atlantic-Mediterranean exchange and tidal structure. NCK Days, Hydrodynamics & Sediment Transport. https://doi.org/10.13140/RG.2.2.30519.48808
- Gomis, D., Tsimplis, M.N., Martín‐Míguez, B., Ratsimandresy, A.W., García‐Lafuente, J., Josey, S.A., 2006. Mediterranean Sea level and barotropic flow through the Strait of Gibraltar for the period 1958–2001 and reconstructed since 1659. J. Geophys. Res. Oceans 111, 2005JC003186. https://doi.org/10.1029/2005JC003186
- Morozov, E., Velarde, M., 2019. Strong Internal Tides in the Strait of Gibraltar: Measurements and Modelling. Proceedings of the 5th International Conference on Geographical Information Systems Theory, Applications and Management. Presented at the Special Session on Observations and Numerical Modeling of the Coastal Ocean Zone Dynamics, SCITEPRESS - Science and Technology Publications, Heraklion, Crete, Greece, pp. 354–357. https://doi.org/10.5220/0007840603540357
- Roustan, J., Bordois, L., Garciá‐Lafuente, J., Dumas, F., Auclair, F., Carton, X., 2024. Evidence of Reflected Internal Solitary Waves in the Strait of Gibraltar. J. Geophys. Res. Oceans 129, e2023JC020152. https://doi.org/10.1029/2023JC020152
- Sverdrup, H.U., Johnson, M.W., Fleming, R.H., 1942. The Oceans: Their Physics, Chemistry, and General Biology. Prentice-Hall, Inc., New York.
- Tanttari, J., Hynninen, A., 2022. Acoustic Source Characterization of Marine Propulsors. J. Mar. Sci. Eng. 10, 1273. https://doi.org/10.3390/jmse10091273
Updated