Multi-Frequency Acoustics for Deep-Sea Sediment Plume Monitoring

Abstract

Deep-sea nodule collection generates suspended sediment plumes that may affect the surrounding benthic environment. Reliable monitoring of suspended sediment concentration (SSC) and particle size near the plume source is therefore important for plume characterization, modelling, and environmental impact assessment. This study evaluates the applicability of multi-frequency acoustic backscatter measurements for determining SSC and particle size estimates under conditions representative of deep-sea nodule collection plumes.

An experimental campaign was conducted in a recirculation tank using two representative deep-sea box-core sediments, referred to as CCZ sediment and Sediment B. Multi-frequency acoustic backscatter measurements were obtained at 1, 2, 4, and 5 MHz frequencies and were evaluated against gravimetric SSC measurements over an experimental suspended sediment concentration range of
0.16 − 32.4 g/L. The recirculation tank was designed to generate a relatively homogeneous suspension of unflocculated deep-sea sediment, thereby simulating conditions within a collector vehicle diffuser. Four acoustic inversion methods were evaluated. Methods 1 and 2 applied commercially available
bin-based inversion approaches to resolve SSC and particle size over range, whereas Methods 3 and 4 applied profile-averaged inversion approaches implemented in a custom Python workflow. The commercially available bin-based inversion methods did not provide reliable results for the investigated fine-grained deep-sea sediments. Method 1 substantially overestimated particle size, by factors of 4.1 and 18.2 for the CCZ sediment and Sediment B, respectively, which corresponded to SSC underestimations of approximately 1.2 and 2.0 orders of magnitude. Method 2 did not yield reliable SSC inversions due to its strong sensitivity to small deviations in the prescribed particle-size parameters (D50
and σ). In contrast, the profile-averaged Methods 3.1 and 4.4 provided quantifiable SSC estimates for both sediments. The inversion performance was governed primarily by the prescribed median particle diameter (D50), while constraining the particle-size distribution width (σ) provided limited additional
benefit. The D50-constrained configurations were therefore selected as the representative cases. Over the common evaluated SSC range of 0.5 − 18.0 g/L, Method 3.1 achieved proportional biases of +36.1 % and +1.2 % for the CCZ sediment and Sediment B, respectively, with corresponding R2 values
of 0.995 and 0.978. Method 4.4 achieved proportional biases of +9.8 % and −15.5 %, with R2 values of 0.944 and 0.867, respectively. For the CCZ sediment, extending the evaluated range to 32.4 g/L resulted in proportional biases of +31.1 % and +19.2 % for Methods 3.1 and 4.4, respectively. Across the selected configurations, the proportional bias remained within approximately ±36 %. Particle-size and distribution width inversions were substantially more sensitive to uncertainties in the acoustic measurements and could not yet be reliably quantified.

For comparison, the optical backscatter sensor (OBS) produced slopes of 0.857 (R2 = 0.983) for Sediment B and 0.546 (R2 = 0.884) for the CCZ sediment. The difference in performance between the sediments demonstrates the dependence of optical measurements on representative sediment-specific calibration. Multi-frequency acoustic backscatter, in contrast, provided range-averaged SSC estimates without requiring sediment-specific calibration, although its effective measurement range significantly decreased with increasing SSC.

Overall, multi-frequency acoustic backscatter shows strong potential for range-averaged SSC monitoring within or close to deep-sea nodule collector diffusers, where particle properties remain relatively stable and theoretical attenuation and scattering models are most applicable. Its ability to provide measurements over a larger representative volume without sediment-specific calibration offers an important advantage over conventional optical point measurements. However, further development is required to improve particle-size inversion and measurement reliability at elevated SSC, followed by validation under representative deep-sea operational conditions.