Simultaneous Inversion of Bio- and Geo-acoustic Parameters in the Yellow Sea

O.I. Diachok and S.C. Wales
Acoustics Division

Introduction: Bio-acoustic absorptivity due to fish with swim bladders can have large (nominally as large as 40 dB at 5 km) frequency selective effects on transmission loss (TL) at frequencies between approximately 0.5 and 5 kHz in shallow waters.^1,2 TL, which is defined as the difference between signal levels measured at one meter and at other ranges, is one of the parameters that controls the detection range of operational sonars. It is the most challenging to predict. TL is determined primarily by the oceanic sound speed profile, the geo-acoustic properties of the bottom, and the bio-acoustic properties of fish layers. Geo-acoustic parameters may be inferred from TL measurements. To date however, the effects of bio-absorptivity on TL have been disregarded in inversions of geo-acoustic parameters in both the scientific and operational literature. As a result of this omission, the scientific value and operational usefulness of inverted geo-acoustic parameters at/near the resonance frequencies of fish swim bladders at fish-dominated sites, such as the Yellow Sea, are highly questionable. We introduce here a new inversion method that permits simultaneous inversion of bio- and geo-acoustic parameters.

Traditional Approach to Observation of Bio-acoustic Effects: The most compelling demonstrations of bio-acoustic effects on TL have relied on long-term (at least 7-hour) broadband measurements between fixed sources and receivers.^1,3 This experimental configuration permits monitoring of changes in resonance frequencies associated with the vertical migrations of fish, such as sardines and anchovies, at twilight. This method, however, is not "robust." Under some environmental circumstances, these species remain near the surface day and night. Furthermore, other species with swim bladders, such as cod, generally remain near the bottom day and night. In addition, this approach has proved to be logistically cumbersome.

Simultaneous Inversion: To overcome these difficulties, we have developed a new inversion method that permits simultaneous inversion of bio- and geo-acoustic parameters from short-term, broadband TL measurements between multiple source and receiver depths. To illustrate the power of this new technique, we summarize the results of simultaneous inversion of bio-acoustic parameters of fish (anchovies) and geo-acoustic parameters of the bottom from transmission loss measurements in the Yellow Sea that were reported by Chinese scientists at a shallow (40-m) water site near Qingdao.² This data set was selected because the bio-absorptivity at the site was extremely large, 40 dB, and their measurements were made between two source depths and two receiving depths. Figure 4 shows the results of the measurements at a range of 5.7 km. Highest losses occur at 1.35 kHz. This is the resonance frequency of 10-cm long anchovies, which are the dominant species in the Yellow Sea. Losses were highest when both source and receiver were placed at 7 m; they were lowest when both source and receiver were placed at 25 m.

Replica fields (theoretically generated acoustic fields for comparison with TL measurements) were calculated with a normal mode model that incorporates the effects of bio-acoustic absorption layers and the geo-acoustic properties of the bottom. The inversion was based on minimizing the root-mean-square (rms) difference Δ between measured and calculated values of TL at multiple ranges and multiple source and receiver depths. It involved a simultaneous search for bio-layer depth d, bio-layer thickness t, bio-alpha α_B, geo-sound speed c_P, and geo-alpha α_P. Figure 5 shows the two-dimensional ambiguity surface of Δ at 1.35 kHz as a function of d and α_B with t, c_P, and α_P held constant. The resultant extremely small value of Δ (1.9 dB) confirmed that the model, which was assumed in replica field calculations, was realistic, and that inverted parameters were meaningful. The inverted value of d, 6.9 ± 0.3 m, is consistent with theoretical calculations of d, 5.8 ± 1 m, of 10-cm long anchovies, the dominant species in the Yellow Sea, and with laboratory measurements of resonance frequencies of anchovies. The inverted value of t, about 0.4 m, is consistent with the nominal thickness of biological layers on continental shelves. The values of c_P (1700 m/s) and α_P (0.2 dB/λ) are consistent with Bowles' review⁴ of previously reported measurements of these parameters.

By contrast, inversion calculations, which assumed that all excess attenuation (above geometrical spreading) at this site was due to the bottom, resulted in an unacceptably large value of Δ(9.5 dB) and an unrealistic value of α_P (3 dB/λ). Figure 6 shows inverted values of α_P from Ref. 2 data with and without α_B, other inversions⁵ of α_P (which disregarded α_B) at a nearby site, and upper and lower bounds according to Ref. 4.

FIGURE 5
Ambiguity surface of the rms difference D between measured and calculated transmission loss at 1.35 kHz vs layer depth and absorption coefficient with the layer. Layer thickness equals 0.4 m. The inverted layer depth is approximately equal to 6.9 m. A relative attenuation of 1 corresponds to 0.3 dB/l in a 1-m thick layer. Contours: 2, 3, and 5 dB.

FIGURE 6
Inversions from data of Ref. 2 with (•) and without (O) bio-alpha, Ref. 5 data without bio-alpha (D), and upper and lower bounds according to Ref. .

Conclusions: We have shown that it is possible to simultaneously invert geo-acoustic parameters of the bottom and bio-acoustic parameters of biological absorbing layers in the water column from TL measurements between multiple source and receiver depths. This method is expected to prove useful for inversion of bio- and geo-acoustic parameters at operationally mandated sites where the concentrations of fish are high.

[Sponsored by ONR]