Influence of bottom trawling on sediment resuspension in the ‘Grande-Vasière’ area (Bay of Biscay, France)

Abstract

Sea trials were performed on two zones with different fishing efforts on the continental shelf of the Bay of Biscay (‘Grande-Vasière’ area of muddy sand) in order to assess particulate matter resuspension and seabed disturbances (i.e., penetration, reworking, grain size changes) induced by different types of trawls. Optical and acoustic measurements made in the water column indicate a significant trawling-induced resuspension mainly due to the scraping action of doors. It manifests as a highly dynamic turbid plume confined near the seabed, where suspended sediment concentrations can reach 200 mg l⁻¹. Concentration levels measured behind an “alternative” configuration (trawls with jumper doors instead of classical doors penetrating the sediment) are significantly lower (around 10–20 mg l⁻¹), which indicates a potential limiting impact regarding the seabed. Grain size analyses of the surficial sediment led to highlight a potential reworking influence of bottom trawling. On the intensively trawled zone, this reworking manifests as an upward coarsening trend in the first 5 cm of the cores. A significant decrease in mud content (30 %) has been also witnessed on this zone between 1967 and 2014, which suggests an influence on the seabed evolution. The geometric analysis of bottom tracks (4–5-cm depth, 20-cm width) observed with a benthic video sledge was used to compute an experimental trawling-induced erosion rate of 0.13 kg m⁻². This erosion rate was combined with fishing effort data, in order to estimate trawling-induced erosion fluxes which were then compared to natural erosion fluxes over the Grande-Vasière at monthly, seasonal and annual scales. Winter storms control the annual resuspended load and trawling contribution to annual resuspension is in the order of 1 %. However, results show that trawling resuspension can become dominant during the fishing high season (i.e., until several times the natural one in summer). In addition, the contribution of trawling-induced resuspension is shown to increase with water depth, because of the rapid decay of wave effects. Finally, the seasonal evolution of the respective contributions for erosion (mainly trawling and waves) could be mapped for the whole study area.

Appendices

Appendix 1 backscatter index provided by ADCP

Tessier et al. (2008b) proposed a processing method for the RDI ADCP 1200 kHz used in this study. All the following coefficients and hypotheses are thus directly extracted from Tessier et al. (2008b). As deduced from the sonar equation (Urick 1975; Lurton 2002), they expressed the backscatter index BI according to the relation:

$$ \mathrm{BI}=\mathrm{R}\mathrm{L}-\mathrm{S}\mathrm{L}+{\mathrm{TL}}_{\mathrm{geo}}+{\mathrm{TL}}_{\mathrm{ws}}-{\mathrm{C}}_{\mathrm{geo}} $$

(7)

where RL is the received level (dB.μPa⁻¹) and is expressed as:

$$ \mathrm{R}\mathrm{L}=B+\mathrm{K}\mathrm{C}\left(\mathrm{E}\mathrm{C}-{\mathrm{EC}}_0\right) $$

(8)

where B and EC₀ are internal transducer noises respectively set at 70 dB and 46 counts; KC is a dB/count conversion coefficient, fixed at 0.423 and EC is the echo received signal in count. SL corresponds to the emitted level and is set to 217 dB μPa⁻¹. TL_geo and TL_ws are terms linked to the transmission loss along the beam path (in dB). TL_geo is the geometrical attenuation for the spherical spreading and is expressed as:

$$ {\mathrm{TL}}_{\mathrm{geo}}=40\times { \log}_{10}\left(\psi R\right) $$

(9)

where R is the distance from the transducer and ψ is the near field correction. TL_ws corresponds to the signal attenuation induced by the water and the particles and is expressed as:

$$ {\mathrm{TL}}_{\mathrm{ws}}=2\left({\alpha}_w+{\alpha}_S\right)R $$

(10)

where α _w is the water attenuation coefficient set to 0.5316 dB m⁻¹ according to the model of Francois and Garrison (1982a, b). When the suspended particulate matter concentrations are generally inferior to 200 mg l⁻¹, the signal attenuation induced by the particles can be neglected (α _S  = 0).

At last, C _geo corresponds to a geometric correction that accounts for the expansion of the backscattering volume with the increasing distance R from the source. It is defined as:

$$ {C}_{\mathrm{geo}}=10\times { \log}_{10}\left(\pi {\left(\frac{\phi }{2}\right)}^2{R}^2\times L\right) $$

(11)

where ϕ is the equivalent opening of the beam and L refers to the half-height cell.

Appendix 2 computation of the bed shear stress under natural forcing

Computation of the wave shear stress

Wave-induced bottom shear stresses have been computed with a wave hindcast database built with the WW3 wave model (Boudière et al. 2013) that provides parameters such as bottom amplitude displacement (A), bottom orbital velocities (U _b ) and the direction of wave propagation. The WW3 code is a phase-averaged wave model that resolves the random phase spectral action density balance equation for a spectrum of wave numbers and directions: in the present application, 24 directions and 32 frequencies were accounted for. The realistic and validated configuration proposed by Boudière et al. (2013) includes all the continental shelf of the Bay of Biscay, using an unstructured grid. The horizontal resolution of the computational grid varies in our case from 200 m (nearshore) to 10 km (open sea). According to Jonsson (1966) formulation, the wave-induced shear stress τ _w is computed as:

$$ {\tau}_w=\frac{1}{2}{f}_w\rho {U_b}^2 $$

(12)

where ρ is the water density and f _w refers to the wave-induced friction factor. This factor is computed according to the Soulsby et al. (1993) formulation:

$$ {f}_w=1.39\ {\left(\frac{A}{z_0}\right)}^{-0.52} $$

(13)

where the bed roughness length z ₀ is equal to 5 × 10⁻⁴ m.

Computation of the current shear stress

The current-induced bottom shear stress estimation is based on the tridimensional hydrodynamic model MARS3D (Lazure and Dumas 2008) which runs on a realistic configuration validated at the regional scale of the Bay of Biscay. This model resolves the Navier-Stokes equations with the classical Boussinesq and hydrostatic hypotheses on Cartesian grid. Regarding the mesh, the horizontal resolution is equal to 2.5 km and the vertical discretisation is characterised by 40 generalised sigma levels. The bed shear stress induced by currents is obtained through the relationship:

$$ {\tau}_c=\rho {u_{*}}^2 $$

(14)

where the friction velocity u _* is computed as:

$$ {u}_{*}=0.4{u}_1/ \ln \left(\frac{z_1}{z_0}\right) $$

(15)

where ρ is the water density, u ₁ is the computed velocity at the elevation z ₁ which corresponds to half the near-bottom cell thickness, and z ₀ is the bottom sediment roughness length set to a constant value of 5 × 10⁻⁴ m.

Computation of the total bottom shear stress

The non-linear interaction between waves and currents is accounted for by computing the total bottom shear stress according to the formulation from Soulsby (1997):

$$ \tau ={\left[{\left({\tau}_m+{\tau}_w| \cos \kern0.1em \phi |\right)}^2+{\left({\tau}_w \sin \kern0.1em \phi \right)}^2\right]}^{1/2} $$

(16)

where ϕ is the angle between the current and the wave propagation directions and τ _m represents the wave-averaged bed shear stress. τ _m is computed from the current- and wave-induced bed shear stresses τ _c and τ _w which correspond to the bed shear stresses due to the current alone and to the wave alone, respectively:

$$ {\tau}_m={\tau}_c\left[1+1.2{\left(\frac{\tau_w}{\tau_c+{\tau}_w}\right)}^{3.2}\right] $$

(17)

The expression of τ is deduced from a vector addition of τ _m and τ _w .