Stochasticity of riverbed backscattering, with implications for acoustical classification of non-cohesive sediment using multibeam sonar

Presented at RiverFlow 2016: the 8th international conference on fluvial hydraulics, At St Louis, Missouri

Transcript

2. Study motivation Acoustic classification of riverbed sediments: high coverage and resolution, quickly and efficiently Physical habitat characterization, sediment availability for transport, sediment patchiness, bed roughness for hydraulic models Acoustic backscatter is a stochastic quantity Few studies characterizing variability in space and time, link to a particular acoustic classification method not made
3. Study motivation Acoustic classification of riverbed sediments: high coverage and resolution, quickly and efficiently Physical habitat characterization, sediment availability for transport, sediment patchiness, bed roughness for hydraulic models Acoustic backscatter is a stochastic quantity Few studies characterizing variability in space and time, link to a particular acoustic classification method not made
4. Study motivation Acoustic classification of riverbed sediments: high coverage and resolution, quickly and efficiently Physical habitat characterization, sediment availability for transport, sediment patchiness, bed roughness for hydraulic models Acoustic backscatter is a stochastic quantity Few studies characterizing variability in space and time, link to a particular acoustic classification method not made
5. Study motivation Acoustic classification of riverbed sediments: high coverage and resolution, quickly and efficiently Physical habitat characterization, sediment availability for transport, sediment patchiness, bed roughness for hydraulic models Acoustic backscatter is a stochastic quantity Few studies characterizing variability in space and time, link to a particular acoustic classification method not made
6. Talk outline Two different flow and sedimentary environments: a highly heterogeneous bed, and a relatively homogeneous bed of migrating sand dunes Method to compute backscatter coefficient from recorded echo amplitudes Variation of maps of acoustic backscatter over time Relationship between backscatter and sediment type Sensitivity of acoustic sediment classifications to backscatter variations
7. Talk outline Two different flow and sedimentary environments: a highly heterogeneous bed, and a relatively homogeneous bed of migrating sand dunes Method to compute backscatter coefficient from recorded echo amplitudes Variation of maps of acoustic backscatter over time Relationship between backscatter and sediment type Sensitivity of acoustic sediment classifications to backscatter variations
8. Talk outline Two different flow and sedimentary environments: a highly heterogeneous bed, and a relatively homogeneous bed of migrating sand dunes Method to compute backscatter coefficient from recorded echo amplitudes Variation of maps of acoustic backscatter over time Relationship between backscatter and sediment type Sensitivity of acoustic sediment classifications to backscatter variations
9. Talk outline Two different flow and sedimentary environments: a highly heterogeneous bed, and a relatively homogeneous bed of migrating sand dunes Method to compute backscatter coefficient from recorded echo amplitudes Variation of maps of acoustic backscatter over time Relationship between backscatter and sediment type Sensitivity of acoustic sediment classifications to backscatter variations
10. Talk outline Two different flow and sedimentary environments: a highly heterogeneous bed, and a relatively homogeneous bed of migrating sand dunes Method to compute backscatter coefficient from recorded echo amplitudes Variation of maps of acoustic backscatter over time Relationship between backscatter and sediment type Sensitivity of acoustic sediment classifications to backscatter variations
11. Site 1. South Canyon, Marble Canyon (May 2013)
12. Site 2. Diamond Creek, Western Grand Canyon (July 2015)
13. Computing backscatter coefficient from echo magnitude Raw echo (what we measure) BS(θ) = EL − SL + 2TL − Af 10 log10 of ratios between a quantity and a reference quantity of acoustic pressure of 1 µ Pa Source level [MEASURED] Transmission losses [ESTIMATED] True area of beam footprint [ESTIMATED] Amiri-Simkooei et al., Journal of the Acoustic Society of America, 2009
14. Computing backscatter coefficient from echo magnitude Raw echo (what we measure) BS(θ) = EL − SL + 2TL − Af ± NL 10 log10 of ratios between a quantity and a reference quantity of acoustic pressure of 1 µ Pa Source level [MEASURED] Transmission losses [ESTIMATED] True area of beam footprint [ESTIMATED] Noise level [IGNORED?] Amiri-Simkooei et al., Journal of the Acoustic Society of America, 2009
15. Change in backscatter: South Canyon (3 surveys, 2 hours)
16. Change in backscatter: Diamond Creek (28 surveys, 4 hours)
17. Backscatter and sediment type: South Canyon
18. Backscatter and sediment type: Both sites
19. Sediment classification method of Buscombe et al. (2014)
20. Sensitivity of sediment classification to backscatter variation
21. Summary Growing popularity of multibeam sonar in rivers Large variations in ”per-pixel” backscatter over minutes to hours Possible to distinguish between substrate types based on backscatter strength Distribution of backscatter values associated with each sediment type limits acoustic sediment classification Sediment classification based on backscatter spectra relatively insensitive to temporal variations in backscatter
22. Summary Growing popularity of multibeam sonar in rivers Large variations in ”per-pixel” backscatter over minutes to hours Possible to distinguish between substrate types based on backscatter strength Distribution of backscatter values associated with each sediment type limits acoustic sediment classification Sediment classification based on backscatter spectra relatively insensitive to temporal variations in backscatter
23. Summary Growing popularity of multibeam sonar in rivers Large variations in ”per-pixel” backscatter over minutes to hours Possible to distinguish between substrate types based on backscatter strength Distribution of backscatter values associated with each sediment type limits acoustic sediment classification Sediment classification based on backscatter spectra relatively insensitive to temporal variations in backscatter
24. Summary Growing popularity of multibeam sonar in rivers Large variations in ”per-pixel” backscatter over minutes to hours Possible to distinguish between substrate types based on backscatter strength Distribution of backscatter values associated with each sediment type limits acoustic sediment classification Sediment classification based on backscatter spectra relatively insensitive to temporal variations in backscatter
25. Ongoing and future work Develop uncertainties for sediment classification estimates More robust estimates of contribution of SSC to transmission losses Investigate backscttering characteristics of submerged aquatic vegetation More, and more complicated, substrate types
26. This slide is intentionally blank.
27. Depth and acoustic characteristics
28. Computing backscatter coefficient from echo magnitude Af = F(aperture, pulse duration) and F(range, grazing angle). Grazing angles are calculated over at least 3 successive beams, therefore for small beams the residual effects of small-scale topography remain Scaling factor relates ‘nominal’ to ‘true’ beam footprint area (log10Af = log10Af + log10 ) Laplacian of bed elevations: log10 ≈ log10 ∇2(x, y) Accounts for increasing surface area due to slope effects
29. Computing backscatter coefficient from echo magnitude Af = F(aperture, pulse duration) and F(range, grazing angle). Grazing angles are calculated over at least 3 successive beams, therefore for small beams the residual effects of small-scale topography remain Scaling factor relates ‘nominal’ to ‘true’ beam footprint area (log10Af = log10Af + log10 ) Laplacian of bed elevations: log10 ≈ log10 ∇2(x, y) Accounts for increasing surface area due to slope effects