Model-based covariance mean variance classification techniques:algorithm development and application to the acoustic classification ofzooplankton

For inversion problems in which the theoretical relationship between observed data and model parameters is well characterized, a promising approach to the classification problem is the application of techniques that capitalize on the predictive power of class-specific models. Theoretical models have been developed for three zooplankton scattering classes (hard elastic-shelled, e.g., pteropods; fluid-like, e.g., euphausiids; and gas-bearing, e.g., siphonophores), providing a sound basis for model-based classification approaches. The covariance mean variance classification (CMVC) techniques classify broad-band echoes from individual zooplankton based on comparisons of observed echo spectra to model space realizations. Three different CMVC algorithms were developed: the integrated score classifier, the pairwise score classifier, and the Bayesian probability classifier; these classifiers assign observations to a class based on similarities in covariance, mean, and variance while accounting for model spare ambiguity and validity. The CMVC techniques were applied to broad-band (~350-750 kHz) echoes acquired from 24 different zooplankton to invert for scatterer class and properties. All three classification algorithms had a high rate of success with high-quality high SNR data. Accurate acoustic classification of zooplankton species has the potential to significantly improve estimates of zooplankton biomass made from ocean acoustic backscatter measurements     

Mine Burial Experiments at the Martha's Vineyard Coastal Observatory

Several experiments to measure postimpact burial of seafloor mines by scour and fill have been conducted near the Woods Hole Oceanographic Institution's Martha's Vineyard Coastal Observatory (MVCO, Edgartown, MA). The sedimentary environment at MVCO consists of a series of rippled scour depressions (RSDs), which are large scale bedforms with alternating areas of coarse and fine sand. This allows simultaneous mine burial experiments in both coarse and fine sand under almost identical hydrodynamic forcing conditions. Two preliminary sets of mine scour burial experiments were conducted during winters 2001-2002 in fine sand and 2002-2003 in coarse sand with a single optically instrumented mine in the field of view of a rotary sidescan sonar. From October 2003 to April of 2004, ten instrumented mines were deployed along with several sonar systems to image mine behavior and to characterize bedform and oceanographic processes. In fine sand, the sonar imagery of the mines revealed that large scour pits form around the mines during energetic wave events. Mines fell into their own scour pits, aligned with the dominant wave crests and became level with the ambient seafloor after several energetic wave events. In quiescent periods, after the energetic wave events, the scour pits episodically infilled with mud. After several scour and infilling events, the scour pits were completely filled and a layer of fine sand covered both the mines and the scour pits, leaving no visible evidence of the mines. In the coarse sand, mines were observed to bury until the exposed height above the ripple crests was approximately the same as the large wave orbital ripple height (wavelengths of 50-125 cm and heights of 10-20 cm). A hypothesis for the physical mechanism responsible for this partial burial in the presence of large bedforms is that the mines bury until they present roughly the same hydrodynamic roughness as the orbital-scale bedforms present in coarse sand.     

Predicting Seabed Burial of Cylinders by Wave-Induced Scour: Application to the Sandy Inner Shelf Off Florida and Massachusetts

A simple parameterized model for wave-induced burial of mine-like cylinders as a function of grain-size, time-varying, wave orbital velocity and mine diameter was implemented and assessed against results from inert instrumented mines placed off the Indian Rocks Beach (IRB, FL), and off the Martha's vineyard coastal observatory (MVCO, Edgartown, MA). The steady flow scour parameters provided by Whitehouse (1998) for self-settling cylinders worked well for predicting burial by depth below the ambient seabed for (0.5 m) diameter mines in fine sand at both sites. By including or excluding scour pit infilling, a range of percent burial by surface area was predicted that was also consistent with observations. Rapid scour pit infilling was often seen at MVCO but never at IRB, suggesting that the environmental presence of fine sediment plays a key role in promoting infilling. Overprediction of mine scour in coarse sand was corrected by assuming a mine within a field of large ripples buries only until it generates no more turbulence than that produced by surrounding bedforms. The feasibility of using a regional wave model to predict mine burial in both hindcast and real-time forecast mode was tested using the National Oceanic and Atmospheric Administration (NOAA, Washington, DC) WaveWatch 3 (WW3) model. Hindcast waves were adequate for useful operational forcing of mine burial predictions, but five-day wave forecasts introduced large errors. This investigation was part of a larger effort to develop simple yet reliable predictions of mine burial suitable for addressing the operational needs of the U.S. Navy.     

River flooding, storm resuspension, and event stratigraphy on the northern California shelf: observations compared with simulations

Coast-hugging surface flood plumes occur on the inner shelf of northern California during the winter season, generating dense, near-bottom suspensions which may attain fluid mud concentrations as particles settle. The period of storm-heightened waves may continue into the flood period, leading to gravity-driven seaward displacement of the bottom suspension; or the wave regime may ameliorate, leaving the suspension to consolidate as a short-lived, inner-shelf flood bed. Such beds tend to be resuspended within days or weeks by subsequent storm events that may recreate the original high concentrations. The sediment is thus dispersed seaward by gravity flows, to be deposited as a muddy flood bed on the central shelf. The locus of deposition of these “high-concentration regimes” is a function of the relative intensities of river discharge and storm wave height. Greater discharge piles thicker storm beds nearer shore, while intense wave regimes allow deposition of the fluid mud further seaward. During events with high values of both parameters, large amount of fluid mud may bypass over the shelf edge. In contrast, “low-concentration regimes” occur during storm periods when there has been no recent flood deposition on the inner shelf. The shelf floor is better consolidated than in the previous case, and the resulting suspended sediment concentrations are lower. As a consequence, low-concentration regimes are winnowing and bypassing regimes, and the beds deposited are thinner and sandier. Algorithms describing deposition by high and low-concentration regimes have been embedded in a probabilistic model. A simulation of a 400-year sequence of beds deposited by winter storms and floods suggests that on the Eel shelf, the Holocene transgressive systems tract consists of back-stepping, seaward-fining event beds, whose timelines (bedding planes) dip more gently than do their gradational facies boundaries. At these longer time scales, flood beds dominate over storm beds.     

Numerical Simulations of the Flow and Sediment Transport Regimes Surrounding a Short Cylinder

The 3-D flow field and bed stress surrounding a short cylinder in response to combined wave and mean-flow forcing events is examined. Model simulations are performed with a 3-D nonhydrostatic computational fluid dynamics model, FLOW-3D. The model is forced with a range of characteristic tidal and wave velocities as observed in 12-15 m of water at the Martha's Vineyard Coastal Observatory (MVCO, Edgartown, MA). The 2.4-m-long and 0.5-m diameter cylinder is buried 10% of the diameter on a flat, fixed bed. Regions of incipient motion are identified through local estimates of the Shields parameter exceeding the critical value. Potential areas of sediment deposition are identified with local estimates of the Rouse parameter exceeding ten. The model predictions of sediment response are in general in agreement with field observations of seabed morphology obtained over a one-week period during the 2003-2004 MVCO mine burial experiment. Both observations and simulations show potential transport occurring at the ends of the mine in wave-dominated events. Mean flows greater than 10 cm/s lead to the formation of larger scour pits upstream of the cylinder. Deposition in both cases tends to occur along the sides, near the center of mass of the mine. However, the fixed-bed assumption prohibits the prediction of full perimeter scour as is observed in nature. Predicted scour and burial regimes for a range of wave and mean-flow combinations are established.     

Sediment transport and trapping in the Hudson River estuary

The Hudson River estuary has a pronounced turbidity maximum zone, in which rapid, short-term deposition of sediment occurs during and following the spring freshet. Water-column measurements of currents and suspended sediment were performed during the spring of 1999 to determine the rate and mechanisms of sediment transport and trapping in the estuary. The net convergence of sediment in the lower estuary was approximately 300,000 tons, consistent with an estimate based on sediment cores. The major input of sediment from the watershed occurred during the spring freshet, as expected. Unexpected, however, was that an even larger quantity of sediment was transported landward into the estuary during the 3-mo observation period. The landward movement was largely accomplished by tidal pumping (i.e., the correlation between concentration and velocity at tidal frequencies) during spring tides, when the concentrations were 5 to 10 times higher than during neap tides. The landward flux is not consistent with the long-term sediment budget, which requires a seaward flux at the mouth to account for the excess input from the watershed relative to net accumulation. The anomalous, landward transport in 1999 occurred in part because the freshet was relatively weak, and the freshet occurred during neapetides when sediment resuspension was minimal. An extreme freshet occurred during 1998, which may have provided a repository of sediment just seaward of the mouth that re-entered the estuary in 1999. The amplitude of the spring freshet and its timing with respect to the spring-neap cycle cause large interannual variations in estuarine sediment flux. These variations can result in the remobilization of previously deposited sediment, the mass of which may exceed the annual inputs from the watershed.