Within the near field, plumes are energetically turbulent with wide variations in instantaneous measurements (Roberts, 1996). Sampling within this area is therefore problematic as results may fluctuate widely and be difficult to interpret. One of the main goals of this work was to develop a reliable in situ method to monitor pollutants discharged in the environment with increased temporal and spatial resolution, making use of an AUV. In order to reduce the uncertainty about plume location and concentrate the monitoring mission only in the hydrodynamic mixing zone, outputs of a near field prediction model, based on effective real time in situ measurements of current speed and direction and density stratification, were opportunistically used to specify in real time the mission transects (Ramos, 2005).
The near field model used was RSB (Roberts et al., 1989). The RSB model, based on the experimental work on multiport diffusers of Roberts et al. (1989) using dimensional analysis and length scale arguments, incorporates the most important hydrodynamic aspects of ocean outfalls. These include, the effects of arbitrary current speed and direction (including parallel currents), stratification, port spacing, source momentum flux, discharges from both sides of the diffuser and the resulting merging of the plumes from both sides, re-entrainment and additional mixing in the spreading layer, direct plume impingement in parallel currents, and lateral gravitational spreading.
The model has been validated against data: (1) from a field tracer and laboratory experiments to the San Francisco outfall (Roberts & Wilson, 1990), (2) from field tests to Whites Point, Los Angeles outfall (Washburn et al., 1992) and Sand Island, Hawaii outfall (Petrenko et al., 1998), and recently (3) from field tracer experiments to Boston outfall (Roberts et al., 2002).
The RSB outputs, including length of hydrodynamic mixing zone, spreading width at the end of the near field, maximum rise height and thickness, in conjunction with current direction, were used to define the sampling transects.
To guarantee the plume observation at the end of near field, the longitudinal distance was actually considered somewhat greater than the initial mixing zone length. The maximum downstream normalized distance used by Roberts et al. (1989) in its laboratory experiments was, in fact, the survey length considered. A minimum safe distance to the outfall was also guaranteed to avoid jets turbulence affect the vehicle navigation. The minimum downstream normalized distance used by Roberts et al. (1989) in its laboratory experiments was used as reference. The wastefield width was increased in 20% to take into account experimental errors, as suggested by Roberts et al. (1989). A minimum distance of 4 m from the sea bed (due to safety navigation requirements), and a minimum distance of 2 m from the surface (due to wave motions interference with the vehicle navigation) was also considered. Eliminating unnecessary measurements, this adaptive sampling approach enables to increase the horizontal and vertical resolutions, that are specially critical to natural tracer tracking in environments with large gradients in background values of the natural tracer.
Yo-yo shaped transects were not particularly useful in this case. With a maximum dive angle of 10°, the vehicle would perform minimum cycles of 159-193 m in yo-yos of 14-17 m amplitude, with a horizontal resolution between 79-96 m at the middle of the water column. If the wastefield width at the end of near field is about 1.5 times the diffuser length (Roberts et al. 1989), i.e. 147 m, the AUV would perform not more than 2 yo-yos in each transect, missing lot of data.
So, horizontal transects at different depths with a minimum vertical spacing of 2 m and a minimum spacing of 20 m between horizontal parallel transects, due to the natural navigation variability in depth of Isurus AUV, seemed to be a right choice for these specific study conditions.
Tracks perpendicular to current direction, instead of parallel, are preferred to minimize temporal aliasing between samples at the same cross section. Cross sections at the end of near field can be comparable to RSB model predictions. Tracks parallel to down current direction are preferred to observe how dilution varies with distance.
To be able to compare field observations with model predictions, density vertical profiles should be collected outside the dispersion area (unaffected by the plume salinity). Those profiles are necessary to run the models.
Two applications were developed to implement easily in the field the sampling strategy adopted. Sea Outfall Monitoring Campaign GUI is used for the automatic specification of the inspection area, and Isurus Mission GUI is used for the automatic definition of the AUV mission file (see their layouts during S. Jacinto outfall campaign in Fig. 3 and Fig. 4).
According to RSB model predictions, obtained in real time during S. Jacinto outfall monitoring campaign, the plume was spreading at the surface, detached from the bottom and forming a two-layer flow. At the end of the near field 42 m downstream from the diffuser, the predicted wastefield width was about 177 m.
The AUV monitoring mission took about 112 minutes, starting approximately at 14:00 GMT. A rectangular area of 200 m x 100 m starting 20 m downstream from the middle point of the diffuser was covered (see the vehicle position estimate in Fig. 5).
As predicted, the vehicle performed six horizontal trajectories at 2, 4, 6, 8, 10, and 12 m depth. In each horizontal trajectory, the vehicle described six parallel transects perpendicular to the water current direction, of 200 m length and spacings of 20 m (labelled as section 1-6 in Fig. 5).
In performing horizontal trajectories, vertical oscillations of the AUV were less than 0.5 m. During the mission the vehicle transited at a fairly constant velocity of 1 m/ s. CTD data were recorded at a rate of 2.4 Hz, so horizontal resolution was about 0.4 m (horizontal resolution is defined here as the approximate distance between consecutive points that are sampled at the same depth, i.e., in the same layer). Vertical resolution varied along the mission due to the influence on the vehicle navigation of natural currents but was almost always between 1-2 m (vertical resolution is defined as the vertical distance between points in the water column that are sampled approximately at the same (X,Y) location but at consecutive depths, i.e., on successive layers).
Comparing to other field studies (Washburn et al., 1992; Wu et al., 1994; Petrenko et al., 1998) this resolution corresponds to a huge improvement in plume tracking surveys.
Was this article helpful?