Mamala Bay
Contact Us
OPL Home
OPL Projects

Mamala Bay Project

Ocean Physics Laboratory

Mamala Bay Publications

Petrenko, A.A., B.H. Jones, and T.D. Dickey, 1998, Shape and near-field dilution of the Sand Island sewage plume: observations compared to model results, J. Hydraul. Eng., 124, 565-571. View highlights of the paper.

Petrenko, A.A., B.H. Jones, T.D. Dickey, M. LeHaitre, and C. Moore, 1997, Effects of a sewage plume on the biology, optical characteristics, and particle size distributions of coastal waters, J. Geophys. Res., 102, 25,061-25,071. View highlights of the paper.

Petrenko, A.A., B.H. Jones, T.D. Dickey, and P. Hamilton, 2000, Internal tidal effects on sewage plume near Sand Island, HI, Cont. Shelf Res., 20, 1-13. View highlights of the paper.

Shape and near-field dilution of the Sand Island sewage plume: observations compared to model results

A.A. Petrenko, B.H. Jones, and T.D. Dickey
J. Hydraul. Eng., 124, 565-571, 1998.

The deeply submerged plume was patchy at its center, thin on its edges and displayed vertically separated layers on 3 out of 5 days of plume mapping. Complexity of the sewage plume shape was due to a combination of factors that include temporal and spatial variations in currents and temperature stratification, and internal tide effects. Equilibrium depth, thickness, and initial dilution of the sewage plume were derived from in situ measurements proximal to the sewage outfall and were compared with simulation results from the Roberts, Snyder, and Baumgartner model. Simulated equilibrium depth and thickness of the plume were within 10 m of their measured counterparts in all but 2 cases for the former and in all but 3 cases for the latter (out of 11 cases). Simulated dilutions were 2.48 greater than values derived from in situ data. Dilution differences are explained by differences between the engineering definition and the oceanographic characterization of initial dilutions.

Select Figures

Figure 4
3-D view of the Sand Island, HI sewage plume. Note the patchiness and spatial complexity. The plume does not remain at a specific depth of equilibrium (see Internal tide paper for explanation on the deepening of the plume with alongshelf distance west of the diffuser. (Click for larger version)

Figure 3
Dilution calculation using T/S diagram and initial mixing lines between the effluent and the ambient waters at diffuser depth. (Click for larger version)

In oceanic environments as variable as Mamala Bay, until synoptic 3-dimension (3D) detection and dilution determination of outfall plumes are technologically feasible, the engineering model dilution and the oceanographic dilution determinations may diverge since they are not defined for the former, or calculated for the latter, at the same location.

Plume behavior is extremely complex in environments in which stratification is variable and current amplitudes and directions change with depth and time. Representation of the oceanographic complexity is beyond the capacity of available models. In the present study, the RSB results captured the gross characteristics, equilibrium depth and thickness, of the observed wastefield. Whether the discrepancies found between Dme and Dmo in this study resulted from limitations in the model, inaccuracies in the model inputs, or uncertainties in Dme cannot be concluded. Studies to investigate model sensitivity to particular inputs may yield additional insights. Near-field models have been incorporated into 3D coupled circulation/water quality models, state-of-the-art tools used for risk assessment analyses (e.g., Blumberg and Connolly 1995). However, the simplified near-field models do not parameterize the complexity of sewage plume shapes and patchiness. Inclusion of in situ sewage plume measurements may be necessary for the 3D models to properly define the entire range of solutions.

At the SITP site, additional field observations should be obtained during periods when temperature stratification is weak or absent in order to determine, first, if the plume is detectable under well-mixed conditions or if use of added tracer is necessary for detection, and, second, whether agreement between plume characteristics derived from in situ data and RSB model outputs is maintained. Validation of the RSB model or its potential modified versions will allow more effective diffuser design in the future, as well as the extension of predictability of plume behavior to a variety of conditions for which in situ data are not available.


Effects of a sewage plume on the biology, optical characteristics, and particle size distributions of coastal waters

A.A. Petrenko, B.H. Jones, T.D. Dickey, M. LeHaitre, and C. Moore
J. Geophys. Res., 102, 25,061-25,071, 1997

The particle field present in waters surrounding the Sand Island sewage outfall at Mamala Bay, HI during the period September 25 - October 1, 1994 was bio-optically characterized using in situ physical and bio-optical measurements and Niskin bottle sampling. The main components of the particle field were associated with shallow and deep phytoplankton, recently discharged ('new') sewage plume waters, and 'old' plume waters. Increased absorption and attenuation of light in combination with a reduced salinity signal characterized plume layers, while increases in absorption, attenuation, and chlorophyll fluorescence, with no change in salinity relative to surrounding waters characterized phytoplankton layers. Sewage plume waters exhibited higher scattering than phytoplankton layers. Both new and old plume layers, but not phytoplankton layers, showed distinct increases in fluorescence for the excitation/emission (Ex/Em) wavelength pair Ex/Em = 228/340 in nm. Throughout new plume waters, chlorophyll fluorescence exhibited either the same trend as outside the plume or a strong increase which was inversely correlated with the Froude number, showing the importance of physical forcing on effluent and phytoplankton interactions. New plume waters showed increases of at least 2.7-fold in particle load, high concentrations of particles larger than 70 10-6m, increases in ammonium, phosphate, and silicate, and high levels of prochlorophytes and heterotrophic bacteria, compared to surrounding waters. Characterizing the particle field allowed not only the mapping of the sewage plume but also showed that plume impacts on the surrounding waters were restricted to local effects during the present study.

Impacts of Sewage Plume on Microorganisms
The impact of the Sand Island sewage outfall on local microorganisms was estimated for the part of the plume detected with natural tracers. Chlorophyll fluorescence within the plume exceeded that in surrounding waters in conditions characterized by low Froude numbers (Fr about 0.1), while chlorophyll fluorescence within the plume was equivalent to that outside the plume when the Froude number was high(Fr about 6) (Petrenko et al., 1996); Froude number is defined as Fr = u3/B, where u is the horizontal velocity and B is the specific wastewater buoyancy flux per diffuser length (e.g., Roberts, 1977). If the change in chlorophyll fluorescence is assumed to reflect a change in phytoplankton concentration and growth, these observations are consistent with previous findings that sewage affects phytoplankton when the time scale of plume dilution is longer than phytoplankton growth response (Thompson and Ho, 1981). Enrichment studies, i.e. adding 24-hour composite sewage samples to water collected close to the diffuser and monitoring microorganism growth, have shown different results than measurements in situ (Laws and Ziemann, 1995). This highlights the difficulty of isolating cultures in bottles to try to reproduce natural conditions and the importance, in this case, of the physical forcing on the plume and phytoplankton interactions.
Flow cytometry and HPLC analyses showed increases of 30% in heterotrophic bacteria (HB) concentration and high ratios of prochlorophytes (PC) to total biomass not only in the plume but also at plume depths offshore. It is not clear whether the high counts of PC and HB at the 55 m offshore station were due to natural distributions of these populations in the water columns or were remnants of plume influence at that depth. The concentration of cyanobacteria (CB) may question the hypothesis of the plume influence offshore (since CB were high offshore at 55 m and low in the plume); unless variations in CB concentrations were due to factors other than the presence or absence of the plume. Horizontal currents, at the depths of the sewage plume and below, were southward earlier that day (during two periods, 12:00 - 2:00 am and 1:00 - 2:00 pm) and could have entrained the plume offshore (Jones et al., 1995).

Around sewage outfalls, flocculation and coagulation result in deposition of metals and dense particles, while flotables and grease rise to the surface. The sample taken in the plume had the volume proportion of medium size (45 - 70 10-6m) particles the closest to that of sewage samples from the treatment plant(1% vs. 2%). In the plume, there were more large particles (> 70 10-6m) than in background waters. Some large particles were present in sewage, while others probably originated from flocculation and coagulation of smaller effluent particles once the wastewater was injected into the water column (e.g., Newman et al., 1990; Johnson et al., 1994; Schulz et al., 1994). These large particles appeared to sink close to the diffuser since they were also detected below the plume at the western end of the diffuser. This could result in deposition of effluent particles and potential accumulation of metals and other contaminants in sediments close to the diffuser (e.g., Klein and Goldberg, 1970). During the cruise, no flotables or changes in surface appearance(grease slicks) were detected, probably because the stratification acted to trap the lighter effluent components at depth.

Only the main body of the plume was detected with salinity and c660. Pathogens have been shown to survive for long time periods especially if not exposed to the sun, as was likely the case here since the plume was deeply trapped (Petrenko et al., 1996). This component of the particle field cannot yet be detected in real-time; but beach sample analyses showed that fecal indicators at the sampled sites of Mamala Bay shorelines were not measurably increased by the outfall discharge (Fujioka and Loh, 1995).

During the period of the cruise, the plume contributed to local increases in heterotrophic bacteria concentration and in the ratio of prochlorophytes to total biomass, and to intermittent increases in chlorophyll fluorescence within the plume. Moreover, the sinking of large particles from the plume was thought to have impacts on sediments near the diffuser. Internal tides were shown to displace the plume vertically (A.A. Petrenko, unpublished data, 1996). Internal tides could move the deeply trapped(hence relatively low diluted) sewage plume to a shallower depth, potentially bringing human pathogens into swimming waters. In this case or in the case of surfacing plumes, biological impacts other than those observed in this study are expected.

Detection of the Plume
Scattering and beam attenuation coefficients and their respective spectral slopes were greater in the sewage plume than in the other components of the particle field. Single- scattering albedo also exhibited an increase in relative contribution of scattering in the plume. With absolute values of a, b and c, spectral algorithms for the various components of the particle field could be established for in-situ and remote detection. Increase in scattering, due to the presence of effluent, is responsible for non-negligible decreases in downwelling apparent optical properties (Petrenko et al., 1994) and could generate increases in water-leaving reflectance for shallow water outfalls.
During the cruise, the plume was trapped at depth, with smaller dilutions(about 1:250) than if it had been surfacing, and hence was detectable by its characteristic signatures of low salinity and high c660. In many cases, such signatures may be undetectable. For example, large gradients in background salinity or c660 could obscure the signature of the plume. A surfacing plume, which could be 5 times as dilute as a submerged plume, would be surrounded with low salinity surface waters, and potentially with high c660 associated with a shallow phytoplankton layer, so that its own weak signals would be further blurred by the background signals. In summary, salinity and c660 were good detectors of the main body of the plume during this cruise, but may not always serve as reliable plume tracers at the Sand Island outfall.

Fluorescence at Ex/Em = 228/340 nm may provide another unique signature for the sewage plume, since it did not correlate with phytoplankton fields. In laboratory experiments, fluorescence for this wavelength pair was associated with the presence of aromatic amino acids, either free or as protein constituents, so this Ex/Em signal is being referred to as protein-like fluorescence (Wolbeis, 1985) . For example, E. coli's tryptophan has been shown to exhibit strong signals at such a wavelength pair (Bronk and Reinisch, 1993). Wastewaters can contain numerous types of chromophores, including humic substances, phenols, steroids, oils, non-volatile acids, and detergents; so that their absorption and fluorescence spectra are likely to contain broad features rather than specific peaks. A monotonic decrease in absorbance from 210 to 350 nm for both raw and treated(at least secondary treatment) sewage sample has been observed previously (Thomas et al., 1993) . Nonetheless, more recent fluorescence spectroscopy studies on sewage samples have indicated sewage maximum absorption at 280 nm, maximum fluorescence in the range 280-340 nm, and, for excitation at 248 nm, maximum emission at about 440 nm (Ahmad et al., 1993; Ahmad and Reynolds, 1995; Reynolds and Ahmad, 1995). In the present study, Ex/Em = 228/340 was a better detector of sewage plume than Ex/Em = 228/365, 228/400, 228/430, 228/470, 265/340, 265/400, 265/430, and 265/470. Additional laboratory experiments on sewage samples from other locations should be pursued and field work with a calibrated SAFIRE should be done in order to define the wavelength pairs most appropriate for particular sewage water detection. Fluorescence at Ex/Em = 228/340 has been observed in marine waters by others (among the most recent works, Coble, 1996; Cowles et al., 1996; Desiderio et al., 1996) , and has been interpreted as recently formed DOC (P. G. Coble, unpublished data, 1996). Any living material containing tryptophan, such as E. coli or recent DOC, is likely to fluoresce at this wavelength pair; but the amplitude of the signal and/or signals at other wavelengths could probably be used to differentiate between fluorescing compounds.

It was paradoxical that Ex/Em = 435/700 detected the plume while the Sea Tech fluorometer(STF), which has neighboring wavelengths (Ex/Em = 425/685), did not exhibit any distinct increase in fluorescence within the plume. Since the emission FWHMs of SAFIRE(20 nm) and of STF (30 nm) were similar, the difference in fluorescence signals could be explained by 1) different optical configurations for the two instruments; 2) the dynamic range of the SAFIRE response allows more precise resolution of small signals; 3) STF broad excitation band (200 nm FWHM), in contrast with the narrow excitation band of the SAFIRE(20 nm FWHM), favored phytoplankton field detection compared to plume detection.

Detection of Dilute Wastewater
The problem with detecting old plume waters with salinity and c660 is that these factors, as explained above, mix with background signals and hence become unreliable for tracking a sewage plume away from the diffuser. Up to now, in situ and real-time detection of old plume waters has only been done by adding tracers (Faisst et al., 1990) or using acoustic methods (Dammann et al., 1991) . But tracer techniques are intrusive and costly methods, which involve intensive preparation(adjustment of tracer buoyancy and coagulation efficiency, dilution of tracer, coordination with cruise time), and acoustic methods can be imprecise due to background signals. In this study, the detection of old plume waters provided the opportunity to check whether instruments such as the SAFIRE, never used before in sewage plume studies, could detect the old sewage plume. Indeed Ex/Em = 228/340 was a good indicator of the old plume; it showed detection comparable to the c660 signal (Fig. 4.4). Once calibrated, Ex/Em = 228/340 fluorescence may prove to be a good tracer of old sewage plumes.

Effluents differ depending on their type of treatment, and, once in the water, on their dilution and the type of background waters they mix with. Further effort should be directed towards characterizing the optical variability associated with such factors in order to test the potentials of SAFIRE fluorescence measurements as a non-invasive real- time technique to detect sewage fields in coastal environments.

Select Figure

Figure 2
Example of physical and bio-optical measurements collected close to the Sand Island sewage diffuser.


Internal tidal effects on sewage plume near Sand Island, HI

A.A. Petrenko, B.H. Jones, T.D. Dickey, and P. Hamilton
Cont. Shelf Res., 20, 1-13, 2000

The sewage plume discharged from the Sand Island, Oahu, HI treatment plant diffuser (70 m depth) was mapped using ship-board profiles and towyo observations during September 25 - October 1, 1994. The plume was generally found west of the diffuser and trapped below 40 m, due to prevailing currents and density stratification. The present study focused on the position of the plume far from the diffuser. The sewage plume generally followed the vertical displacements of isothermal and isopycnal surfaces. Large semi-diurnal vertical oscillations of the isotherms (up to 40 m) occurred in the lower part of the water column when water column stratification was strong. The semi-diurnal baroclinic component of the horizontal currents exhibited vertical shear characteristic of internal tides. In addition to the forcing of this semi-diurnal internal tide, other superinertial motions may also have influenced vertical displacements of the plume. These large-scale vertical oscillations of relatively dense sewage plume water could cause plume contaminants to reach bottom sediments or to mix upward into the surface layer, with potential negative consequences for the environment and water quality.

Select Figures

Example of the plume being displaced downward by the internal tide (towyo 27, 09/28/94).

During towyo 27 (shaped deep to shallow from west to east), the isotherms are rising at the mooring D2 located close to the sewage diffuser. This indicates that the semi-diurnal internal tide is propagating westward.

Regardless of their origins, the vertical isopycnal (isothermal) oscillations have a strong influence on the plume's position within the water column. This vertical forcing is a component of plume dispersion models, only in as much as the plume position near the diffuser is an input in these models; but the process itself is generally not modeled (e.g. Blumberg and Connolly, 1995) . The omission of processes contributing to vertical isopycnal displacements may account for some of the discrepancies between model results and in situ observations of vertical plume locations. The plume could, if displaced downward, reach and contaminate sediments and benthic life or, if displaced upward and in combination with vigorous upper layer mixing, reach surface waters. Potential contamination of beaches could occur either through resuspension and upwelling of the sediments in the first case, or simply by entrainment by the currents in the second case. Future models of plume dynamics should include internal tide forcing mechanisms in order to provide more accurate predictions of sewage plume fates.