CHAPTER 4. SUBTIDAL CIRCULATION PATHWAYS

Home

Chapter 1. Introduction and Background

Chapter 2. Methodology

Chapter 3. Surfzone Bacteria Patterns

Chapter 4. Subtidal Circulation Pathways

Chapter 5. Newport Canyon Transport Pathway

Chapter 6. Sea Breeze

Chapter 7. Tidal Transport Pathways

Chapter 8. Sediment Resuspension and Transport near the OCSD Outfall

Chapter 9. Nearshore Circulation and Transport Pathways

Chapter 10. Spatial and Temporal Patterns of Plume Tracers

Chapter 11. Temporal and Spatial Patterns for Surfzone Bacteria before and after Disinfection of the OCSD Outfall

Acknowledgements

4.1.1. Historical Flow Patterns

Previous studies have shown that the subtidal flow on the San Pedro shelf is dominated by sustained strong alongshore currents that are unrelated to the local winds (Hamilton et al., 2001). Characteristic periods range from about 10 to 30 days with downcoast (i.e., towards San Diego) and upcoast currents strongly sheared vertically and more depth independent, respectively. In the summer of 1999, downcoast flows dominated surface currents on the outer shelf near the outfall, where the vertical shears often allow upcoast flows in the lower water column. Thus, the plume can be dispersed in both the up- and downcoast directions with similar probability. On the 15-m isobath, the 1999 study showed that the flows were similar to the outer shelf, but had less magnitude. It was concluded that the subtidal flows were remotely forced, probably by the current regime over the San Pedro slope. The origin of the strong downcoast flows in summer and at other times of the year is not known, and is opposite to the view that the slope regions of the southern California bight are dominated by the poleward flows of the Southern California Eddy, the recirculation limb of the southward flowing California Current system (Hickey, 1992).

The present study allows the investigation of the patterns of current flow from the outer shelf to nearshore regions of the San Pedro Shelf with coverage of the middle part of the shelf that has not hitherto been available. Besides the dominant influence of the low-frequency currents on the dispersion of outfall plume, the flow regime controls the distribution of the temperature and salinity fields across the shelf that in turn modulate the higher frequency internal tide and sea-breeze-forced flows. Subtidal changes in nearshore temperatures are directly related to the distribution of low-frequency flows across the width of the shelf.

This section discusses the basic statistical characteristics of the subtidal current and temperature fields, the geostrophic balance of the alongshore currents, and the time-variable changes in the flow patterns through empirical orthogonal function (EOF) analysis. The ability of the subtidal current field to transport material across the shelf is evaluated.

4.1.2. Basic Statistics

Basic statistics for the velocity and temperature 40-hour low-pass-filtered (40-HLP) time series were calculated for the longest common period that accommodated most of the nearshore and main-array moorings (Figures 4-1 and 4-2). Calculations were made for means of the along- and cross-shore velocities and the standard deviations along the major and minor principal axes directions. The directions, relative to true north, of the major principal axes were also calculated. The results are given as vertical section plots for the main and south transects in Figures 4-3 and 4-4, respectively. The most important features are the strong downcoast, surface-intensified, alongshore flows. Mooring HB06 has the highest near-surface means and variances in the alongshore direction. These means and variances decrease with depth and towards shallow water. Corresponding to the mean downcoast flows, the mean isotherms slope up towards the coast, which is consistent with geostrophic balance of the alongshore vertical shears with the cross-shelf density gradients. The cross-shore velocity means suggest onshore flow at depth with offshore flow in the surface layer. These mean cross-shore flows, however, have large uncertainties because the alongshore currents have much larger magnitudes than the cross-shore, and, thus, small changes in the direction of the axis can have large effects on the calculated values. The minor axis standard deviations show the same pattern as those for the major axis, with decreases towards the coast and with depth. The direction of the principal axes shows a relatively rapid change between upper and lower layers where the directions are quite constant. This is exaggerated in the plots because the axis chosen for the direction switches quadrants. However, even if the same quadrant is used there is a still a significant change across the layers of between 10 and 30°. The section plots suggest that if the thin layer where the axis direction changes represents the separation between upper and lower layers, then the lower layer is below about 20 to 30 m, but extends into the nearshore where it is very thin. As will be discussed in subsequent sections, this is consistent with the behavior of the higher frequency components of the circulation.

The main difference in the south transect statistics (Figure 4-4) is that, compared to the main transect, there is more deep, upcoast mean flow at the outer shelf moorings (HB11 and HB10) than at the main transect (HB07 and HB06; Figure 4-3). This is reflected in the downslope of the lower-layer mean 12 and 13°C isotherms towards the coast. Poleward mean flow is clearly observed over the slope (HB08) below about 50 m (Figure 4-5). The largest poleward flows occur at about 100 to 120 m depth, which is below the depth of the shelf break (~60 m). These deep currents are part of the poleward undercurrent that is a persistent feature of slope circulation along the U.S. west coast. These poleward flows have a seasonal signal and can surface in the winter and spring (Hickey, 1992). Short-term (~10 days) surfacing of the poleward flow occurs in the summer of 2001, causing upcoast flow at all depths over the middle and outer shelf.

Another view of the subtidal statistics is given by the mean current and standard deviation ellipses for the near-surface and lower layers (Figure 4-6). The near-surface means are generally comparable or greater than the along-shelf fluctuations that show the persistence and strength of the downcoast currents. The directions of the near-surface major axes and mean current vectors tend to parallel the coast and show little influence of the changing isobath directions on the outer shelf. The lower-layer means, on the other hand, are less than the major axis standard deviations, except at 100-m depth over the slope. The means remain downcoast in the nearshore and become upcoast seaward of the 30-m isobath. As noted before, the lower-layer upcoast means are larger, east of the main transect where the isobaths turn eastward and the shelf width narrows. Thus, it appears that in the upper part of the slope, poleward undercurrent overruns the outer part of the shelf after it encounters the slope to the west side of the Newport Canyon. Variances in the lower layer are considerably smaller than in the near-surface layer (Figure 4-6). The map view of the change in principal axis direction between the upper and lower layers is given in Figure 4-7. The degree of rotation increases from inner to outer shelf, and the rotations are larger on the south than the main transect. The lower-layer current fluctuations on the middle and outer shelf follow the trend of the isobaths, indicating that topography is controlling the flows, while the upper-layers are more parallel to the coastline. The anticlockwise rotation of the axes, with depth, is consistent with Ekman turning, and up- and down-welling cross-shelf flows, forced by alongshore down- and upcoast current fluctuations, respectively. This is even true in the shallow water where the isobaths and the coastline are in the same direction (HBN2 and HBN3).

4.1.3. The Influence of Alongshore Currents on Nearshore Cold Water Intrusions

The persistent alongshore downcoast currents are consistent with the upsloping of isotherms towards the coast through geostrophy or the thermal wind relation (Pond and Pickard, 1978). Thus, the raising of the isotherms near the shore and the consequent occurrence of colder water should be closely related to the strength of the downcoast currents. Similarly, the lowering of isotherms on the inner shelf and the occurrence of warmer water deeper in the water column should accompany upcoast current events. To test the degree that the thermal wind relation holds, the cross-shelf temperature differences are compared with alongshore currents between pairs of moorings on the main transect. Figure 4-8 shows the temperature gradients at mid-depth between alternate moorings and the alongshore velocities at moorings in between for middle shelf. There is a clear relation between the velocity and temperature difference curves as the correlation coefficients (R) confirm. Particularly noteworthy is the correspondence between the positive temperature differences (warmer inshore) and the reversal of the currents to upcoast.

At HB03 (Figure 4-9), the correlation between the temperature difference from HB01 to HB05 and the near-surface currents is even higher than offshore. The current fluctuations at HB03 are relatively stronger at shorter periods than further offshore and the fluctuations are not always similar to those at moorings 5 and 6. Similarly, there are differences in the temperature gradients between HB01 and HB05 and those further offshore. This indicates the temperature gradients are not uniform across the shelf, and the inner shelf has some differences in response when compared to the middle and outer shelf. This will influence the occurrence of colder water in the nearshore regions, and will be explored in more detail in the next section. At the shallowest, 10-m mooring (HBNC), the currents are clearly similar to the flows at HB03 on the 15-m isobath, but they have no relation to the local temperature gradients between HBN6 and HB01. The distance between these two stations is only 0.5 km and near-surface temperatures were used. Therefore, the calculations could be less precise than with moorings further apart. However, it seems that the thermal wind balance breaks down in the nearshore, probably because of mixing and boundary layer effects.

4.1.4. Characteristic Circulation Patterns

To find the dominant spatial and temporal patterns in the subtidal currents, complex empirical orthogonal function (CEOF) analysis was used for the longest common period. CEOF analysis does not separate out the U (cross-shelf) and V (along-shelf) components of the velocity vectors so that they are not separately weighted in the covariance matrix. This accounts for the large differences in cross- and alongshelf variances without discounting their connections. This can occur in EOF analyses where the U and V components are treated as separate variables. The approach in this study follows the analysis methods used in Munchow and Chant (2000). The mean vectors were removed from each velocity record on all available moorings. Because the moorings had different degrees of vertical resolution, the number of velocity positions that were used at each ADCP mooring was made proportional to the number of whole degree isotherms present in the mean temperature profile at each site (Figures 4-3 and 4-4). This resulted in using velocity bins at ~4-m and ~2-m intervals on the outer and inner shelf moorings, respectively. Thus, for the CEOF analysis, the number of velocity positions used at each mooring was roughly weighted according to water depth and stratification. The moorings are fairly evenly spaced so no weighting by the area that an individual record represents was used. As in Munchow and Chant (2000), the spatial eigenvectors show the variance accounted for at each position by the mode. In CEOF analysis the direction of mode eigenvector is arbitrary, therefore to facilitate physical interpretation, the eigenvectors are rotated into the principal axis coordinates of the corresponding complex amplitude time series (Merrifield and Winant, 1989). The rotated amplitude time series were also normalized to unit variance.

The mean vectors and the complex eigenvector components for the first two modes are given in Figures 4-10 and 4-11, respectively. Because the eigenvector components represent the fluctuations about the means, approximately 50% of the time the spatial patterns will be reversed. Therefore, adding and subtracting the mode eigenvectors to the means can be used to construct characteristic flow patterns. However, since the means are strongly downcoast, upcoast flows occur for only relatively short periods over most of the shelf. The mean vectors show the same patterns as discussed above with the poleward undercurrent present at depth on the slope and outer shelf (Figure 4-10). Mode 1 fluctuations are all in the same direction and fairly uniform across the shelf and slope waters above ~60 m. There is some decrease in magnitude towards the shore and with depth. Thus, this mode has a similar pattern to the distribution of major axis velocity variance in Figures 4-3 and 4-4, and indicates that up- and downcoast flows, relative to the mean currents, occur quite uniformly across the shelf and the upper slope above the shelf break. The second significant mode shows maximum response at the inshore moorings HBN2, HBN3, HB03, and HB13. At the 10-m isobath moorings, the second mode response has larger amplitude than the first mode, and since the modes are uncorrelated, this indicates that nearshore is dominated by fluctuations of a different character to the middle and outer shelf. The second mode also has a significant response in the opposite direction to the nearshore at deeper depths, extending into the core of the undercurrent, over the slope. This seems to indicate that fluctuations of the undercurrent, which are largely uncorrelated with the shallower layers over the slope and outer shelf, are coherent with nearshore fluctuations of the opposite sign. Therefore, increasing poleward flow at depth over the slope corresponds to increasing downcoast flows in shallow water near the coast. Dynamical explanations of this phenomenon have not yet been made. Similar connections between oppositely directed fluctuations at the outer shelf and the nearshore (15-m isobath) were found by Hamilton et al. (2001) in the analysis of the Phase II current data (June 1999 to June 2000). In that study, the reversed flow pattern was not so prominent, because no very shallow moorings were deployed.

The common period is from the end of July to the beginning of October. The coldest water was present in the nearshore in July, and, therefore, another CEOF analysis of the flow patterns that included the July period was performed. This long-period CEOF analysis excluded the slope mooring (HB08) and the nearshore SIO moorings (HBN2 and HBN3), replacing them with the 10-m AES ADCP mooring at HBNC. The mean currents for the long period are given in Figure 4-10 and only show minor differences with the shorter common period means for the same moorings. The first two modes, from the long-period analysis, account for similar percentages of the total variances as those from the common period (Figure 4-12). The first mode has the same pattern as Figure 4-11, with the vectors all in the same direction and a decrease in the amplitudes with depth and towards the nearshore. The second mode, again, has its largest amplitudes in the nearshore (HBNC and HB03) and indications of a correlation with reversed vectors on the outer shelf (HB07). This is very similar to mode 2 for the common period, and indicates that this mode is a robust feature of the circulation. In Figure 4-13, the component amplitudes and coherence squared of the velocities with the mode are shown for two moorings extracted from the complete analysis. The latter is equivalent to the fractional variance explained by a mode for a particular measurement position. According to Figure 4-12, HB06 and HBNC are strongly dominated by modes 1 and 2, respectively. The coherence squared in Figure 4-13 also shows that more than half the observed variance at the nearshore mooring HBNC is explained by mode 2, and nearly all the observed variance at HB06 is explained by mode 1. This implies that the nearshore current fluctuations are somewhat decoupled from the circulations on the middle and outer shelf (Figures 4-8 and 4-9). This also implies that the behavior of the amplitude time series of the two modes can be used to characterize different regimes in the cross-shelf current and temperature fields that occurred during the summer 2001 experiment.

The normalized amplitude time series of the modes from the long-period analysis are given in Figure 4-14. Mode 1 is characterized by relatively long periods of sustained flows with both up and downcoast fluctuations, whilst mode 2 has generally shorter periods. This figure compares the mode 1 time series with the 5-m depth velocities (mean removed) at HB06, which are almost totally explained by this mode. It can be seen that the observed fluctuations are well characterized by this mode. Similarly, mode 2 is compared with the 2-m depth velocities at HBNC, and again this mode accounts for most of the observed fluctuations. Since the modes are uncorrelated, the distinction between the nearshore subtidal currents and the majority of the flows over the shelf appears to be valid.

Hamilton et al. (2001) and Hickey (1992) could find very little evidence of relationship of the alongshore shelf currents to the local wind in San Pedro and Santa Monica Bays. At periods longer than a day, local winds at the coast tend to be weak and variable. This study deployed a meteorological buoy at the shelf break (HB07) that showed persistent downcoast winds modulated by a strong sea-breeze signal (Chapter 6). The kinetic energy spectra, in variance preserving form, of the wind from the meteorological buoy and the normalized complex mode amplitudes (Figure 4-15) show that mode 2 and the wind have a similar peak at about 7 to 9 day periods. The major peak for mode 1 is at about 15 days where the wind has little energy. The difference in the characteristic periods of the mode fluctuations is visually apparent in Figure 4-14. The coherence squared of the wind alongshore component with the V components of the modes (Figure 4-15) show that mode 1 is not significantly correlated, but mode 2 is correlated at periods longer than 5 days with a lag of about 12 to 18 hours. Thus, it appears that the nearshore currents, associated with mode 2, are forced by local alongshore winds over the shelf, but the majority of the subtidal flow fluctuations over the middle and outer shelf are unrelated to the local winds. Unfortunately, there are no reliable wind measurements in the nearshore zone, so it is not known how the winds at the beach differ from those at the shelf break.

Figure 4-14 also shows time series of the estimated offshore position, along the main transect, of the 14°C isotherm where it intersects the bottom. The 14°C isotherm characterizes the lower thermocline, and is associated with the equilibrium depth of the top of the plume, immediately after discharge from the outfall. The large shoreward excursions of this isotherm in July are accompanied by strong (mode 1) downcoast flows. Similarly, the cold event at the end of September was preceded by strong downcoast flow. When mode 1 switches to upcoast at the beginning of August, the isotherm moves offshore, consistent with thermal wind and the leveling and depression of the isotherms. Therefore, based on the directions of the fluctuating mode amplitudes, four periods have been selected (Table 4-1) where the density field across the shelf could be expected to have different average characteristics. These periods, separated by the vertical lines in Figure 4-14, have been selected with tidal and sea-breeze analysis in mind, because the characteristics of the internal tide and sea-breeze-forced flows are modified by changing low-frequency density fields.

The main transect mean temperature fields for each period in Table 4-1 are given in Figure 4-16. These fields should be examined along with the current mode time series in Figure 4-14. Thus, in period 1, enhanced downcoast flows over most of the shelf result in up-sloped isotherms over most of the transect. In period 2, at the beginning of August, the upcoast mode 1 is consistent with the leveling of the isotherms and increased stratification above the thermocline. The down-slope of the deeper isotherms shows that there is upcoast flow at depth (Figure 4-8). In the latter half of August (period 3), mode 1 is weakly downcoast, or upcoast, towards the end of the period, and mode 2 is upcoast. This results in a mean temperature field that has fairly level isotherms. The thermocline has deepened and surface layer has become less stratified, presumably because of larger scale seasonal changes. In the last period, both modes 1 and 2 are downcoast, and therefore reinforce each other. This results in the mean temperature field having the largest up-slope of the isotherms of any of the previous periods. The surface layer in the offshore waters remains only weakly stratified.

Large-scale changes in the temperature distribution across the shelf, over periods of two to four weeks, have been qualitatively related to the behavior of the current modes. It might be expected that the temperature fluctuations can be closely related to the velocity-mode time series. However, the offshore distance of the 14°C isotherm, plotted in Figure 4-14, is not significantly correlated with either velocity mode, and this suggests that more than one process is influencing this signal. To investigate this further, a time-domain EOF analysis was performed just using the 40-HLP temperature records (with means removed) from the main transect. The record lengths were exactly the same as for the CEOF analysis of the velocities. The amplitudes of the eigenvectors of the first two modes are given in Figure 4-17. The first mode amplitudes are all positive with maxima at the depths of the thermocline (10 to 25 m). This can be interpreted as changes in the depth of the thermocline that are imposed on the shelf by the larger southern California Bight circulation processes. This effect is amplified across the shelf so that the largest amplitudes of this mode are in the nearshore at stations with between 10- and 20-m water depth. Mode 2 amplitudes are of both signs with the node at about 10-m depth. This can be interpreted as being caused by tilting of the isotherms, such that when temperatures increase below 20 m in the offshore, there is a corresponding decrease in the nearshore and in the surface layer (and vice versa). Maximum amplitudes of both temperature modes occur in the nearshore.

The time series of the amplitudes of the temperature modes are given in the top panel of Figure 4-14. It can be seen that mode 2 is closely related to the velocity mode 1, such that downcoast flows correspond to positive temperature amplitudes, and therefore, increasing tilts to the isotherms. The mode 1 temperature amplitudes seem to represent the external effects on the shelf, and it can be seen that the time series of the offshore distance of the 14°C isotherm is a combination of both modes. Thus, mode 1 is responsible for the general warming through the month of August. The correlation (R) of the mode 2 temperatures with the velocity mode 1 time series is -0.78, and shows the inverse relation of the tilting isotherm mode with the primary mode of the middle and outer shelf velocity fluctuations that is dynamically consistent with the thermal wind relation. There are no other significant correlations of the temperature and velocity modes, and this implies that the nearshore velocity mode 2 has little influence on appearance of cold water at the coast, though it is clear that there are times when mode 2 can reinforce mode 1. For example, in period 4, the reinforcing downcoast modes produce stronger isotherm tilts. This does not apparently apply throughout the summer. The maximum amplitudes of the mode 2 currents also occur close to shore where thermal wind breaks down (Figure 4-9), and, therefore, it should not be expected that mode 2 velocities be closely related to cross-shelf temperature gradients.

4.1.5. Discussion

The analysis of the spatial patterns and time evolution of the current and temperature modes have produced some new information on the dominant modes of variability of the subtidal circulation on the San Pedro shelf. Important features are the separation of the alongshore flow into uncorrelated shelf and nearshore modes. The nearshore currents are closely related to the alongshore wind fluctuations over the shelf. The majority of the shelf current fluctuations are not directly forced by local winds but appear to be part of larger-scale California Bight circulation processes that are not well characterized. The shelf mode is correlated with the up and down tilts of the isotherms through thermal wind, but the temperature field also is strongly controlled by external processes that impose isotherm depth changes across the width of the shelf. Both these processes can move colder water into the nearshore.

Cross-shelf transport processes appear to be complex. Changing positions of isotherms implies cross-shelf circulations arising from changes in the alongshore currents through “geostrophic adjustment” and current-induced upwelling. The former is likely transitory, and the latter, because of the downcoast mean flow, is likely to result in very small cross-shelf transports. The limited current measurements in the bottom boundary layer generally show small means and weak subtidal fluctuations, and there is little evidence of cross-shelf flows in a distinct bottom Ekman layer. The on- and offshore excursions of the 14°C isotherm (Figure 4-14) show that the maximum displacement over a day is about 2 km, which results in a cross-shelf transport velocity of 2.3 cm/s. This kind of transport is difficult to extract from the current measurements when the alongshore currents have magnitudes of 10 to 20 cm/s. An attempt was made to calculate the cross-shelf transport velocity in the lower layer, using the area under the 14°C isotherm on the main transect, using:

U = 1/h * dA/dt (4.1)

where A = the cross-sectional area under the 14°C isotherm;

h = the lower layer depth between 14°C and the bottom at mooring HB07;

U = the cross-shelf transport between 14°C and the bottom.

This assumes that 14°C is an impermeable surface and that the cross-shelf transport is uniform in the alongshore direction. The calculated U was compared with the same quantity estimated from the ADCP at HB07, but no significant correlation was found for several estimations using different coordinate axes for the observations (e.g., the minor principal axis, isobath coordinates, etc.). This seems to indicate that cross-shelf subtidal transport is a three-dimensional process.

In the nearshore, the wind-forced (mode 2) currents imply onshore transport in the lower layer and offshore in the upper layer for downcoast directed winds. However, because of the shallow depths (~<10 m), the top and bottom Ekman layers merge and the cross-shelf transports are suppressed. Only very small transport velocities (<1 cm/s) are expected from this process. Therefore, subtidal cross-shelf transport is very difficult to quantify, and, as will be discussed in later sections, sea-breeze- and internal-tide-forced cross-shelf transports have larger magnitudes and may be more effective at short-term movement of material across the shelf. The best guide to subtidal cross-shelf transport appears to be the on- and offshore excursions of the isotherms.

4.2. Typical Diffusion Time on the Shelf

Surveys of the outfall plume (Chapter 10) have shown that the distribution of material can be quite patchy with areas of relatively higher concentrations interspersed with lower concentration regions. This is a quite common occurrence for pollutant discharges into the ocean and results from stirring by the sheared currents and turbulent mixing (Eckart, 1948). The time- and space-dependent current field stretches and distorts the plume and may separate the plume into patches. Turbulent mixing processes act most effectively to smooth large concentration gradients created by the current shears. Eckart’s famous analogy was of stirring cream into a coffee cup. It is useful to make rough estimates of how long such a large separated patch or “rogue blob” may be expected to exist with concentrations substantially above background. This rough estimate of the dispersal time assumes that a section of the plume with the typical widths and heights of the established plume has become detached from the main discharge.

It is assumed that the initial cloud is a patch of plume water with volume = p l02d/4, where

l0 ~ 10 m is the horizontal diameter, and d ~ 5 m is the depth of the cloud. Typical maximum concentrations of sewage, after initial mixing by the outfall discharges, are about 200 mg/L. Background concentrations of particles in the coastal ocean, away from bottom boundary layers, are generally less than 10 mg/L. Assuming that a dilution of 50:1 is required and vertical diffusion is negligible, then the diameter that this dilution achieves is given by L2 = 50 l02 , thus

L ~ 70 m. The resulting standard deviation, s = L/3 ~ 23 m, and the relation between s and the elapsed diffusion time, t, is (Csanady, 1973):

s2 = 4 K t

Representative values of K from the Okubo diffusion diagram or dye diffusion studies in coastal waters, for 10 to 100 m horizontal scales, are 0.01 to 0.1 m2/s (102 - 103 cm2/s). This gives a diffusion time, t ~ 220 to 22 minutes (i.e., between 0.5 and 4 hr).

A dilution of 500:1 gives a range of diffusion times of 4 to 40 hours. The shorter time is probably more relevant because K increases with length scale. Therefore, a large rogue blob of the plume would be expected to disperse to background levels in less than 6 hours or about half a tidal cycle.

4.3. Transport Patterns: Summer 1999 versus 2001

From June 1999 to June 2000, SAIC deployed and maintained four current meter moorings and a number of bottom-mounted thermistors on the San Pedro shelf in the vicinity of the outfall. Three conventional moorings, with two current meters and two thermistors, were deployed on the 15-m isobath between the Newport Canyon and the approximate position of HB03. The offshore mooring was in the same position as HB12 and consisted of a real-time surface mooring with current measurements at 1- and 45-m depth and temperature measurements at 5 m depth, and an associated bottom mounted ADCP. These measurements have been discussed in Hamilton et al. (2001). Comparison of these measurements with the present experiment allows a limited assessment of the inter-annual variability of the circulation. The summer of 1999 was noted for extensive bacterial contamination (Chapter 3), and, therefore, it is of interest to see if the circulation differed substantially from that of the summer of 2001.

The 40-HLP current vectors and temperature records from instruments in essentially the same positions on the 60-m isobath (moorings P and HB12) and on the 15-m isobath (moorings Q and HB03 in 1999 and 2001, respectively) are shown in Figure 4-18. In 2001, the outer shelf currents are stronger than in 1999, but with fewer major upcoast reversals. In the upper layer, the temperatures are generally cooler in 1999. From the middle of August through September, the outer-shelf upper layer was substantially cooler in 1999 compared to 2001. On the other hand, the bottom waters were generally warmer and this indicates that the outer shelf was less stratified, particularly in August and September, in 1999 than in 2001. In the nearshore, comparing the two years, mooring Q, at 10-m depth, was warmer in early July, but was substantially cooler (similar to P) in the second half of the summer. Thus, for most of the summer, there was less of a cross-shelf temperature gradient with weaker up-tilts of the isotherms towards shore than in 2001. This is consistent with the relatively weaker fluctuating flows at Q in 1999, than at HB03 in 2001. Therefore, despite being relatively colder in July 1999, the 14°C isotherm does not penetrate quite as far into the coast as in July 2001. This is shown in the top panel of Figure 4-18 where the position of the isotherm in 1999 is estimated using the 15-m mooring (R) and the bottom thermistors (U, V, W, X and P) deployed along the cross-shelf path of the outfall pipe. Resolution, above 5-m depth and nearshore, is not as good as in 2001. In August and September 1999, the cooler water offshore, relative to the warmer water in 2001, does allow the 14°C isotherm to penetrate much closer to the coast than in the same period in 2001, and, thus, sub-thermocline water would have had a more direct connection with the nearshore, similar to the earlier summer.

Basic statistics of the time series presented in Figure 4-18 are given in Table 4-2. The same length 99-day, 40-HLP time series are used for all the statistics. The statistics for 1999 are given in red, and the U and V velocity components are across- and along-shelf, respectively. At the surface, the 2001 data show much larger downcoast mean flows with temperatures 1.5 to 2°C warmer than in 1999. In the lower layer, the 1999 data show higher variances with temperatures about 0.5°C warmer than in 2001. Therefore, at plume depths, the dispersion of material should be greater in 1999 than in 2001. In the nearshore, the larger means and higher variances of the currents in 2001 are quite clear. Since the analysis of Chapter 6 indicates that on the 15-m isobath the flows are partially accounted for by local winds, it is possible that offshore downcoast winds were weaker in 1999 compared to 2001. Unfortunately, wind measurements were not taken on the outer shelf in the summer of 1999, and this speculation cannot be confirmed. Table 4-2 also shows the statistics of the depths below the surface of the top and middle of the plume, calculated from the RSB initial dilution model for outfall diffusers. The middle of the plume corresponds to the depth of minimum dilution. A description of the use of this model for the 1999 period is given in SAIC (2002), and the model was applied in exactly the same way for the summer of 2001. The mean and minimum depths are about 2 and 5 m higher, respectively, in the water column in 1999 as compared to 2001. Since the currents at plume depths are more energetic in the earlier period, the higher rise heights are primarily caused by weaker stratification below the thermocline.

The summer of 1999 had colder water, less energetic currents on the inner shelf, was less stratified, and had more energetic flows at plume depth near the outfall, when compared to the summer of 2001. The consequences for pollutant dispersion from the outfall were that, in 2001 the plume was generally higher in the water column, and was probably more dispersed by the offshore flows in 1999 versus 2001. The generally colder conditions in 1999 means that sub-thermocline water was more often in contact with the nearshore, and any plume material reaching the 15-m isobath would be less likely to be flushed because of the weaker alongshore currents.

4.4. References

Csanady, G.T., 1973. Turbulent diffusion in the environment. Reidel, Holland, 248 p.

Eckart, C., 1948. An analysis of stirring and mixing processes in incompressible fluids. Journal of Marine Research, v. 7, p. 265-275.

Hamilton, P., J.J. Singer, E. Waddell, and G. Robertson, 2001. Circulation processes on the San Pedro Shelf. Proceeding, MTS 2001 Conference, Nov. 2001, Honolulu, Hawaii, 8 p.

Hickey, B.M., 1992. Circulation over the Santa Monica-San Pedro basin and shelf. Pro. Oceanography, v. 30, p. 37-115.

Merrifield, M.A. and C.D. Winant, 1989. Shelf circulation in the Gulf of California: A description of the variability. Journal of Geophysical Research, v. 94, p. 18133-18160.

Munchow, A. and R.J. Chant, 2000. Kinematics of inner shelf motions during the summer stratified season off New Jersey. Journal of Physical Oceanography, v. 30, p. 247-268.

Pond, S. and G.L. Pickard, 1978. Introductory dynamic oceanography. Pergamon Press, Oxford, 241 p.

SAIC, 2002. Strategic Process Study #1: Plume tracking–plume modeling. Final Report prepared for Orange County Sanitation District (OCSD), Fountain Valley, CA.