A Mass Budget Analysis of Fecal Indicator Bacteria ... - ACS Publications

Mar 24, 2004 - of the mass budget analysis and its application to measure- ments of fecal indicator bacteria in the surf zone at. Huntington State Bea...
5 downloads 0 Views 719KB Size
Environ. Sci. Technol. 2004, 38, 2626-2636

Locating Sources of Surf Zone Pollution: A Mass Budget Analysis of Fecal Indicator Bacteria at Huntington Beach, California J O O N H A K I M , † S T A N L E Y B . G R A N T , * ,† CHARLES D. MCGEE,‡ BRETT F. SANDERS,| AND JOHN L. LARGIER§ Department of Chemical Engineering and Materials Science and Department of Civil and Environmental Engineering, Henry Samueli School of Engineering, University of California, Irvine, California 92697, Orange County Sanitation District, 10844 Ellis Avenue, Fountain Valley, California 92708, and Scripps Institution of Oceanography, University of CaliforniasSan Diego, 8602 La Jolla Shores Drive, La Jolla, California 93027

The surf zone is the unique environment where ocean meets land and a place of critical ecological, economic, and recreational importance. In the United States, this natural resource is increasingly off-limits to the public due to elevated concentrations of fecal indicator bacteria and other contaminants, the sources of which are often unknown. In this paper, we describe an approach for calculating mass budgets of pollutants in the surf zone from shoreline monitoring data. The analysis reveals that fecal indicator bacteria pollution in the surf zone at several contiguous beaches in Orange County, California, originates from welldefined locations along the shore, including the tidal outlets of the Santa Ana River and Talbert Marsh. Fecal pollution flows into the ocean from the Santa Ana River and Talbert Marsh outlets during ebb tides and from there is transported parallel to the shoreline by wave-driven surf zone currents and/or offshore tidal currents, frequently contaminating >5 km of the surf zone. The methodology developed here for locating and quantifying sources of surf zone pollution should be applicable to a wide array of contaminants and coastal settings.

Introduction The coastal ocean is both a critical natural resource and a final repository for all manner of human waste. The latter inexorably diminishes the former, as evidenced by a wide spectrum of coastal ills, including frequent postings and closures of popular swimming beaches (1-3). Coastal pollution often comes to light during the course of routine surf zone monitoring programs, in which samples are periodically collected from a series of monitoring stations and analyzed for one or more pollutants. In this paper, we * Corresponding author e-mail: [email protected]; phone: (949)824-7320; fax: (949)824-2541. † Department of Chemical Engineering and Materials Science, University of CaliforniasIrvine. ‡ Orange County Sanitation District. | Department of Civil and Environmental Engineering, University of CaliforniasIrvine. § University of CaliforniasSan Diego. 2626

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 9, 2004

demonstrate that the loading of pollutants into the surf zone (rate of input in pollutant mass per time) can be computed from a mass budget analysis of surf zone monitoring data, thereby providing critical information about the location, magnitude, and timing of specific inputs of pollution to the coastal ocean. The mass budget analysis described here is complementary to chemical and microbiological pollutant source tracking methodologies (4), for which the goal is typically to characterize the relative contributions of specific types of fecal pollution; for example, from human, bird, cow, etc. The approach is also complementary to other modeling approaches; for example, forward modeling of pollutant fate and transport in the ocean (5-7) and surf zone (8-10). The latter approaches rely on a microscopic mass balance modeling of pollutant fate and transport in which one or more differential equations are derived and solved (either exactly or numerically) subject to suitable initial and boundary conditions. In contrast, the approach described in this paper relies on a macroscopic mass balance, in which application of the Reynolds Transport Theorem leads to algebraic equations that can be employed to calculate the loading of pollutants in to and out of the surf zone. The paper is organized as follows. Nearshore currents affecting the transport and mixing of pollutants in the surf zone are reviewed, followed by observations of fecal indicator bacteria pollutant transport in the surf zone at Huntington State Beach. These observations motivate the development of the mass budget analysis and its application to measurements of fecal indicator bacteria in the surf zone at Huntington State Beach. To make the results and analysis accessible to a broad audience, each section begins with a question around which the section is focused together with an answer supported by the data and modeling.

Nearshore Currents Question: What are the dominant nearshore currents that affect the fate and transport of pollutants introduced into the surf zone? Answer: Pollutants entrained in the surf zone are carried parallel to shore by wind and wave-driven currents and crossshore by rip currents, vertically stratified shear flows, and offshore forced currents (e.g., internal waves). Just offshore of the surf zone, pollutants are transported by coastal currents that respond to tides, winds, and remote forcing. The combination of longshore and cross-shore advective transport enhances the longshore spreading of nearshore pollution by dispersion. The region of the surf zone impacted by coastal pollution can be influenced by the tidal phasing of pollutant input to the surf zone (e.g., from tidal outlets along the shoreline), the tidal phasing of coastal currents, and the prevailing wave climate. The fate and transport of pollutants in the surf zone is controlled by a set of highly dynamic currents that determine longshore transport and exchange with offshore. These can be broadly classified as surf zone currents, which are primarily wave-driven, and coastal currents offshore of the wave-driven surf zone, which are primarily driven by tides, winds, and remote forcing (Figure 1). Wave-driven surf zone currents flow parallel to the shoreline (so-called “longshore drift”) in a direction controlled by the angle at which waves approach the shore. In the simple case of a straight sandy beach bordering a uniformly sloping submarine shelf, the surf zone current takes on the direction of the approaching deepwater waves (11, 12). At beaches bordering more complex bathymetry (e.g., with submarine headlands and canyons), refraction of the deep-water waves can generate surf zone 10.1021/es034831r CCC: $27.50

 2004 American Chemical Society Published on Web 03/24/2004

FIGURE 1. Photograph of the Huntington State Beach surf zone taken July 15, 2002, from the Huntington Beach pier (surf zone station 18N, see map in Figure 2), looking to the southeast. The dashed box in the photograph represents the top of the imaginary surf zone prism illustrated at the bottom of the figure. The points a and b in the surf zone prism correspond to hypothetical surf zone monitoring stations (indicated by the solid blue circles) where water samples are collected periodically from ankle-depth water. Note that these two blue circles do not correspond to actual surf zone sampling stations at Huntington Beach. The variables indicated in the surf zone prism are described in the text. currents that converge and diverge at different locations along the beach (13). In the photograph of the Huntington State Beach surf zone in Figure 1, for example, waves approaching from the south strike the beach at a roughly 35-40° angle and generate a net upcoast (toward northwest) flow in the surf zone, as represented by the large arrow oriented parallel to the shore (the orientation of the Huntington State Beach shoreline is indicated in Figure 2). Within the surf zone, the currents are highly variable in time because of tidal effects, infragravity waves, longshore shear instability, effects of coastal currents, and wind effects (14). These effects also induce spatial structure to surf zone flow, in addition to the spatial patterns arising from rip current circulation (see below) and interactions of flow with shoreline morphology. For the straight and smooth approximately planar Huntington Beach shoreline, longshore drift is generally in the direction of wave forcing with rip current spacing at scales of 1001000 m (15). Coastal currents (indicated by the arrow on the right side of the photograph in Figure 1) are often uncoupled from the wave-driven currents in the surf zone (16). The direction and magnitude of coastal currents are controlled by prevailing winds, tides, and regional-scale circulation (17). Off Huntington State Beach, for example, the coastal currents are oriented mostly parallel to shore, with upcoast flow occurring primarily during rising tides and downcoast flow occurring primarily during falling tides. Based on velocity measurements at station B, located just outside of the surf zone at Huntington State Beach (see Figure 2), these two patterns (i.e., upcoast directed coastal currents during rising tides and downcoast directed coastal currents during falling tides) collectively account for two-thirds of current observations (see Figure S1 in Supporting Information). At Huntington State Beach, peak longshore tidal velocities are in the range 0.2-0.3m/s during spring tides, with implied

longshore tidal excursions of 4-6 km. The phasing of the longshore tidal transport is such that water flowing into the ocean from a lagoon or river during an ebb tide will be transported upcoast (i.e., to the northwest) during the following flood tide. Cross-shore currents connect the wave-driven surf zone with the coastal waters offshore of the surf zone. These currents may be associated with surf zone or offshore forcing (e.g., rip currents and internal waves, respectively (15)) and can be structured vertically (e.g., with near-surface onshore flow and near-bottom offshore flow) or horizontally in the form of feeder and rip currents (18). Cross-shore currents facilitate the export of pollutants from the surf zone and the dilution of pollutants that remain in the surf zone; consequently, the dilution is scaled by the vigor of the cross-shore exchange. When combined with wave-driven longshore currents, cross-shore currents exert a major control on the longshore spreading and mixing of nearshore pollutants by dispersion (19). In the following sections, we describe observations and modeling of fecal pollution in the surf zone at Huntington State Beach, Huntington City Beach, and Newport Beach.

Observations of Surf Zone Pollution at Huntington State Beach Question: How are fecal indicator bacteria in the surf zone at Huntington State Beach (and adjacent beaches) distributed in space and time? Are fecal indicator bacteria consistently higher during specific phases of the tides or in certain regions of the surf zone? Are these spatial and temporal patterns the same for all three groups of fecal indicator bacteria, including total coliform (TC), Eschericia coli (EC), and Enterococci bacteria (ENT)? Answer: The occurrence of fecal indicator bacteria in the surf zone is highly variable in space and time. Two different VOL. 38, NO. 9, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2627

FIGURE 2. Map of the field area showing the location of surf zone monitoring stations (solid blue circles), which are designated in thousands of feet north or south of the Santa Ana River outlet; e.g., surf zone stations 15N and 15S are 15 000 ft north and south, respectively, of the Santa Ana River outlet. Also shown are several known point sources of fecal indicator bacteria including a wastewater outfall, a power plant thermal outfall, and the tidal outlets of the Santa Ana River and the Talbert Marsh (between surf zone stations 0 and 3S and 0 and 3N, respectively). The location of an Acoustic Doppler Current Profiler (ADCP) is also shown (station B, red filled circle). The map was downloaded from the U.S. Geological Survey website (http://www.usgs.gov) and modified. patterns can be identified. The concentration of TC is highest during the transition from the nighttime falling to rising tide and mostly confined to the region of the surf zone upcoast of the Santa Ana River and Talbert Marsh outlets. This pattern suggests that TC originates in ebb flow from the two tidal outlets and, once entrained in the surf zone, is carried upcoast by wave-driven surf zone currents and/or tidally driven coastal currents. The concentration of EC and ENT, on the other hand, is highest in a region 2 km upcoast of the Santa Ana River and Talbert Marsh tidal outlets and exhibits a more complex tidal signature. Possible explanations for the different occurrence patterns of TC, on one hand and EC and ENT on the other hand include the following: (1) The existence of multiple shoreline sources of fecal pollution, each with different relative abundance of TC, EC, and ENT and/or (2) differential fate and transport of the different fecal indicator bacteria groups in the ocean. Although these two occurrence patterns appear to be reproducible, there is significant study-to-study variability, and the data were acquired over a limited number of environmental conditions (e.g., all studies were conducted during spring tides and during summer dry weather). Hence, additional occurrence patterns cannot be ruled out based on the existing data. Description of Field Site. In this study, we investigated surf zone pollution at a set of contiguous beaches (including Newport Beach, Huntington State Beach, and Huntington City Beach, see Figure 2) located along a NW-SE striking section of the Pacific coastline, between Los Angeles and 2628

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 9, 2004

San Diego, in Orange County, California. Several areas of Huntington State Beach have suffered chronic beach postings and closures over the past several years due to elevated concentrations of fecal indicator bacteria in the surf zone (20). This beach is also very popular (more than 5 million visitors per year), and the combination of surf zone pollution and significant beach usage implies that a large number of people (perhaps as many as 50 000) may acquire highly credible gastroenteritis from swimming and surfing in this area each year (21). It appears that the fecal indicator bacteria pollution at Huntington State Beach originates from a combination of sources, including bird droppings deposited in the Talbert Marsh, regrowth of bacteria on vegetation and marsh sediments, and dry and wet weather runoff from the surrounding community (22, 23). Additional potential sources of fecal indicator bacteria include the offshore discharge of partially treated sewage effluent (24), the offshore discharge of power plant cooling water that contains fecal indicator bacteria from plant wash-down and other activities (25), bather shedding (26), the accumulation of bird droppings along the shoreline and offshore (27), the exfiltration of sewage-contaminated groundwater (28), and contributions from watershed outlets located north and south of the study area including the Los Angeles River, the San Gabriel River, and outlets for Huntington Harbor and Newport Bay (29). The thermal boil generated by the power plant cooling water discharge also provides a potential mechanism for transporting contaminants from offshore into the surf zone, as documented in a separate study (28). Several of these potential sources of fecal indicator bacteria pollution (including the wastewater treatment outfall and the power plant cooling water outfall) are noted on the map in Figure 2. Methods. To better constrain sources of fecal pollution in the surf zone at our study site, four separate field experiments were carried out during the summer of 2001. During each experiment, water samples were collected every hour for 48 consecutive hours from 11 different surf zone monitoring stations (528 samples per study for a total of 2132 samples). The locations of the surf zone monitoring stations are indicated on the map in Figure 2 (blue circles). In this figure, the stations are designated according to their distance (in thousands of feet) north or south of the Santa Ana River outlet (e.g., 15 N is located approximately 15 000 ft, approximately 5 km, north of the Santa Ana River outlet). Each station was manned by a separate team of samplers who collected the surf zone samples (100 mL total volume) every hour from ankle-depth water on an incoming wave. The water samples were immediately placed on ice and transported to the Orange County Sanitation District laboratory (Fountain Valley, CA) where they were analyzed within 6 h of collection for total coliform (TC), E. coli (EC, a subset of fecal coliform), and Enterococci bacteria (ENT) using defined substrate tests known commercially as Colilert-18 and Enterolert (IDEXX, Westbrook, MN) implemented in a 97-well quanti-tray format. The four studies were carried out on the following dates in 2001: June 19-21, July 5-7, July 19-21, and August 19-21. The timing of the studies coincided with spring tides, when surf zone water quality at Huntington State Beach is frequently impaired (30); all studies were conducted during dry weather periods. Additional special surf zone studies were conducted during the summer of 2001, as described elsewhere (28). These additional studies are not included here because of their relatively short duration (99% of the fecal pollution shed from the local watershed is delivered to the surf zone during storm events. Furthermore, a recent study of historical shoreline water quality data reported that fecal indicator bacteria concentrations in the surf zone in Orange County are strongly correlated with storm flow from nearby watershed outlets (35). Collectively, these other reports suggest that rainstorms can significantly affect nearshore water quality, quite possibly in a manner that is qualitatively and quantitatively different than observed in the dry weather studies plotted in Figure 3. As a final point, we note that individual fecal indicator bacteria events usually persist at a specific shoreline station for no more than 6 h (the nominal time of a single ebb or flood tide, see Figure 3). Hence, any monitoring program with a sampling interval longer than this time period will not detect the underlying fecal indicator bacteria signal (i.e., the underlying fecal indicator bacteria signal will be aliased by the sampling program). By way of comparison, sampling frequencies for typical shoreline monitoring programs vary from once per day to once per week. In California and elsewhere, warning signs are posted on the beach if the concentration of fecal indicator bacteria in a single sample exceeds state standards. As described in a companion paper (20), the highly variable water quality signals together with relatively infrequent sampling schedules result in the situation where water quality warning signs are frequently posted in error.

Surf Zone Mass Budget Analysis Question: Can a mass budget analysis be formulated to estimate the loading of fecal indicator bacteria into the surf zone from offshore and/or onshore sources and estimate the export of pollution from the surf zone by cross-shore exchange and other processes (e.g., coagulation/sedimentation)? Answer: An algebraic equation for calculating loading (and export) rates can be formulated from a surf zone mass budget analysis, but its application to real systems requires assumptions regarding the magnitude of advective and turbulent diffusive longshore transport rates, assumptions about the nature of die-off kinetics, and relatively high-frequency and spatially intensive measurements of pollutant concentrations in the surf zone. Uncertainty in the loading estimates can be calculated from the uncertainty of key variables appearing in the loading equation. Given the highly variable nature of water quality signals documented in the last section it is easy to understand how sources of coastal pollution might be difficult to find and mitigate. In this section, we derive an algebraic equation that can be used, in conjunction with surf zone monitoring data, to define the location and magnitude of pollution sources in the surf zone. The derivation begins with a mass budget analysis over the imaginary surf zone prism included below the picture in Figure 1. The surf zone prism is oriented such that points a and b coincide precisely with two adjacent surf zone monitoring stations located a distance L apart. Because surf zone samples are typically collected from ankledepth water, the entire surf zone prism will move up the beach during rising tides and down the beach during falling tides. However, the prism is fixed relative to the y-coordinate (i.e., the surf zone prism does not move upcoast or downcoast). Pollutants can enter and leave the surf zone prism by several different pathways. Pollutants can traverse: (1) faces

add′a′ and bcc′b′ by wave-driven surf zone currents and turbulent mixing; (2) face abb′a′ by, for example, tidal exchange at an estuary outlet; (3) face dcc′d′ by cross-shore currents and turbulent mixing; (4) face a′b′c′d′ by exchange with the sediment bed; and (5) face abcd by exchange with the atmosphere. Conservation of mass requires that accumulation of pollutant mass inside the surf zone prism must equal the net flux of pollutants across all boundaries of the surf zone prism plus any loss of pollutant inside the surf zone prism by reaction, as expressed quantitatively through the Reynolds Transport Theorem (36):

∫ C(x,t) dV ) -∫ q(x,t)‚n dA - ∫ R(x,t) dV

∂ ∂t

V

A

V

(1)

The variables in eq 1 represent time (t), spatial coordinate (x), pollutant concentration (C(x,t), units of mass per volume), pollutant flux vector (q, units of mass per area per time), pollutant reaction rate (R(x,t), units of mass per volume per time), and the outward facing unit vector oriented normal to the surface of the surf zone prism (n, unitless). The integrals in eq 1 are evaluated over the volume (V) or surface area (A) of the surf zone prism. The flux of pollutants across the surface of the surf zone prism (first term on the right-hand side, RHS, of eq 1) can be divided into three parts: the part that crosses face add′a′, the part that crosses face bcc′b′, and the part S(t) crossing all other faces (i.e., pollutant pathways 2-5 above):

∫ q(x,t)‚n dA ) ∫

-

A

add′a′

qy(x,t) dA -



q (x,t) bcc′b′ y

dA + S(t) (2)

where qy represents the component of the pollutant flux oriented parallel to the shoreline. The magnitude and sign of the variable S(t) provides critical information about the nature of contaminant loading into the surf zone prism from onshore and offshore sources of pollution. Specifically, S(t) is positive if the rate of pollutant addition by pathways 2-5 exceeds loss by rip current export or other processes; conversely, S(t) is negative if pollutant loss exceeds input from pathways 2-5. Combining eqs 1 and 2 and solving for S(t), we arrive at the following expression for pollutant loading into the surf zone in terms of flux across boundaries add′a′ and bcc′b′ (first and second terms on RHS) and accumulation and reaction of pollutant in the surf zone prism (third and fourth terms on RHS):

S(t) )



q (x,t) bcc′b′ y

dA -



add′a′

∂ ∂t

qy(x,t) dA +

∫ C(x,t) dV + ∫ R(x,t) dV V

V

TABLE 1. Relative Uncertainties of Variables Used To Estimate Fecal Indicator Bacteria Loading Rates quantity, Xi

relative uncertainty, UR (Xi)

evaluation of uncertainty by

I C kFIB L w ∆t 〈vy〉

0.05 0.05 0.10 0.05 0.10 0.10 0.10

manufacturer’s specification manufacturer’s specification ref 42 personal communication scientific judgment personal communication scientific judgment

station (located at y ) 0 and L) together with an estimate for the longshore surf zone current (averaged over the cross section add′a′ and bcc′b′)〈vy〉, the fecal indicator bacteria dieoff rate constant kFIB, and hourly measurements of sunlight intensity I(t). The result is the following approximate expression for pollutant loading (see Supporting Information):

S(t) ≈

[

Lwh 1 (C(0,t + ∆t) + C(L,t + ∆t) - C(0,t) - C(L,t)) + 2 2∆t 〈vy〉(C(L,t) - C(0,t)) C(0,t) + C(L,t) (4) + kFIBI(t) L 2

]

where L, w, and h represent the length, width, and timeaveraged water depth at the ocean-ward edge of the surf zone prism (see Figure 1), and ∆t represents the time interval between fecal indicator bacteria concentration measurements. As detailed in the Supporting Information, eq 4 is only valid in cases where the pollutant flux across faces add′a′ and bcc′b′ is dominated by longshore advection. A more complex expression is included in the Supporting Information for cases when both advection and turbulent diffusion contribute to the longshore flux of pollutants. For the range of die-off rates utilized and solar insolation values measured, the T90 (i.e., time for 90% die-off) range from no die-off (at night when I(t) ) 0) to 1.3, 1.4, and 2.4 h (for TC, EC, and ENT, respectively) when the insolation values are maximal (I(t) ) 1000 W m-2). It is important to consider how estimates of the loading rate (i.e., the magnitude and sign of S(t)) will be affected by uncertainty in the magnitude of variables that appear on the RHS of eq 4. The variance in S(t) (denoted here as u(S)) can be related to the variance of any variable Xi (denoted u(Xi)) appearing on the RHS of eq 4 through the Law of Propagation of Uncertainty (37), also known as a First-Order Uncertainty Analysis (38-40): N

(3)

Apart from the conceptualization of surf zone transport illustrated schematically in Figure 1, no significant approximations have been employed in deriving eq 3. However, to estimate the magnitude of terms appearing on the RHS side of this last equation will require, in practice, assumptions regarding the mathematical formulation of each term. Clearly, a tradeoff exists between how well the transport processes at a particular site are characterized by field measurements and the number and nature of assumptions required to translate the field measurements into estimates for the magnitude of terms on the RHS of eq 3. In our case, we employed a set of physically reasonable assumptions that yield estimates for the terms on the RHS of eq 3 based only on hourly measurements of fecal indicator bacteria concentration C(y,t) in the surf zone at two adjacent shoreline

u2(S) )

[]

∑u (X ) ∂X

∂S

2

i

i)1

2

+

i

N-1 N

2

∑∑

u(Xi)u(Xj)r(Xi,Xj)

i)1 j)i+1

[ ][ ] ∂S

∂S

∂Xi ∂Xj

(5)

where Xi represents the set of N ) 10 variables {L, w, ∆t, 〈vy〉, kFIB, I, C(0,t), C(L,t), C(0,t + ∆t), C(L,t + ∆t)} and r(Xi,Xj) represents the correlation coefficient for any two variables Xi and Xj. The variance associated with each variable Xi was estimated as follows:

u(Xi) ) UR(Xi)|Xi|

(6)

where UR(Xi) represents the relative variance of the variable Xi and |Xi| represents the magnitude of the variable in question (which may vary with time) (41). The relative variances UR(Xi) VOL. 38, NO. 9, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2631

were estimated from manufacturers specifications (for the variables I and C), previously published estimates (for the variable kFIB) (42), and scientific judgment or personal communication (for the variables L, w, ∆t, and 〈vy〉) (37) (see Table 1). Application of eqs 4 and 5 to surf zone monitoring data yields estimates for the magnitude, sign, and variance of the loading of pollutants in to or out of a given region of the surf zone. For practical purposes, we assumed that any loading estimate with a relative variance of UR(S) < 0.317 was meaningful; this choice of relative variance is equivalent to saying that, within a confidence level of 68% (one standard deviation), the true value of the loading lies within the range S ( 0.317|S| (37).

Test of the Surf Zone Mass Budget Analysis Question: Can the loading rate equation derived in the last section accurately predict the loading of fecal pollution into the surf zone from specific shoreline sources? Answer: Loading rates estimated from eq 4 closely track direct measurements of the loading of TC, EC, and ENT into the surf zone from the Talbert Marsh outlet and the loading of TC from the Santa Ana River outlet. The comparison is not favorable for the loading of EC and ENT from the Santa Ana River outlet, perhaps because there were other sources (i.e., other than the Santa Ana River) for the last two fecal indicator bacteria groups during this particular field experiment. As a first test of our methodology for calculating pollutant budgets for the surf zone, we compared fecal indicator bacteria loading rates calculated from surf zone monitoring data (eq 4) with independent estimates of the load of fecal indicator bacteria flowing into the surf zone from the Santa Ana River and Talbert Marsh outlets (eq 7):

Soutlet(t) ) Coutlet(t)Qoutlet(t)

(7)

In this last equation, Soutlet(t) represents the loading (in bacteria per time) flowing into the surf zone from either the Santa Ana River and Talbert Marsh outlets during ebb tides (S > 0) or flowing into the outlets from the surf zone during flood tides (S < 0) (the flow at the Santa Ana River and Talbert Marsh outlets is tidally forced during dry weather periods). The variables Coutlet(t) and Qoutlet(t) represent hourly measurements of the fecal indicator bacteria concentration and volumetric flow rate, respectively, at the Santa Ana River or Talbert Marsh outlets. Details of how these variables were estimated can be found in the Supporting Information. Direct comparison of loading rates estimated from eqs 4 and 7 was only possible during the second surf zone study (July 5-7, 2001) because this was the only study when hourly measurements were available for the concentration of fecal indicator bacteria at all surf zone stations and in the outlets of the Santa Ana River and Talbert Marsh. In addition, estimates for the volumetric flow rate (i.e., Qoutlet(t)) of water flowing in to and out of the Santa Ana River and Talbert Marsh outlets were measured only during this single study. TC loading rates estimated from the surf zone monitoring data generally agree with independent measurements of TC flowing into the surf zone from the two watershed outlets (compare dashed and solid lines in the TC graphs in panels B and C, Figure 4). TC loading rates peak around 0.5-1 trillion MPN/h during falling tides when water from the Talbert Marsh and Santa Ana River outlets flows into the ocean (compare loading rates with tide level curves in panel A); in some cases, the loading of TC from the Santa Ana River extends into the next flood tide because the change in flow direction at the outlets (from ebb to flood) sometimes lags the change from falling to rising tide (compare velocity measurement at the outlet of the Santa Ana River in panel B to water level curves in panel A). It is noteworthy that the loading of TC into the surf zone from the two outlets often 2632

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 9, 2004

peak at the end of the ebb tide or slightly into the next flood tide, when the coastal current is often switching to an upcoast direction (see Figure S1). The observation that peak pollutant input from the Santa Ana River and Talbert Marsh outlets immediately precedes upcoast-directed coastal currents may explain why much of the TC pollution appears in the surf zone upcoast of the Santa Ana River and Talbert Marsh outlets, even during periods of weak westerly waves when the surf zone current might be directed downcoast (see discussion of TC occurrence patterns). In any case, the results presented in Figure 4 are consistent with the idea that TC loading in the surf zone between stations 3S and 3N is dominated by ebb flow from the Santa Ana River and Talbert Marsh outlets. Furthermore, the comparability of the two loading estimates (i.e., from eqs 4 and 7) lends credibility to the methodology described here for calculating pollutant loadings from surf zone monitoring data (i.e., eq 4). The loading of EC and ENT calculated from the Talbert Marsh outlet data (eq 7) also agrees fairly well with the loading of EC and ENT calculated from the surf zone monitoring data at stations 0 and 3N (eq 4), particularly during the second large ebb event (panel C in Figure 4). However, the loading of EC and ENT calculated from the Santa Ana River outlet data does not correlate with EC and ENT loading calculated from the surf zone monitoring data at stations 3S and 0 (panel B in Figure 4). Indeed, on at least two occasions, ENT loading calculated from eq 7 is negative (S < 0) while the ENT loading calculated from eq 4 is positive (S > 0), suggesting that ENT was flowing into the surf zone from some other (i.e., not Santa Ana River) source. Other potential sources of surf zone pollution are explored in the next section.

Shoreline Sources and Tidal Phasing of Fecal Bacteria Loading Question: Can the loading equation (eq 4) derived in this study be used to identify the location and tidal phasing of sources of fecal indicator bacteria in the surf zone? Answer: When applied to fecal bacteria monitoring data collected during the first two studies (June 19-21 and July 5-7, 2001), the loading equation indicates that TC loading in the surf zone is greatest around the mouths of the Santa Ana River and Talbert Marsh outlets and flows into the surf zone primarily during the nighttime falling tides. While these outlets also contribute EC and ENT, a larger source of these latter fecal indicator bacteria groups is located approximately 2 km upcoast of the outlets, around surf zone stations 6N9N. This last observation is consistent with an earlier sanitary survey, which implicated subsurface sewage collection pipes as a source of sewage contamination in the surf zone around surf zone station 6N-9N. As a next step in our analysis, we used the loading equation derived in the last section (eq 4) to estimate the loading of fecal indicator bacteria into a 10-km stretch of the surf zone (from 15S to 21N surf zone stations) centered around the Santa Ana River and Talbert Marsh outlets. The loading estimates were calculated from measurements of fecal indicator bacteria in the surf zone (as described earlier) together with estimates for specific parameters (e.g., longshore drift velocity) as described in the Supporting Information for this paper. Of the four surf zone studies conducted, only the first two (June 19-21 and July 5-7, 2001) coincided with wave conditions (i.e., southerly swells) that were likely to yield upcoast-directed longshore drift in the surf zone. Furthermore, during these first two studies, TC originating from the Santa Ana River and Talbert Marsh outlets appears to be transported in an upcoast direction at a fairly constant rate (implied longshore drift of approximately 0.3 m/s, see black lines in the two upper left panels in Figure 3). Indeed, the tilt of the TC events can be used to estimate the longshore

FIGURE 4. (A) Tide level and solar insolation curves for the 48-h study conducted from July 5 to July 7, 2001. (B) Velocity measurements recorded at the Santa Ana River outlet and fecal indicator bacteria loading rates calculated from the surf zone monitoring data at surf zone stations 3S and 0 (dashed lines, eq 4) and from measurements of fecal indicator bacteria and flow in the Santa Ana River outlet (solid lines, eq 7). (C) Velocity measurements recorded at the Talbert Marsh outlet, and fecal indicator bacteria loading rates calculated from the surf zone monitoring data at surf zone stations 0 and 3N (dashed lines, eq 4) and from measurements of the fecal indicator bacteria and flow in the Talbert Marsh outlet (solid lines, eq 7). Pink and blue vertical stripes denote falling and rising tides, respectively. The abbreviations TM and SAR represent Talbert Marsh and Santa Ana River, respectively. drift velocity, which is difficult to assess by any other method (43). Because the longshore drift was best constrained during these first two studies, they were chosen for the loading analysis presented below. The color contour plots in Figure 5 depict the spatial and temporal distribution of fecal indicator bacteria loading rates calculated from the surf zone monitoring data using eq 4. The color in these plots ranges from red for large positive loading rates to blue for large negative loading rates. For the purposes of the discussion below, we refer to positive S(t) values as fecal indicator bacteria source events and negative S(t) values as fecal indicator bacteria removal events. Many

of the fecal indicator bacteria source events depicted in Figure 5 (as indicated by the blotches of red color) are paired with fecal indicator bacteria removal events (blotches of blue color). In the figure we placed arrows on source/removal pairs for which the loading estimates had relative uncertainties UR(S) < 0.317. The tilt of the arrows (relative to vertical) indicates the longshore transport velocity 〈vy〉 estimated from the tilt of the TC events in Figure 3; the length of the arrows in Figure 5 has no significance. For several of the source/ removal pairs with relatively low uncertainty (i.e., those marked by an arrow), the removal events appear just upcoast of the source events and are displaced by the travel time VOL. 38, NO. 9, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2633

FIGURE 5. Spatial distribution of fecal indicator bacteria loads calculated from hourly surf zone monitoring data on two separate occasions (both 2001): June 19-21 and July 5-7. Regions of the surf zone that are a net source of fecal indicator bacteria appear red (S > 0), and regions that are a net sink of fecal indicator bacteria (e.g., by rip current export) appear blue (S < 0). Arrows indicate source/removal pairs for which the loading estimates had relative uncertainties UR(S) < 0.317. The orientation of the arrow represents the long-shore drift velocity 〈vy〉 estimated from the tilt of the TC events (see Figure 3). The abbreviations DR, NF, NR, and DF correspond to the beginning of daytime rising, nighttime falling, nighttime rising, and daytime falling tides, respectively. associated with the wave-driven surf zone current (i.e., the source/removal pairs are oriented parallel to the arrows). Our interpretation of these source/removal patterns is that fecal indicator bacteria enter the surf zone at relatively welldefined locations along the shore (over shoreline distances of