Environ. Sci. Technol. 2007, 41, 695-702
Size Distribution, Sources, and Seasonality of Suspended Particles in Southern California Marine Bathing Waters JONG HO AHN AND STANLEY B. GRANT* Department of Chemical Engineering and Materials Science, Henry Samueli School of Engineering, University of California, Irvine, California 92697
In this paper we define seasonal and along-shore variations in suspended particle size distributions (PSDs) at two marine bathing beaches in southern California, using a low-angle light scattering instrument (LISST). Empirical Orthogonal Function (EOF) analysis of the LISST data set (n ) 55 651) identified three particle size modes that collectively account for >90% of the variance in the de-meaned PSD data at six sites along the shoreline at Huntington Beach and Newport Beach: a dinoflagellate mode, a large particle mode, and a small particle mode. These three modes exhibit distinct seasonal patterns, and along-shore distributions, reflecting both the sources of particles and environmental factors that trigger their occurrence. Comparison of volume-based PSDs generated from the LISST and from image analysis of optical micrographs indicates that the LISST performs well when measuring the size distribution of particles associated with dinoflagellate blooms. However, LISST measurements on stormwater-impacted samples consistently yield a rising tail at small particle sizes that may be an artifact arising from the non-spherical nature of inorganic particles in terrestrial runoff. The results presented here demonstrate that PSDs measured by light scattering instruments such as the LISST represent a new data resource for assessing water quality, and managing human health risk, at marine bathing beaches.
Introduction Coastal marine bathing beaches represent an important economic and recreational resource in California, hosting upward of 400 million visits per year and providing state and local economic benefits running into the billions of dollars annually (1). At present, water quality at marine bathing beaches in California, and throughout most of the world, is evaluated using fecal indicator bacteria (FIB), a class of enteric bacteria that are not usually themselves pathogenic, but may indicate the presence of disease-causing organisms from sewage and runoff. The use of FIB as a measure of water quality is supported by epidemiological studies conducted over the last 30 years that demonstrate a dose-response relationship between swimming in marine bathing waters with high concentrations of FIB (in particular, high concentrations of enterococci bacteria) and increased incidence of gastrointestinal illness (2). The dose-response relationships * Corresponding author phone: (949) 824-7320; fax: (949) 8242541; e-mail:
[email protected] (949) 824-7320. 10.1021/es061960+ CCC: $37.00 Published on Web 12/29/2006
2007 American Chemical Society
were established in marine settings where the likely source of FIB was partially treated human sewage (3), and recently extended to marine beaches impacted by urban runoff with no known sewage inputs (4). The use of FIB as an indicator of marine bathing water quality and human health risk has a number of inherent drawbacks, including (1) the long turn-around-time associated with culture based assays of FIB (>24 h) frequently exceed the time over which FIB concentrations in marine waters change (200 µm) and small ( 0.7) and significantly (p < 0.01) correlated with particle volume concentrations in the 20-40 µm size range (red line in Figure 3B). These results,
together with the optical microscopy results presented above and field notes on the timing of bloom events, confirm that the sequence of peaks at 40 µm correspond to blooms of L. polyedrum. Peaks at the small end of the size spectrum (200 µm size range are not significantly correlated (p < 0.01) with either chlorophyll fluorescence or rainfall, suggesting an alternate (i.e., not bloom nor runoff) origin, and/or larger measurement error (see above), for these larger particles. EOF Analysis of the LISST PSDs. Particle Size Modes. The top-five particle size modes identified by the EOF analysis are presented in Figure S10. EOF results for data collected at the Balboa Pier are shown in Figure 4A. The top three EOF modes correspond closely to the three primary categories of particles described above: mode A corresponds to blooms VOL. 41, NO. 3, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
699
FIGURE 4. Top three EOF modes calculated from LISST PSD measurements on samples collected from the Balboa Pier. (A) Particle size modes, (B) Seasonal patterns of temporal eigenvectors.
TABLE 1. Percent of Variance Captured by the Top Three EOF Modes at Each Sampling Site
Na mode A mode B mode C a
Newport Bay Outlet
Balboa Pier
Newport Pier
Santa Ana River Outlet
Talbert Marsh Outlet
Huntington Pier
296 16 69 5
340 56 29 6
340 59 25 9
303 59 20 13
300 60 19 13
340 73 17 4
The number of samples analyzed.
of L. polyedrum, mode B signals the presence of large (>200 µm) particles, and mode C captures the rising tail at small particle sizes. These three modes are present, to a greater or lesser extent, at all sampling sites (see Figure S10), where they capture between 91 and 94% of the data variance. Higherorder EOF modes exhibit significant site-to-site variability, and appear to capture low-variance and site-specific features of the PSDs (Figure S10). Seasonal Patterns. Each EOF mode has an associated temporal eigenvector that indicates how the mode increases and decreases with time. Temporal eigenvectors for the top five modes at each sampling site are presented in Figures S11-S16. As illustrated in Figure 4B, the top three modes at the Balboa Pier exhibit distinct seasonal patterns. The bloom mode (mode A) is most pronounced during the summer season, from June through October. The large particle mode (mode B) is relatively weak during the summer and more pronounced from November through May. The rising tail mode (mode C) peaks during storms in January and February. Mode A follows a nearly identical seasonal pattern at all 700
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 3, 2007
sampling sites (Figure S17). Seasonal patterns for modes B and C are also fairly consistent across sampling sites, although some site-to-site variability is evident (Figures S18 and S19). Spatial (Along-Shore) Trends. The amount of variance captured by the top three modes exhibit distinct alongshore trends. These alongshore trends appear to indicate where particles originate along the shore. For example, the variance captured by the bloom mode (mode A) decreases in a downcoast direction, from 73% at the Huntington Pier to 16% at the Newport Bay Outlet (Table 1), consistent with probability distributions of chlorophyll fluorescence that also exhibit a declining down-coast trend (Figure S20A). The variance captured by the large particle mode (mode B), on the other hand, increases in a down-coast direction, from 17% at the Huntington Pier to 69% at the Newport Bay Outlet (Table 1). The underlying causes for the observed spatial gradients in modes A and B are not clear, although a few possibilities should be mentioned. The decreasing southward trend associated with the phytoplankton mode (mode A) may be caused by changes in the phytoplankton production rate
and dominant transport mechanisms associated with the southward narrowing of the coastal shelf at our field site, from greater than 10 km at the northernmost station to less than 2 km at the southernmost station (see bathymetric contours in Figure 1) (34). The tidal transport of large biological debris out of Newport Bay is probably responsible for the dominance of mode B at the Newport Bay Outlet. Newport Bay is the second largest tidal embayment in California, and includes a large pleasure craft marina (and associated commercial activities) and a protected tidal saltwater marsh with seagrass beds, unvegetated tidal flats, and a saltmarsh dominated by Spartina foliosa and Salicornia virgnica (35). The amount of variance captured by the small particle mode (mode C) is highest at the Santa Ana River and Talbert Marsh outlets (approximately 13% of the variance at each site, Table 1), suggesting that this mode is associated with stormwater runoff from these two outlets. Probability distributions of salinity at the Santa Ana River and Talbert marsh sites have long low-salinity tails (Figure S20B), reflecting the impact of stormwater runoff on ocean water collected from these two sites. Probability distributions of all FIB measurements collected during this study (Figure S21) reveal that FIB concentrations were also consistently higher at the Santa Ana River and Talbert marsh sampling sites. Correlation between FIB and LISST Measurements. At the Santa Ana River and Talbert Marsh outlets, the concentrations of all three FIB groups (TC, EC, and ENT) are significantly (p < 0.05) and moderately (Sp ) 0.2-0.45) correlated with LISST measurements of the volume concentration of small particles (
