Monitoring Southern California Water Quality Using Synthetic Aperture Radar-NASA DEVELOP

The Concern

 

The water quality of the coastal ocean and the safety of beaches for public use are potentially threatened by pollution from urban runoff. Urban stormwater runoff is currently the most significant pollution hazard for coastal waters in the Southern California Bight (SCB) coastal region, and the greatest single source of water pollution in the country [1].  In the SCB, more than 95% of the annual runoff volume and pollutant load comes from episodic storm events, typically in the fall and spring. Stormwater runoff rates in Los Angeles are increasing due to the growing population and proliferation of impervious surfaces [2]. Stormwater plumes in the Santa Monica Bay (SMB) have elevated toxicity and nutrient concentrations, leading to increased primary productivity and the potential for harmful algal blooms [3]. Epidemiological evidence of adverse symptoms in swimmers exposed to stormwater runoffs in the SMB confirmed the public health hazards posed by these coastal pollution sources [4].

 

The city of Los Angeles is dedicating over $100 million of Proposition O funds to improve the safety and cleanliness of beaches, and the city of Santa Monica is seeking to minimize urban runoff into the bay (Santa Monica Office of Sustainability and the Environment ). 

 

Currently Los Angeles County health officials routinely issue a public warning to avoid recreational water contact for three days following a storm larger than 2.5 millimeter (mm) [5]. The warnings, based principally on precipitation data from the Los Angeles International Airport (LAX), result in the closure of beaches for the entire county and may not reflect extremely local conditions. The DEVELOP team had two goals: to provide an innovative approach to managing beach closures by prototyping Synthetic Aperture Radar (SAR) products to (a) facilitate the identification of stormwater plumes at specific beaches and (b) provide estimated Enterococci levels within the plumes, given that Enterococci level is the most frequently failed water quality standard at beaches. 

 

The Science

 

During the fall of 2012, NASA DEVELOP at the Jet Propulsion Laboratory partnered with the Southern California Waters Research Project (SCCWRP) and Heal the Bay Santa Monica to examine stormwater plumes in the Southern California Bight (Figure 1). The team combined beach contamination data, provided by Heal the Bay, with historical SAR imagery (ENVISAT, UAVSAR, ALOS, ERS-1, and ERS-2) from the rainy seasons’ fall and winter months between 1992 and 2011 in Ballona Creek and Los Angeles Harbor to investigate the capabilities of SAR to detect individual stormwater plumes and serve as a proxy for the bacterium Enterococci.

 

Figure 1. Southern California Bight (SCB)

 

SAR was the preferred remote sensing tool (as opposed to optical products) because of its ability to detect backscatter regardless of cloud cover conditions, allowing imagery to be acquired even during precipitation events.  SAR detected stormwater plumes as reduced backscatter areas due to the reduction of surface roughness by surfactants associated with the plumes [1]. Plume extent and size were determined from these reduced backscatter areas.

 

Before analysis of the plumes, the raw SAR imagery underwent preliminary processing that included radio calibration and georectification in Next ESA SAR Toolbox (NEST). The goal of the image analysis was to identify areas with potential runoff plumes, determine the extent and direction, and calculate statistics such as plume area and mean backscatter intensity. Low backscatter areas, such as in plumes, typically have negative values. Areas of low backscatter connected and emanating from a mouth of a coastal watershed were identified as stormwater runoff plumes. Regions of interest (ROI) were then identified, and NEST was used to compute statistics, such as mean, maximum, minimum, and median backscatter intensity, in units of decibel (dB). Plume backscatters were compared to the SAR backscatter in an adjacent patch of clean water, which served as the control.

 

To measure and analyze the extent and distribution of the plumes, the ROIs were converted from vector to raster in ArcMap, and a weighted sum overlay was performed.  The output of that analysis was a raster image with values representing the number of plumes that covered a certain area. The weighted sum overlay was produced for both Los Angeles Harbor and Ballona Creek to visualize the frequency and distribution of the plumes.

 

Time series of the all the plumes show the different behavior of the plumes in the two regions. For Ballona Creek, the largest plumes occur less frequently, while the most frequent plumes are concentrated near the waterway’s mouth (Figure 2). The effects of the breakwater split the flow of the plumes north and south, with 55% of the total plume areas moving north and 44% south. In the LA Harbor, the breakwater has the effect of trapping most plumes within the harbor, with occasional, rarer spillage farther offshore in the bay (Figure 3).

 

Figure 2. Frequency and distribution of nine Ballona Creek plumes, ranging from 1992-2011. Maximum plume extent offshore for each percent class is measured in kilometers.

 

 

Figure 3. Frequency and distribution of 11 plumes in the Los Angeles Harbor from 1992-2011. Plumes are from both the Los Angeles River and San Gabriel River, which flow into the harbor.

 

Our analysis found that SAR was able to detect significant plume signatures in almost all plumes examined (Figure 4). A clear relationship between Enterococci and plume backscatter is required in order to develop a SAR proxy for Enterococci. However, the data were not conclusive. The DEVELOP team’s hypothesis was that higher Enterococci concentrations were expected to correspond with lower (more negative) plume backscatter values, since lower backscatter values characterize more intense plumes and are related to the concentration of organic oils in the plume that reduce surface roughness. No clear trend was found in the data. In the LA Harbor (Figure 5) most of the plumes had unexpectedly low Enterococci values. In fact, only one of the plumes exceeded the enterococcus standard: the highest amounts of Enterococci corresponded with a mean plume backscatter of approximately -14 dB, which was the second highest backscatter among the cases observed.

 

Figure 4. Mean SAR backscatter versus plume area for all SAR images, with the plume and control backscatters plotted, along with the absolute value of their difference. Error bars show the standard deviation of the mean backscatters. The purple line denotes the threshold at a difference of 3.5 dB, above which all plume and control data points are separated by at least 1 standard deviation.

 

Figure 5. Box plot of Enterococci versus mean plume backscatter for LA Harbor stations matching with five plumes from 2008-2009.Whiskers are one times the interquartile range, and outliers are shown by the red crosses.

 

 

The Outcome

 

The team’s results demonstrate the potential for SAR to be incorporated into Southern California’s decision-making process regarding beach water quality. For the LA Harbor region, SAR was capable of detecting plume signals associated with precipitation and river discharge events. At the operational level, the SAR product would need to identify a backscatter threshold above which a significant percentage of beach stations would exceed the Enterococci tolerance level. At present, the data cannot define such a threshold, but additional study is warranted given the limited number of cases evaluated.

 

Acknowledgements

Benjamin Holt, Jet Propulsion Laboratory (Science Advisor)

Katrina Laygo, Daniel Cusworth, Edgar Vargas (Past Contributors)

 

References

 

[1] DiGiacomo, P.M., Washburn, L., Holt, B., Jones, B.H. (2004). Coastal pollution hazards in southern California observed by SAR imagery: stormwater plumes, wastewater plumes, and natural hydrocarbon seeps. Marine Pollution Bulletin. 49, 1013-1024.

[2] Schiff, K.C., Allen, M.J., Zeng, E.Y., Bay, S.M. (2000) Southern California. Marine Pollution Bulletin. 41 (1-6), 76-93.

[3] Corocan, A.A., Riefel, K.M., Jones, B.H., Shipe, R.F. (2010) Spatiotemporal development of physical, chemical, and biological characteristics of stormwater plumes in Santa Monica Bay, California (USA). Journal of Sea Research. 63, 129-142.

[4] Haile et al. (1996). An epidemiological study of possible adverse health effects on swimming in Santa Monica Bay. Monterey Park, CA. Report to the Santa Monica Bay Restoration Project.

[5] Ackerman , D. and Weisberg, S.B. (2003). Relationship between rainfall and beach bacterial concentrations on Santa Monica Bay beaches. Journal of Water and Health.

 

 

Editor’s Note: The DEVELOP National Program is a capacity building internship sponsored by NASA’s Applied Sciences Program that provides interns the opportunity to learn about NASA Earth Science and the practical applications of Earth observations.