Article pubs.acs.org/est
Shallow Groundwater Conveyance of Geologically Derived Contaminants to Urban Creeks in Southern California Audra I. Bardsley,*,† Douglas E. Hammond,† Theodore von Bitner,‡ Nikolaus H. Buenning,† and Amy Townsend-Small§ †
Department of Earth Sciences, University of Southern California, 3651 Trousdale Parkway, Los Angeles, California 90089-0740, United States ‡ OC Watersheds, County of Orange Public Works, 2301 Glassell Street, Orange, California 92865, United States § Department of Earth System Science, University of California, Irvine, California 92697, United States S Supporting Information *
ABSTRACT: In California alone, there are currently over 200 instances on the EPA’s list of impaired water bodies with unknown sources of excessive salinity or trace contaminants. This investigation focuses on Orange County, CA, a region that has undergone extensive hydrological modification, relies heavily on imported water for municipal supply, and has come under regulatory scrutiny for elevated TDS, sulfate, Cd, Ni, and Se. A survey of shallow groundwater weeps and springs, discharging directly to urban creeks, reveals high concentrations of TDS, sulfate, Cd, Ni, Zn, Cu, and Se that are often far in excess of water quality standards. Isotopic (δ34S and δ18O) and geochemical evidence indicate that the source of sulfate and TDS is weathering of sulfide minerals in the Capistrano Formation marine mudstone and dissolution of secondary minerals formed during past periods of sulfide oxidation, rather than anthropogenic inputs. The relative availability of carbonate minerals along the flow path appears to control pH, which then influences trace metal mobility to surface waters. Stable isotopes of H2O indicate that despite widespread use of imported water, meteoric recharge dominates shallow groundwater inputs with municipal sources contributing only 13−29% of discharge. These findings highlight the importance of understanding the hydrogeological setting to properly apportion contaminant sources and conveyances.
■
INTRODUCTION Properly apportioning contaminant sources and conveyance mechanisms, especially between natural and anthropogenic contributions, is essential for effective water quality management. This is of particular concern in densely populated southern California, where healthy streams are a scarce resource1 but causes of impairment are often challenging to identify.2 The region has undergone extensive hydrologic modification from drainage system engineering, construction of impermeable surface cover, and use of imported water for municipal supplies. Among the most conspicuous results has been an upsurge in dry weather discharge, with annual water fluxes from impacted urban creeks increasing between 250 and 500% over the past 50 years.3 Numerous factors likely contribute to increased dry weather discharge. Though not important for the region investigated in this study,4 in many areas treated wastewater effluent is discharged to creeks. In addition, municipal storm drain systems, constructed to reduce flood damage by efficiently moving precipitation away from developed areas, have had the unintended consequence of routing dry season urban runoff from outdoor application of © 2015 American Chemical Society
imported municipal water to nearby creeks. As municipal water from landscape irrigation, car washing, and other maintenance activities flows across urban surfaces, it picks up and transports contaminants, including trace metals, nutrients, salts, pesticides, and bacteria to urban creeks and ultimately coastal zones. Previous studies of dry weather surface water quality in southern California and the current regulatory framework have mostly focused on impacts from this dry weather urban runoff.5−7 Another less well characterized factor that can increase dry weather creek discharge is input from shallow groundwater. The combined effects of evapotranspiration suppression by removal of riparian vegetation and shallow water table modification by flood control channel incision can enhance shallow groundwater contributions.8 Leakage from pipes conveying flows of potable water or sewage, coupled with infiltration of irrigation water, can introduce imported water Received: Revised: Accepted: Published: 9610
February 25, 2015 May 30, 2015 July 13, 2015 August 4, 2015 DOI: 10.1021/acs.est.5b01006 Environ. Sci. Technol. 2015, 49, 9610−9619
Article
Environmental Science & Technology and raise the level of local water tables.9 Unlike surface runoff, shallow groundwater can interact with aquifer sediments and local geology before discharging to urban creeks and may consist of a mixture of water derived from local meteoric sources and imported municipal supplies. Catchment geology is often overlooked during examination of urban surface water quality problems, yet many contaminants associated with anthropogenic inputs including sulfate, excess salinity, cadmium, nickel, zinc, selenium, and copper are also mobilized during weathering of sulfide rich deposits, including marine shales prevalent in large portions of California.5−7,10−16 The complexity of urban hydrologic systems makes identifying key source and conveyance mechanisms challenging for government agencies tasked with water quality management. Mandated monitoring programs focus on measuring only specific contaminants in creeks and urban runoff. The large amount of data generated is useful for identifying the existence of problems but not necessarily the key causes. Additionally, in urban catchments, source and conveyance may become decoupled. For example, naturally occurring sulfate from geologic sources may be carried to creeks by imported municipal water or local meteoric water may mobilize anthropogenic contaminants. Management strategies for each scenario are different and may be incompatible with the current regulatory framework. Aqueous geochemistry offers many tools that can help illuminate key contaminant transport processes and fingerprint sources, even in highly altered urbanized settings. This investigation utilizes several of these tools. The study area encompasses southeastern Orange County, CA, USA (SE OC), where several urban creeks have come under state and federal regulatory scrutiny for elevated concentrations of sulfate, total dissolved solids (TDS), and trace elements including Se, Cd, and Ni.2 EPA cites no known sources for these contaminants. Routine compliance monitoring in SE OC revealed springs and weeps with high-salinity and often low pH, discharging shallow groundwater directly to urban creeks. Focusing on these shallow groundwater inputs, this investigation addresses three issues. First, stable isotopes of water (δD, δ18O) are used to determine what fraction of shallow groundwater conveying high TDS is derived from local meteoric sources versus being derived from imported municipal water. Second, major and minor ion composition, pH, alkalinity, and mineral saturation indices are used to identify which contaminants covary and are conveyed by shallow groundwater, and to evaluate the potential importance of mineral weathering processes in contaminant mobilization. Finally, stable isotopes of aqueous sulfate and sulfur bearing minerals (δ34S, δ18O) are used in conjunction with previously described geochemical parameters to evaluate the importance of sulfide mineral oxidation and secondary mineral dissolution as sources of sulfate and drivers of contaminant release.
Figure 1. Spring and weep locations sampled in southeast Orange County mapped against surficial deposits of the Capistrano Formation.19 Outcrop sampled for secondary minerals is denoted with a yellow star; locations of reference streams are represented with blue squares.
Orange County (MWDOC) for public supply. Except for two small retailers that incorporate some local water, SE OC relies exclusively on imported municipal water (81−95%) and recycled wastewater (5−19%) to meet demand.18 A diverse mixture of geology underlies Orange County, including igneous volcanic rocks in the Santa Ana Mountains, a large alluvial plain straddling the Santa Ana River, and varied marine and terrestrial sedimentary formations in the southeastern portion of the county.19 The Capistrano Formation in particular outcrops extensively in SE OC but is nearly absent in northwestern Orange County (NW OC, Figure 1). As described by White,20 this formation consists of mudstone interbedded with some shale and interspersed with limestone nodules in lower units, whereas upper units consist of somewhat coarser mudstone and sandy mudstones deposited near the edge of the Los Angeles Basin.20,21 The Capistrano Formation spans the late Miocene to early Pliocene and conformably overlies the Monterey Shale, a better-known organic-rich petroleum source rock. Access to unweathered portions of the Capistrano Formation in the course of this study was limited because of widespread development in areas of interest. A rapidly eroding, moderately weathered outcrop was identified along the Pacific Coast Highway near Capistrano Beach in the city of Dana Point (Figure 1). This location is characterized by poorly consolidated medium-brown mudstone with some interbedding of lighter, sandy mudstones. It is rich in secondary minerals associated with extensive pyrite oxidation in sulfide-rich sedimentary deposits, including widespread jarosite crusts and gypsum precipitation along bedding planes.22−26 All of these phases were targeted for isotopic analysis. To evaluate the potential influence of shallow groundwater on urban creek contamination, seven high-salinity weep and spring locations were selected across five watersheds in SE OC to serve as sampling sites for this investigation (Table S2). These locations constitute a diverse set of groundwater
■
SITE DESCRIPTION The study area is located in SE OC and consists of five small coastal watersheds that cover a total area of about 630 km2, ranging from the base of the Santa Ana Mountains to the Pacific Ocean (Figure 1). Although headwaters for these drainages consist mostly of open space, lower elevations are highly urbanized with residential and commercial land uses dominant.17 The semiarid Mediterranean climate, lack of large viable aquifers, and dense population results in heavy reliance on imported water distributed by Municipal Water District of 9611
DOI: 10.1021/acs.est.5b01006 Environ. Sci. Technol. 2015, 49, 9610−9619
Article
Environmental Science & Technology
coupled to a Thermo Finnigan (Delta Plus XP) IRMS with precisions of 0.23‰ (δ18O) and 1.1‰ (δD). An evaporation-corrected mixing model was used to estimate the fraction of imported municipal water contributing to weeps and springs investigated in this study. Coastal watersheds in SE OC receive water from three sources: municipal water, recycled water, and precipitation. Because recycled water should be essentially a mixture of the other two sources, it is not explicitly considered as an endmember. It is composed of mostly treated sewage (e.g., used imported municipal water) and is used to a limited degree in the watersheds studied (5−19% of total water use).18 Average tap water was used to represent the imported municipal water endmember. To characterize the precipitation endmember used for modeling, the geographic setting of each study site was considered. Results from reference creek samples were used to locally calibrate an empirically based equation developed by Williams and Rodoni27 to describe native, meteorically derived baseflow in southern California streams. The calibrated equation was then used with specific elevation and distance inland measurements to define a meteoric water isotope endmember signature for each weep and spring site. For simplicity when calculating contribution of municipal water to springs and weeps, it was assumed that endmembers mixed and then underwent evaporation (Figure 2B). On the basis of
discharge types, from perennial artesian springs to water issuing from concrete flood control infrastructure along concrete seams or at weepholes that were installed during construction to allow for groundwater drainage. All sites are adjacent to urban areas, in close proximity to the Capistrano Formation, and discharge directly into a nearby creek.
■
METHODS
Water Samples. Water was collected at targeted groundwater sites during three sampling efforts: in late summer 2012, after onset of the rainy season in winter 2012, and in summer of 2013. Water was not collected for every site during every sampling effort because new springs were identified over the course of the investigation and site access changed. In the field, YSI Sonde multiparameter probe measurements were made for pH, specific conductance, temperature, and dissolved oxygen. For one sample collected at site 1S, field pH was inconsistent with previous measurements and with a later determination in the lab. We assumed that there was an error reading the instrument and used the average field pH of two other visits to this site for calculations. To characterize potential endmembers for sources feeding high-salinity groundwater springs and weeps, imported municipal water, local rainfall, and undeveloped reference creek samples were also collected. Imported municipal water from the two major water retailers was taken at public faucets that supply potable water to areas upstream of our targeted groundwater sample sites. Both retailers also operate recycling programs that provide water used in select areas for irrigation, dust control, industrial processes, and commercial toilets,4,18 so samples were also collected from designated recycled water sprinklers within each jurisdiction. Precipitation was collected at OC Watershed’s headquarters located in Orange, CA for nine large storms during 2011−2014 and monthly at University of Southern California (USC) in Los Angeles, CA during 2011−2013. Five undeveloped reference creeks were also sampled for water isotopes in 2007 and 2011. General chemistry of groundwater and selected municipal water samples, including major ions, nutrients, alkalinity, trace metals, and total dissolved solids, was determined by research partner OC Watershed’s contractor Weck Laboratories located in City of Industry, CA, according to EPA and standard method protocols. (Additional details are provided in the Supporting Information.) Isotopic analyses were carried out at a number of facilities. Zymax Isotope Laboratory in Escondido, CA, conducted analyses for stable isotopes of dissolved sulfate and most water isotopes. All isotope analyses at Zymax were carried out using a EuroVector Elemental Analyzer (EuroEA 3028HT) coupled to an Isoprime isotope-ratio mass spectrometer (IRMS) with precisions of 0.2‰ (δ 34S-sulfate and δ18Osulfate), 0.3‰ (δD), and 0.1‰ (δ18O). OC rainwater samples collected during 2013−2014 and all USC rainwater samples were analyzed at USC using a Picarro Liquid Water Isotope Analyzer (L1102-i) cavity ringdown spectrometer (CRDS) with precisions of 0.2‰ (δ18O) and 0.34‰ (δD). When possible, groundwater samples were also analyzed for δ13C of dissolved inorganic carbon (DIC) at USC using a Picarro CO2 Isotope Analyzer (G2121-i) CRDS with precision of 0.4‰. Rainfall and reference creek samples were only analyzed for stable isotopes of water. Two reference creek samples collected in 2007 were measured at the University of California, Irvine using a Thermo-Chemical Elemental Analyzer (TC/EA)
Figure 2. (A) Stable isotopic composition of various water sources in Orange County and rainfall from University of Southern California campus located in Los Angeles, CA. (B) Expanded scale of A, simplified to show averages for endmembers and the range of evaporation trajectories expected to create water observed at site 7W.
previous modeling and field investigations, it is reasonable to assume that evaporation trend lines in SE OC should have slopes between 3 and 5.28−30 More details of these calculations and equations are available in the Supporting Information section. To aid in understanding aquifer weathering processes that influence high-salinity groundwater geochemistry and contaminant mobility, major and minor ion results from water analyses 9612
DOI: 10.1021/acs.est.5b01006 Environ. Sci. Technol. 2015, 49, 9610−9619
Article
Environmental Science & Technology
Figure 3. Dissolved trace element concentrations and mineral saturation indices in spring and weeps plotted as a function of pH. (A and B) Many groundwater samples have acidic pH and have elevated levels of Cd, Ni, Zn, and Cu. Regulatory limits are indicated by dashed lines. (C) In contrast, trace elements As and Se are highest in concentration at circumneutral pH. (D) Acidic waters are undersaturated in calcite and dolomite, whereas neutral and more basic waters are supersaturated. Gypsum is near saturation at all sites.
the global meteoric water line (GMWL)33 and are highly variable with a 46.3‰ range in δD and 6.7‰ range in δ18O (Figure 2A). The spread of values agrees well with monthly precipitation data collected approximately 65 km northwest at USC. Data for water collected from undeveloped reference creeks plots within the range of precipitation but is less variable, which is consistent with native baseflow that reflects an integrated rainfall signal.34 Variability that does exist among reference creek signatures likely reflects differences in elevation, distance inland from the coast, and precipitation dynamics.27,35,37 Imported municipal waters are also distinct from local precipitation and undeveloped reference creeks (Figure 2A). In contrast to the high variability observed for rainwater isotopic measurements, imported municipal tap water samples are closely clustered, lying below the GMWL. This offset reflects evaporative alteration, consistent with expected behavior as it undergoes surface transport and storage in California’s semiarid climate.35,37 Recycled water is slightly more enriched than tap water samples. This shift is likely due to infiltration of native groundwater into sewage pipes that run below the water table during transport to the treatment plant and some additional evaporative alteration.8,38 Because this recycled water appears to be a mix of imported water and native meteoric water, the assumption of considering only two endmembers in the mixing model is reasonable. One recycled water sample was more enriched than any other recycled or municipal sample collected, most likely because of a brief periodic inclusion of captured creek flows as an additional source for recycled water for this water district. The δD values for high-salinity weep and spring samples plot within the range of local precipitation and reference stream measurements, though notably below the GMWL, and have consistent isotopic compositions across seasons (Table S4). Using the two-endmember mixing model described above, influence from imported municipal water is evident at all weep
were evaluated in conjunction with alkalinity, pH, temperature and DO at each site using the USGS modeling software PHREEQC (V2, LLNL database). Saturation indices were calculated for major mineral sources of alkalinity (calcite and dolomite) and for the two secondary sulfate minerals most commonly observed in SE OC (gypsum and jarosite). Because aqueous speciation is important for As and Se mobility but is not within the analytical scope of this investigation, most probable valence states were calculated for these two trace elements using PHREEQC. Mineral Samples. Hand samples were collected from beneath the top layers of exposed rock at the rapidly eroding cliffside described in the previous section. Secondary minerals gypsum and jarosite were separated from weathered mudstone by hand and identified using a Hitachi SEM (TM-100) with a Swift ED-TM EDS attachment at USC to assess bulk composition. To segregate any remaining pyrite and other sulfide minerals from the weathered Capistrano outcrop, a reduced inorganic sulfur extraction was carried out. This was achieved using a chromium reduction technique31 modified to include an additional filtration step for the reduced chromium solution.32 Gypsum precipitates were also collected from an underground box culvert, adjacent to low-pH groundwater discharges corresponding to sample site 5W (Figure 1). To evaluate the stable isotopic signature of sulfate in gypsum, jarosite, and extracted sulfide, samples were sent to the University of Arizona Environmental Isotope Laboratory. Samples were combusted at high temperature to produce SO2 and CO gas, and then δ34S and δ18O were measured using a ThermoQuest Finnigan (Delta PlusXL) continuous-flow IRMS with precisions of 0.15‰ and 0.3‰, respectively.
■
RESULTS AND DISCUSSION Imported Municipal Water Supplies 13−29% of Weep and Spring Discharges. Data from OC rainwater fall along 9613
DOI: 10.1021/acs.est.5b01006 Environ. Sci. Technol. 2015, 49, 9610−9619
Article
Environmental Science & Technology and spring study locations. However, simply mixing imported and local meteoric water cannot explain the observed isotopic signatures. Water at all locations must have also undergone evaporation to move it to the right of the mixing line (Figure 2B). It is unclear if mixing or evaporation occurred first, but in either case, a similar proportion of water should be sourced from each endmember to achieve the observed isotopic signatures. Interestingly, calculated imported municipal water contributions are not the same at all sites. There is an association between the density of nearby development and the amount of imported municipal water contributing to shallow groundwater. Sites 3W, 4S, and 7W have calculated imported municipal water contributions of 29 ± 6, 26 ± 6, and 25 ± 6%, respectively. Sites 1S, 2S, and 6S have more nearby open space and calculated imported municipal water contributions of 13 ± 8, 13 ± 7, and 19 ± 7%, respectively. Site 5W is unique because it is the only site situated in a water district that blends local and imported water for municipal supply. It has an imported water contribution of 14 ± 7%, which is lower than similar sites (3W, 4S, and 7W) located in water districts that use no local water for municipal supply. (See the Supporting Information for more details.) The results from SE OC suggest that in addition to enhancing dry weather creek discharge by way of surface runoff municipal water input is also impacting shallow groundwater and may be entering channels in part via weeps and springs especially in more densely urbanized areas. Despite the clear influence of imported water, these findings emphasize the continued importance of native meteorically derived water in driving discharge from high-salinity weeps and springs, even in urban settings with engineered drainage networks and substantial impervious cover, under protracted drought conditions. This paints a nuanced picture of conveyance, in which both native and imported municipal water transport dissolved contaminants to urban creeks. Weeps and Springs Supply High Concentrations of Sulfate, TDS, and Trace Contaminants to Urban Creeks Compared to Regulatory Standards. Results for weep and spring samples reveal that shallow groundwater discharges are elevated in sulfate, TDS, Cd, Ni, Cu, Zn, Se, and to a lesser extent As, compared to ambient water quality standards (Figures 3A−C and 4A). Sulfate concentrations span 20−115 mM and account for approximately half of TDS by mass, ranging from 4000−22 000 mg/L (Table S3). All samples exceed the ambient water quality thresholds of 500 mg/L for TDS and 2.6 mM for sulfate by at least one order of magnitude.39 As expected from the high TDS measurements, other major ions including calcium, sodium, chloride, potassium, and magnesium are also high relative to concentrations typically observed in surface waters (Table S3). Ni and Zn concentrations range from 0.06−27.5 μM, Se from 0.02−2.79 μM, As from
