Vegetated Treatment Systems for Removing Contaminants Associated with Surface Water Toxicity in Agriculture and Urban Runoff

Summary

This article summarizes the design attributes and the effectiveness of treatment systems that treat urban stormwater and agriculture irrigation runoff to remove pesticides and other contaminants associated with aquatic toxicity.

Abstract

Urban stormwater and agriculture irrigation runoff contain a complex mixture of contaminants that are often toxic to adjacent receiving waters. Runoff may be treated with simple systems designed to promote sorption of contaminants to vegetation and soils and promote infiltration. Two example systems are described: a bioswale treatment system for urban stormwater treatment, and a vegetated drainage ditch for treating agriculture irrigation runoff. Both have similar attributes that reduce contaminant loading in runoff: vegetation that results in sorption of the contaminants to the soil and plant surfaces, and water infiltration. These systems may also include the integration of granulated activated carbon as a polishing step to remove residual contaminants. Implementation of these systems in agriculture and urban watersheds requires system monitoring to verify treatment efficacy. This includes chemical monitoring for specific contaminants responsible for toxicity. The current paper emphasizes monitoring of current use pesticides since these are responsible for surface water toxicity to aquatic invertebrates.

Introduction

Surface water toxicity is prevalent in California watersheds and decades of monitoring have shown that toxicity is often due to pesticides and other contaminants¹. The primary sources of surface water contamination are stormwater and irrigation runoff from urban and agriculture sources. As waterbodies are listed as degraded due to contaminants and the toxicity is identified from urban and agricultural sources, water quality regulators partner with state and federal funding sources to implement practices to reduce contaminant loading. Green infrastructure is being promoted in California urban watersheds to reduce flooding and increase the recovery of stormwater through infiltration and storage. While Low Impact Development (LID) designs are being mandated for new construction in many regions, few studies have monitored the efficacy of these systems beyond measurements of conventional contaminants like dissolved solids, metals, and hydrocarbons. More intensive monitoring has recently evaluated reductions in chemical concentrations and chemical loading responsible for surface water toxicity, and to directly determine whether bioswales reduce toxicity of runoff. This has shown that bioswales are effective at removing toxicity associated with some contaminant classes², but additional research is required for emerging chemicals of concern.

Vegetated treatment systems are also being implemented in agriculture watersheds in California, and these have been shown to be effective at reducing pesticides and other contaminants in agriculture irrigation runoff^3,4. These systems represent components of a suite of approaches to reduce contaminant loading to surface waters. Because they are intended to mitigate contaminants responsible for surface water toxicity, a key component of the implementation process is monitoring to ensure their long-term effectiveness. Monitoring includes both chemical analyses of chemicals of concern, as well as toxicity testing with sensitive indicator species. This article describes protocols and monitoring results for an urban parking lot bioswale and an agricultural vegetated drainage ditch system.

The design attributes of a typical parking lot bioswale, such as may be used to treat storm runoff in a typical mixed-use urban shopping parking area depend on the area being treated. In the example described here, 53,286 square feet of asphalt create an impervious surface area that drains to a swale, which consists of 4,683 square feet of landscaping. To accommodate runoff from this surface area, a 215 feet long flat-bottom, semi-V shape channel comprises the swale with a side slope less than 50% and a longitudinal slope of 1% (Figure 1). This swale comprises three layers including native bunch grass planted in 6 inches of topsoil, layered over 2.5 feet of compacted subgrade. Stormwater flows from parking areas to multiple entry points along the swale. The water infiltrates the vegetated area, then permeates the subgrade and drains into a 4-inch perforated drain. This system drains water through a system plumbed to an adjacent wetland that eventually drains into a local creek.

Protocol

1. Urban Bioswale Efficacy Monitoring

1. Storm Water Sampling
   1. Sample 4 L of pre-treatment stormwater leaving the parking lot as it enters the bioswale inlet, and then 4 L of post-treatment stormwater as it leaves the bioswale through the 4" outlet drain.
   2. Using local weather predictions, collect samples at the beginning, middle, and end of the storm's hydrograph. Composite the samples to characterize runoff variability during the storm event.
   3. Collect 1.3 L samples by hand and composite them in a 4 L amber bottle. Collect inlet samples at several curb openings where storm water flows into the bioswale.
   4. Collect 1.3 L outlet samples from the flow meter attached to the outlet drain (described below) and composite them in a 4 L amber bottle.
   5. Store the composited samples on ice until the final hydrograph sample is collected. Then transport them to the laboratory and hold in a refrigerator at 4 °C prior to subsampling for chemistry and toxicity testing. Ship samples to the chemistry lab within 48 h of sample collection.
2. Load Calculation
   1. Prior to the storm, install a tipping-bucket digital logger rain gauge by attaching it to a light or other pole adjacent to the bioswale site. Use the rain data to indicate instantaneous and total rainfall for the site.
   2. Install a mechanical geared pulse flow meter on the outlet drains of the bioswale. Record total flow exiting the bioswale.
      NOTE: Reduction in runoff volume is presumed to reduce overall loading of contaminants in LID designs.
   3. Model the volume of water falling on the parking lot catchment area during the rain event by extrapolation using the inches of rain recorded by the rain gauge. Use these data to determine the volume entering the treatment system based on parking lot surface area.
   4. Use the total flow recorded by the outlet flow meter to calculate infiltration percentage. Calculate the difference between inlet and outlet volume to determine stormwater infiltration.
   5. Calculate contaminant loading and load reduction percentages during the storm using inlet and outlet volume in conjunction with contaminant analytical measurements.
   6. Measure chemical analytes that are relevant to surface water toxicity (as discussed below). Total chemical groups to simplify load calculations and base on their similar toxic modes of action (e.g., total polynuclear aromatic hydrocarbons [PAHs], total pyrethroids, and total fipronil and degradates).
3. Chemistry
   1. Analyze all samples for the following parameters: total suspended solids (TSS), trace metals (USEPA method 200.8⁵; inductively coupled plasma-mass spectrometry [ICP/MS]), and PAHs (USEPA method 625⁶).
   2. Analyze samples for current-use urban pesticides, including 9 pyrethroids (USEPA method SW846 8270 modified⁷; bifenthrin, cypermethrin, fenvalerate/esfenvalerate, permethrin, tetramethrin, L-cyhalothrin, cyfluthrin, and allethrin), and fipronil and its three primary degradates (fipronil sulfide, fipronil sulfone, fipronil desulfinyl).
   3. Analyze pyrethroids using gas chromatography-mass spectrometry (GC/MS) using negative chemical ionization or other appropriate method to provide adequate detection limits. Since most current-use pesticides are highly toxic at low concentrations, their analyses require low chemical reporting limits to be relevant for environmental risk assessment. The method reporting limits for pyrethroids are from 0.5 ng/L to 1.0 ng/L for all pyrethroids except permethrin (reporting limit = 10 ng/L).
   4. Use an analytical procedure for fipronil that provides a method reporting limit of 1.0 ng/L. Organophosphate pesticides do not need to be measured depending on local use patterns for example in urban areas in California^8,9.
   5. Measure neonicotinoid pesticides (e.g., imidacloprid) using ultra performance liquid chromatography coupled to a triple quadrupole mass spectrometer, which has a reporting limit for imidacloprid of 50 ng/L.
4. Toxicity Testing
   1. Conduct toxicity tests on the composited inlet and outlet stormwater samples using 3 test species, following modified US Environmental Protection Agency (USEPA) acute test protocols¹⁰. The test with the cladoceran Ceriodaphnia dubia measures survival after 96 h. The test with the amphipod Hyalella azteca measures survival after 10 days. The test with the midge Chironomus dilutus measures survival and growth after 10 days.
   2. Conduct acute 96 h survival tests with the cladoceran C. dubia following U.S. EPA guidance.
      1. Expose five C. dubia neonates in each of five replicates of inlet and outlet stormwater samples. Replicates consist of 20-mL scintillation vials containing 15 mL of test solution.
      2. Feed neonates a mixture of yeast, cerophyll, trout chow (=YCT; following U.S. EPA guidance) and Selenastrum algae 2 h prior to daily 100% renewal of the stormwater test solutions. Record total number of surviving neonates daily.
      3. Compare final C. dubia survival after 96 h exposure to inlet and outlet stormwater samples to survival in moderately hard control water using a t-test. Follow statistical procedures recommended by U.S. EPA.
   3. Conduct acute 10 d survival tests with the amphipod H. azteca following U.S. EPA guidance.
      1. Expose 10, 9 days to 15 days old amphipods in each of five replicates. Replicates consist of 300-mL glass beakers containing 200 mL of test solution.
      2. Conduct amphipod tests for 10 days, count the number of surviving amphipods daily, and renew 50% of the test solution every 48 h. Feed each beaker every 48 h with 1.5 mL of YCT after the renewal.
      3. Compare final survival of amphipods in stormwater samples to 10 days survival in laboratory well water as described above.
   4. Conduct chronic 10 d survival and growth tests with the midge C. dilutus following U.S. EPA guidance.
      1. Expose 12, 7-d old animals in each of four replicates. Replicates consist of 300 mL glass beakers containing 200 mL of test solution. Supply each midge test container with 5 mL of sand as substrate for tube building by the larvae.
      2. Conduct tests for 10 d, and renew 50% of the test solution every 48 h for each beaker daily with an increasing amount of fish food slurry (4 g/L), as follows: days 0 to 3, 0.5 mL/day; days 4 to 6, 1.0 mL/day; days 7 to 10, 1.5 mL/day.
      3. Compare final survival and growth in stormwater samples to 10-d survival in laboratory well water as described above. Measure growth of surviving animals as ash-free dry weight at 10 d compared to initial weight of the test organisms.
   5. For all toxicity tests, measure dissolved oxygen, pH, and conductivity using appropriate meters and electrodes. Measure un-ionized ammonia using a spectrophotometer.
      1. Measure water hardness and alkalinity at initiation and termination of tests.¹⁰
      2. Record water temperature with a continuous recording thermometer.

2. Integrated Vegetated Agricultural Drainage Ditch Efficacy Monitoring

1. Integrated Ditch Construction
   NOTE: The agricultural drainage ditch used in the current example is 152 m long and has a semi-V-shaped cross-section width of 5 m at the top and 1 m depth. The ditch vegetation is a combination of native grass species primarily seeded with red fescue (Festuca rubra). In this example, integrated vegetative ditch trials consisted of granulated activated carbon (GAC) and compost filter treatments integrated with the vegetated ditch.
   1. Construct two compost filters and six carbon filters and install them in three different sections of the vegetated ditch (Figure 2). Use 2 m long 20 cm diameter sleeves filled with either carbon or compost.
   2. Fill six sleeves with 30 L of granulated activated carbon and place these across the ditch at the 146 m point, near the end of 152 m vegetated ditch. Anchor the GAC-filled sleeves to the ditch bottom with wire stakes on the upstream edge.
   3. Place a 2.5 m long 6" wide section of pine board on the downstream edge of each of the GAC sleeves. Dig the pine boards into the two sides and bottom of the channel to minimize water bypassing and undercutting the carbon sleeves. The boards will also provide vertical support to maximize water contact time with the carbon.
   4. Fill the compost sleeves with approximately 15 kg each of partially decomposed yard waste from any clean source, such as a local landfill. Position two 2 m long compost sleeves across the vegetated channel at 64 m and at 123 m along the length of the 152 m vegetated ditch (Figure 2).
2. Runoff Simulation and Sampling
   NOTE: This protocol describes methods for conducting simulated agricultural runoff trials and associated monitoring to evaluate treatment efficacy using the integrated vegetative treatment system. In the current example, the integrated vegetated-compost-carbon system was evaluated at two flow rates that represented rates typical off-field discharge from commercial farms in the Salinas Valley, 3.2 L/s and 6.3 L/s. The organophosphate pesticide chlorpyrifos was used as a model pesticide in these trials because it has a moderate solubility, and therefore represents the mid-range of solubility of representative pesticides commonly used in pest management. Chlorpyrifos is also the subject of on-going regulatory actions in central California because of its impacts on agriculture watersheds. The target chlorpyrifos dose was approximately 2,600 ng/L. Flow rates and target chlorpyrifos concentrations were within the ranges previously measured in local irrigation runoff ^3,11. The hydraulic residence time for a pulse of water transiting the vegetated ditch was not monitored in the example given here. The residence time in these systems vary with water inflow rate, the degree of soil saturation due to previous irrigation and rain, the presence of structures to impede flow such as weirs and sedimentation basins, and the amount of surface area covered by vegetation. Previous studies have demonstrated residence times of several hours for small scale ditch systems in the Salinas Valley ^3,4. Visual observations indicated the residence time for the GAC filters was one or two minutes.
   1. Create simulated agricultural runoff using ground water mixed with suspended sediment. For trials with the model pesticide, chlorpyrifos, prepare a fresh stock solution of 10 mg/L for every 3.2 L/s trial by adding certified stock solution to a known volume of distilled water. Prepare a fresh chlorpyrifos stock solution of 20 mg/L for every 6.3 L/s trial.
      1. Use a metering pump to provide a consistent volume of stock solution to the runoff water before it enters the vegetated treatment ditch inlet. Use the metering pump to deliver stock solution at 50 mL/min to the flow of simulated irrigation water.
   2. Monitor the inlet flow rate with a digital meter and use these data to quantify total volume of runoff water applied to the ditch inlet.
   3. Construct a weir at the outlet of the ditch and plumb this with an outlet pipe connected to a digital flow meter. Use this meter to record the volume of runoff exiting the ditch.
   4. Use data loggers connected to the digital meters to record flow at 5 min intervals. Program the data loggers to activate peristaltic pumps located at the inlet and at various stations (e.g., 23 m, 45 m, and 68 m) below the inlet of the ditch to collect composite subsamples of runoff into stainless steel containers at 5 min intervals.
3. Chemistry
   1. Transfer composite samples of runoff water from trials into amber glass bottles at the end of each runoff trial and maintain the samples on ice at 4 °C for later toxicity and chemical analyses.
   2. Analyze the composite samples for total suspended solids (TSS), and chlorpyrifos using GC-MS or enzyme-linked immunosorbent assays (ELISA).
   3. Compare "inlet" composite samples (pre-treatment) to "outlet" composite samples (post-treatment) to evaluate efficacy of the integrated ditch system to reduce TSS and pesticide loads.
4. Toxicity Testing
   1. Determine water column toxicity was in composite samples from the inlet (pre-treatment) and outlet (post-treatment) of each trial using 96 h Ceriodaphnia dubia toxicity tests¹⁰, as described above for the bioswale monitoring. C. dubia is an appropriate monitoring species for agriculture runoff toxicity due to its sensitivity to chlorpyrifos (median lethal concentration (LC50) = 53 ng/L¹²).

Results

Urban Bioswale Efficacy

During the 18.5 h of the storm, 1.52" of rain was recorded by the rain gauge, and this resulted in 50,490 gallons of water flowing from the parking lots into the bioswale. Of this total volume, 5,248 gallons were recorded by the outlet flow meter, resulting in a total infiltration of 90% of the stormwater that flowed into the bioswale. The bioswale reduced all of the chemicals monitored. Total suspended ...

Discussion

The practices described in this protocol are intended as final steps in an overall strategy to remove pollutants in agriculture irrigation and stormwater runoff. Use of bioswales and other urban green-infrastructure LID practices are intended as a final piece of the puzzle to remove contaminants in runoff before they reach adjacent receiving waters. This protocol emphasizes methods to monitor urban bioswales to determine treatment efficacy for removing toxicity associated with urban contaminants, with emphasis on current...

Materials

Name	Company	Catalog Number	Comments
HOBO tipping-bucket digital logger rain gauge	Onset Computer Co., Bourne MA, USA)	Onset RG3	Rain gauge
Mechanical geared pulse flow meter	Seametrics Inc., Kent WA	Seametrics MJ-R	Flow meter for measuring bioswale outlet flow
Filtrexx SafteySoxx	Filtrexx Co. - info@filtrexx.com	SafetySoxx	perforated synthetic cloth for granulated activated carbon and compost
Granulated activated carbon	Evoqua - Siemens Corp., Oakland CA	AC380	GAC for agriculture irrigation water treatment
Digital flow meters	Seametrics Inc. Kent WA	Ag2000; WMP101	Flow meters for agriculture irrigation treatment system monitoring
Data Loggers	Campbell Scientific Inc., Logan, UT	CR1000	Data loggers for recording flow data
Peristaltic pumps for composite sampling	Omega Engineering Inc. Stamford CT	Omegaflex FPU-122-12VDC	Pumps for composite sampling