Applying Whole Effluent Toxicity Testing to Aircraft Deicing Runoff (2015)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

22 4.1 Collection of a Representative Sample 4.1.1 Background Many of the airport discharge permits reviewed with require- ments for aquatic toxicity testing required the collection of grab samples. However, given the changes in flow rate and chemical concentration, collection of a single grab sample is unlikely to be representative of the discharge of stormwater associated with airport deicing operations. Thus, the initial focus in the devel- opment of a WET testing program should be on how to collect a representative sample. A “representative sample” is defined by one online source as “a subset of a statistical population that accurately reflects the members of the entire population. A representative sample should be an unbiased indication of what the population is like.” (http://www.investopedia.com/terms/r/representative-sample. asp). Similarly, representative sampling is defined as “sampling in which the relative sizes of sub-population samples are cho- sen equal to the relative sizes of the sub-populations” (http:// www.merriam-webster.com/dictionary/representative% 20sampling). Stormwater varies in 2 aspects, flow and chemi- cal composition. Flow can be both observed and measured in the field. However, chemical variability usually requires more sophisticated analytical methods. Thus, stormwater discharge sampling should reflect and account for changes in flow such that changes in chemical composition are proportionally weighted in the final sample. 4.1.2 Recommendation If both the variability in chemical composition of the storm- water is known not to vary by more than 10% over time, a single grab sample may be collected and be considered representative of the stormwater discharge. However, as described above, this condition is unlikely to exist at most airport stormwater dis- charge points and only those airports with large, well-mixed Provided below is guidance on the collection of environmen- tally representative samples for airport stormwater discharges as well as suggestions for the conduct of aquatic toxicity tests such that test exposures are more consistent with environmen- tal exposure conditions. Note that sample type and sampling requirements for aquatic toxicity testing may be specifically defined in the NPDES discharge permit. In this case, the sam- pling procedures must comply with the permit. However, many permits and permit writers allow flexibility in sampling and other aspects of testing to ensure that environmentally repre- sentative results are obtained. It should be clearly recognized that the cornerstone of environmental impact assessment is the collection of rep- resentative data reflective of field exposure conditions. This starts with the collection of a representative sample(s). How- ever, to be truly environmentally representative requires the conduct of onsite flow-through tests in which a portion of the stormwater discharge is diverted to an onsite laboratory/ testing system such that test organisms are exposed to the stormwater in real time. In this manner, changes in storm- water exposure concentrations are identical to field condi- tions. However, the cost of this type of testing as a compliance tool is excessive. Thus, EPA has identified that stormwater sampling may consist of either grab or composite samples; however, limited information is available to allow for the determination of how best to collect a sample from an inter- mittent stormwater discharge. The guidance below primarily focuses on collecting a repre- sentative sample. Because the test methods have been codified into 40 CFR 136, changes to those protocols are not recom- mended at this time and comments are limited to those changes that can be incorporated into the testing protocol without changing the basic requirements of the tests. General recom- mendations are summarized in Table 4-1 and are discussed in detail below and in Appendix A. Specifically, for each issue, a general background is provided followed by a recommendation or series of recommendations. Guidance for Environmentally Representative Sampling and Testing S E C T I O N 4

23 (continued on next page) Tool Purpose Method When Does This Apply? Collection of Representative Sample (See Section 4.1) Single grab sample Provide a cost- effective and time- efficient sample representative of non-variable discharge A grab sample is collected in an open container from a single point at the required sampling point. Grab samples can be collected with a suspended or hand-held polypropylene container, disposable bailer, or narrow, open-mouth bottle. The sample should be collected from the centroid of the flow by immersion of the bottle into the flow. • Discharge is not variable over time • Sampling resources are limited • Toxicity tests that are <24 hour duration Multiple grab samples Provide a cost- effective and time- efficient sample representative of a variable discharge Sampling consists of grab samples collected at specific time intervals. Samples may be combined in proportion to flow (i.e., flow-proportional composite, if data are available) or may be combined without regard to flow rates (i.e., time proportional composite). • Variable discharge but not overly complex • Limited sampling resources • Toxicity tests that are >24 hour duration Composite sampling Provide a representative sample of variable discharge Samples may be collected manually or automatically. Automatic sampling is preferred and consists of 2 strategies: • Discharge is variable in terms of flow and chemical characteristics. • Discharge has a flow meter or there is a means to measure flow in real time Constant Time—Volume Proportional to Flow Rate—samples are collected at equal time intervals; however, the volume of sample collected is proportional to the flow rate at the time of collection. Either manual or automated collection technologies can be utilized. However, the volume of sample collected is proportional to the flow rate. Thus, a portable flow meter (impeller or electromagnetic type) that provides an instantaneous flow velocity can be utilized to determine the volume of sample to collect. Alternatively, a fixed sample volume can be collected and, once flow data are retrieved, the volume of each sample to be added to the final sample can be determined. Constant Volume—Time Proportional to Flow Volume Increment—the volume of sample collected is uniform; however, the frequency of sample collection is dependent upon the volume of flow. At higher flow rates, samples will be collected more frequently. This type of sample is best collected automatically and requires the use of a flow meter connected to the automatic sampler. Table 4-1. General recommendations for the conduct of aquatic toxicity tests for stormwater discharges from airport operations.

24 Table 4-1. (Continued). Tool Purpose Method When Does This Apply? Toxicity Testing Renewals (see section 4.2) Provide test exposures that are representative of variable discharges Renew the test solution with effluent collected. If there is no discharge, test solutions should be renewed with laboratory control dilution water or receiving water. Prior agreement with the regulatory authority should be obtained as to how samples are to be collected and utilized for test renewal prior to test initiation. • Test duration > 24 hours • Variable flow discharge anticipated • Short-term discharge from detention pond • Delayed release due to freeze-thaw Temperature (see section 4.3) Provide test exposures that are representative of discharges to cold receiving waters Conduct toxicity tests at temperatures that are at or near receiving water temperatures using acceptable cold-adapted species. Investigate and obtain agreement from regulatory authority to allow for testing using cold-water species for deicing events. • Receiving waters that are likely to remain cold (<10 °C) throughout discharge • Acceptable cold- adapted test species are available for testing Dissolved oxygen (see section 4.4) Ensure that appropriate test conditions are maintained throughout testing Notify the laboratory that the sample may contain elevated levels of oxygen demanding substances. Request that the laboratory monitor DO frequently, providing aeration if DO falls below recommended limits. Review resulting test data to ensure that dissolved oxygen concentrations were maintained at acceptable concentrations. • BOD/COD is expected to be elevated • Fish tests or invertebrates that are fed during testing (decaying food may decrease DO) Pavement and aircraft deicing material application rates and time of application (see section 4.7) These data further allow the characterization of the storm event relative to deicing operations. Depending on the timing of the storm event, deicing operations may or may not be occurring Information should be collected regarding the time of application of deicing materials and the location of application. Data should be analyzed by drainage basin with a focus on those basins that are being sampled for aquatic toxicity. • Data should be collected for each discharge event to allow characterization of the discharge event Data Review and Application Concurrent monitoring (see section 4.5) Providing supporting data to understand flow characteristics to support data interpretation WET testing typically only requires the collection of DO, pH, temperature, conductivity, hardness, alkalinity, and chlorine and ammonia concentrations for the test sample. Data interpretation can be significantly enhanced through the collection of the following constituents: • COD • BOD • Ethylene and propylene glycol concentration • Calcium, sodium, potassium, and magnesium. • Data should be collected for every test to allow for establishment of a baseline condition • Data should be plotted such that unusual conditions can be identified Receiving water and discharge flow records (see section 4.6) Allow determination of whether receiving water was at critical low flow conditions and if discharge or storm event met design conditions (i.e., 24-hour, 10-year storm event) If available, receiving water flow data can be obtained from United States Geological Survey (USGS) monitoring stations. However, stations may not be located on all receiving waters. Weather event information can be obtained from local National Oceanic and Atmospheric Administration (NOAA) weather station. Stormwater flow rates can be utilized to determine instream concentration after mixing in the receiving water. • Data should be collected for each discharge event to allow characterization of the discharge event

25 Toxicity test data review (see section 4.8) Ensure that test results are defensible and meet QA/QC requirements Conduct a review of the following test conditions: • Is sample hold time acceptable (<36 hours)? • Are test temperatures within acceptable ranges and do not vary by more than 3°C? • Are DO levels maintained above 4 mg/L (warm-water test species) and 6 mg/L (cold-water test species)? • Is the age of test organisms within acceptable standards? • Is the control survival greater than 90%? • Are test solutions renewed at least every 48 hours? • Does the dose-response curve demonstrate an expected response in which higher concentrations exhibit a higher response? • Does the reference toxicity test fall within acceptable laboratory levels? Data Review and Application (Continued) Toxicity identification and evaluation (see section 4.9) To identify toxicants contributing to observed aquatic toxicity Utilize historical data collected for toxic and non-toxic discharges to characterize differences between samples. Screening level testing should be conducted for each sample to identify those useful for toxicity identification and evaluation (TIE) procedures. Utilize EPA Methods for Aquatic Toxicity Identification Evaluations: Phase I Toxicity Characterization Procedures, second edition (EPA-600-R-91-003) (EPA 1991a). When toxicity is consistently observed, TIE procedures should be implemented Tool Purpose Method When Does This Apply? Table 4-1. (Continued). stormwater management ponds are likely to meet these crite- ria. Thus, composite stormwater sampling technologies should be utilized to collect a sample representative of the stormwater discharge over a 24-hour period. Ideally, sampling should be weighted based on flow using either Constant Time— Volume Proportional to Flow Volume Increment or Constant Volume—Time Proportional to Flow Volume Increment (See Appendix A, Section 2.1 for a description of sampling approaches). Both of these methods provide the best estima- tion of the event mean discharge concentration. If capital resources are limited, constant time—constant volume meth- ods of discharge compositing may be utilized, however, these are less likely to be representative of the event mean discharge concentration under highly variable flow conditions. Critical information necessary to develop a sampling pro- tocol and program the sampler are 1) volume of sample to collect and 2) estimated stormwater flow. With respect to the volume of sample required, acute aquatic toxicity tests using C. dubia and P. promelas require approximately 1- and 2-L sample volume for each test, respectively. Thus, a minimum of 3 L (~1 gallon) of sample is required for aquatic toxicity testing. Note that additional analyses should be conducted (described below) to characterize the sample and facilitate in data interpretation. Thus, sample volume will likely be greater than 1 gallon to accommodate these additional analy- ses. Also note that if the test is renewed on a daily basis and a fresh sample is not collected, a larger sample volume will also be required. Depending on the method of sample composite collection, the volume of the individual sample may vary. For the Con- stant Time—Volume Proportional to Flow Volume Increment method, a variable sample volume is collected at a specific time interval based on the amount of flow discharged from the previously collected sample. For the Constant Volume— Time Proportional to Flow Volume Increment method, a fixed sample volume is collected for every gallon of discharge. For example, if 4 gallons (~15 L) of sample are to be col- lected using the Constant Time—Volume Proportional to Flow

26 Volume Increment method, the following calculations and assumptions are made: 1 sample is to be collected every hour for 24 hours based on the flow during that hour and The total storm event discharge volume is estimated at 5 million gallons. The sampler should be set to collect 0.3 L for every 100,000 gallons of flow (4 gal/5,000,000 gal * 100,000 gal * 3.78 L/gal). At a flow of 5 million gallons per day (MGD), the average hourly flow will be 208,333 gallons/hour and will result in an average of 0.642 L/hour of sample collected. This will result in a final sample volume of 15 L. However, if the stormwater flow volume is overestimated and is actually 2 million gallons, the average hourly flow will only be 83,333 gallons. Because the average flow volume is lower, a smaller sample aliquot will be collected each hour. This lower flow rate results in 0.25 L of sample collected per hour pro- viding a total sample volume of only 6 L, which is only slightly higher than that required for aquatic toxicity testing using C. dubia and P. promelas. Thus, critical to the successful col- lection of a stormwater composite sample is the estimation of total stormwater volume. Similarly, for the Constant Volume—Time Proportional to Flow Volume Increment method, a fixed sample amount is collected for every X gallons of stormwater discharge. Simi- Automatic Sampler Programming Critical to the successful collection of a composite sample is the estimation of the total discharge flow. However, there is likely to be a large variability associated with the estimated volume due to temperature effects (freeze/ thaw) and storage of precipitation on the airfield. Thus, an estimated minimum and maximum discharge volume should be utilized to determine sampler programming parameters. Once the maximum and minimum stormwater discharge volumes have been estimated, the average volume is utilized to calculate the volume of each discrete sample. For example, if the maximum, minimum, and aver- age discharge volumes are 5, 2, and 3.5 million gallons respectively, then the volume of sample to collect is calculated as the desired sample volume (e.g., 3 gal, 11.3 L) divided by the total estimated discharge volume (3.5 million gallons) to derive 0.32 L per 100,000 gallons of flow. Note that the maximum capacity of most automatic, portable samplers is 4 gallons (15 L); however, a sample volume of 3 gal was utilized to allow for uncertainty. Using this information, the sampler would be programmed to collect 0.64 L of sample for every 200,000 gallons of discharge. To determine if this setting will provide adequate sample volume under the low discharge estimate or if it will overflow the sample bottle under the high discharge estimate, calculations are made under both scenarios. At a low flow of 2 million gallons, the sampler will collect 10 samples for a total volume of 6.4 L (1.7 gallons). This is sufficient for aquatic toxicity testing but may be insufficient for other chemical analyses. At a high flow of 5 million gallons, the sampler will collect 25 samples for a total volume of 16 L which will overfill the sample bottle. Based on this, the sample volume should be decreased to 0.6 L per 200,000 gallons of flow resulting in sample volumes of 6 L (1.5 gal) and 15 L (4 gal), both of which are acceptable. lar to the above method, critical information in developing a sampling protocol is 1) volume of total sample required and 2) estimated stormwater flow for sampling period. Under this sampling regime 0.3 L are collected for every 100,000 gallons of discharge. Thus, under average flow conditions (208,333 gph), the sampler will collect 0.34 L of sample every 28.8 minutes. Similar to the previous method, if stormwater flows are under or overestimated, then either too much sample will be col- lected (flows are underestimated and the sample bottle is overfilled) or insufficient sample will be collected. 4.2 Test Solution Renewal 4.2.1 Background Aquatic toxicity test protocols require that sample test solu- tions be renewed at a minimum of every 48 hours for the dura- tion of the test. Samples can be renewed with the existing original sample or can be renewed with a freshly collected sample. Daily test renewal with freshly collected sample is recommended based on the following: Unless demonstrated otherwise, it should be assumed that stormwater flow rates and concentrations are variable. Continued exposure of the test organisms to the originally collected stormwater at the 24-hour period is unlikely to

27 test species avoids uncertainty regarding changes in toxicity relative to differences in test temperatures and field exposure temperatures. 4.3.2 Recommendation If toxicity is observed using warm-water test species, con- sider acclimating test organisms to lower temperatures prior to conducting successive tests or consider testing with a cold- water test species to confirm the potential for instream toxic- ity. Prior to initiating tests, the proposed approach should be discussed with the permit writer to obtain consensus on the acclimation procedure, test method, and test species. 4.4 Dissolved Oxygen Monitoring 4.4.1 Background Discharges of stormwater impacted by deicing operations may contain elevated levels of oxygen demanding substances (measured as either BOD or COD). As these substances degrade, DO is consumed from the water solution. Because toxicity tests are conducted at an elevated temperature com- pared to field conditions, the degradation rate of these sub- stances is increased. Thus, DO levels may decrease rapidly in the test solutions. EPA protocols require that the DO concentration be main- tained above 4.0 mg/L for warm-water species and 6.0 mg/L for cold-water species. The laboratory is to monitor DO concen- trations during the first several hours of the test to determine if the test solution DO concentrations are likely to decrease below the required concentration. If this occurs, the laboratory is to aerate the samples. 4.4.2 Recommendation Often testing laboratories have minimal information regarding the potential chemical concentrations of samples collected for aquatic toxicity testing. Thus, the laboratory should be notified that the sample may contain elevated con- centrations of oxygen demanding substances and increased DO monitoring is required. Further, the laboratory should be notified that if DO monitoring indicates that DO concentra- tions may decrease below acceptable levels, the EPA protocol is to be followed regarding aeration of samples. In addition to the above, daily test solution renewal can minimize decreases in DO concentrations. Upon receipt of testing data, the raw data should be reviewed to determine if DO levels were maintained at acceptable con- centrations. Should dissolved concentrations fall below accept- able values, the test should be considered invalid because it does not meet quality assurance requirements. be representative of discharge conditions. Further, renewal of the test solution after 48 hours with the originally col- lected sample is also unlikely to be representative due to changes in stormwater composition over time as materials are washed from the airfield, diluted, and degraded. Test solution renewal minimizes the potential for decreases in DO during the test, which could disqualify the test or stress the test organisms. Stormwater discharges may cease shortly after the end of the precipitation event. Under this condition, the test could be renewed with the original sample or the test can be renewed with laboratory or receiving water. If the objective of the test is to characterize the potential for toxicity within the receiv- ing water, then the continued use of the originally collected sample when there is no discharge is not environmentally representative of field exposure conditions. The use of labo- ratory or dilution water for renewals when the stormwater discharge has ceased should be discussed with and agreed upon by the regulatory agency. 4.2.2 Recommendation Samples of stormwater should be collected on a daily basis and utilized to renew the test solution. If there is no discharge from the stormwater outfall, then, to be environmentally rep- resentative of field exposure conditions, the test should be renewed with laboratory dilution water or with a freshly col- lected sample of the receiving water. Note, however, state regu- latory agencies such as California’s SWRCB (2011) may have specific requirements regarding the renewal of toxicity tests on stormwater discharges. Thus, a clear understanding of how samples will be collected and utilized to renew the toxicity test solutions, especially when there is no discharge, should be agreed upon and documented. 4.3 Temperature 4.3.1 Background Aquatic toxicity tests using warm-water species are con- ducted at temperatures between 20°C and 25°C, however, receiving water temperatures under deicing conditions may approach 0°C. Thus, there is a significant difference between test conditions and instream exposure conditions. Limited testing of warm-water species under standard (20°C and 25°C) and reduced (6°C–15°C) temperature conditions indi- cate that toxicity for C. dubia may be unchanged or slightly reduced at lower test temperatures compared to standard test temperatures. EPA guidance allows the use of cold-water spe- cies for toxicity testing. No testing has been conducted to date indicating whether the cold-water species exhibit differences in toxicity based on temperature. Yet, the use of cold-water

28 While the initial collection of data may not provide a sig- nificant amount of information, establishing a baseline of stormwater composition and its associated aquatic toxic- ity will provide information for comparison when a sample is toxic. Specifically, comparison of constituent concentrations for non-toxic and toxic samples can allow some constituents to be ruled out and others to be identified as potential toxicants. 4.6 Receiving Water and Discharge Flow Analysis 4.6.1 Background If aquatic toxicity tests indicate that the sample was acutely or chronically toxic and failed to meet the permit limits, the test results should be reported as required. However, additional data should be collected to evaluate the potential for environmental impact. As discussed in the previous sections, the permit writer should consider the discharge flow, receiving water flow, and presence of regulatory mixing zones in calculating permit lim- its. If no mixing zones are allowed by state regulations, then all limitations must be met at the end of the pipe. However, if mix- ing zones are allowed, the volume of water discharged and the volume of water available in the receiving water are important. In the development of permit limits, the discharge and receiv- ing water flows are typically established at critical levels, such as the 10-year, 24-hour precipitation event for the discharge flow and the 7Q10 flow for the receiving water flow. Under these con- ditions of a low receiving water flow and a high discharge flow, dilution within the receiving water will be minimal, resulting in high exposure conditions. In the event of a toxic discharge, actual flow data will allow an analysis of predicted exposure conditions at the edge of the mixing zones. These exposures can be compared to the resulting data to determine if instream toxicity would be predicted. While this does not negate a permit violation, it allows instream impacts to be estimated. 4.6.2 Recommendation If available, receiving water flow data can be obtained from the nearest USGS monitoring stations. This data can typically be downloaded directly from the Internet (http://waterwatch. usgs.gov/?id=ww_current). However, flow monitoring stations may not be located on all receiving waters. Weather event infor- mation can be obtained from local NOAA weather stations. This data can be utilized to characterize the storm in terms of inten- sity and precipitation type and estimate the volume of discharge from stormwater outfalls based on watershed basin character- istics. If flow rates are measured, actual stormwater flow rates can be utilized and compared to those flow rates utilized to establish permit limits. The combination of this data can be utilized to determine instream concentration after mixing in the receiving water. 4.5 Concurrent Monitoring 4.5.1 Background The conduct of aquatic toxicity testing on stormwater dis- charges provides an indication of whether the sample is acutely or chronically toxic to aquatic organisms. However, the test does not provide an indication of what may be contributing to aquatic toxicity. Should the stormwater be consistently toxic, it is likely that the discharger will be required to implement a TIE study. To facilitate test interpretation, additional data should be collected for each sample submitted to the laboratory for toxic- ity testing. Upon receipt of results, the data should be analyzed to identify common trends. For example, concentrations of BOD can be evaluated to identify concentrations of BOD that are always associated with toxicity and these concentrations of BOD that are always associated with a non-toxic sample. While this does not identify the toxicant, it can provide a “fingerprint” of a toxic sample. Further, this information can be utilized to identify corrective actions and formulate a basis of design for stormwater management systems. 4.5.2 Recommendation Water chemistry data collected as part of the WET testing protocol consist of DO, pH, temperature, conductivity, hard- ness, alkalinity, chlorine, and ammonia. These data are typically collected as part of the aquatic toxicity test protocol. Other data may or may not be required to be collected in the permit as part of the WET testing program. Additional data recommended for collection consist of the following constituents: COD. This is the amount of oxygen required to chemically oxidize organic material such as glycol, acetate and formate as well as ammonia and nitrate nitrogen. Biochemical oxygen demand (BOD5). This is the amount of DO required to biologically degrade the sample and is indic- ative of organic pollutants present in the stormwater runoff. BOD is a component of COD. Ethylene and propylene glycol. These are the active ingredients in both Type I and Type IV ADFs. Stormwater impacted by aircraft deicing operations is likely to contain residual con- centrations of these chemicals. Calcium, sodium, potassium, magnesium. These compounds are conservative pollutants (meaning that they do not degrade) associated with pavement deicers. Elevated con- centrations can be directly toxic to aquatic organisms as well as exert osmotic stress. Conductivity. This is a measure of the ion concentration of a water sample. Conductivity values greater than 3,000 µS/cm may contribute to aquatic toxicity of sensitive freshwater organisms.

29 Are test solutions renewed at least every 48 hours? Does the dose-response curve demonstrate an expected response in which higher concentrations exhibit a higher response? Does the reference toxicity test fall within acceptable labora- tory levels? 4.9 TIE Procedures 4.9.1 Background NPDES permits typically require the initiation of a TIE study should the discharge exhibit toxicity for multiple samples. Spe- cifically, if a stormwater sample fails to meet permit limits (e.g., exhibits toxicity in excess of permit limits), the permit typi- cally requires retesting within several weeks. Should the second sample exhibit toxicity, a third retest may be required. If the third or subsequent samples also exhibit toxicity, a TIE is typi- cally required to be implemented. A TIE is a series of tests designed to alter or render biologi- cally unavailable a group of toxicants coupled with aquatic toxicity testing to monitor changes in toxicity associated with modified samples. Although the specific chemical toxicant may not be identified using this methodology, the chemical and physical characteristics of the toxicant can be sufficiently described such that treatment or control technologies can be identified. Guidance on the conduct of TIE studies can be found in the following documents: EPA. 1991a. Methods for Aquatic Toxicity Identification Eval- uations: Phase I Toxicity Characterization Procedures, Sec- ond Edition (EPA-600-R-91-003). EPA. 1989. Generalized Methodology for Conducting Indus- trial Toxicity Reduction Evaluations (TREs). EPA-600-2- 88-070. EPA. 1993a. Methods for Aquatic Toxicity Identification Evaluations: Phase II Toxicity Identification Procedures for Samples Exhibiting Acute and Chronic Toxicity. EPA- 600-R-92-080. EPA. 1993b. Methods for Aquatic Toxicity Identification Evalua- tions: Phase III Toxicity Confirmation Procedures for Samples Exhibiting Acute and Chronic Toxicity. EPA-600-R-92-081. Challenges associated with the conduct of TIEs on storm- water consist of the following: Variability in storm event conditions that affect stormwater quality. As a result of differences in storm events, the chemi- cal constituents may change. The changing chemical char- acteristics of a stormwater will increase the time and cost required for successful completion of a TIE. 4.7 Material Application Rates 4.7.1 Background The composition of stormwater impacted by deicing oper- ations is a function of the location, types, and amounts of material applied on the airport. To facilitate data interpreta- tion, the amount and type of deicing applied and the loca- tion of application of the material should be documented. This data, in conjunction with collected concurrent monitor- ing data (Section 4.5) will facilitate data interpretation and provide insight as to why some samples or drainage basins may have stormwater discharges that are toxic whereas other basins have discharges that are not toxic. In addition to application data for each drainage basin, information regarding collection of materials within each basin will allow estimates of total material discharged from the drainage basin. 4.7.2 Recommendation Information should be collected regarding the time of application of deicing materials, the location of application, and quantities of materials collected. Data should be analyzed by drainage basin with a focus on those basins that are being sampled for aquatic toxicity. 4.8 Toxicity Test Data Review 4.8.1 Background Toxicity tests are to be conducted in accordance with protocols established in 40 CFR 136. Although the aquatic toxicity testing laboratory should have a QA/QC program, it is recommended that all data and particularly data in which toxicity is identified should be reviewed by the air- port environmental manager prior to acceptance of the test data. Points of deviation should be identified and discussed with the laboratory prior to reporting the data to regulatory agencies. 4.8.2 Recommendation The following information and test conditions should be reviewed: Is sample hold time acceptable (<36 hours)? Are test temperatures within acceptable ranges and do not vary by more than 3°C? Are DO levels maintained above 4 mg/L (warm-water test species) and 6 mg/L (cold-water test species)? Is the age of test organisms within acceptable standards? Is the control survival greater than 90%?

30 cally expressed in terms of mass of product per volume of water (i.e., mg/L). However, the toxicity values can be expressed in terms of BOD, propylene glycol, or other constituents using data from the SDS. For example, if the BOD of the product is known, the LC50 value can be converted from mg of product per L to mg of BOD associated with the product per L. The revised product toxicity values can then be compared to val- ues monitored in the stormwater discharge to determine if the chemical product has the potential to contribute to toxicity. For example, if the LC50 for the product expressed in terms of BOD is 5,000 mg/L and the sample tested for aquatic toxicity contained a concentration of more than 5,000 mg/L of BOD, then the product can be considered a potential toxicant in the stormwater sample. The conclusions of the above data evaluations will influ- ence the design of the TIE studies. For example, targeted sampling of stormwater with a COD or propylene glycol concentration greater than a specific amount can be insti- tuted to further confirm the relationship between the moni- tored analyte and toxicity. In addition, these analyses may indicate levels of COD (or BOD) that must be achieved to minimize the potential for the discharge of a toxic effluent. Finally, as noted above, EPA has developed extensive guid- ance for the conduct of a TIE. However, these studies should be site-specific in nature and tailored to a specific site based on historical data. Stormwater consists of a mixture of a large number of poten- tial chemical contaminants. As noted above, these contam- inants may change from storm event to storm event. 4.9.2 Recommendation Procedures for the conduct of a TIE are well established. However, the first phase of a TIE should consist of a thorough evaluation of historical data collected with respect to airport discharges and a review of the chemical safety data sheets for products utilized or applied on the airfield. Specifically, historical data collected as described in Section 4.6 above will provide an initial first step in identifying differences between the chemical composition of toxic and non-toxic samples. Further, differences in toxicity and chemical composition between stormwater draining into different stormwater basins can be investigated. In addition to comparisons between samples and drainage basins, an evaluation of chemical safety data sheets (SDS), previously known as material safety data sheets (MSDS), for products utilized on the airfield should be conducted to deter- mine product toxicity and the potential for each product to be discharged in concentrations that would contribute to toxicity observed in the stormwater discharge. With respect to deicing materials, the SDS provides a summary of the toxicity of the product to aquatic organisms. These toxicity values are typi-