Guidance for Treatment of Airport Stormwater Containing Deicers (2013)

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.

D-1 Using Appendix D Appendix D contains summaries of the deicer treatment experiences at 15 airports, listed in the following, that have used the 11 deicer treatment technologies referenced in the fact sheets. Bradley International Airport (reverse osmosis) Nashville International Airport (aerated lagoon) Buffalo Niagara International Airport (aerated gravel bed) Akron–Canton Airport (anaerobic fluidized bed reactor) Westover Air Force Reserve Base (passive facultative treatment) Cincinnati/Northern Kentucky Airport (activated sludge, mechanical vapor recompression) Denver International Airport (mechanical vapor recompression, distillation, public wastewater treatment system) Detroit Metropolitan International Airport (private off-site recycling, public wastewater treatment system) Wilmington Airpark (aerated gravel beds) London Heathrow (passive facultative treatment, aerated gravel beds) Oslo Gardermoen (moving bed biofilm reactor) Portland International Airport (anaerobic fluidized bed reactor, public wastewater treatment system) Edmonton International Airport (passive facultative treatment, aerated gravel beds) Halifax International Airport (mechanical vapor recompression) Zurich International Airport (passive facultative biological treatment) The Appendix D airport summaries contain the following categories of information. A P P E N D I X D Airport Deicer Treatment System Summaries SUMMARY SECTION CONTENT DESCRIPTION Treatment Technology Category One or more of the 11 treatment technology categories for which treatment technology fact sheets are prepared. Years Operated The number of years that the treatment system has been operational. Deicer Management System Description A brief overview of the entire deicer management system used at the airport, including deicer collection, conveyance, storage, treatment, and disposal. Deicer Treatment Technology Selection Considerations Summary of the history of treatment technologies used at the airport, including a description of the considerations the airport used in selecting its current technology.

D-2 Guidance for Treatment of Airport Stormwater Containing Deicers When assessing the actual performance of specific deicer treatment systems included in the air- port summaries, an attempt was made by the research team to consider the numeric performance indicators in the context of the system operation conditions. It is recommended that guidebook users who seek out information from other airports’ systems use similar caution. Many factors can affect treatment system performance on any given day or in any given season, and these con- ditions may be important for correctly interpreting operating data for the purposes of assessing a particular treatment technology’s capabilities and suitability for other applications. Some of these factors are: • Stormwater characteristics at the time of treatment (e.g., temperature, pH, nutrient content), • Effluent limits the airport is trying to meet, • Operational decisions on process settings, • Maintenance issues, • Whether the system is in the start-up portion of the season or in mid-season, and • The actual loading or flow rate compared to the system load or flow rate capacity (i.e., if the treatment system is being underloaded or overloaded). It is suggested that when an airport team uses information from other airport deicer treat- ment systems to supplement its own evaluation, the stakeholders should come to a common understanding of the basis for the performance assessments for the other airports’ treatment systems. Deicer Treatment Technology Description A technical description of the treatment technology as used at the airport, including sizing information and a description of the treatment technology’s support systems. Treatment System Performance Numeric description of the design and actual performance of the treatment system at the airport. Cost Assessment for Treatment System Presentation of the capital and operations and maintenance costs for the portion of the system associated with the treatment technology. In some cases, capital costs for the treatment portion of the system were not specifically determined by the airport and costs for the broader deicer management system are presented. Conclusions on Performance for Treatment System A discussion of the treatment system performance in relation to its design intent. Lessons Learned for Potential Implementation of Treatment System at Other Airports Lessons learned from the operation of the system that may be applicable to others.

Airport Deicer Treatment System Summaries D-3 Airport Treatment Summary No. 1 Airport: Bradley International Airport—Windsor Locks, CT (BDL) Treatment Technology: Reverse Osmosis and Mechanical Vapor Recompression Years Operated: 2006–2012 (Currently Operational) Deicer Management System Description The BDL deicer management system uses a passive and active collection system, collection basins, piping, pump stations, a recycling facility, and a POTW for discharging wastewater. The recycling facility uses RO and MVR treatment technologies. The passive and active collection system in place is used for the capture of spent aircraft deicing fluid. Deicing operations are conducted at the terminal gates, freight/remote parking areas, and the remote deicing facil- ity (RDF). Active collection involves the use of glycol recovery vehicles (GRVs) at designated gate areas. Passive collection involves the use of dedicated glycol collection drainage systems for both the terminal gate areas and the RDF. Ultimately, all spent ADF captured is sent to two storage tanks (each with 1 million gallons of capacity) located at the on-site recycling facility. The two storage tanks act as the interim storage and feed reservoir for the glycol processing activities. Spent ADF is segregated according to glycol concentration. One storage tank is designated for high-concentration propylene glycol that is 4% and higher, while the other million-gallon tank is designated for low-concentration propylene glycol of less than 4%. All of the spent-ADF processing equipment is housed in two buildings. One building houses the MVR equipment, while the other houses the chemical pretreatment and membrane systems. The membrane systems include ultrafiltration (UF) and RO. The entire spent-ADF management system is operated to ensure that unpermitted levels of glycol do not enter the stormwater system and to comply with Consent Order #WC5727 that was issued in 1998 by the Connecticut Department of Environmental Protection. All wastewater generated from the on-site treatment systems is discharged to an off-site wastewater treatment plant called the Metropolitan District Poquonock Water Pollution Control Facility. Figure 1 demonstrates the deicer management system at BDL. Figure 1. BDL spent-ADF management system process flow diagram.

D-4 Guidance for Treatment of Airport Stormwater Containing Deicers Deicer Treatment Technology Selection Considerations The on-site recycling system as a whole at BDL was designed, installed, and implemented to meet the following requirements: • Compliance with federal and state environmental regulations, Connecticut Department of Energy and Environment orders, and wastewater discharge/pretreatment permits at the airport. • On-site treatment that generated wastewater requires compliance with the following: maxi- mum discharge of 288,000 gallons of wastewater per day, limitations of 125-mg/L propylene glycol, 200-mg/L BOD5, 600-mg/L COD, a pH of 6.0–10.0, and 125-mg/L TSS. • Equipment with the ability to conduct glycol processing at an average production/removal rate of 600,000 gallons per month when spent-ADF volumes are present. Minimum volume of 100,000 gallons in each of the high- and low-concentration storage tanks before systems have to be started. • Empty storage tanks on or before September 1 each year. The RO and MVR treatment technologies were specifically selected because: • The combined technologies are able to handle fluctuating glycol concentrations in spent ADF that occur with each weather-related deicing event. • The systems could be separated into two independent processing trains capable of recycling: – Propylene glycol of 0.1% to 4% concentration through the one tank. – Propylene glycol of greater than 4% concentration through the second tank. • All glycol captured above 0.1% in concentration could be recycled. • The glycol that is reclaimed from the system is sold, and the revenues generated are used to offset program costs to provide glycol management services. Deicer Treatment Technology Description The BDL treatment system employs both the RO and MVR treatment processes. The deic- ing treatment system was designed to operate both of these systems simultaneously. Descrip- tions of the MVR and RO treatment technologies can be found in Fact Sheet 106 and Fact Sheet 111, respectively. See Figure 2 for a photograph of the low-concentration processing building. Figure 2. Low-concentration processing building.

Airport Deicer Treatment System Summaries D-5 Description of Support Systems The support systems at BDL for the RO and MVR treatment technology are diversion and storage, pumping systems, chemical pretreatment, and a storage tank for the recycled glycol. The passive system for ADF collection at BDL includes a diversion structure and pump sta- tion to move fluid from terminal areas to the recycling facility. Incorporated in this conveyance system is underground piping to allow testing of fluid so that spent ADF can be appropriately directed to recycling storage tanks based on glycol concentration. The main storage reservoirs for glycol recycling activities are two 1-million-gallon tanks. A pretreatment system was installed prior to the UF and RO membrane systems to treat all diluted spent-ADF fluid. The constituents in the feed are analyzed to determine which chemi- cal additives will perform best. In the pretreatment tank, chemical pretreatment is carried out to remove undesirable constituents from the waste fluid. The tank consists of a water-softening system and a mixed-reaction tank with pH control and chemical addition. After the influent is treated with the chemicals, it is transferred to a series of settlement tanks, where the chemically precipitated constituents of the waste stream are allowed to precipitate and settle. This material is removed from the system prior to passing from the pretreatment tank to the UF system. The UF system is used to remove constituents that may foul the RO membrane. All recycled glycol at a 50% concentration is temporarily stored in two double-walled 20,000-gallon storage tanks. The 50% glycol is shipped to an off-site centralized distillation system where it is recycled to a 99%+ concentration before it is sold. All solid waste and membrane wash fluid is temporarily stored on-site and then shipped to an approved waste disposal facility Key Treatment Sizing Parameters See Table 1 through Table 3 for system sizing parameters. Treatment System Performance Although the RO and MVR systems operate simultaneously, they have different design requirements. Therefore, Table 4 and Table 5 reflect the design parameters for RO and MVR separately. The membrane systems were designed to meet the required removal rates in Table 4. The membrane systems were specifically configured to accomplish two tasks: 1. Treat influent streams with glycol concentrations from 0.1% to 4% 2. Treat all water produced from both MVR and membrane operations to ensure water quality levels meet sanitary discharge permit requirements. Component/Parameter Size/Capacity of Treatment Units Number of Treatment Units Total Capacity Stormwater storage capacity 1 million gallons 1 million gallons 2 2.0 million gallons Treatment unit volume RO MVR 1,060 ft³ 980 ft³ 1 4 1,060 ft³ 3,920 ft³ Treatment unit dimensions RO MVR 22-ft L x 6-ft W 20-ft L x 6-ft W 1 4 Total area: 612 ft2 Treatment facility footprint 0.11-acre building 0.04-acre building 2 0.15 acres Table 1. Treatment system size and capacity parameters.

D-6 Guidance for Treatment of Airport Stormwater Containing Deicers Table 5 summarizes the treatment design for a single MVR treatment unit. Typically, a treat- ment system that uses MVR technology would use sufficient MVR units to meet the needs of the airport system. Additional design parameters for the RO treatment processes are provided in Table 6 and Table 7. The MVR concentrators were designed to meet the required removal rates in Table 5. The primary function of these systems is to concentrate all collected spent glycol to a minimum of concentration of 50%. See Table 8 for additional MVR system design parameters. Component/Parameter Size/Description of Treatment Units Number of Treatment Units Total Value Stormwater storage capacity (low-concentration PG tank) 1 million gallons 1 1 million gallons Annual chemical pretreatment rate 7.956 million gallons 1 7.956 million gallons Annual UF treatment rate 2 >10.5 million per year Annual RO treatment rate >10.5 million gallon 1 >10.5 million gallon Support system dimensions: Chemical pretreatment: UF 1 UF 2 Process tank UF poly tank RO poly tank 9’8” L x 6’ W x 6’ H 11’L x 6’ W x 7’ H 9’ L x 7’ H x 5’ W 20’ L x 9’ H x 10’ W 4,000 gallons 2,000 gallons 1 1 1 6 1 1 10,800 ft3 4,000 gallons 2,000 gallons Treatment building footprint (RO building) 100’ W x 50’ L x 22’ H 1 5,000 ft2 Wastewater discharge tanks 20’ L x 9’ H x 10’ W (13,000 gallons) 2 3,600 ft3 (26,000) gallons Table 2. Additional system sizing parameters for RO (low-concentration treatment facility). Component/Parameter Size/Description of Treatment Units Number of Treatment Units or Capacity Total Value Stormwater storage capacity (high-concentration PG tank) 1 million gallons 1 1 million gallons Annual MVR treatment rate 1.2 million gallons 4 4.8 million gallons Support system dimensions: MVR feed tanks (spent-ADF storage) Product storage tanks (recycled glycol) 13,000 gallons 20,000 gallons 2 2 26,000 gallons 40,000 gallons MVR treatment unit dimensions L = 20’, W = 6’, H= 8’2 with scrubber 22’H 4 Treatment facility footprint (MVR building) L= 60’, W= 32’, H = 22’ 1 1,920 ft2 Table 3. Additional system sizing parameters for MVR (high-concentration treatment facility).

Airport Deicer Treatment System Summaries D-7 Parameter Value Unit Design flow rates - Minimum - Average - Maximum 40 Not available 50 Gallons per minute Design treatment load capacity 51,000* 30,000** 30,000 lbs COD/day lbs BOD5/day lbs PG/day Design influent concentration - Range 0~87,000* 0~50,000** 0~50,000 mg COD/L mg BOD5/L mg PG/L Design effluent concentration (average) 50~450 Not available 0~1000 mg COD/L mg BOD5/L mg PG/L Design treatment efficiency 99.5 % influent COD load treated *Data based on conversion: [COD] = 1.7 [BOD5]. **Data based on conversion: [PG] = [BOD5]. Table 4. Design basis for RO system performance. Parameter Value Unit Design flow rates - Minimum - Average - Maximum 2 Not available 4 Gallons per minute Design treatment load capacity 16,500*† 9,700** 9,700 lbs COD/d lbs BOD5/day lbs PG/day Design influent concentration - Range 17,000~459,000* 10,000~270,000** 10,000~270,000 mg COD/L mg BOD5/L mg PG/L Design effluent concentration (average) <50~1000 Not available <50~1000 mg COD/L mg BOD5/L mg PG/L Design treatment efficiency 94.1~99.7 % influent COD load treated *Data based on conversion: [COD] = 1.7 [BOD5]. **Data based on conversion: [PG] = [BOD5]. † Data reflect absolute maximum. Typical maximum loads are 12,000-lbs COD/day. Table 5. Design basis for MVR system performance. System Design Criteria UF Unit #1 UF Unit #2 Influent flow rate range (gpm) 10 to 16 20 to 60 Influent glycol concentration range (% PG) 0%–5% 0%–5% Influent temperature range (°F) 40 to 100 40 to 100 Influent TSS (NTU) 200 200 Influent pH 3 to 11 2 to 13 Effluent process fluid TSS (NTU) <15 NTU <40 NTU Effluent process fluid flow rate range (gpm) <16 <60 Effluent temperature range (°F) 40 to 100 40 to 100 Note: gpm = gallons per minute; NTU = nephelometric turbidity units. Table 6. Additional design basis for UF system.

Number of stages 2 Constant flow or batch Constant flow Influent pressure range (psi) 575 to 625 Influent flow rate range for stage 1 (gpm) <120 Influent glycol concentration range for stage 1 0% to 5% Influent temperature range (°F) 40 to 100 Effluent stage 1 permeate (PG) 0% to 0.5% Effluent stage 1 permeate (pH) 3 to 7 Effluent stage 1 permeate flow rate range (gpm) 40 to 50 Effluent stage 1 reject (PG) 1% to 5.0% Effluent stage 1 reject (pH) 6 to 7 Effluent stage 1 reject flow rate range 7 to 15 Influent stage 2 flow rate range (gpm) 40 to 50 Effluent stage 2 permeate (PG) 0% to 0.1% Effluent stage 2 permeate (pH) 6 to 7 Effluent stage 2 permeate (COD mg/L) 50 to 450 Effluent stage 2 permeate (TSS) <10 NTU Effluent stage 2 permeate flow rate range (gpm) 15 to 40 Effluent stage 2 reject (PG) 0.1% to 0.4 % Effluent stage 2 reject (pH) 6 to 7 Effluent stage 2 reject (COD) N/A Effluent stage 2 reject (TSS) <10 NTU Effluent stage 2 reject flow rate range 18–25 Estimated of waste produced (per gal) N/A Temperature range requirement 40°F–100°F Table 7. Additional design basis for RO system. Table 8. Additional design basis for MVR system. Parameter Single-Stage Production Two-Stage Production Stage 1 Stage 2 Influent flow rate range (gallons per hour) 150 to 200 170 to 230 130 to 170 Influent glycol concentration range (% glycol) 4 to 27 1 to 4 13 to 27 Influent temperature range (F or C) Ambient Ambient Ambient Number of effluent streams produced 2 streams– distillate and concentrate 2 streams– distillate and concentrate 2 streams– distillate and concentrate Distillate effluent flow rate range (gallons per hour) 60 to 184 136 to 219 52 to 126 Distillate effluent water quality (COD range in mg/L) <50 to 1,000 <50 to 1,000 <50 to 1,000 Distillate effluent water quality (pH range) 3 to 8 3 to 8 3 to 8 Concentrate effluent flow rate range (gallons per hour) 12 to 120 8.5 to 61 33 to 102 Concentrate effluent concentration (% glycol range) 50 to 55 15 to 20 50 to 55 Heat source Electric-powered steam compression Control system PLC Energy consumption information 0.4Kw per gal feed Estimate of waste to be produced Sludge and solids negligible, and glycol in overheads less than 0.1% Anticipated frequency of maintenance activities Duty cycle of 95% expected, depending on influent quality Footprint, dimensions, etc. Each MVR unit is 20’ (L) x 6’ (W) x 8’ 2” (H), with scrubber 13’ (H) or 22’ (H) Other support systems Feed preheater heat exchanger, electric air compressor, cold and hot filter systems, piping for feed, distillate, concentrate, and storage tanks for feed, distillate, and concentrate

Airport Deicer Treatment System Summaries D-9 Parameter Value Unit Flow rates - Minimum - Average - Maximum 0.3 9.2 31 Gallons per minute Actual COD treatment load rate - Average - Maximum 2,700* 8,100* lbs/day Actual BOD5 treatment load rate - Average - Maximum 1,600** 4,800** lbs/day Actual PG treatment load rate - Average - Maximum 1,600 4,800 lbs/day Influent COD concentration - Minimum - Average - Maximum 6,300* 23,800* 34,200* mg/L Influent BOD5 concentration - Minimum - Average - Maximum 3,700** 14,000** 20,100** mg/L Influent PG concentration - Minimum - Average - Maximum 3,700 14,000 20,100 mg/L Effluent COD Concentration - Minimum - Average - Maximum 5 157 430 mg/L Effluent BOD5 concentration - Minimum - Average - Maximum 2 81 230 mg/L Effluent PG concentration - Minimum - Average - Maximum 1 27 120 mg/L Treatment efficiency - Minimum - Average - Maximum 95.4 99.0 99.9 % influent COD load treated *Data based on conversion: [COD] = 1.7 [BOD5]. **Data based on conversion: [PG] = [BOD5]. Table 9. Actual RO system performance. The information in Table 9 on actual system performance was derived from monthly average data collected at the facility between 2009 and 2012. The information in Table 10 on actual system performance was derived from monthly average data collected at the facility between 2009 and 2012. Cost Assessment for RO and MVR Treatment System The Connecticut DOT paid for and installed the RO system, the low-concentration process- ing building, the two 1-million-gallon storage tanks, associated pumping stations, and the 11 process tanks. The recycling vendor installed the chemical pretreatment system, the two UF units, the MVR building, four ADF concentrators, and two concentrate product storage tanks, and upgraded the RO system (state owned). The recycling vendor is responsible for the maintenance and operation of all equipment associated with the processing of spent ADF.

D-10 Guidance for Treatment of Airport Stormwater Containing Deicers Conclusions on Performance of BDL RO and MVR Treatment System Influent Deicer Concentrations The influent deicer concentration is a primary factor in the design and operation of RO and MVR treatment systems. While dilute concentrations of PG-affected stormwater can be treated by an RO/MVR treatment system, the RO/MVR treatment system performs better with higher influent deicer concentrations. Therefore, it is beneficial to operate the collection system in a manner that provides high influent deicer concentrations to the RO/MVR treatment process. Influent deicer concentrations of less than 40,000-mg PG/L are concentrated using the RO process. Concentrate from the RO process and influent deicer concentrations of greater than 40,000-mg PG/L are treated by the MVR. The RO/MVR treatment system at BDL has treated concentrations as low as 3,700-mg PG/L and as high as 105,000-mg PG/L. These concentrations fall well within the design concentrations of 0-mg to 270,000-mg PG/L. To prevent fouling of the RO membrane, the RO process requires pretreatment of the influ- ent deicing-affected stormwater by the UF processes. The UF systems are a very important part Parameter Value Unit Flow rates - Minimum - Average - Maximum 1.7 9.0 18 Gallons per minute Actual COD treatment load rate - Average - Maximum 3,000* 13,000* lbs/day Actual BOD5 treatment load rate - Average - Maximum 1,800** 7,650** lbs/day Actual PG treatment load rate - Average - Maximum 1,800 7,650 lbs/day Influent COD concentration - Minimum - Average - Maximum 45,900* 88,700* 178,500* mg/L Influent BOD5 concentration - Minimum - Average - Maximum 27,000** 52,000** 105,000** mg/L Influent PG concentration - Minimum - Average - Maximum 27,000 52,000 105,000 mg/L Effluent COD concentration*** - Minimum - Average - Maximum Not available Not available Not available mg/L Effluent BOD5 concentration*** - Minimum - Average - Maximum Not available Not available Not available mg/L Effluent PG concentration*** - Minimum - Average - Maximum Not available Not available Not available mg/L Treatment efficiency*** - Minimum - Average - Maximum Not available Not available Not available % influent COD load treated *Data based on conversion: [COD] = 1.7 [BOD5]. **Data based on conversion: [PG] = [BOD5]. ***Effluent from the MVR is not monitored since it is sent to the RO treatment system. Table 10. Actual MVR system performance.

Stage 1 Oct Nov Dec Jan Feb Mar Apr May Jun Jul Total or Average for Season Influent volume processed (gal) 20,273 75,812 625,786 602,007 300,774 1,326,694 141,721 0 139,206 29,738 3,262,011 Average influent glycol concentration (% PG) 0.40% 0.37% 0.75% 1.30% 1.30% 1.30% 1.30% 0 1.36% 1.40% 1.17% Average influent (pH) 5.6 5.8 4.85 4.22 6.8 9.79 6.8 0 6.24 6.38 5.6 Average influent temperature (Fahrenheit) 65.8 71.5 60.77 60.18 58.8 66.36 76.2 0 76.11 85.00 60.43 Effluent volume of permeate produced (gal) 18,597 69,544 547,608 490,290 241,904 1,018,574 127,558 0 116,040 26,006 2,656,121 Average effluent permeate (pH) 5.65 8.33 7.07 7.48 8.60 9.56 7.53 0 7.93 6.18 7.6 Average effluent permeate (% PG) <0.5% <0.5% <0.5% <0.5% <0.5% <0.5% <0.5% 0 <0.5% <0.5% <0.5% Effluent volume of reject produced (gal) 1,676 6,268 78,178 111,717 58,870 308,120 14,163 0 23,166 3,732 605,890 Average effluent reject (% PG) 5.28% 5.86% 5.32% 5.30% 5.70% 5.64% 5.38% 0 0.00% 5.82% 5.32% Stage 2 Influent volume processed (gal) 18,597 69,544 547,608 490,290 241,904 1,018,574 127,558 0 116,040 26,006 2,656,121 Average influent glycol concentration (% PG) <0.5% <0.5% <0.5% <0.5% <0.5% <0.5% <0.5% 0% <0.5% <0.5% <0.5% Average influent (pH) 5.65 8.33 7.07 7.48 8.6 9.56 7.53 0 7.93 6.18 7.6 Average influent temperature (Fahrenheit) 63 67.3 56.2 57.7 55.6 64 70.7 0 78.1 80.8 65.9 Effluent volume of permeate produced (gal) sanitary 9,287 42,807 325,140 297,524 137,577 639,978 82,057 0 68,937 16,205 1,619,512 Average effluent permeate (pH) 8.2 8.5 7.42 7.78 8.46 9.02 7.96 0 7.27 6.3 7.9 Average effluent permeate (PG ppm) 1 1 3.2 1 1 16.8 23 0 16 27 9.7 Average effluent permeate (BOD mg/L) 3 2 163 64.5 86.25 29 100 0 94.5 230 74.8 Average effluent permeate (COD mg/L) 230 290 88.3 5 130 44 107.5 0 110 320 69.4 Average effluent reject (% PG) 0.5%–1.0 % 0.5%– 1.0 % 0.5% - 1.0 % 0.5%–1.0 % 0.5%–1.0 % 0.5%–1.0 % 0.5% - 1.0 % 0 0.5%– 1.0 % 0.5%– 1.0 % 0.5%– 1.0 % Effluent volume of reject produced (gal) 9,310 26,737 222,468 192,766 104,327 378,596 45,501 0 47,103 9,801 1,036,609 Overall ratio of pure PG removed by both stages 98.85% 98.47% 97.78% 99.62% 99.65% 93.77% 89.76% 0 94.17% 89.49% 95.89% Average amount of waste produced month (gal) 5,000 5,000 10,000 15,000 0 15,000 10,000 10,000 5,000 5,000 80,000 Table 11. Actual BDL RO data for 2009–2010.

9-28 to 10-25 10-26 to 11-22 11-23 to 12-20 12-21 to 1-17 1-18 to 02-14 2-15 to 03- 14 03-15 to 04-11 04-12 to 05-09 5-10 to 06-06 06-07 to 07-04 07-05 to 08-01 Stage 1 (1/4 – MVRs) 1 MVRs 2 MVRs 2 MVRs 3 MVRs 2 MVRs 4 MVRs 3 MVRs 1 MVRs 0 MVRs 1 MVRs 1 MVRs Season Total Influent volume processed (gallons) 2,074 23,471 20,192 195,621 103,384 211,179 129,853 9,656 11,116 8,378 714,924 Average influent glycol concentration (%PG) 10.5 6.4 6.5 6.1 6.6 7.0 6.4 5.5 5.7 6.0 6.51 Volume of 100% PG in influent (gallons) 218 1,502 1,312 11,933 6,823 14,783 8,311 531 634 503 46,549 Average influent temperature (°C) 70 75 77 76 75 73 72 81 89 92 78 Average influent flow rate (gph) 296.3 357 421.0 584.5 379.2 897.6 562.1 205.4 138.5 209.5 405 Effluent volume of distillate produced (gallons) 1,289 18,598 15,478 152,504 79,417 152,595 100,023 7,297 8,869 5,792 541,862 Effluent volume of concentrate produced (gallons) 785 4,873 4,714 43,117 23,967 58,584 29,830 2,359 2,247 2,586 173,062 Average effluent concentration of concentrate (% PG) 24.4 23.3 24.0 23.5 25.3 24.2 26.9 23.0 22.3 20.0 24.51 Volume of 100% PG in concentrate (gallons) 192 1,135 1,131 10,132 6,064 14,177 8,024 543 501 517 42,417 Stage 2/Single Stage (1/4 MVRs) 1 MVRs 3 MVRs 0 MVRs 2 MVRs 2 MVRs 2 MVRs 3 MVRs 1 MVRs 1 MVRs 1 MVRs 1 MVRs Influent volume processed (gallons) 2,076 2,279 38,843 21,912 77,237 169,872 2,290 9,890 1,347 5,003 314,509 Average influent glycol concentration (%PG) 12.8 20.5 22.8 25.0 18.5 11.4 24.5 11.2 20.5 23.4 15.67 Volume of 100% PG in influent (gallons) 266 467 8,856 5,478 14,289 19,365 561 1,108 276 1,171 49,282 Average influent temperature (°C) 69 68 79 76 80 74 67 83 80 91 73 Average influent flow rate (gph) 143.2 483 301.6 317.6 366.6 597.8 229.0 133.2 244.9 178.7 348 Effluent volume of distillate produced (gallons) 1,545 1,479 23,160 12,725 48,476 121,754 1,074 7,098 778 2,755 210,213 Effluent volume of concentrate produced (gallons) 531 800 15,683 9,187 28,761 48,118 1,216 2,792 569 2,248 104,296 Average effluent concentration of concentrate (% PG) 52.5 52.5 54.5 55.5 48.8 37.8 53.0 43.0 52.2 51.5 45.27 Volume of 100% PG in concentrate (gallons) 279 420 8,547 5,099 14,035 18,189 644 1,201 297 1,158 47,213 % ratio of glycol produced versus infeed 104.9 89.9 0.0 96.5 93.1 98.2 93.9 114.9 108.4 107.6 98.9 95.8 Table 12. Actual BDL MVR data for 2009–2010 deicing season.

Airport Deicer Treatment System Summaries D-13 of the low-concentration treatment system. The processed fluid from the UF systems must be less than 200 NTU before being fed to the RO system. The temperature of the influent is closely monitored to maximize the flow rate through the membranes. Both UF systems are monitored continuously for influent temperature and show symptoms of fouling when flow rates fall below desired parameters. At that point the units are flushed with a mild cleaning solution to clean the membranes. Flow Rate The treated flow rate is not a parameter used to demonstrate performance of the BDL RO/MVR treatment system, but it is useful in interpreting other performance parameters and in establishing the potential range of treatment capabilities. It was anticipated that flow rates would average 40 gallons per minute (gpm) to 50 gpm through the RO treatment system and average 2 gpm to 3 gpm for the MVR treatment process. In practice, the flow rates have averaged 9.2 gpm for RO and 9.0 gpm for MVR. The lower-than-anticipated flow rates are a reflection of the system not operating continuously (i.e., 24 hours per day). Instead, the RO system operates only partial days since it can easily handle the volume being fed from both the UF and concentrator systems. The UF effluent output capacity, the MVR distillate output, and the overall availability of low- concentration spent ADF are the reverse osmosis system’s limiting factors for the flow rate. Treated Load Rate The rate at which PG is removed from the system (treated load rate) is a key measure of the performance in RO and MVR technologies. This is because RO and MVR treatment processes are used to recover PG for reuse and recycling. Therefore, a high treatment load rate, such as at BDL, is an indicator that the system is recovering PG at a significant rate. Correspondingly, this demonstrates that the treatment system is meeting its design expectations. The two-stage RO installed at BDL performs as intended. It has more than enough capacity to meet required processing removal rates. The RO does not run continuously (i.e., 24 hours per day) and runs only partial days since it can easily handle the volume fed from both the UF and concentrator systems. The RO was originally designed to handle up to 219,700 gallons per day, so 40,000 gallons to 60,000 gallons can easily be processed per day. Based on the historical operation of the system at BDL, it can be concluded that the system has performed for the needs of the airport, and its full potential has yet to be demonstrated. Effluent Concentrations The effluent PG concentration of the RO system effluent is a key performance indicator for removal efficiencies. The BDL RO and MVR were designed to concentrate the PG for recycling and reuse. By concentrating the PG into one stream, PG is removed from the distillate stream. Cost Category Projected at Initial Implementation Actual Capital cost Not provided $3.4 million in 2006 RO: $150,000* UF: $300,000 MVR: $2,050,000 Subsystems:$900,000 Annual operating cost - Utilities - Chemicals - Analysis - Material handling Total operating cost Not provided Not provided Not provided Not provided Not provided Not provided Not provided Not provided Not provided $500,000 *Upgrade to existing system. Table 13. Costs for the treatment system.

D-14 Guidance for Treatment of Airport Stormwater Containing Deicers The distillate stream discharges or is sent through the RO treatment system again. The RO sys- tem is operated so that the distillate stream contains concentrations below the permitted concen- tration. The average PG concentration in the RO distillate stream has been 27 mg/L. The MVR also has a concentrate and a distillate stream. However, the concentrate from the MVR is trucked off-site at BDL for PG reuse. The MVR distillate stream is sent back to the RO system for further treatment. Treatment Efficiencies Based on the data, the average influent concentration of glycol was approximately 10,400-mg PG/L during the 3-year span. The RO reject produced yielded glycol concentrations averaging 3% to 5%. This indicates that at least 50% of the glycol flowing through RO is removed and sent to the MVR systems for recycling. The remaining glycol that carries over in the Stage 1 permeate of the MVR eventually becomes Stage 2 influent and averages between 0.5% to 1% glycol. The remaining glycol is removed, and the fluid quality consistently meets all discharge requirements. Overall, the RO system 3-year data indicate a 92% average removal rate of glycol. The data suggest that the unit is capable of removing 99% of the glycol that is processed, but according to the recycling vendor, the unit is set to continually meet the permit requirements while maximizing flow rates and in turn maximizing removal rates to maximize spent-ADF storage capacity at any given time. For this reason, the main focus when adjusting parameters on the RO system is not to reclaim all glycol but to maximize production flow rates while main- taining permit compliance. Cost Cost is another key indicator of performance. High-volume seasons increase costs since more consumables are used and labor is extended into the summer to monitor equipment before shutdown. The costs also typically increase as influent has a higher concentration of TSS since these require more chemical pretreatment and typically more operational shutdowns for clean- ing the system. If the treated fluid is very diluted and has glycol levels of less than 1% PG, then the system becomes less cost-effective since the volume of PG recovered per unit volume treated is typically lower. Average annual operating costs of utilities, chemicals, analyses, and solids management are approximately $500,000 annually. The BDL system uses one full-time supervisor, two full-time operators, and seasonal operators as necessary for the system. Most maintenance activities at BDL are performed by the glycol recycling contractor as part of its duties. A cost model developed in Task 5 of this research was used to relate required RO and MVR technology COD loading (lbs/day) to cost. Considerations from analysis of RO and MVR cost data application to the model include: • Actual treatment capacity/RO volume compared to nominal design capacity. • Effect of treatment efficiency on caustic demand. • Chemical use data per pound of COD treated, which may vary with the concentration of influent soluble COD. • Electrical costs per cubic foot of membrane, and • Solids generation rates per pound of COD treated. Lessons Learned for Potential Implementation of the RO and MVR Technology at Other Airports Several factors have proven critical to effective and efficient performance in the RO and MVR systems at BDL. These factors are:

Airport Deicer Treatment System Summaries D-15 1. Adequate filtration methods prior to treatment; 2. Maintaining process variables such as temperature, turbidity, flow rate, and pressures at con- sistent set points; 3. The ability to adjust the UF/RO and MVR systems to respond to variability in influent glycol concentrations; 4. The ability to meet desired effluent concentrations, which affects influent processing rate; 5. The need to integrate daily preventative maintenance into operations in order to optimize equipment performance; and 6. The membrane systems should be treated with biocide when the processing systems sit idle for extended periods of time to eliminate potential biological growth. The RO system is pH sensitive, so caustic injection systems continually run to ensure that the pH is maintained at an optimal level. The RO system is continually monitored for pressure read- ings and permeate quality. This gives an indication when fouling is occurring and the system needs to be stopped for flushing. Pressures gradually climbing, coupled with increasing COD on the permeate discharge, are typical indications that the RO unit needs to be shut down for wash- ing. The system is flushed with a mild cleaning solution to clean the membranes. A small volume of waste is produced each year from slops disposal. The 3-year data indicate that the volume of wastewater generated by the system and trucked off-site is approximately 1% to 2% of the volume that is fed through the system. By employing the use of both membrane and MVR technologies, Bradley International Airport is able to handle a large range of influent concentrations. This includes treatment of spent ADF of as low as 0.1% in concentration to as high as 25% in concentration. With the ability to use mem- brane systems, this also allows the airport to meet very stringent discharge limitations. By installing both types of recycling technologies, the airport was able to maintain a relatively small footprint with a significant amount of treatment capacity. All of the units installed are modular in design, and as a result, additional systems were able to be added in 2008 to meet an increase in volumes of spent ADF collected with the terminal gates being tied into the existing collection system. The type of technologies used at BDL could be effective for airports that generate a substantial volume of spent ADF at generally low concentrations. Each MVR at BDL can be adjusted to produce a desired glycol concentration product. The MVR units produce two effluent streams, and the desired concentration set points in each effluent stream directly affect the performance of the concentrators. The glycol concentration is continually moni- tored to balance the parameters on the machine to increase the processing rate. The effluent glycol level is crucial since the recycling contractor has a goal to produce effluent with a concentration of 50% PG. At this level and higher, the contractor trucks the fluid off-site so that it can be distilled to the 99.1% and higher concentration level. The second effluent stream produced from the MVR units is the distillate. This is the distilled water and is not continuously monitored since this fluid is sent to an interim storage tank where it is comingled with the other low-concentration spent ADF to be processed through the mem- brane systems. The quality of distillate is clean enough to be fed directly through the RO system. The RO system will remove any fugitive glycol to meet discharge permit levels. Based on the data, 94.6% of the glycol that was fed through the MVR systems was reclaimed. The remaining glycol was reclaimed through the RO system and the balance discharged through the effluent stream to the POTW. Conclusions from operation of the RO and MVR at BDL that can be used by other airports considering this technology include: 1. The MVR technology is excellent for enabling recycling of the concentrated PG for offsetting costs.

D-16 Guidance for Treatment of Airport Stormwater Containing Deicers 2. The RO and MVR technologies are excellent for treating high propylene glycol concentrations. 3. The effluent concentrations can be minimized through optimizing turbidity, pressure, tem- perature, and pH. 4. The system performs very consistently and predictably once a constant concentration, pres- sure, temperature, and pH are obtained in the influent. 5. The system can start and stop as required with little impact to the influent loading rates or effluent concentrations. 6. Cost recovery from the recycled PG is dependent on the market value of PG and the amount of PG available for capture and recycling. 7. Sufficient ability to control flow rates is important, especially if influent concentrations are high, resulting in higher chemical dosing and maintenance. Documents and Information Review in Development of Airport Summary 1. Bradley International Airport. Treatment System Operational Records, 2010. 2. Svedruzic, Michael and Arendt, Tim. Deicer Treatment Options and Considerations for ELG, 22 July 2010.

Airport Deicer Treatment System Summaries D-17 Airport Treatment Summary No. 2 Airport: Nashville International Airport—Nashville, TN (BNA) Treatment Technology: Aerated Lagoon, POTW Discharge Years Operated: 1997–2012 (Currently Operational) Deicer Management System Description At BNA, aircraft deicing fluid is applied primarily on dedicated deicing areas (pads). All storm- water runoff from the pads and runoff from selected non-deicing areas around the gates flows by gravity to the north and south storage ponds. The inclusion of non-deicing area runoff results in lower COD concentrations than would otherwise be expected from a deicing pad operation. Each storage pond has a pump station for conveying runoff to an aerated lagoon for treatment. The stor- age pond pumps are operated based only on the water level in the ponds; as a result, the influent flow rates to the aerated lagoon are variable and not controlled based on treatment system needs. Treated effluent from the aerated lagoon discharges to Sims Branch via an effluent weir at the north end of the lagoon. With permission from the POTW, the treated effluent can also be discharged to the sanitary sewer under special circumstances. Overflows from the storage ponds are conveyed to Sims Branch. Figure 1 shows a diagram of the system. Deicer Treatment Technology Selection Considerations In the early 1990s, the Metropolitan Nashville Airport Authority (MNAA) experienced issues with low dissolved oxygen and bacterial growth in Sims Branch as a result of discharge of deicer- affected stormwater. MNAA considered biological treatment as a potential treatment methodol- ogy, and in 1992 a treatability study was performed to assess the feasibility of treating the runoff with aerated lagoon and activated sludge technologies. The objectives of the study were to define treatment process design and sizing criteria and to assess potential system performance. The estimated effluent quality using the potential technologies was compared to anticipated effluent limitations for BOD and TSS. The treatability study was conducted at 9°C (48°F). The treat- ability evaluated a first-stage aerated lagoon technology with a 20-day hydraulic detention time, which was projected to achieve 97% removal efficiency at that temperature, plus a second-stage activated sludge technology with a 1-day detention time, which would achieve 96% removal efficiency and achieve the desired BOD effluent concentrations. Figure 1. BNA deicing-affected stormwater management system.

D-18 Guidance for Treatment of Airport Stormwater Containing Deicers Because of budgetary constraints, MNAA implemented the aerated lagoon technology but not the activated sludge system. A plan was also implemented to monitor performance of the aerated lagoon on an ongoing basis to assess the need to add the potential second-stage activated sludge treatment in the future. The activated sludge portion was ultimately not added, although later a system for recycling biosolids back into the aerated lagoon was implemented in an effort to boost system efficiency. Deicer Treatment Technology Description Aerated Lagoon See the aerated lagoon fact sheet (Fact Sheet 103) for a description of the treatment technology used at BNA. The aerated lagoon at BNA (see Figure 2) is a lined structure, approximately 18 ft in depth at its deepest point. It was originally portioned into three sections of aerated treatment, with a final section for solids settling and sludge stabilization. A total of 13 surface aerators with draft tubes were used in the three aerated sections to provide mixing and oxygen transfer. At the present time, the liner-based baffles that segregated the basin into four sections are not in place, but MNAA plans to again segregate the basin in an upcoming system upgrade. Biological solids settle and partially degrade in the last section of the basin. Settled solids are removed occasion- ally, but typically no more than once per year. MNAA has experimented with several methods for removal of the solids from the basin, but the process of removing the solids without damaging the lagoon remains a challenge. Description of Support Systems Two ponds are located upstream of the treatment system for storage of stormwater runoff from the deicing pad and concourse areas. The capacity of the north pond is 1.09 million gallons, and the capacity of the south pond is 2.0 million gallons. The ponds are designed to hold run- off from the first flush, or up to 1.5 in. of rainfall. When the ponds are full, collected runoff will overflow at a weir in each pond and discharge directly to Sims Branch. The north pond has two 125-gpm pumps. The south pond has two 250-gpm pumps. The maximum hydraulic retention time at average flows is 5 days for the north pond and 7 days for the south pond, although in practice water can stay in the ponds for much longer during dry periods. Figure 2. BNA treatment system.

Airport Deicer Treatment System Summaries D-19 The system operators manually feed nutrients to supplement growth of the bacteria in the treatment system. Originally, solid nutrients were fed into the influent end of the lagoon until operators noted during lagoon maintenance that the solid nutrients were building up on the lagoon bottom, apparently due to a lack of dissolution in the water. The system operators now manually feed a liquid form of the nutrients. Effluent from the settling section discharges over a two-sided V-notch weir. The baffles were removed from the lagoon in 2010 because of deteriorating condition and operational concerns with solids buildup in the aerated portions of the lagoon. Key System Sizing Parameters (Original Design) See Table 1 for system sizing parameters and Table 2 for other sizing parameters. Treatment System Performance Table 3 data on the intended design performance were derived from the 1995 BNA Opera- tions Manual prepared for the original system design. Table 4 contains data on actual system performance. Component/Parameter Size/Capacity of Treatment Units Number of Treatment Units Total Capacity Stormwater storage capacity North pond: South pond: 1.09 million gallons 2 million gallons 2 3.09 million gallons Treatment unit (lagoon) volume 5.75 million gallons 1 5.75 million gallons Treatment unit (lagoon) dimensions Length Width at bottom Width at top of berm Depth Side slopes 685 ft 35 ft 110 ft 18 ft 2.2 to 1 1 890,000 ft3 Treatment facility footprint (treatment lagoon only) 2-acre total site 1 2 acres Design biomass suspended solids concentrations in lagoon from treatability study 1,200 mg/L 3 N/A Note: Data based on 1995 Operations and Maintenance Manual. Table 1. Key system sizing parameters. Component/Parameter Value Unit )muminim( 5.4 emulov tinu tnemtaerT 5.75 (maximum) Million gallons Aeration system (when baffles in place): Cell 1 Cell 2 Cell 3 Cell 4 6 4 3 0 Aerators Hydraulic retention time at average flow 25 Days teeF 301 htgnel elffab lortnoc wolF niar fo sehcnI 5.1 sisab ngised dnop egarotS Aerated lagoon discharge weir length (v-notch) 35 Feet Note: Data based on 1995 Operations and Maintenance Manual. Table 2. Additional system sizing parameters (original design for aerated lagoon).

D-20 Guidance for Treatment of Airport Stormwater Containing Deicers Cost Assessment for the BNA Aerated Lagoon Treatment System See Table 5 for treatment system costs. Conclusions on Performance of BNA Aerated Lagoon System Influent Deicer Concentrations The influent concentration is variable and cannot be controlled in the system as it is operated today. As a result, fluctuations in concentration can affect treatment. The long detention times typical in the storage ponds help to dampen the range of influent concentrations sent to the treatment lagoon. Parameter Value Unit Design flow rates - Average 130 Gallons per minute Design treatment load capacity 15,800* 9,300 9,300** lbs COD/day lbs BOD5/day lbs PG/day Design influent concentration - Range 8,500~17,000 5,000~10,000 5,000~10,000** mg COD/L mg BOD5/L mg PG/L Biomass suspended solids concentrations in lagoon 123 mg/L Design effluent concentration (average) 578 167 Not provided mg COD/L mg BOD5/L mg PG/L Design treatment efficiency (average) 95 % influent COD load treated Note: Design data based on 1995 Operations and Maintenance Manual and 1999 CDM Review of five MNAA environmental programs and response strategies. The design effluent concentrations shown were the intended design effluent if an activated sludge system were added for polishing. (The activated sludge system was not ultimately installed.) *Data based on conversion: [COD] = 1.7 [BOD5]. **Data based on conversion: [PG] = [BOD5]. Table 3. Design basis for system performance. tinU eulaV retemaraP Flow rates - Minimum - Maximum 70 244 Gallons per minute Actual COD treatment load rate - Maximum 600 lbs/day Influent COD concentration - Average - Maximum 800 3,000 mg/L Effluent BOD5 concentration 20–1,600 mg/L tub ,nwonk toN ycneiciffe tnemtaerT estimated to be <50% % influent COD load treated Table 4. Actual system performance. Cost Category Projected at Initial Implementation Actual )etamixorppa( M5.2$ elbaliava toN tsoc latipaC Annual operating cost Not available Not available Table 5. Costs for the treatment system.

Airport Deicer Treatment System Summaries D-21 Flow Rate The flow rates into the system fluctuate significantly, as different pumps from the ponds are activated, with flow rates fluctuating between 70 gpm and 244 gpm when the pumps are running. This fluctuation in flow rates may not have been accounted for in the original system treatability studies, which assumed a more steady-state operation. Treated Load Rate The deicer mass loading rate is likely highly variable as both flow rates and COD concentra- tions from the two storage ponds vary. Flow rates into the treatment system cannot presently be controlled. The COD loads that the system can treat are significantly lower than the design capability due to challenges in controlling mixed liquor suspended solids concentrations. Effluent Concentrations Effluent concentrations are highly variable, with numerous periods below 50 mg/L, but also with spikes into the hundreds of mg/L COD nearly every year as large deicing events occur. Over time, the treatment system has occasionally experienced effluent concentrations that exceed permit limits. Treatment Efficiencies The ability to calculate treatment efficiencies is limited by the fluctuation of the influent flows, influent COD concentrations, and to a lesser degree, variations in effluent flows and concentra- tions. As a result, insufficient data have been collected to calculate treatment efficiency with any accuracy. It is believed that the current ability to remove COD in treatment is limited by chal- lenges in maintaining a viable biomass in the aerated lagoon. Cost Operational costs are not cumulatively tracked for the treatment system, but the airport is in the process of assisting in gathering data. Operating costs include power cost for the pumps and aerators, chemical cost for nutrient addition, analytical costs, and operator labor costs. One full-time operator is employed to operate the system. Lessons Learned for Potential Implementation of the Aerated Lagoon Technology at Other Airports The following factors have proven critical to effective and efficient performance of the performance of the BNA aerated lagoon: 1. Ability to maintain a healthy and sufficiently concentrated biomass. 2. Adapting to cold temperatures, leading to reduction to treatment performance. 3. Managing variable COD loadings, leading to reduction to treatment performance. 4. Ability to add nutrients regularly and sufficiently for biomass needs. 5. Ability to add sufficient aeration and achieve adequate mixing throughout extent of lagoon. 6. Separation of treatment from solids settling processes. 7. The large volume of stormwater that must be processed due to collection of runoff not affected by deicing activities. Critical Performance Factors Critical to the performance of the aerated lagoon is the ability to maintain a biomass population that is healthy, settles properly, is stable, and is present in sufficient quantities to meet the treatment needs. The lack of control of influent COD loading and the significant variations in loading make it difficult to sustain an appropriate biomass in the BNA system even with sufficient management

D-22 Guidance for Treatment of Airport Stormwater Containing Deicers of other factors. Other factors that may influence the biomass are appropriate nutrient loading, sufficient mixing, sufficient aeration, and control of returned sludge to bolster the biomass con- centration. Appropriate addition of nutrients based on the influent load is one key to keeping the biomass healthy. Maintaining a low velocity through the lagoon will prevent the biomass from being washed out with the effluent. Reinstallation of the baffles that were removed several years ago could also help maintain the biomass, and addition of a solids removal system to remove solids from the effluent and pump them back into the system could also help maintain the population. The temperature of the water in the system also has a significant impact on the performance of the system. Since it is exposed to the elements, it is very difficult to control the temperature in the lagoon. In Nashville, Tennessee, the climate is more temperate than at some other airports, although in many years water temperatures can drop below 40°F, which will significantly slow biomass growth. Based on performance limitations, MNAA has initiated a new project to improve the existing treatment system that will include: • Reduction in the volume of water from non-deicing area runoff; • Online monitoring of TOC, flow rates, and temperature; • Control of mass loading rates into the treatment system; • New nutrient feed system; • Upgraded aeration system; • Segregation of the treatment and solids removal components; • Solids removal system for biological solids; and • Improved maintenance of existing equipment and infrastructure to reduce downtime. Documents and Information Review in Development of Airport Summary 1. CDM. BNA Deicer Management System Review Report, 1998. 2. Nashville International Airport. Treatment System Operational Records, 2012. 3. Ogden. BNA Deicer Treatment System Operations and Maintenance Manual, 1995. 4. GS&P. Feasibility Study Report for Improvements to Deicing Fluid Collection and Treatment Systems, 2012.

Airport Deicer Treatment System Summaries D-23 Airport Treatment Summary No. 3 Airport: Buffalo Niagara International Airport—Buffalo, NY (BUF) Treatment Technology: Aerated Gravel Bed Years Operated: 2009–2012 (Currently Operational) Deicer Management System Description In 2009, BUF implemented a system for collection of deicer and on-site treatment of deicer prior to discharge to an adjacent stream, Scajaquada Creek. All stormwater from the southeast side of the airport, which includes the main terminal and air cargo, is captured and treated year round. There are 3 million gallons of equalization storage prior to treatment. The treatment sys- tem was installed in response to New York State SPDES permit limits on BOD5 (30-mg/L), glycol, and flow rate (154 cubic feet per sec) discharges to Scajaquada Creek. On-site treatment consists of a subsurface aerated gravel bed located on the airside of the airport facility. Figure 1 shows a process flow diagram of the system. Deicer Treatment Technology Selection Considerations It was found that the sanitary sewer had limited hydraulic capacity (4-in. pipe was limiting), and on-site treatment with discharge to the surface waters was necessary to comply with the permit limits. Factors that were considered in the selection of the on-site aerated gravel bed treatment technology included: 1. Budget: 10 million USD available for construction. 2. Limited land was available. 3. A low-profile system was desired such that it could be placed on available land on the airside of the airport. Figure 1. BUF deicer management system process flow diagram.

D-24 Guidance for Treatment of Airport Stormwater Containing Deicers 4. The treatment system needed to be retrofitted into the existing stormwater management system. 5. The nature of the weather and deicing at BUF resulted in high variability in flow and strength of water to be treated, which had to be controlled for effective treatment. Treatability testing and pilot-scale testing were performed prior to design of the BUF aerated gravel bed system to establish technology capabilities and design parameters. Deicer Treatment Technology Description Description of Aerated Gravel Bed BUF uses an aerated gravel bed treatment technology (see Fact Sheet 102 for technology details). The BUF aerated gravel bed uses a vertical flow configuration in which stormwater is uniformly distributed with infiltration chambers buried near the surface of the substrate over the gravel beds. The applied water percolates downward through the gravel substrate to an under-drain system. A relatively large gravel size, 1/2 in.–3/4 in. in diameter, provides a surface for growth of a bacterial biofilm. The biofilm grows during the deicing season and degrades in the summer when no deicer is applied. The BUF treatment system consists of four discrete gravel beds excavated from an existing open area near the airport’s main runway. The gravel beds are vegetated with grasses growing in a mulch surface. The mulch surface helps to contain heat in the winter months within the bed. See Figure 2 for a photograph of the aerated gravel bed during construction. Description of Support Systems The aerated gravel bed at BUF includes the following support systems for the treatment sys- tem: aeration system, dosing system, nutrient feed system, and analytical system. The aeration system uses four blowers, a manifold system, and a forced bed aeration system. The dosing system for supplying the deicer-affected stormwater to the treatment system includes four dosing pumps and a dosing tank. The objective of the dosing system is to provide a uniform mass loading to the treatment cells to stabilize the biological population and pro- vide for efficient treatment. The dosing system is a means of counteracting the swings in deicer concentrations inherent in collected runoff during winter. The aerated gravel bed is monitored with influent and effluent TOC meters, an effluent ammonia meter, and four influent flow Figure 2. Construction of the BUF aerated gravel bed.

Airport Deicer Treatment System Summaries D-25 meters (1 meter per bed). The nutrient feed system uses three chemical feed pumps and one 500-gallon batch feed tank to supply nutrients to support the aerated gravel bed with pH. Key System Sizing Parameters Table 1 shows key sizing parameters. Treatment System Performance Table 2 indicates the intended design performance of the BUF system, and Table 3 shows actual system performance. Table 4 provides field-measured weekly average performance data for the 2010–2011 deicing season. Monitoring to date indicates that influent CBOD5 is roughly two times influent TOC values, and effluent CBOD5 is roughly half of effluent TOC values. Cost Assessment for the BUF Aerated Gravel Bed Treatment System Table 5 shows the treatment system costs. Conclusions on Performance of BUF Aerated Gravel Bed System Influent Deicer Concentrations The influent deicer concentration is not a direct operational factor in the aerated gravel bed treatment system. However, variations in the influent concentration affect the treatment loading rate operational limit. Influent TOC values range from 11 mg/L to 6,909 mg/L. These TOC values would correspond to roughly 22-mg/L to 13,818-mg/L influent for CBOD5 using BUF sampling data. The concentrations indicate that the aerated gravel bed system is capable of treating con- centrations as high as 130% of the design concentration maximum. Component/Parameter Size/Capacity of Treatment Units Number of Treatment Units Total Capacity Stormwater storage capacity 3,000,000 gallons 1 3,000,000 gallons Treatment unit volume 0.70 million gallons 4 2.83 million gallons Treatment unit dimensions 166-ft L x 300-ft W x 5-ft D 4 N/A Treatment facility footprint 1.1 acres 4 4.5 acres Table 1. Key system sizing parameters. Parameter Value Unit Design flow rates - Maximum 6,152 Gallons per minute Design treatment load capacity 17,000* 10,000 10,000** lbs COD/day lbs BOD5/day lbs PG/day Design influent concentration - Range Not provided mg BOD5/L Design effluent concentration (average) 30 Not provided mg BOD5/L mg PG/L Design treatment efficiency (average) Not provided % influent COD load treated *Data based on conversion: [COD] = 1.7 [BOD5]. **Data based on conversion: [PG] = [BOD5]. Table 2. Design basis for system performance.

D-26 Guidance for Treatment of Airport Stormwater Containing Deicers Flow Rate The flow rate is not a direct operational factor in the aerated gravel bed treatment system. However, the flow rate does vary based on the treatment loading rate and treatment system detention time. Treated Load Rate The aerated gravel bed system is designed and operated for 10,000-lbs BOD5/day. The design treatment load rate of 10,000-lbs BOD5/day is a limiting factor in the operation of the aerated gravel bed system. The system is loaded and monitored each day beginning at 8 a.m. The system continues to receive loads until the load limit of 10,000-lbs BOD5/day is reached, at which time no additional influent is sent to the aerated gravel bed until 8 a.m. This batch loading system of feeding the aerated gravel bed works well due to the long detention time. The system was able to withstand periodic high loadings that were four times design capacity. Effluent Concentrations Effluent TOC values range from 48 mg/L to 357 mg/L. This would correspond to roughly 24-mg/L to 179-mg/L effluent for CBOD5 using BUF sampling data. Treatment Efficiencies Treatment efficiency was routinely above 90% for 2010–2011. Treatment performance ramped up quickly with an increase in TOC concentrations in the influent. Cost The extent of the aerated gravel bed system installed at BUF to date is a function of the avail- able budget. tinU eulaV retemaraP Flow rates - Minimum - Average - Maximum 187.5 402.8 1416.7 Gallons per minute Actual COD treatment load rate - Average - Maximum 19,000* 72,000* lbs/day Actual BOD5 treatment load rate - Average - Maximum 11,200 42,400 lbs/day Actual PG treatment load rate - Average - Maximum 11,200** 42,400** lbs/day Influent COD concentration - Minimum - Maximum 38* 23,500* mg/L Influent BOD5 concentration - Minimum - Maximum 22 13,818 mg/L Influent PG concentration - Minimum - Maximum 22** 13,818** mg/L Effluent BOD5 concentration - Minimum - Average - Maximum 25 73 178 mg/L detaert daol DOC tneulfni % dedivorp toN ycneiciffe tnemtaerT *Data based on conversion: [COD] = 1.7 [BOD5]. **Data based on conversion: [PG] = [BOD5]. Table 3. Actual system performance.

Airport Deicer Treatment System Summaries D-27 Lessons Learned for Potential Implementation of the Aerated Gravel Bed Technology at Other Airports The factors listed in the following have proven critical to effective and efficient performance in the BUF aerated gravel bed system: 1. Control of flow and TOC loading rates. 2. Management of nutrient loadings at start-up. 3. Routine clean out of dosing lines. 4. Adequate aeration through the use of forced bed aeration. 5. Use of plants has no effect on performance or treatment efficiency. Date Average Flow (gpd) Average Load (lbs TOC/day) Average Influent TOC (mg/L) Average Effluent TOC (mg/L) 10/4/2010 1,046,362 640 74 54 10/11/2010 317,894 385 78 61 10/18/2010 543,039 108 18 56 10/25/2010 485,733 75 11 48 11/1/2010 331,446 119 13 53 11/8/2010 38,095 16 33 48 11/15/2010 695,874 814 132 54 11/22/2010 1,470,369 2,885 178 64 11/29/2010 988,104 6,270 458 211 12/6/2010 988,104 6,270 458 211 12/13/2010 827,512 5,567 1,731 116 12/20/2010 827,512 5,567 1,731 116 12/27/2010 248,767 19,335 1,828 163 1/3/2011 556,360 9,974 3,370 141 1/10/2011 41,714 4,804 6,714 122 1/17/2011 163,205 13,866 6,204 255 1/24/2011 38,889 4,587 6,909 138 1/31/2011 27,661 3,014 6,318 115 2/7/2011 46,886 4,572 4,670 134 2/14/2011 1,177,909 15,103 1,738 277 2/21/2011 122,705 3,287 1,400 142 2/28/2011 1,181,498 7,390 798 162 3/7/2011 2,039,782 21,232 581 357 3/14/2011 289,177 2,331 426 176 3/21/2011 656,285 11,993 1,392 318 3/28/2011 507,051 5,269 515 242 4/4/2011 527,261 2,032 294 142 4/11/2011 186,215 163 387 118 Maximum 2,039,782 21,232 6,909 357 Average 584,693 5,631 1,731 146 Minimum 27,661 16 11 48 Table 4. BUF weekly performance data 2010–2011 season. Cost Category Projected at Initial Implementation Actual M01$ 8002 ni M01$ *tsoc latipaC Annual operating cost Not provided *Capital costs are for the treatment system only. Costs do not include site-specific costs for collection, storage, and discharge. Table 5. Costs for the treatment system.

D-28 Guidance for Treatment of Airport Stormwater Containing Deicers The system went online in spring of 2009 and performed as expected until late December 2009. In late December 2009, the formation of polysaccharides (slime) was observed within the treatment bed. To remedy the reduced treatment performance and remove the polysaccharides, the aeration and nutrient addition was increased. After 2 months, the system began operating at design performance. The nutrient addition levels have since increased to match a high rate of bacterial growth. Since the aerated gravel bed went online in 2009, the sampling pumps burned out and required replacement and the SCADA system was upgraded to provide operator interface and off-site data access. Ammonia meter readings are erratic and the unit has undergone repeated troubleshooting. It was determined that the influent dosing lines require regular cleaning to prevent clogging by floating debris (plastics). Regular cleaning of dosing lines is required to maximize flow through the system. Conclusions from operation of the aerated gravel bed at BUF that can be used by other air- ports considering this technology include: 1. The AGB technology is excellent for isolating treatment from the effects of the weather and cold water temperatures because of heat generated in the cells during treatment. 2. The AGB technology is excellent for achieving effluent limits for propylene glycol and BOD concentrations. Some TOC remains, indicating that there are some organics that are not readily biodegradable. 3. The system performs very consistently and predictably over a wide range of influent concen- trations because loading into the treatment system is controlled. 4. It is unclear as to the need to initially seed the system during start-up. Off-season operation at low concentrations appears to develop an acclimated culture prior to onset of the deicing season. 5. Providing sufficient nutrient balance is critical. 6. Sufficient ability to control flow rates is important, especially if influent concentrations are high, resulting in lower flow rates. 7. If treating high concentrations, such as is the case with flows from deicing pads, consider the potential impacts of the lower flow rates that are needed to maintain a consistent TOC loading, including effects on storage, effects on solids removal, and effects on effluent concentrations. Documents and Information Review in Development of Airport Summary 1. BUF. Construction Drawings (Plans and Specs–2008). 2. BUF Weekly Reports 2010–2011 (summarized operating data). 3. Austin, D. C., Maciolek, D. J., Davis, B. M., Wallace, S. D., “Damköhler number design method to avoid clogging of subsurface flow constructed wetlands by heterotrophic biofilms.” Water Science and Technology. 56.3 (2007): 7–14. 4. Higgins, J. P., Maclean, J., “The use of a very large constructed sub-surface flow wetland to treat glycol contaminated stormwater from aircraft de-icing operations.” Water Quality Research Journal of Canada. 37.4 (2002): 785–792. 5. Kadlec, R. H., Wallace, S. D., Treatment Wetlands, Second Edition. Boca Raton, Florida: CRC Press, 2009. Print. 6. Wallace, S. D., Higgins, J., Liner, M. O., Diebold, J., “Degradation of aircraft deicing runoff in aerated engineered wetlands.” In: Multi Functions of Wetland Systems: An International Confer- ence, 26–29. University of Padova and International Water Association: Padova, Italy. 2007.

Airport Deicer Treatment System Summaries D-29 Airport Treatment Summary No. 4 Airport: Akron–Canton Airport—North Canton, OH (CAK) Treatment Technology: Anaerobic Fluidized Bed Reactor Years Operated: 2007–2012 (Currently Operational) Deicer Management System Description In response to effluent limits for PG in its 2004 NPDES permit, CAK initiated a study to assess means for deicer application, runoff collection, conveyance, treatment, and disposal of aircraft deicer that would provide the ability to meet its permit limits for Outfall 003. The implemented deicer management system includes two deicing pads, gravity drainage from the pads to two storage tanks, an anaerobic fluidized bed reactor biological treatment system, and discharge of treated effluent to the existing detention basin upstream of Outfall 003. The discharge from Outfall 003 is to the city of Green municipal separate storm sewer, with subsequent discharge to Zimber Ditch, a regulated surface water. Operators have the ability to route diluted flows from the deicing pads around the treatment system directly to the Outfall 003 detention basin without treatment. Figure 1 shows a process flow diagram of the system. Deicer Treatment Technology Selection Considerations In its decision-making process, CAK placed a high premium on minimizing project costs to support its goal of being the local low-cost provider of air services. Since the capital portion of the treatment system installation was covered by a federal grant, decisions on treatment technol- ogies were primarily driven by two factors: (1) minimizing annual operating and maintenance costs and (2) the ability to consistently and predictably achieve compliance with the NPDES permit effluent limits. Construction of the deicing pads was an important element in reducing costs because of the reduction in the volumes of water that would need to be stored, conveyed, and potentially heated. When considering treatment alternatives, CAK considered it important to minimize the footprint of the treatment operations to reduce expenditures associated with facility buildings. Based on an assessment of a wide range of treatment options in 2005, it was determined that discharge to the local sanitary sewer (a POTW), recycling, and two types of on-site biological Figure 1. CAK deicing-affected stormwater management system.

D-30 Guidance for Treatment of Airport Stormwater Containing Deicers treatment were the most applicable potential treatment and disposal options. The following conclusions on treatment technologies were reached before the AFBR technology was selected: • POTW (sanitary) discharge was eliminated as a possible treatment and disposal technology due to insufficient capacity at two local wastewater treatment plants. The POTWs were not interested in modifying their plants to accommodate the increased seasonal loading. • Treatment using membrane filtration or evaporation units, with ultimate transport of the moderately concentrated glycol off-site for recycling, was considered carefully, but was elimi- nated for several reasons: – The additional units needed to reach the PG concentrations in the dilute effluent stream drove up operating costs in relation to biological treatment. – CAK had concerns about relying on an outside entity for treatment services. – CAK preferred not to be dependent on potentially fluctuating market conditions for recycled glycol. • Both aerobic and anaerobic biological treatment methods were considered. Since minimizing the treatment footprint was important, the aerobic membrane bioreactor system and the anaerobic fluidized bed reactor system were considered. The AFBR was selected over the MBR based on the following criteria: – Lower operating costs. – Built-in means for isolating the treatment effectiveness from weather concerns through the use of off-gassed methane as fuel to heat the runoff. – Proven success at another airport (Albany International). Deicer Treatment Technology Description Anaerobic Fluidized Bed Reactor See the AFBR treatment technology fact sheet (Fact Sheet 104). Figure 2 shows reactor units used in the AFBR system. Description of Support Systems The AFBR at CAK includes the following support systems for the treatment reactor-separator unit: storage (two 750,000-gallon concrete tanks), influent pumping system, heat generation Figure 2. Biological reactor units (at right) in the CAK AFBR system.

Airport Deicer Treatment System Summaries D-31 and exchange loop, chemical feed for nutrient addition and pH control, biogas handling, and biological solids removal and handling. Collected runoff water from the storage tanks is pumped at a flow rate set by the system operators to achieve a constant COD loading as influent COD concentrations change. The cold influent water is heated first by passing it by warm effluent water in a heat exchanger and then by passing it by hot water from a boiler in a second heat exchanger. The hot water is obtained by heating potable water in a boiler using biogas captured from the reactor. The biogas is approximately 70% methane and 30% carbon dioxide and is used similarly to natural gas. For the CAK system, the heating system burns exclusively self- generated biogas for the entire deicing season, except for initial yearly start-up when natural gas is used. Any excess biogas is burned in a flare external to the building. The AFBR technology requires addition of a base chemical (sodium hydroxide) to keep pH in the reactors neutral, as well as addition of various chemical nutrients to support growth of the bacteria. Biological sol- ids exiting the reactor-separator unit with the treated effluent are removed with a dissolved air flotation clarifier. Treated effluent is discharged to CAK’s Outfall 003 detention basin. Biological solids are disposed of in a landfill. Key Treatment System Sizing Parameters Table 1 shows system sizing parameters. Treatment System Performance The data in Table 2 on the intended design performance of the system were derived from the Engineering Report and Permit-to-Install Application to Ohio EPA. Component/Parameter Size/Capacity of Treatment Units Number of Treatment Units Total Capacity Stormwater storage capacity 750,000 gallons 2 1.5 million gallons Treatment unit volume 2,500 ft³ 2 5,000 ft³ 37,400 gallons Treatment unit dimensions Reactors: 10-ft diameter 2 N/A Treatment facility footprint 0.1-acre building, 0.2-acre total site 1 0.2 acre Table 1. Key system sizing parameters. Parameter Value Unit Design flow rates - Minimum - Average - Maximum 5 20 50 Gallons per minute Design treatment load capacity 3,400 2,000* 2,000** lbs COD/day lbs BOD5/day lbs PG/day Design influent concentration - Range 50~32,000 30~18,800* 30~18,800** mg COD/L mg BOD5/L mg PG/L Design effluent concentration (average) 340 200* <35 mg COD/L mg BOD5/L mg PG/L Design treatment efficiency (average) 98% % influent COD load treated Design data based on 2006 Engineering Report. Design values based on two-reactor system. *Data based on conversion: [COD] = 1.7 [BOD5]. **Data based on conversion: [PG] = [BOD5]. Table 2. Design basis for system performance.

D-32 Guidance for Treatment of Airport Stormwater Containing Deicers The information in Table 3 on actual system performance was derived from daily data col- lected at the facility between 2008 and 2011. Table 4 presents performance data for each of the system’s five operating seasons. Figure 3 presents daily COD loading data for a typical season. Figure 4 presents typical daily COD effluent concentrations from the AFBR reactor units. COD effluent concentrations are further reduced in the plant’s dissolved air flotation solids removal unit. Cost Assessment for the CAK AFBR Treatment System See Table 5 for treatment system costs. Conclusions on Performance of CAK AFBR System Influent Deicer Concentrations Influent deicer concentrations are not a parameter used to demonstrate performance of the CAK AFBR, but they are useful in interpreting other performance parameters and in establishing Parameter Value Unit Flow rates - Minimum - Average - Maximum 2.8 5.9 10.2 Gallons per minute Actual COD treatment load rate - Average - Maximum 3,715 4,381 lbs/day Actual BOD5 treatment load rate - Average - Maximum 2,185* 2,580* lbs/day Actual PG treatment load rate - Average - Maximum 2,185** 2,580** lbs/day Influent COD concentration - Minimum - Average - Maximum 34,065 62,016 85,775 mg/L Influent BOD5 concentration - Minimum - Average - Maximum 20,040 36,480 50,455 mg/L Influent PG concentration - Minimum - Average - Maximum 20,040 36,480 50,455 mg/L Effluent COD concentration - Minimum - Average - Maximum 36 94 280 mg/L Effluent BOD5 concentration TBD mg/L Effluent PG concentration Not detected mg/L Treatment efficiency 99.54% % influent COD load treated The following data was excluded from the data set used to assess performance documented in Table 3. a. 2007–2008 season: data between system commissioning and introduction of suitable bioseed on 12/4/07; start-up period (12/4/07–1/11/08); data after 2/11/08 when phosphorus addition was stopped, resulting in reduced performance. b. 2008–2009 season: period prior to 12/1/08 when phosphorus addition resumed; start-up period. c. 2009–2010, 2010–2011 seasons: start-up periods. d. 2011–2012 season: No data used as all deicer was used up in start-up period. *Data based on conversion: [COD] = 1.7 [BOD5]. **Data based on conversion: [PG] = [BOD5]. Table 3. Actual system performance.

Airport Deicer Treatment System Summaries D-33 Year Unit 2 007–2008 (Start-up Season) 2008–2009 2009–2010 2010– 2011 2011– 2012 TREATMENT TIME Full season Treatment system start- up date 12/4/2007 12/1/2008 1/23/2010 1/11/2011 1/11/2012 Treatment system end date 7/24/2008 7/17/2009 6/30/2010 5/17/2011 2/17/2012 Treatment days – season Days 233 228 158 126 37 Start-up period Start-up end date 1/11/2008 2/7/2009 2/14/2010 2/15/2011 2/17/2012 Treatment days – start- up Days 38 68 22 35 37 Period for performance assessment Start date 12/24/2007 2/7/2009 2/14/2010 2/15/2011 1/11/2012 End date 2/11/2008 7/17/2009 6/30/2010 5/17/2011 2/17/2012 Treatment days – performance assessment Days 49 160 136 91 37 Hydraulic retention time average Days 3.7 7.3 5.0 5.6 8.2 TOTAL SYSTEM PERFORMANCE Flow rate maximum gpm 14.7 7.1 10.7 8.3 9.6 Flow rate average gpm 8.4 3.9 5.9 5.4 4.4 Flow rate minimum gpm 4.6 0.9 2.8 3.1 1.4 Influent COD load maximum lbs/day 4,727 4,156 4,597 4044 3236 Influent COD load average lbs/day 3,390 3,654 4,151 3665 1648 Effluent COD load maximum lbs/day 146 32 46 65 21 Effluent COD load average lbs/day 34 8 10 14 3 Influent COD concentration maximum mg/L 55,900 105,100 97,600 84,500 53,340 Influent COD concentration average mg/L 39,195 84,129 64,382 60,359 33,783 Influent COD concentration minimum mg/L 25,350 45,290 29,500 36,120 18,540 Reactor effluent COD concentration average mg/L 288 162 141 231 62 Clarifier effluent COD concentration maximum mg/L 475 105 190 351 28 Clarifier effluent COD concentration average mg/L 176 41 37 128 5 Load removed % 98.99% 99.77% 99.76% 99.62% 99.78% Table 4. CAK season-by-season performance assessment.

D-34 Guidance for Treatment of Airport Stormwater Containing Deicers Note: sCOD = soluble COD. 0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000 2100 2200 2300 2400 2500 1/ 1/ 10 1/ 15 /1 0 1/ 29 /1 0 2/ 12 /1 0 2/ 26 /1 0 3/ 12 /1 0 3/ 26 /1 0 4/ 9/ 10 4/ 23 /1 0 5/ 7/ 10 5/ 21 /1 0 6/ 4/ 10 6/ 18 /1 0 7/ 2/ 10 CO D Re m ov ed (p ou nd s/d ay ) Date CAK AFBR TREATMENT SYSTEM Daily COD Removal 2009-2010 R-601 Total sCOD Load Removed in the Reactor (Lbs/d) R-602 Total sCOD Load Removed in the Reactor (Lbs/d) Design Loading Operational Target Loading Figure 3. Typical season daily loading curve. Figure 4. Typical season system effluent concentrations.

Airport Deicer Treatment System Summaries D-35 the potential range of treatment capabilities. Soluble influent PG and COD concentrations, as dic- tated by deicing activity and runoff volumes from the deicing pad, have been higher than projected primarily due to less rainfall in winter than anticipated. Average influent soluble COD concentra- tions have been 62,000 mg/L, and concentrations have been as high as 106,000 mg/L, compared to the 19,000 mg/L projected during conceptual design. Since the system is operated as a constant COD load system, there are no effects from the higher concentrations on treatment, other than the flow rate adjustments described in the following. There is some evidence that treatment removal efficiency (percentage of incoming COD load removed) is actually better at higher concentrations, possibly due to longer residence times. The operators simply reduce flow rates as influent concen- trations rise to keep the constant COD loading that is desired. Flow Rate As with influent deicer concentrations, the treated flow rate is not a parameter used to dem- onstrate performance of the CAK AFBR, but it is useful in interpreting other performance parameters and in establishing the potential range of treatment capabilities. It was anticipated that flow rates would average 20 gpm through the treatment plant. In practice, the flow rates have averaged 5.9 gpm and have been consistently below 10 gpm. The lower-than-anticipated flow rates are a direct response to higher concentrations of COD from the deicing pad runoff and the operational goal of maintaining constant COD loading. System programming and set points were adjusted after the first year of operation to allow sufficient flow control at lower flow rates. The lower flow rates processed by the system have not had any negative effects on the ability to drain the storage tanks well before the next deicing season (The latest date for completion of treatment and emptying of tanks has been mid-July.) Treated Load Rate The rate at which soluble COD is removed from the system (treated load rate) is a key measure of performance of the AFBR system. Soluble COD is used to manage flow rate as opposed to total COD (soluble COD plus insoluble solids-based COD) because (a) the influent has very little COD from solids as primary deicer constituents are in a dissolved state in runoff, and (b) considering the COD from biological solids in the effluent does not provide a true indicator of treatment efficiency. COD from biological solids is managed through solids removal in the system’s clarification unit. The Cost Category Projected at Initial Implementation Actual Capital cost* $3.2M in 2007 $3.2M $0.3M added in upgrades to solids and chemical handling since start-up from 2008 to 2011 Annual operating cost** - Utilities - Chemicals - Analysis - Material handling Total operating cost $30,000 $38,000 $2,000 $9,000 $75,000 $35,000 $4,500 $5,400 $11,000 $55,900 Notes: Source of capital cost data: design (Engineering Report 2006), actual (airport cost records). Costs are for the treatment system and the building in which it is housed. Excluded are costs for deicing pads, storage tanks, and conveyance piping/structures external to treatment system. Source of operating cost data: design (Engineering Report 2006), actual (operating logs for quantities, vendor prices for material,utility records). Costs exclude the costs of the two system operators. *Capital costs are for the treatment system only, including the building and basic building infrastructure. Costs do not include site-specific costs for collection, storage, and discharge. **Operating costs do not include labor costs, but equate to approximately two full-time operators. Maintenance costs, which vary, are not included. Table 5. Costs for the treatment system.

D-36 Guidance for Treatment of Airport Stormwater Containing Deicers soluble COD loading rate is the primary means by which the system is controlled and provides a good comparison of whether the treatment system meets its design expectations. The CAK AFBR system has been able to treat at a higher throughput than envisioned during design. In the 5 years of operation (not including yearly start-up periods), it has averaged 3,715-lbs soluble COD per day removed through treatment, which is higher than the 3,400-lbs COD/day design target. Operators have been typically running at approximately 18% above design capacity at 4,000-lbs COD/day during normal operational periods and have pushed system capacity as high as 4,440-lbs COD/day on occasion. At 4,440-lbs COD/day, the gas production from bacteria in the reactors begins to produce some instabilities in the biomass sludge layer in the reactor, so this loading (which is 30% higher than the design maximum loading) is viewed as the true system capacity. Based on these data, it can be concluded that the CAK AFBR system performed better than expected from a treated load capacity standpoint. Effluent Concentrations The effluent concentration of primary deicer constituents is another key performance indica- tor for the AFBR system. The CAK AFBR was designed primarily to reduce PG concentrations to meet the PG limits in the CAK NPDES permit. All PG measurements taken in the plant effluent have been below detection limits. Low or nonexistent PG concentrations were expected since PG degrades quickly and easily to other organic chemicals during biological treatment. The operators use soluble COD as the primary means of monitoring and controlling the plant. The next iteration of the NPDES permit will contain COD limits as well. Soluble COD concentrations in the effluent are obtained by measuring the COD of the filtrate from the TSS method. The anticipated average soluble COD effluent concentration from the AFBR facility at design was 340 mg/L. The average soluble COD concentration in practice for the plant as a whole has been 94 mg/L. The vast majority of the soluble COD is removed in the treatment reactors. These units have averaged approximately 180-mg/L soluble COD in their effluent. The average soluble COD con- centration is further reduced to 94 mg/L, on average over a 4-year period, as the treated water passes through a dissolved air flotation unit designed to remove suspended biological solids. In three of the five operating seasons, the total system soluble COD has averaged less than 40 mg/L for the entire season. The recent addition of the dissolved air flotation unit for removal of solids indicated that soluble effluent CODs averaged 5 mg/L in the unit effluent. The operators have only taken occasional BOD5 concentration measurements. A series of weekly effluent BOD5 analyses were performed on the treated effluent in the 2011–2012 seasons as part of this study. Both effluent COD concentrations (design 340 mg/L versus actual 180 mg/L in reactor efflu- ent and 94 mg/L average in system effluent) and effluent PG concentrations (design <35 mg/L versus actual below detection limits) have been lower than anticipated at design, indicating that the CAK AFBR has outperformed its design conditions for effluent quality. Treatment Efficiencies The third key indicator of performance for AFBR systems is the percentage of the influent soluble COD load that is removed through treatment (i.e., treatment efficiency). The removal efficiency for COD has been higher than expected at 99.54% for the five operational seasons between 2008 and 2012, compared to the anticipated 98%. Cost Cost is another key indicator of performance. Capital costs following construction were within $0.1 million (3%) of the design cost estimates. Approximately $0.3 million in additional capital has been expended since start-up to upgrade the biological solids removal system, isolate the air

Airport Deicer Treatment System Summaries D-37 compressor from the main building for noise control purposes, modify the sampling system, and provide better heat management in the building. Average annual operating costs are approximately $55,900, including utilities, chemicals, analy- ses, and solids management. Overall operating costs have been approximately 30% ($20,000) lower than what was projected at design. Costs for labor were intentionally excluded from the cost calculation by the researchers because the means by which different airports account for labor costs vary. The CAK system uses two full-time operators. Maintenance costs are also not included in the annual operating cost calculation because (1) there were difficulties in obtaining consistent and all-inclusive maintenance operation from all airports, and (2) maintenance costs vary con- siderably by year and are influenced by the age of the system. Most maintenance activities at CAK are performed by system operators as part of their duties. The highest maintenance costs to date were in the first year following operation, when adjustments were made to the system. Utility costs, primarily power, have been near what was projected at design. Chemical use costs have been lower than projected due to less caustic demand. The reduced caustic demand is related to higher treatment efficiency and consequent lower concentrations of breakdown prod- ucts in the effluent. Lower concentrations of effluent breakdown products equates to less acid in the effluent, which reduces the caustic demand. The operators also found that purchasing chemi- cals in the summer months is less expensive; therefore, they get full seasonal loads of chemicals at that time of year. Biosolids generation rates and costs were approximately 25% higher than expected at design, resulting in higher costs than expected for solids disposal. A cost model was developed in Task 5 of this research to relate required AFBR technology COD loading (lbs/day) to cost. Considerations from analysis of CAK cost data application to the model include: • Actual treatment capacity/reactor volume compared to nominal design capacity, • Effect of treatment efficiency on caustic demand, • Chemical use data per pound of COD treated, which may vary with the concentration of influent soluble COD, • Electrical costs per cubic foot of reactor, and • Sludge and biogas generation rates per pound of COD treated. Lessons Learned for Potential Implementation of the AFBR Technology at Other Airports The following factors have proven critical to effective and efficient performance in the CAK AFBR: 1. Ability to adequately control flow and therefore COD loading rates. 2. Consistent nutrient loadings and understanding of differences in nutrient uptake at start-up versus standard operation. 3. Understanding which forms of phosphorus are and are not available to the anaerobic bacteria. 4. Adequate control of the reactor temperature. 5. Adequate control of the reactor pH. 6. Management of sludge bed levels in reactors. 7. Management of biological solids wasted from the reactors using dissolved air flotation, rather than reliance on gravity settling. CAK experienced firsthand the importance of nutrient addition in keeping the biomass healthy and the treatment effectiveness high during the first year’s start-up. During that period, phosphorus was left out of the nutrient mix for several months because it was found in the influent samples at sufficient concentrations. It was ultimately determined, however, that the

D-38 Guidance for Treatment of Airport Stormwater Containing Deicers phosphorous in the influent was not bio-available. The lack of bio-available phosphorus limited biomass growth and temporarily reduced the treatment capacity to approximately 25% of design capacity. Once phosphorus was added to the nutrient solutions, treatment rates reached 100% of design capacity and beyond within a matter of days. Because the runoff entering the treatment system is heated with biogas generated from the biological reaction, the temperature of the influent runoff has had no effect on treatment perfor- mance. The pH is also controlled to near neutral for an optimized environment for the bacteria. Varying influent COD concentrations do not affect treatment because the mass loading of the COD into the reactor is controlled, which leads to a stable bacterial population. The CAK AFBR has demonstrated the ability to treat non-PG constituents, including pave- ment deicer constituents (acetates, formates) and PG breakdown products (acetate, propionic acid, acetic acid). PG breakdown products were found in the storage tanks at CAK during the start-up season when the collected deicer had to be stored in the storage tanks while the treat- ment plant’s construction was completed. Initially, it was thought the AFBR system had sig- nificant difficulty in treating those breakdown products; however, it was later determined that the lower treatment rates experienced during portions of the first year were due to the lack of bio-available phosphorus, as noted previously. If PG breakdown products or pavement deicer constituents form a majority of the COD in the runoff to be treated, however, the biology in the system may not be completely acclimated to that mix. Therefore, system operators are advised to more carefully monitor system effluent for volatile fatty acid concentrations and more closely control system influent loading. This is because the various types of bacteria in the reactors grow in proportion to the rate at which the PG is degraded when it is fed as the primary food source. If an intermediate compound, say propionic acid or acetic acid, is fed in significant concentra- tion, the feed must be temporarily reduced so that the biological population can adjust its level to degrade the additional load. Conclusions from operation of the AFBR at CAK that can be used by other airports considering this technology include: 1. The AFBR technology is excellent for isolating treatment from the effects of the weather and cold water temperatures. 2. The AFBR technology is excellent for achieving propylene glycol and BOD5 concentrations that are near or below detection limits. Some COD remains, indicating that there are some organics that are not readily biodegradable. COD in the effluent can be minimized through biosolids removal from the effluent. 3. The system performs very consistently and predictably once a constant COD influent load- ing is achieved each year. 4. At initial start-up of the system after construction, acquisition of the appropriate type of healthy bioseed is critical. The bioseed must be obtained from a similar type of anaerobic operation. CAK has not had to obtain outside bioseed after the first year start-up. Enough anaerobic bac- teria survive the 6- to 8-month shutdown period to start up the system each season. 5. At each year’s plant start-up, 1 to 2 weeks are needed to restart the equipment, flush out dead biomass, and get system operational settings to desired levels before increases in COD loading rates can be made. 6. Based on historical data, the system has approximately 25% of its design capacity at the initial seasonal start-up. Approximately 40 days of continuous ramping up of COD load- ing are needed to get to design operational loadings. During that start-up time, treatment efficiencies remain over 99%, and effluent concentrations are sufficiently low to discharge. 7. Providing sufficient nutrient balance is critical. 8. Sufficient ability to control flow rates is important, especially if influent concentrations are high, resulting in lower flow rates.

Airport Deicer Treatment System Summaries D-39 9. If treating high concentrations, such as is the case with flows from deicing pads, consider the potential impacts of the lower flow rates that are needed to maintain a consistent COD loading, including effects on storage, effects on solids removal, and effects on effluent concentrations. 10. Due consideration of the means of removing and dewatering anaerobic biosolids from the AFBR is critical in design as that sludge has its own unique characteristics. Use of dissolved air flotation for biosolids removal from the treated effluent stream has proved the most suc- cessful method for CAK. Documents and Information Review in Development of Airport Summary 1. CAK. Construction drawings (Plans and Specs, March 2006). 2. CAK. Operational and performance records. 2012. 3. Dvirka and Bartilucci. Anaerobic Fluidized Bed Design Basis Report for Akron-Canton Regional Airport. November 2005. 4. Facility Site Visit. 2012. 5. Gresham, Smith and Partners, Dvirka and Bartilucci. Summary of Akron-Canton Airport Deicer Management System. May 2006. 6. Gresham, Smith and Partners. Conceptual Engineering and Analysis of On-Site Treatment Technologies. May 2005. 7. McQueen, Rick. Akron-Canton Airport Deicer Management System Costs. 2007. E-mail.

D-40 Guidance for Treatment of Airport Stormwater Containing Deicers Airport Treatment Summary No. 5 Airport: Westover Air Reserve Base—Chicopee, MA (CEF) Treatment Technology: Passive Facultative Treatment (Subsurface Flow Wetland) Years Operated: 2001–2012 (Currently Operational) Deicer Management System Description Winter stormwater is collected at Westover Air Reserve Base (CEF) in storm drains adjacent to the deicing areas. The stormwater is directed through a constructed subsurface wetland system for treatment prior to discharge to Cooley Brook south of the base. The subsurface wetland is planted with reeds and is considered a passive facultative treatment technology. The reeds planted in the wetland system are fully grown. The CEF wetland system is currently in operation. There is, however, no ongoing monitoring program of water quality or quantity. Figure 1 shows an aerial photograph of the wetland. Deicer Treatment Technology Selection Considerations The constructed subsurface wetland system was installed to demonstrate the efficacy of the technology with respect to treatment of stormwater from deicing activity. A principal consideration Figure 1. Aerial photograph of the horizontal subsurface flow wetland at the Westover Air Reserve Base, Chicopee, Massachusetts, from NAVFAC Technical Report TR-2251-ENV.

Airport Deicer Treatment System Summaries D-41 for selecting a subsurface flow wetland (SSFW) system was limitation of free water surfaces to minimize attraction of birds and the potential threat to aircraft safety. A passive treatment technology was also desired—that is, one without pumps or other mechanical equipment. Deicer Treatment Technology Description Constructed Subsurface Flow Wetland See the passive facultative treatment systems fact sheet (Fact Sheet 108). Figure 2 shows plan and profile views of the SSFW. Description of Support Systems The SSFW at CEF requires oil/water separation, uniform distribution on the front of the system, an aggregate bed, and collection drains on the downstream side. Water elevation is con- trolled to partially submerge the gravel. Stormwater is diverted to the constructed treatment wetland after the oil/water separator adjacent to the outfall. The influent enters a splitter struc- ture that supplies two perforated inlet pipes on the front of the bed. The uniform distribution of stormwater is used to maintain equal distribution of the load through the entire SSFW system. The gravel bed serves as a medium for promoting anaerobic attached growth. The gravel layer is approximately 2-ft thick. The gravel bed is planted with reeds (phragmites) to assist in pro- moting growth of an attached biological treatment and as a means of exchanging nitrogen and phosphorous from the SSFW. Perforated outlet pipes on the opposite side collect the flow and discharge to the outflow structure. Figure 2. Plan and profile views of constructed subsurface wetland from NAVFAC Technical Report TR-2251-ENV.

D-42 Guidance for Treatment of Airport Stormwater Containing Deicers Key Treatment System Sizing Parameters1 See Table 1 for sizing parameters. Treatment System Performance The data in Table 2 presents the intended design performance of the system. The data in Table 3 on actual performance of the treatment system are from the 2002–2003 season. Although the system still operates, performance data have not been collected in recent years. Table 4 shows analytical results, and Table 5 shows a summary of BOD loadings. Cost Assessment for the CEF SSFW Treatment System See Table 6 for treatment system costs. Conclusions on Performance of CEF SSFW System Influent Deicer Concentrations Influent deicer concentrations are not a parameter used to demonstrate performance of the CEF SSFW since the system is slug loaded. Flow Rate The SSFW has a maximum flow rate that requires some flow to bypass the treatment system during events with high volumes. The flow rate is determined based on the hydraulic loading rate for the SSFW system. 1 From NAVFAC Technical Report TR-2251-ENV Component/Parameter Size/Capacity of Treatment Units Number of Treatment Units Total Capacity Stormwater storage capacity Not provided Not provided Not provided Treatment unit volume 120,000 gal 1 120,000 gal Treatment unit dimensions 0.6-acre x 2-ft deep 1 52,200 ft3 Treatment facility footprint 0.6 acre total site 1 0.6 acre Table 1. Key system sizing parameters. Parameter Value Unit Design flow rates - Minimum - Average - Maximum 0 70 280 Gallons per minute Design treatment load capacity 200* 120 120** lbs COD/day lbs BOD5/day lbs PG/day Design effluent concentration (average) <30 mg BOD5/L Design treatment efficiency (average) >80% % influent COD load treated *Data based on conversion: [COD] = 1.7 [BOD5]. **Data based on conversion: [PG] = [BOD5]. Table 2. Design basis for system performance.

Airport Deicer Treatment System Summaries D-43 Parameter Value Unit Flow rates - Minimum - Average - Maximum 0 118 354 Gallons per minute Actual COD treatment load rate - Maximum 1,550 lbs/day Actual BOD5 treatment load rate - Maximum 910 lbs/day Actual PG treatment load rate - Average - Maximum 2,185** 2,580** lbs/day Influent COD concentration - Minimum - Average - Maximum 100 1,335 23,100 mg/L Influent BOD5 concentration - Minimum - Average - Maximum 16 2,226 12,900 mg/L Influent PG concentration - Minimum - Average - Maximum 16** 2,226** 12,900** mg/L Effluent BOD5 concentration - Minimum - Average - Maximum 51 2,094 12,900 mg/L Treatment efficiency 10~80% % influent COD load treated *Data based on conversion: [COD] = 1.7 [BOD5]. **Data based on conversion: [PG] = [BOD5]. Table 3. Actual system performance. Parameter Units Statistic Wetland Inflow Wetland Outflow BOD mg/L Average 2,226 2,094 Max 12,100 12,900 Min 16.2 50.8 COD mg/L Average 1,883 1,335 Max 37,900 23,100 Min 3 100 MeBT mg/L Average 0.68 0.72 Max 20.93 4.77 Min 0.02 0.02 DO % Average 52.20 47.70 Max 103.90 69.80 Min 8.8 8.8 pH SU Average 7.58 9.54 Max 8.95 13.92 Min 5.61 6.54 Redox mV Average 391.00 172.00 Max 596.00 518.00 Min 235 -272 Temp C Average 17.30 18.90 Max 26.80 32.70 Min 10.8 12.1 Turbidity NTU Average 5.22 4.61 Max 10.70 7.06 Min 0.88 1.16 Table 4. Summary of analytical results from the Westover subsurface flow constructed wetland, 2002–2003 deicing season.

D-44 Guidance for Treatment of Airport Stormwater Containing Deicers Treated Load Rate The SSFW treatment system is not supported by a storage system. Therefore, the treatment load rates are slug loaded when a deicing event occurs. As a result, loading of the system was in excess of design due to the unrestricted flow, slug concentrations, and relatively small footprint. The observed treatment load rates were as low as baseline conditions and as high as 910-lbs BOD5/day. Effluent Concentrations Effluent concentrations from the SSFW were highly variable and, therefore, demonstrate that the current treatment system at CEF does not produce consistent concentration reductions. Treatment Efficiencies Like effluent concentrations, the treatment efficiencies from the SSFW at CEF are inconsistent. Although treatment efficiencies as high as 80% were achieved, generally the removal of organics as measured by BOD were low and erratic. The range of treatment efficiencies indicates that the SSFW at CEF is not a viable treatment technology for application at other airports. Cost The SSFW has an attractively low capital cost of approximately $350,000/acre (2002 dollars). Additionally, the annual operating costs are negligible. However, when evaluating the cost per pound of COD removed, or cost per effluent limit exceedance, the SSFW treatment system quickly becomes an unattractive economic investment. Lessons Learned for Potential Implementation of the SSFW Technology at Other Airports Conclusions from operation of the SSFW at CEF that can be used by other airports considering this technology include: 1. Although some treatment removals were achieved, the demonstrated removals of less than 20% are far too low for practical and successful application. Parameter Unit Dec. 2002 Feb. 2003 Mar. 2003 Apr. 2003 SSFW inflow (kg/d) --- 414 109 334 SSFW outflow (kg/d) --- 360 113 264 SSFW removed (kg/d) --- 54 -3 69 SSFW removed (%) --- 13 -3 20.8 Bypass --- 26 130 122 Combined Outfall 001 (kg/d) --- 386 243 386 Table 5. Summary of BOD loadings from the Westover subsurface flow constructed wetland, 2002–2003 deicing season. Cost Category Projected at Initial Implementation Actual Capital cost* $286,000 in 2002 $326,000 Annual operating cost - Utilities - Analysis Total operating cost Not provided Not provided Not provided $900 $7,000 $7,900 *Capital costs are for the treatment system only, including the building and basic building infrastructure. Costs do not include site-specific costs for collection, storage, and discharge. Table 6. Costs for the treatment system.

Airport Deicer Treatment System Summaries D-45 2. Treatment effectiveness might be improved by use of equalization up front to reduce slug loads. 3. No theoretical or empirical model provides guidance on the sizing criteria required to con- sistently achieve target effluent values. 4. The results corroborate results from the Wilmington Air Park passive subsurface treatment system pilot study performed from 1997 to 1999. Although the passive treatment technology is attractive since the management and cost of this system would be minimal, the SSFW at CEF indicates that passive systems are not credible means of treating deicer-affected stormwater. Documents and Information Review in Development of Airport Summary 1. Naval Facilities Engineering Command (NAVFAC). Cost and Performance Report Enhanced Biological Attenuation of Aircraft Deicing Fluid Runoff Using Constructed Wetlands, April 2005. 2. Jack Moriarty, Environmental Engineer of Westover Air Reserve Base. Personal Interview. February 1, 2012. 3. U.S. EPA. Federal Remediation Technologies Roundtable. Technology Cost and Perfor- mance Report Summary: Enhanced Biological Attenuation of Aircraft Deicing Fluid Runoff using Subsurface Flow Constructed Wetlands at the Westover Air Reserve Base, Chicopee, Massachusetts. 2005.

D-46 Guidance for Treatment of Airport Stormwater Containing Deicers Airport Treatment Summary No. 6 Airport: Cincinnati/Northern Kentucky International Airport—Kenton County, KY (CVG) Treatment Technologies: Activated Sludge, Mechanical Vapor Recompression Years Operated: 2003–2012 (Currently Operational) Deicer Management System Description The Kenton County Airport Board (KCAB), operators of CVG, began to address deicer man- agement issues in the early 1990s. At that time, local regulators made the airport aware of elevated BOD5 concentrations in the airport’s receiving streams. The regulator developed a TMDL for the receiving stream and subsequently incorporated effluent limitations in the airport’s NPDES per- mit. Implementation of a deicer management system was initiated with the construction of deic- ing pads in the 1990s, which are still in use. Subsequent deicer management elements included construction of pump stations near the airport outfalls to collect stormwater from much of the airfield, storage tanks, mechanical vapor recompression for treatment of higher-concentration runoff from the deicing pads, a sequencing batch reactor system, an extended aeration activated sludge treatment system, and aeration of stormwater storage basins just upstream of the airport discharges to the surface waters. Currently, CVG operates the deicing management system so that concentrations of glycol of >1% are sent to the glycol processing recycling facility (GPRF), and concentrations of glycol of <1% are sent to the stormwater treatment plant (SWTP). The GPRF uses MVR treatment to concentrate spent aircraft deicing fluid. The SWTP uses biological activated sludge and support systems to remove glycol from the deicing-affected stormwater. The pump stations at the airport outfalls are designed to collect a maximum of 5.8 million gallons per day. Figure 1 shows a process flow diagram of the system, and Figure 2 shows an aerial view of the system. Figure 1. CVG deicing-affected stormwater management system process flow diagram.

Airport Deicer Treatment System Summaries D-47 Deicer Treatment Technology Selection Considerations Deicer management at CVG has evolved over time to meet changing permit requirements, deicer use changes, and performance needs. The initial implementation consisted of deicing pads, storage, and subsequent recycling using mechanical vapor recompression, plus discharge of runoff from other areas of the airport to the local POTW. Deicer concentrations routed to the mechanical vapor recompression recycling system average approximately 6% propylene glycol concentration. The recycling system continues to operate today. After the initial implementation of deicing pads and the recycling operation, aerators were installed in large ponds upstream of the receiving streams. The ponds had been designed primar- ily to manage the quantity and timing of stormwater discharges, and the aeration function was added after initial pond construction. The objective of adding aerators was to polish the water to reduce BOD5 concentrations prior to discharge to the receiving streams. However, because of multiple influent sources and variable flow volumes entering the aerated basins, it has been dif- ficult to characterize what effect the aeration is actually having on BOD5 concentrations. As the airport operations and deicer use changed over time, the local POTW began having issues with the BOD5 load from airport deicers swamping the POTW’s treatment capacity. This led KCAB to seek methods for on-site treatment of the deicer-affected flows from large portions of the airport rather than relying on a discharge to the POTW. One of the principal challenges that KCAB faced was the large drainage area that needed to be collected from to meet regula- tory requirements, at least partially driven by concerns with biological nuisance growth in the streams. Collection of the stormwater necessitated installation of pump stations first for the discharges to Gunpowder Creek and later for the discharges to Elijah Creek. Finding appropri- ate and effective means to treat high flow volumes with highly variable BOD5 concentrations is Figure 2. Aerial photo of CVG activated sludge treatment system (Photo courtesy of KCAB).

D-48 Guidance for Treatment of Airport Stormwater Containing Deicers one of the more significant deicer management challenges that an airport can face. Several years of study were performed prior to selection and design of the initial biological treatment system. The first phase of on-site treatment at the airport included installation of SBR-based aero- bic biological treatment. The SBR technology is a variation on the activated sludge technology where the treatment is provided by aerobic bacteria suspended in a reactor. SBRs are operated in a fill-and-draw batch mode. Testing of the SBR technology was performed by KCAB and led to the design of a full-scale system with three SBR reactors. The SBRs were sized based on two key assumptions: (1) the deicer applied on aircraft was ethylene glycol based, and (2) only the col- lected flow from the Gunpowder Creek drainage areas would be treated. During the construction and initial operation of the SBR systems, the airlines switched to propylene-glycol–based deicers, and the Elijah Creek drainage area runoff was added to the volume that needed to be treated. Both of these changes led to a lack of capacity with the SBR units, and it was clear additional treatment capacity would be needed. The second phase of biological treatment implementation at CVG led to installation of a treatment system based on extended aeration activated sludge treatment technology. The new system had a greater BOD5 treatment capacity (30,000 lbs/day) and was designed as the principal means for degrading the deicer-affected stormwater. As part of the implementation of this sys- tem upgrade, the SBR units were converted into two sludge digesters and a sludge storage tank for the activated sludge system’s effluent solids. The airport has progressively been improving the efficiency and effectiveness of the activated sludge treatment system, focusing in particular on ways to manage the high and variable volumes of deicer-affected runoff from the airport. Deicer Treatment Technology Description Activated Sludge Biological Treatment System Fact Sheet 101 contains a general description of the activated sludge technology. This airport’s particular activated sludge system is designed to promote an extended detention time for the deicer in the treatment system to help manage the variability in deicer loadings and create a more stable bacterial population than conventional activated sludge technology. The system is aerated using mechanical blowers supplying air through fine-bubble diffusers suspended from floating aeration chains in three aeration basins (see Figure 3). Air delivery to the basin can be Figure 3. The airport’s activated sludge treatment system aeration basins (Photo courtesy of KCAB).

Airport Deicer Treatment System Summaries D-49 reduced during periods of low loading while still maintaining the contact between the deicer and reducing the risk of solids settling out of the water. A clarifier for removal of solids is integral to the aerated basin unit. The CVG system is designed to treat a maximum of 30,000-lbs BOD/day over an extended period, although over short periods of less than 1 to 2 days, it can treat up to 45,000 lbs/day. Mechanical Vapor Recompression The concentrated deicer collected from the deicing pads at the airport is processed in mechan- ical vapor recompression units housed at the airport for the purpose of creating recycled propyl- ene glycol. See Fact Sheet 106 for a description of this technology. Description of Support Systems CVG’s deicer management system has 10 pump stations for collecting runoff from the airfield. The pump stations are designed to collect all flows up to the maximum capacity of the pumps, sending those flows to the activated sludge treatment system. Higher-concentrate deicer that is collected from the deicing pads is routed to three aboveground storage tanks totaling 8 million gallons in capacity. The deicer in the storage tanks can be routed to the MVR system or can be bled into the activated sludge treatment system to help sustain the biomass. Chemical nutrients are added to the aeration basins in the activated sludge system to pro- mote healthy growth of the microorganisms living in the activated sludge aeration basins. Add- ing sufficient nutrients is a key element of maintaining an effective biomass in the aeration basins and getting good treatment. Blowers are used to supply air to the three aeration basins. The aeration of the three basins is independently controlled, and CVG may not always operate all of the basins. The extended aeration activated sludge process produces biological solids. A portion of these are routed back into the aeration basin to help maintain a high level of mixed liquor suspended solids in the aeration basins. The remaining solids are wasted from the treatment system. A clarifier integral to the aeration basin is used to settle biological solids contained in the treated effluent. Management of the wasted sludge and the solids removed in the clarifier is a significant com- ponent of management of the treatment system. CVG has an advanced system for the dewatering and disposal of the biological solids. The dewatered solids are blended with soil and land applied on the airport property to save disposal costs. Key System Sizing Parameters Parameters that quantify the size and capacity of the key components of the airport’s deicer management system related to treatment are provided in Tables 1 and 2. Component/Parameter Number of Units Total Capacity Activated sludge treatment unit volume 3 14 million gallons Activated sludge aeration basin footprint 3 11 acres Tanks for spent aircraft deicing fluid 2 6 million gallons Mechanical vapor recompression system building footprint 3 acres Table 1. Key system sizing parameters.

D-50 Guidance for Treatment of Airport Stormwater Containing Deicers Treatment System Performance2 The data in Table 3 on the intended design performance of the activated sludge system were derived from airport records and publicly available presentations. The information in Table 4 and Table 5 on actual system performance was derived from daily data collected at the facility for the 2011–2012 deicing season. The average TSS effluent concen- tration was 2 mg/L. Cost Assessment for CVG’s Activated Sludge Treatment System Table 6 shows treatment system costs, and Table 7 shows improvement costs. Conclusions on Performance of Airport’s Activated Sludge System Influent Deicer Concentrations Concentrations of influent to the CVG extended aeration activated sludge treatment system average 1,300-mg/L COD, with a peak of 5,600 mg/L. While these are not unusual concentra- tions for deicer treatment, they are higher than what activated sludge systems experience in their typical municipal sanitary wastewater applications. The variability of concentrations at CVG is relatively significant compared to non-deicer applications. CVG does not have storage for the diluted stormwater collected from the drainage areas. The large potential capacity of the 2 Data from outlier periods of operation were excluded from the performance analysis, including data prior to installation of the activated sludge system, 2 weeks of a pump malfunction starting on February 12, 2010, and various periods with main- tenance or operational issues. These outlier data represent periods where circumstances outside of the treatment technology operation affected the treatment results. Parameter Value Unit Air delivery rate ranges - Activated sludge aeration basin 1 - Activate sludge aeration basins 2 and 3 1,600–2,950 2,900–5,300 scfm Total storage (three tanks) 8.0 Million gallons Note: scfm = standard cubic feet per minute. Table 2. Additional system sizing parameters. Parameter Value Unit Design flow rates - Average - Maximum 3,470 5,200 Gallons per minute Design treatment load capacity 61,200* 36,000 36,000** lbs COD/day lbs BOD5/day lbs PG/day Design influent concentration - Range 1220* 719.5 719.5** mg COD/L mg BOD5/L mg PG/L Design effluent concentration (average) 50 30 mg BOD5/L mg TSS/L Design treatment efficiency (average) 93% % influent COD load treated Design data courtesy of Parkson. *Data based on conversion: [COD] = 1.7 [BOD5]. **Data based on conversion: [PG] = [BOD5]. Table 3. Design basis for system performance.

Airport Deicer Treatment System Summaries D-51 aeration basins aids in managing the variable concentrations. Maintaining a constant and healthy biomass in response to variable concentrations can be challenging at times because flows into the treatment are not controlled. Flow Rate The activated sludge treatment system at the airport has treated flow rates of as high as 1,950 gpm. The system has the capacity to treat up to 7 mgd (approximately 4,700 gpm). KCAB indicated that there is additional hydraulic capacity available in its system. Treated Load Rate The CVG activated sludge technology has significant capacity to receive BOD loads while meeting water quality goals. This type of biological treatment technology is most efficient and effective where influent BOD loads do not vary significantly because a constant BOD load facili- tates a less variable biomass population. During low deicing periods, the airport has had to Parameter Value Unit Flow rates - Minimum - Average - Maximum 278 1,950 4,700 Gallons per minute Actual COD treatment load rate - Average - Maximum 20,000* 81,000* lbs/day Actual BOD5 treatment load rate - Average - Maximum 11,800 47,600 lbs/day Actual PG treatment load rate - Average - Maximum 11,800** 47,600** lbs/day Influent COD concentration - Minimum - Average - Maximum 28 1,300 5,600 mg/L Influent BOD5 concentration - Minimum - Average - Maximum 16* 760* 3,300* mg/L Influent PG concentration - Minimum - Average - Maximum 16** 760** 3,300** mg/L Effluent COD concentration - Minimum - Average - Maximum 1 16 81 mg/L Treatment efficiency 98.7% % influent COD load treated *Data based on conversion: [COD] = 1.7 [BOD5]. **Data based on conversion: [PG] = [BOD5]. Table 4. Actual system performance. Parameter Value Number of collection pump stations 10 Nitrogen addition 130 gallons of urea ammonium nitrate per day (31% urea ammonium nitrate) Phosphorus addition 22 gallons of phosphoric acid per day (75%) Table 5. Actual performance of treatment system support systems.

D-52 Guidance for Treatment of Airport Stormwater Containing Deicers Cost Category Projected at Initial Implementation Actual Capital cost Not provided SWTP (2001) Pilot: $393,425 SWTP (2005) Engineering/design: $2,016,184 SWTP (2007) Gunpowder Creek: $9,627,798 SWTP (2008) Elijah Creek: $16,688,626 SWTP Total: $28,726,033 GPRF (2010) Treatment system: $9,038,217 Capital improvements: $76,302 GPRF (2011) Rehabilitation of existing storage tanks: $1,315,978 GPRF total: $10,561,595 Annual operating cost* $0.01/gal treated $304,000 Source of capital cost data: Spent Aircraft Deicing Fluid Management System Letter Report (April 20, 2012). *Operating costs based on 2008–2009 deicing season. Table 6. Costs for the treatment system. Completion Project Description Total Cost Category Dec 1993 Northwest Environmental Collection System $1,591,378 Sep 2001 SWTP – Pilot Project 393,425 Dec 2002 South Detention Basin Closure 3,215,322 May 2003 Runway 36R Large Hold Pad and Deicing Recovery System 2,860,726 Dec 2003 Southwest Detention Facility – Land and Easements 4,100,151 Dec 2003 Southwest Detention Facility – Construction 3,072,641 Dec 2003 Deicing System Enhancements: Increase Storage Capacity 7,505,566 South Airfield Glycol Dispensing/Storage Facility 1,447,076 Oct 2005 SWTP – Instream Engineering/Design 2,016,184 Sep 2006 Deice Pad 8 Reconstruction 252,971 Jan 2007 SWTP – Gunpowder Creek 9,627,798 Jun 2008 SWTP – Elijah Creek 16,688,626 Jun 2008 Detention Facilities – Design 402,178 Jun 2008 Detention Facilities – All Other Costs 3,497,960 Sep 2008 Runway 17/35 (Future 18R/36L)-SWTP Elijah Creek 2,563,062 Sep 2008 Runway 17/35 (Future 18R/36L)-Gunpowder Creek 332,007 Sep 2010 GPRF – Instream Treatment 9,038,217 GPRF – Misc. Cap Improvements 76,302 Sep 2011 Rehab Existing Glycol Concrete Storage Tanks 1,315,978 Grand Total – All Projects $69,997,568 Notes: Categories: C = collection; S = storage/containment; P = processing/treatment. Table 7. Costs for improvements to the spent aircraft deicing fluid management system.

Airport Deicer Treatment System Summaries D-53 supplement the food (BOD) source to the treatment system to maintain an active biomass, resulting in additional costs incurred. Although the BOD load to the treatment system varies, the treatment system does operate at its 30,000-lbs/day limit at times. Effluent Concentrations The effluent concentrations achieved by the CVG treatment system routinely meet the water quality goals as long as the BOD load entering the system is not greater than the treatment capacity of the system. To mitigate for the risk of insufficient treatment capacity, the operators maintain a high biomass population and high treatment capacity through regular supplements of high-concentrate deicing pad runoff stored in the aboveground storage tanks. Treatment Efficiency The overall removal efficiency for COD has been over 98%, which has been sufficient to meet water quality goals. The actual treatment efficiency is similar to the treatment efficiency that was expected. Operational adjustments have been made to address unusual system conditions such as low temperature and excessive loading that can cause short-term decreases in treatment efficiency. Treatment in cold weather temperatures has not been a significant issue at CVG due to several factors, including (1) the large biological population that is maintained in the aeration basins and (2) heat generated from the biological degradation. Cost The capital costs for collection, storage, and treatment at the airport were approximately $70 million. The total costs for the SWTP and the GPRF were $39.3 million. The anticipated annual labor costs for the collection storage and treatment system were $0.01/gallon treated. Actual annual costs were approximately $304,000. Tables 6 and 7 provide additional information on the historical improvements at CVG and the associated capital costs. Lessons Learned for Potential Implementation of the Activated Sludge Technology at Other Airports The following parameters are critical to the success of the CVG activated sludge system: 1. Monitoring of influent characteristics. 2. Monitoring and management of nutrient loadings. 3. Maintaining sufficient and uniform dissolved oxygen concentrations. 4. Measuring and calculating food-to-microorganism ratios. 5. Avoiding spike loadings from storage. 6. Management of wasted and clarified biological solids. 7. Effectively managing the redevelopment of the biological population at the start of each new season. General insights from the operation of this system that could be applied elsewhere include: 1. At CVG, the operators have experimented with various means of re-establishing the desired biological population for the start of the new deicing seasons. Experiments have included: a. Stopping treatment activity and reseeding the system in the fall with sludge from a local POTW. b. Keeping the system active the entire summer through the feeding of BOD sources, nutri- ents, and oxygen. c. Providing only aeration in the summer with no feeding of a BOD source or nutrients, followed by a fall start-up without seeding, initial introduction of nutrients, and minimal initial BOD sources. The operators feel this method works well.

D-54 Guidance for Treatment of Airport Stormwater Containing Deicers 2. It is necessary to add nutrients to the aeration basin on a regular basis. In the early years of operation, nutrient addition was less regular. The operators found that without nutrients, the health of the biological population decreases and treatment effectiveness decreases signifi- cantly. The operators have also noted that while concentrations of nutrients, in the form of total phosphorus and total nitrogen, are regularly measured in the treated effluent, the ana- lytical tests on effluent samples measure both residual chemical concentration from added nutrients and nutrients released from dying bacteria. Since nutrients released from dying bacteria may not necessarily be in a form that can be absorbed by the living bacteria, the effluent nutrient measurements may overestimate the available nutrients. Therefore, while the operators try to maintain total phosphorus concentrations in the effluent of 1 mg/L and total nitrogen concentrations of 1–1.5 mg/L, they also add nutrients regularly even if residual concentrations are measured in the effluent characterization. 3. When the change was made from an EG-based deicer to a PG-based deicer, the biological treatment system reacted differently. It appears that treating the EG provided a somewhat more stable biological population. 4. During one season, due to a malfunction, the treatment system was significantly overloaded with propylene glycol, causing much of the biomass in the treatment system to die, greatly reducing the treatment capacity for a time. From this episode, it became apparent that: a. Once the bacteria population is lost, weeks may be required to regain full treatment capacity. b. The treatment system is resilient and flexible within a range of influent concentrations but not for all conditions. c. Appropriate monitoring and checks can be incorporated to quickly identify or prevent such operational issues. Documents and Information Review in Development of Airport Summary 1. As-builts (record drawings). 2. Dietrich, Tom. Facility from Site Visits. 2011. 3. KCAB. Photographs. 2012. 4. CVG. Record drawings (plans and specs). 5. CVG. Operational logs and daily laboratory worksheets. 2012. 6. CVG. Airport correspondence and communications. 2011. 7. CVG. Treatment System Operation and Maintenance Manual. 8. CVG. Spent Aircraft Deicing Fluid Management System Letter Report by Leigh Fischer. April 20, 2012. 9. U.S. EPA. NPDES Permit No. KY0082864. 2005.

Airport Deicer Treatment System Summaries D-55 Airport Treatment Summary No. 7 Airport: Denver International Airport—Denver, CO (DEN) Treatment Technology: Mechanical Vapor Recompression and Distillation Years Operated: 2004–2012 (Currently Operational) Deicer Management System Description DEN is owned, operated, and maintained by the city and county of Denver (collectively, “the city”). The airport was built during the early 1990s and was specifically designed with infrastruc- ture meant to reduce or control the potential for spent ADF to contribute pollutants to storm- water discharges. Deicer-affected stormwater is managed as part of DEN’s airport deicing system (ADS). Components of the ADS include dedicated deicing pads, a deicing waste stormwater collection system, low-flow stormwater runoff diversion from the clean stormwater system into the deicing waste stormwater system, storage, a spent deicing fluid recycling plant, and discharge of lower-concentration runoff to the POTW. The city contracts with an operator to maintain, operate, and manage the ADS. Figure 1 shows a process flow diagram of the system. Currently, full deicing at DEN is permitted on deicing pads and some aprons, while deicing on gates and concourses is limited. Each deicing area has a dedicated collection system that conveys stormwater runoff contaminated with ADF to temporary storage ponds or tanks. Conveyance of the runoff is managed through a system of valves and underground piping. Runoff is segregated based on glycol concentration for recycling (high concentration) and discharge to the POTW (low concentration), as described in the following. Figure 1. DEN deicing-affected stormwater management system.

D-56 Guidance for Treatment of Airport Stormwater Containing Deicers High-concentrate runoff of greater than 1% concentration (10,000-mg/L PG) is managed for recycling-based treatment as follows. Deicer-affected runoff from the four central deicing pads is collected and conveyed to one of five 420,000-gallon storage tanks. In addition, a 4-million-gallon pond (Figure 2) can be used as contingency storage for high-concentrate runoff. Runoff from independent deicing pads is collected separately in storage tanks of 835,000 gallons and 420,000 gallons. Collected higher-concentrate runoff is conveyed from storage to the recycling plant via a pump and piping system or via truck. On deicing areas that are not part of the ADS, the operator uses a GRV in order to recover all fluids of over 1% glycol concentration. See Figure 3 for a photo- graph of the MVR building. The recycling equipment includes eight MVR units (Figure 4) and a vacuum distillation sys- tem to produce 99%+ PG. The MVR systems in use at DEN are called “aircraft deicing fluid concentrators.” Condensate from the MVR units and distillation systems contain BOD5 concen- trations that require monitoring and storage at the Western Airfield Diversion System (WADS) prior to metering to the Metro Wastewater Reclamation District’s wastewater treatment plant, the local POTW. The residual waste, which contains additives and contaminants, removed from the deicing fluid by the vacuum distillation system is sent off-site for disposal at an approved waste handling facility. The 99%-PG product is sold. Figure 2. Pond 003A at the glycol recycling facility. Figure 3. MVR building.

Airport Deicer Treatment System Summaries D-57 Lower-concentrate runoff from concourses and ramps of less than 10,000-mg/L PG is directed to lined retention ponds and sent to WADS for metering to the sanitary sewer system and sub- sequent treatment at the local POTW. The effluent discharge limitations to the POTW from recycling operations are: • Daily maximum BOD load: 0.5 tons, • Instantaneous maximum concentration COD: 2,500 mg/L, and • Maximum daily flow volume: 0.288 mgd. Deicer Treatment Technology Selection Considerations As part of the original ADS system when the airport was constructed, the city built an on-site distillation plant to recycle spent ADF. Later, the city added a preconcentrator evaporator system, which was operated until 2004. Subsequently, MVR technology was installed at DEN to replace the preconcentrator system after the operator determined that the MVR technology was more energy efficient and economical than the preconcentrator system. Currently, both MVR and distillation technologies are used to recycle spent ADF at DEN. The current operator owns the MVR technology. Based on high ADF usage and local climate characteristics, recycling systems were considered ideal treatment technologies for DEN. Spent ADF collected at DEN is generally higher in average propylene glycol concentration since the area’s snow moisture content is typically low, and full deicing is only allowed on dedicated pads with collection capability. This yields a larger volume of spent ADF containing above-average glycol concentration that is considered ideally suited for recycling. Since the majority of captured spent ADF can be recycled, large volumes of propylene glycol can be reclaimed and sold in secondary industrial markets. This generates higher revenues related to the sale of recycled glycol, which offsets overall ADS management costs. It is not economical to recycle runoff with PG concentrations of less than 10,000 mg/L due to the large water content that must be evaporated. This requires a separate treatment technology, Figure 4. MVR unit.

D-58 Guidance for Treatment of Airport Stormwater Containing Deicers which at this time is metering to the local POTW. DEN has a user permit for this discharge and pays fees based on volume and BOD load. Deicer Treatment Technology Description Descriptions of the vapor recompression and distillation treatment technologies can be found in Fact Sheet 106 and Fact Sheet 105, respectively. The MVR systems at DEN were designed to treat all spent ADF with glycol concentrations between 1% and 25% and concentrate to a mini- mum concentration of between 38% and 55% (Figure 4). The MVR concentrate glycol is routed into intermediate storage tanks and then sent through the distillation system (Figure 5), which generates a distillate of >99% PG. All ADF at DEN is PG-based. Description of Support Systems The deicer treatment technology at DEN includes support systems for the MVR and distillation systems such as influent filtration systems and effluent glycol polishing units. Each ADF concentrator includes the following support systems: blowers, main plate heat exchanger, stainless-steel tanks and piping, and scrubber-absorber. The instrumentation includes pressure, temperature, and flow transmitters and gauges, and a control panel with PLC. Stainless-steel hot filter vessels with 1-micron filter bags are used on each MVR prior to the feed entering the unit. This allows the influent to be filtered while it is hot in an effort to remove as much TSS as possible, thus maximizing production throughput and minimizing stoppages due to premature maintenance and cleaning requirements of the MVR heat exchangers. Figure 5. Distillation columns.

Airport Deicer Treatment System Summaries D-59 An activated carbon filtration step was added to the influent of the distillation process to decrease the amount of solids and particulate matter that normally would build up in the heat exchangers and cause loss of heat exchange. Without this filtration step, the system would require frequent shutdowns to perform tedious maintenance. The overall positive result is an increase in performance and productivity of the distillation system. The distillation system includes the following equipment: • Numerous pumps and motors. • Instrumentation: pressure, temperature, and flow transmitters and gauges. • Control panel with PLC equipment. • Various motor controllers. The final step in the recycling process after the fluid has been sent through the MVR systems and distilled to 99%+ PG concentration is a product value-added step called “polishing.” The glycol polisher is a carbon filtration, deionization, and demineralization process that was developed to remove trace airfield contaminants left in the 99% glycol after distillation. An operator control room is located in the distillation facility. This room is equipped with a computer control station with an interactive system to monitor operating conditions. The automation system uses a PLC to provide the operator access to operating data and monitor alarms. This location serves as the main monitoring area where most of the system can be operated. Key Treatment Sizing Parameters Table 1 and Table 2 show size parameters for the system. Treatment System Performance Table 3 through Table 6 show system design performance information, and Table 7 and Table 8 show actual values. Cost Assessment for Treatment System The costs indicated in the following reflect the MVR and distillation treatment technolo- gies. At DEN, the airport was responsible for the capital cost of the recycling building, facility infrastructure, ADF distribution system, ADF distribution tanks, collection piping, collection tanks, and distillation system, and the recycling/treatment equipment. The vendor operating the system was responsible for the capital costs of the MVR treatment units. Component/Parameter Size/Capacity of Treatment Units Number of Treatment Units Total Capacity Treatment unit dimensions MVR Distillation 6-ft L x 20-ft W 45-ft L x 40-ft W x 23-ft H 8 1 960 ft2 1,800 ft2 Treatment facility footprint MVR Distillation 0.05-acre building 0.07-acre building (34-ft H) 1 1 0.12 acres Table 1. Treatment system size and capacity parameters.

D-60 Guidance for Treatment of Airport Stormwater Containing Deicers Component/Parameter Size/Capacity of Treatment Units Number of Treatment Units Total Capacity Recycling system stormwater storage capacity 420,000 gallons 835,000 gallons 3,200,000 gallons 5 1 1 6.135 million gallons Concentrated recycled product storage tanks 7 tanks; each tank is 12’ D x 25’ H 20,000 gallons each 140,000 gallons Low-concentrate system (POTW discharge) stormwater storage Ponds 001, 002, 004, 005 and 009 3.5–30 million gallons 60.5 million gallons Table 2. Treatment support systems size and capacity parameters. Parameter Value Unit Design flow rates - Minimum - Average - Maximum 2 34 Not provided Gallons per minute Design treatment load capacity 16,500* 9,700** 9,700 lbs COD/day lbs BOD5/day lbs PG/day Design influent concentration - Range 17,000~459,000* 10,000~270,000** 10,000~270,000 mg COD/L mg BOD5/L mg PG/L Design concentrate (glycol) stream concentration - Range 35–55% % PG Design effluent (condensate) concentration - Range <50~1000 Not available <50~1000 mg COD/L mg BOD5/L mg PG/L Design treatment efficiency 94.1~99.7 % influent COD load removed *Data based on conversion: [COD] = 1.7 [BOD5]. **Data based on conversion: [PG] = [BOD5]. Table 3. Design basis for MVR system performance. Parameter Value Unit Design flow rates - Minimum - Average - Maximum 9.7 10 12.5 Gallons per minute Design treatment load capacity 107,100* 63,000** 63,000 lbs COD/day lbs BOD5/day lbs PG/day Design influent concentration - Range 646,000~714,000* 380,000~420,000** 380,000~420,000 mg COD/L mg BOD5/L mg PG/L Design concentrate (product) stream concentration - Range 99-99.5 % PG Design effluent (condensate) Concentration*** Range 8,000~15,000 Not available 5,000~10,000 mg COD/L mg BOD5/L mg PG/L Design treatment efficiency Not provided % influent COD load removed *Data based on conversion: [COD] = 1.7 [BOD5]. **Data based on conversion: [PG] = [BOD5]. ***Condensate from the distillation unit is sent to the MVR for additional treatment. Table 4. Design basis for distillation system performance.

Airport Deicer Treatment System Summaries D-61 Parameter Single-Stage Production Two-Stage Production Stage 1 Stage 2 Influent flow rate range (gallons per hour) 150 to 200 170 to 230 130 to 170 Influent glycol concentration range (% glycol) 4 to 27 1 to 4 13 to 27 Influent temperature range (F or C) Ambient Ambient Ambient Number of effluent streams produced Two streams –distillate and concentrate Two streams – distillate and concentrate Two streams – distillate and concentrate Distillate effluent flow rate range (gallons per hour) 60 to 184 136 to 219 52 to 126 Distillate effluent water quality (COD range in mg/L) <50 to 1,000 <50 to 1,000 <50 to 1,000 Distillate effluent water quality (pH range) 3 to 8 3 to 8 3 to 8 Concentrate effluent flow rate range (gallons per hour) 12 to 120 8.5 to 61 33 to 102 Concentrate effluent concentration (% glycol range) 50 to 55 15 to 20 50 to 55 Heat source Electric-powered steam compression Control system PLC Energy consumption information 0.4Kw per gal feed Estimate of waste to be produced Sludge and solids negligible and glycol in overheads less than 0.1% Anticipated frequency of maintenance activities Duty cycle of 95% expected depending on influent quality Footprint, dimensions, etc. Each MVR unit is 20’ (L) x 6’ (W) x 8’ 2” (H), with scrubber 13’ (H) or 22’ (H). Other support systems Feed preheater heat exchanger, electric air compressor, cold and hot filter systems, piping for feed, distillate, concentrate, and storage tanks for feed, distillate, concentrate Note: DEN has eight MVR units installed. Table 5. Additional design basis for MVR system. Parameter Value Unit Influent flow rate range 14,000 to 18,000 gpd Influent flow rate average 9.7 to 12.5 gpm Influent glycol concentration range* 38 to 42 % propylene glycol Influent temperature range 40–50 °F Number of effluent streams produced 2 Distillate and concentrate Distillate effluent flow rate range 5.5 to 7.5 gpm Distillate effluent water quality range (COD) 8,000 to 15,000 mg/L Distillate effluent water quality range 0.5 to 1 % propylene glycol Distillate effluent water quality 3 to 8 pH Product effluent flow rate range 4 to 5 gpm Product effluent glycol concentration 99–99.5 % propylene glycol *Criteria provided based on design change in 2004. These are not the original specifications when the unit was built. Table 6. Additional design basis for distillation system performance.

9-28 to 10-25 10-26 to 11- 22 11-23 to 12- 20 12-21 to 01-17 1-18 to 2-14 2-15 to 3- 14 3-15 to 4-11 4-12 to 5-9 5-10 to 6-6 6-7 to 7-4 7-5 to 8-1 8-2 to 8-29 08-30 to 9-26 Total or Avg for Season Number of MVR units 6 MVRs 6 MVRs 6 MVRs 6 MVRs 8 MVRs 8 MVRs 8 MVRs 8 MVRs 8 MVRs 8 MVRs 8 MVRs 8 MVRs 8 MVRs Influent vol processed (gallons) 570,777 542,268 555,587 586,543 832,951 723,507 704,000 750,366 804,766 770,043 800,906 766,551 716,590 9,124,855 Avg influent glycol concentration (% PG) 5.0 10.5 12.0 16.0 10.3 17.0 22.5 19.4 16.3 13.5 7.5 5.5 5.5 12.4 Vol of 100% PG in influent (gallons) 28,539 56,938 66,670 93,847 85,377 122,996 158,400 145,885 130,774 103,956 60,068 42,160 39,412 1,135,024 Avg influent temperature °C 65 64 61 61 62 67 69 66 72 75 77 76 74 68 Avg influent flow rate (GPH)* 972.5 845 855.1 903.9 1,431.7 1,121.3 1,150.8 1,163.6 1,270.0 1,204.7 1,228.7 1,180.9 1,132.4 1,112 Hours of operation 3,566 3,850 3,898 3,891 4,698 5,158 4,883 5,156 5,072 5,110 5,215 5,189 5,062 60,748 Duty cycle (%)** 94.0 95.0 97.0 97.0 97.0 96.0 91.0 96.0 94.0 95.0 97.0 97.0 94.0 95 Effluent vol of Distillate produced (gallons) 467,599 436,260 442,675 394,482 566,113 472,762 332,578 412,846 504,306 523,080 654,118 664,969 624,052 6,495,840 Avg effluent distillate per MVR (COD in mg/L) 6,044 6,516 4,981 4,201 5,592 4,070 3,085 3,779 4,241 6,478 6,906 6,298 6,079 5,252 Effluent vol of concentrate produced (gallons) 103,178 106,008 112,912 192,061 266,838 250,745 371,422 337,520 300,460 246,963 146,788 101,582 92,538 2,629,015 Avg effluent concentration of concentrate (% PG) 26.0 48.0 48.0 42.0 29.5 45.5 42.5 43.0 43.5 42.0 40.5 40.5 40.0 41.1 Vol of 100% PG in concentrate (gallons) 26,826 50,884 54,198 80,666 78,717 114,089 157,854 145,134 130,700 103,724 59,449 41,141 37,015 1,080,397 % ratio of glycol reclaimed vs. infeed 94.0 89.4 90.3 91.7 92.2 92.8 99.7 99.5 99.9 99.8 99.0 97.6 93.9 95.2 Data compiled per 28-day period while MVR systems were running. All treatment data provided by operational logs provided by Inland Technologies Int’l Ltd., DEN staff. *Average flow rate of all machines running during this time period. **Average % hours operation calculated by comparing how many hours the MVRs ran against theoretical hours possible for the time period the machines were running. ***Balance of PG discharged in distillate effluent to POTW. Table 7. Actual DEN MVR data for 2009–2010 deicing season.

Airport Deicer Treatment System Summaries D-63 Feb Mar Apr May Jun Jul Aug Sep Total or Avg for Season Influent volume processed (gallons) 143,874 281,391 341,698 215,092 246,194 206,286 213,827 66,847 1,715,209 Average concentration of influent glycol (% PG) 42.00% 40.00% 41.00% 41.00% 40.00% 37.50% 35.00% 37.00% 39.45% Volume of 100% PG in influent (gallons) 60,427 112,556 140,096 88,188 98,478 77,357 74,839 24,733 676,675 Average influent temperature (°F) 50 50 50 50 50 50 50 50 50 Average influent flow rate (GPH) 630 600 570 570 600 570 600 600 592.5 Average hours of operation (HPD) 24 24 24 23.5 24 24 24 23.5 23.875 Effluent volume of distillate produced (gallons) 86,661 164,601 201,305 146,981 142,674 120,550 131,610 45,355 1,039,737 Avg concentration of effluent distillate (% PG) 1.0% 1.0% 1.0% 1.0% 1.0% 1.0% 1.0% 1.1% 1.00% Average COD of effluent distillate (mg/L) 11,890 10,262 12,310 12,980 11,860 12,370 13,070 14,250 8,624 Effluent volume of product produced (gallons) 57,213 116,790 140,393 68,111 103,520 85,736 82,217 21,492 675,472 Avg concentration of effluent product (% PG) 99.2% 99.1% 99.2% 99.2% 99.2% 99.2% 99.2% 99.2% 99.18% Volume of 100% PG in effluent product (gallons) 56,755 115,739 139,270 67,566 102,692 85,050 81,559 21,320 669,951 % ratio of glycol produced vs. glycol in feed* 93.92% 102.83% 99.41% 76.62% 104.28% 109.94% 108.98% 86.20% 99.01% *Variability per month due to timing of when first and last processing data were recorded. Table 8. Actual DEN distillation data for 2009–2010 deicing season. The installation cost of the recycling building, facility infrastructure, ADF distribution sys- tem, ADF distribution tanks, collection piping, collection tanks, and distillation system was $14.6 million. Details were not available to determine the capital costs associated only with recycling/treatment equipment associated with the distillation system. The MVR facility was installed at a cost of $1 million. The cost of eight MVR treatment units and associated support equipment was $2.8 million. At DEN, the operating costs to recycle spent ADF are borne by the recycling vendor. The airport only covers capital replacement costs for the distillation system when the major com- ponents fail. The airport pays the operating costs to manage and discharge all spent ADF of less than 1% glycol concentration. Cost Category Projected at Initial Implementation Actual Capital Cost Collection system, treatment building, and distillation system Treatment building and system Not provided Not provided $14.6M $2.8M Annual Operating Cost - Utilities - Chemicals - Analysis - Material handling Total operating cost Not provided Not provided Not provided Not provided Not provided Not provided Not provided Not provided Not provided $1.5M~2.0M *The annual operating costs are typically offset by the sale of the recovered PG. Table 9. Costs for the treatment system.

D-64 Guidance for Treatment of Airport Stormwater Containing Deicers Conclusions on Performance of DEN MVR and Distillation Treatment System Influent Deicer Concentrations Influent glycol concentration is a primary parameter used to demonstrate the performance of the MVR systems. With each deicing event, concentrations of spent ADF fluctuate. Collection during storm events can generate influent concentrations that range from 1% to 25% during any single deicing event. The MVRs are capable of handling these concentrations without any major setbacks. The effluent from the MVR is the influent for the distillation treatment system. Therefore, the concentration of glycol influent sent to the distillation system is directly influenced by the glycol produced by the MVR systems. The distillation system was fed approximately 39% concentra- tion glycol during the 2009–2010 and 2010–2011 deicing seasons and 45% during the 2011–2012 season. Per experimentation in previous years, where 8% to 20% glycol concentrations were fed through the distillation system, 99%+ product concentrations could not be achieved, and a greater quantity of natural gas was consumed. On average, with the unit being fed 39% to 45% glycol concentrations, the unit performs at 100% of redesign expectations. Flow Rate In an effort to sustain good flow rates, the MVR units are run 24 hours per day, 7 days per week. They have the ability to be adjusted based on influent PG. Based on analysis of the data, there is a clear correlation showing that PG concentration affects the processing rate. Over the course of the three seasons of data, the MVR systems processed at an average rate of 185 gpm. It appears the units perform best when influent concentrations are between 6.5% and 19.5% gly- col. Individually, an MVR unit processed 2.3 gpm, which is 91.7% of the 2.5-gpm design speci- fication. During the 2010–2011 deicing season, the MVR average was 19.9 gpm for all machines or 2.48 gpm individually, which equates to 99.3% of design. Adjustments are made on the PLC, and the influent and effluent streams are measured on an hourly basis. Operating flow rates between 2.6 gpm and 2.8 gpm are achievable per MVR unit, with influent concentrations of between 12% and 15%, but other factors such as quality of feed and desired product output also affect processing rate. Treated Load Rate The rate at which soluble or total COD is removed from the system is not a key measure of performance for the MVR system. The MVR units were designed to separate glycol from water. Effluent Concentrations Effluent wastewater generated from recycling activities at Denver International Airport is not allowed to be discharged to surface waters. The effluent streams produced by both the MVR and distillation systems are comingled and transferred to the WADS. In general, all low-concentra- tion spent ADF (typically of less than 1%) captured from the dedicated collection system as well as the recycling facility wastewater is managed at WADS and then discharged to the Metro Waste- water Reclamation District’s Central Treatment Plant (CTP). The airport’s wastewater contribu- tion permit with CTP defines a maximum allowable BOD load per day and an allowable monthly average. This typically ranges between 9 and 12 tons per day of BOD loading, depending on the time of year. On average, the airport manages and discharges 100 to 150 million gallons of low- concentration spent ADF to CTP. The MVR units have the ability to produce up to 55% PG concentration. The units on aver- age have produced concentrations between 39.6% and 42.5% PG. Influent processing flow rates tend to decrease the higher the concentration being achieved while the amount of PG produced

Airport Deicer Treatment System Summaries D-65 per gallon in the effluent increases. In addition, the effluent concentration of the fluid being produced becomes the influent for the distillation system. The system operator indicated that the 40% concentration target rate is ideal to maximize the influent processing rate of the MVR so as to not jeopardize spent-ADF storage and is also considered an ideal concentration to maximize distillation processing rates. The distillation system can be adjusted to produce a desired glycol concentration product. The higher the glycol content produced, the greater the value of the product for resale. The average concentration of effluent product made over the course of 3 years was 99.13% PG. Although the system has the ability to produce up to 99.5% PG concentration, the quality and color of the product can be jeopardized when exposed to additional heat. As a result, the operator of the facility maintains a 99.1% concentration target with specific product quality requirements. Treatment Efficiencies Based on the data, over 91.4% (or 93.1% with adjustment accounting for meter error) of the glycol that was fed through the MVR systems was reclaimed. The remaining balance of glycol was discharged through the effluent distillate stream to the POTW. At a peak, the ratio of glycol produced from the MVR for reuse compared to the amount of glycol fed through the system reached over 95.2% during the 2009–2010 season. In an effort to increase the amount reclaimed from the MVR systems, adjustments can be made to the scrubber system on the MVR units to reduce the amount of glycol in the distillate stream. The issue is that influent processing rates have to be slowed down and each machine balanced individually for this to occur. Going forward, the recycling vendor has plans to install a reverse osmosis system after the MVR units to treat the distillate effluent and reclaim a larger percentage of the glycol. Based on the 3 years of data, over 97.4% of the glycol that was fed through the distillation sys- tem was reclaimed for sales. The remaining balance of glycol was discharged through the effluent distillate stream to the POTW. As reported by the recycling vendor, 99% is considered typical, but there were maintenance issues experienced with some of the heat exchangers, which caused uncharacteristically high discharges of glycol in the effluent water for discharge to the POTW. The recycling vendor plans to install a reverse osmosis system after the distillation unit to treat the distillate effluent and reclaim a larger percentage of the glycol. Cost Prior to 2004, the city paid an annual fixed fee for a subcontractor to operate the entire ADS system. During this time, only 8% and higher glycol concentrations were removed from the airport’s diversion system and recycled. The city incurred a significant expense to discharge all of the less than 8% concentration fluid to an off-site POTW. In addition, the city faced discharge restrictions and storage issues in managing all of the spent ADF that was not designated for recycling. In 2004, the MVR technology was introduced to recycle all spent ADF that was 1% or higher in propylene glycol concentration. The addition of this technology allowed the vendor to assume all costs to recycle the spent ADF. Denver is unique in that there is a large enough volume of ADF applied each year (approximately 1.4 million gallons) so that over 50% of that volume can be reclaimed from the system and processed to a final product to be sold in indus- trial markets. With this ability, the vendor can cover all expenses to recover and recycle the 1% and higher PG with the revenues generated by the sale of recycled glycol. The city does incur expenses if less than a predetermined amount of ADF is applied annually during a mild season in an effort to cover a portion of the subcontractor’s costs that cannot be recouped with limited glycol volumes or if a major component needs replacement on the city-owned equipment. In addition, the city covers expenses to manage and discharge all of the fluid that is less than 1%

D-66 Guidance for Treatment of Airport Stormwater Containing Deicers in glycol concentration. Overall, the recycling process saves a significant amount of money for the city each year. The city reported savings of $2 million during the 2009–2010 deicing season, $1 million of which was from disposal fees.3 Lessons Learned for Potential Implementation of the MVR and Distillation Technologies at Other Airports The following factors have proven critical to the effective and efficient performance in the DEN MVR and distillation systems: 1. Influent glycol concentrations. 2. Quality of influent improved by mechanical filtration methods prior to treatment by both systems. 3. Desired effluent concentration of product produced affects influent processing rate. 4. Daily preventative maintenance is integrated into operations in order to optimize equipment performance. 5. Maintaining process variables such as temperature, flow rate, and pressures at consistent set points improves production rates. The effluent concentrations in the distillate have spiked occasionally during the last three deicing seasons. The distillation system has experienced multiple heat exchanger tube failures due to the age of the components, the incompatibility between feed/effluent mixture, and the tube material composition. This caused the effluent (distillate) to entrap more and more residue from the decaying exchanger tubes, thus driving the overall COD levels higher, as the data indicate. At the same time, influent from the feed/steam heat exchangers will enter the distillate stream, bringing the 0.4% to 0.6% PG up to 1.0% to 1.2%. A combination of these two factors has caused the effluent being removed from distillation to be temporarily sent back through the MVR units in order to remove the remainder of glycol while the exchangers are replaced. Conclusions from operation of the MVR and distillation systems at DEN that can be used by other airports considering these technologies include: 1. MVR technologies are very viable and most applicable with airports that generate spent-ADF concentrations of 1% and higher. 2. MVR technologies are modular, which means they can be installed in a relatively small foot- print and can be adjusted to deal with varying influent concentrations. 3. MVR technologies are also scalable, and additional units can be added without the need to significantly expand infrastructure, as in the case of DEN, where three additional units have been added over the past 5 years to keep up with the airport’s growth. 4. It is also important to note that MVR systems are typically installed at airports where there is an outlet for the effluent water produced such as a POTW or other type of system to treat low levels of COD and glycol. Generally, MVR units are more economical the greater the volume of ADF sprayed at the airport and, more importantly, the more the glycol that can be captured at the airport for recy- cling. The greater the volume reclaimed, the larger the volume of product that can be sold to generate revenues to offset capital and operating expenses. If the volume is less than 200,000 to 300,000 gallons a year of influent that is greater than 1% glycol concentration, then another treatment technology may be more cost-effective than installation of an on-site MVR system. 3 Financial data provided from September 2011 edition of Airport Improvement Magazine, “Denver International Airport and Portland Jetport Stand Ready for New Glycol Regs.”

Airport Deicer Treatment System Summaries D-67 There are few airports that spray and recover enough ADF to justify installation of an on-site distillation system. Although this model has been extremely successful in Denver, many airports could not generate enough glycol to offset the capital and operating expenses. Instead, many airports that have MVR or other recycling systems typically transport partially recycled glycol to centralized distillation plants. Technology has now been developed where modular distilla- tion systems can be installed at smaller airports, and then that airport can serve as a centralized distillation outlet for other airports in the region. Documents and Information Review in Development of Airport Summary 1. All treatment data provided by operational logs provided by DEN Inland Technologies staff. 2. Financial data provided from September 2011 edition of Airport Improvement Magazine, “Denver International Airport and Portland Jetport Stand Ready for New Glycol Regs.”

D-68 Guidance for Treatment of Airport Stormwater Containing Deicers Airport Treatment Summary No. 8 Airport: Detroit Metropolitan International Airport—Detroit, MI (DTW) Treatment Technology: Industrial Recycling and POTW Discharge Years Operated: Early 1990s–2012 (Currently Operational) Deicer Management System Description In the early 1990s, DTW made efforts to reduce the PG load it sent to the local POTW. To reduce the load, DTW concentrated application of PG-based aircraft deicers onto several deicing pads. DTW contracted with a local private deicer management and recycling firm to manage all runoff with a concentration of greater than 2% PG from the deicing pads. In addition, DTW uses a GRV to collect high-concentration glycol from the gates from frost deicing activities or miscellaneous deicing activities beyond the pad areas. Since the late 1990s, DTW has been recycling runoff with PG concentrations of greater than 2%. Figure 1 shows a process flow diagram of the system. Deicer Treatment Technology Selection Considerations In the early 1990s, DTW realized that an improved deicer management program was necessary. The existing program of gate deicing, stormwater capture in ponds (several millions of gallons), and discharge to the local county-owned Downriver Wastewater Treatment Facility (WWTF) in Note: SADR = spent aircraft deicing runoff. Figure 1. DTW deicing-affected stormwater flow diagram.

Airport Deicer Treatment System Summaries D-69 Wyandotte, Michigan, experienced challenges. Gate deicing by tenants was affecting very large volumes of stormwater runoff, and occasional overflows of stormwater management ponds were occurring more and more often. Not only was DTW experiencing environmental challenges at the time, but the Wyandotte WWTF, a 50- to 60-million-gallon per day facility, was also experi- encing treatment issues due to court-mandated construction projects. As a result, DTW’s ability to discharge to the WWTF was at risk. To reduce this risk (i.e., the amount of high-concentration runoff being discharged to the WWTF), DTW embarked on a program to change how and where aircraft deicing was being performed. DTW started working with tenants and developed a pad deicing operation. The pad deicing operation provided a much smaller area of pavement to be affected by aircraft deicer, resulting in a much higher concentration of stormwater runoff. This high-concentration runoff was a valuable commodity for DTW that could be handled by alternative means. DTW contracted with a local deicer management and recycling firm to manage all runoff with a concentration of greater than 2% PG from the deicing pads. DTW investigated other treatment processes for deicing-affected stormwater with concen- trations of less than 2% PG, including on-site and off-site methods. DTW determined that a new force main connecting to the city of Detroit’s WWTF would provide the needed improved compliance margin of safety. DTW constructed a new 5-mile-long force main that connected to the city of Detroit sanitary sewer system at a cost of approximately $11 million. This discharge location could accept up to 30,000 pounds per day with a flow limitation of 1 mgd. It too, how- ever, is susceptible to precipitation events, and limits to DTW discharges can be imposed by the Detroit WWTF if inflows to the WWTF exceed 900 mgd. Deicer Treatment Technology Description Industrial Recycling See the private recycling facilities fact sheet (Fact Sheet 110). Description of Support Systems The industrial recycling treatment system does not require the airport to maintain or operate support systems for treatment. Key Treatment System Sizing Parameters See Table 1 for sizing parameters. Treatment System Performance No normal basis-of-design treatment criteria were identified in the review of the project. Design data for the treatment system are not a parameter of concern for the airport since the deicing-affected stormwater is sent to a POTW or to an off-site industrial recycling facility. Component/Parameter Size/Capacity of Treatment Units Number of Treatment Units Total Capacity Stormwater storage capacity 100 million gallons 106 million gallons 216 million gallons 70 million gallons 0.7 million gallons 0.75 million gallons 2 493 million gallons Table 1. Key system sizing parameters.

D-70 Guidance for Treatment of Airport Stormwater Containing Deicers No treatment performance data were routinely collected or readily available for review by the research team since the POTW or off-site recycler manages all treatment. Cost Assessment for the DTW POTW Discharges and Recycling The rates paid for discharge of deicer-affected stormwater to the POTW are shown in Table 2. Cost terms of the contract for managing and recycling fluid from the deicing pads include: • If greater than 600,000 gallons of fluid of greater than 2% PG are collected, DTW receives a $0.10/gallon rebate. • If greater than 450,000 gallons and less than 600,000 gallons (2% PG) are collected, DTW breaks even and pays no fee and receives no rebate. • If less than 450,000 gallons of 2% PG are collected, DTW pays $1.50 per gallon to the recycling firm. Conclusions on Performance of DTW Deicer Treatment System Influent Deicer Concentrations At DTW, influent deicer concentrations are used to determine which treatment system receives deicer-affected stormwater. Concentrations of greater than 2% PG are sent to the industrial recy- cling treatment facility, and concentrations of less than 2% are sent to the POTW. Flow Rate The flow rate of the treatment systems is not a concern for DTW. However, the rates at which flows can be sent to the POTW are limited to the hydraulic limits of the pipe and delivery. The rate at which the industrial recycler removes deicing-affected stormwater is limited. At DTW, the industrial recycler removes deicing-affected stormwater by truck. Treated Load Rate The DTW discharges to POTWs capable of accepting up to 30,000-lbs BOD/day and 12,000~15,000 lbs BOD/day. Treatment load rate in excess of this limit must be stored on-site or will exact a fine from the POTW. Effluent Concentrations Effluent concentrations are not a concern for the airport since all deicing-affected stormwater is managed by a separate management facility. Treatment Efficiencies Treatment efficiencies are not a concern for the airport since all deicing-affected stormwater is managed by a separate management facility. Cost The deicer management firm contracted by the airport supplies all personnel, vehicles, frac- tanks, pumps, transportation, and recycling services as part of the contract. The runoff is taken BOD (lbs/day) Flow (million ft3) Downriver WWTF $0.271 $18.44 Detroit WWTF $0.269 $10.69 Table 2. Costs for discharges to the POTW.

Airport Deicer Treatment System Summaries D-71 to the firm’s facility a few miles away from the airport for further processing and purification. The industrial recycling operation has been very economical for DTW since the contract provides for a payback to DTW depending on the volume of fluid collected. Lessons Learned for Potential Implementation of the Industrial Recycling and POTW Technologies at Other Airports Only the rate of sending flow from DTW to the industrial recycling system and the POTW are factors critical to the effective and efficient performance of deicing treatment. Other factors criti- cal to the treatment of the deicing-affected stormwater are managed by the industrial recycling firm and POTW operator. Although the industrial recycling worked well for several years for DTW, the winter of 2011–2012 had two significant changes at DTW that affected the treatment system. First, the largest airline ten- ant invested in a new ADF bulk blend-to-temperature system and six new deicer application trucks equipped with real-time blending to temperature capabilities. Second, a very mild winter greatly reduced the amount of deicer applied. These events caused a significant shift in the volumes of water being recycled versus the vol- ume of water being discharged to the WWTF. Significantly more water was being discharged to the WWTF, while less was being recycled. It also changed the economic model for the airport in that for the first time in several years, the airport was required to pay the industrial recycling firm for recycling services, rather than breaking even or receiving a rebate. Through 2012, DTW has invested over $125 million in stormwater separation, storage, and support infrastructure and continues to investigate the impacts of recent changes on their com- pliance status and costs. DTW is considering its various contracts and may be investigating additional alternatives for disposal as future seasons reveal new challenges. Conclusions from operation of the industrial recycling and use of the POTW at DTW that can be used by other airports considering these technologies include: 1. Industrial technologies are most applicable with airports that generate spent-ADF concentra- tions of 2% and higher. 2. A significant amount of spent PG is required to make industrial recycling profitable. 3. Changes to deicing operations can affect the economies associated with industrial recycling and POTW discharges. 4. POTWs have permit limits that may affect the flow, concentration, or load sent from an air- port. Changes to these permit limits may require an airport to add a treatment technology or change operations.

D-72 Guidance for Treatment of Airport Stormwater Containing Deicers Airport Treatment Summary No. 9 Airport: Wilmington Air Park—Wilmington, OH (ILN) Treatment Technology: Reciprocating Aerated Gravel Bed Years Operated: 2000–2009 Deicer Management System Description From 1980 through 2009, ILN was operated as a hub for the cargo shipping and distribution businesses of Airborne Express and DHL, with a peak of 250 airport operations a day in 2007. The 2,200-acre ILN airport is a former military base with two 10,000-ft runways and over 200 acres of concrete ramps for parking, loading, and deicing of aircraft. During its peak years of operation, 200,000 to 300,000 gallons of pure PG-based aircraft deicer were applied each year. Three of the airport’s 15 outfalls drain the main deicing areas and discharge stormwater to the airport’s two receiving streams, Lytle Creek and Indian Run. These outfalls also discharge the majority of stormwater flow from the site. Outfall flow rates vary widely, from less than 50 gpm in dry weather conditions to over 150,000 gpm in wet weather. Figure 1 shows a process flow diagram of the system, and Figure 2 shows the system itself. Figure 3 shows the bacterial film used in the system. In 1996, effluent limitations in a new NPDES stormwater discharge permit for the facility trig- gered the implementation of a deicer management system for collection, conveyance, storage, treatment, and discharge of deicer-affected stormwater from hundreds of acres of aircraft park- ing areas. Because of the vast deicing and collection area, the ILN deicer management system was designed as two separate systems, located on the north and south sides of the airport. Each system included collection of deicer-affected stormwater from existing airport outfalls, temporary storage in lined open basins, treatment of stormwater using the reciprocating aerated gravel bed technol- ogy, and discharge of treated effluent to the surface waters. DHL purchased the facility in 2003 and expanded airport operations. Additional conveyance and storage were added to the deicer manage- ment system in 2004 to account for the corresponding increased deicing operations. When DHL Figure 1. ILN deicer management system.

Airport Deicer Treatment System Summaries D-73 eliminated most of its domestic cargo operations and moved its international operations to the Cincinnati/Northern Kentucky Airport in 2009, the deicer management system was mothballed. Currently, there is an aircraft maintenance operation at ILN. Should aircraft deicing activities grow in the future, the deicer management system can be restarted. Deicer Treatment Technology Selection Considerations In 1995, ABX Air, owner of the airport at the time, began preliminary evaluation of alterna- tive deicer management measures to achieve compliance with conditions in its NPDES permit. The initial work included evaluation of stormwater characteristics, evaluation of existing deicing practices, assessment of deicer collection methods, and evaluation of deicer disposal alternatives. A review of deicer application options resulted in ABX Air deciding that deicing on deicing pads was not a feasible operations scenario. As a result, runoff from over 800 acres of airfield surface Figure 2. ILN reciprocating gravel bed system. Figure 3. Example of bacterial film on reciprocating bed gravel.

D-74 Guidance for Treatment of Airport Stormwater Containing Deicers had to be collected and managed. Being a cargo-only facility, each aircraft was deiced at its own unique location, once per day. This deicer application system led to vast quantities of highly vari- able deicer-affected stormwater runoff with wildly varying COD concentrations. A series of preliminary assessments were conducted into means for collecting, storing, treat- ing, and disposing of the runoff. The sheer quantity of stormwater, in addition to the huge COD load, made discharge to the POTW in the small town of Wilmington infeasible. Once it was determined that on-site treatment was necessary, several methods for treatment were considered. Recycling technologies were evaluated but were deemed not cost-effective because the volumes of water resulted in average concentrations too low for economical recycling. Several biological treatment options were considered, including AFBRs, trickling filters, and activated sludge. The volume of water to treat and the size range of COD concentrations led to size and cost challenges with the AFBR technology. More conventional biological technologies like activated sludge were considered to be overly affected by weather. After preliminary evaluation of conventional bio- logical treatment processes, ABX pursued development of a subsurface, attached-growth gravel bed treatment system through implementation of a pilot study. From 1997 to 1999, two on-site biological treatment technologies were pilot tested: a traditional subsurface gravel bed wetland technology and a reciprocating gravel bed technology. The pilot-study results indicated that the constructed gravel bed wetland did not sufficiently degrade the deicer. The failure of the constructed wetland to treat sufficiently was associated with the lack of an aerobic environment to develop bacteria suitable for large treatment loads and with cold temperatures. It was also discovered that the addition of wetland plants provided no treatment value to either the con- structed wetland or the reciprocating gravel bed system. The reciprocating subsurface gravel bed technology without plants was successfully pilot tested. The results of the pilot study were used to design and size two full-scale treatment systems at ILN that were based on the reciprocating technology. Deicer Treatment Technology Description Reciprocating Gravel Bed, a variant of Aerated Gravel Beds See aerated gravel bed treatment technology fact sheet (Fact Sheet 102). The reciprocating gravel bed technology has been applied for treatment of various types of stormwater and waste- water applications, but ILN is the only deicer treatment application. The ILN gravel beds are constructed as a series of cell pairs in which water is pumped between the two gravel beds in the cell pair, alternately exposing one gravel bed to contaminated water in the cell for 30 to 60 minutes at a time and its partner cell to atmospheric oxygen for the same period. After the 30- to 60-minute period, water is pumped from the full gravel bed to the empty gravel bed, resulting in a reversal of the exposure to water and atmospheric oxygen. The benefit of this type of a system is the lack of a need for blowers and aeration piping buried in the gravel bed. Description of Support Systems The two reciprocating subsurface gravel bed systems at ILN each ultimately included over 14 million gallons in up-front storage, means for adding ammonia (for nitrogen) and phosphate to individual basin pairs at the pump stations, a Parshall flume for measuring effluent flow, a recir- culation line back to storage for situations when effluent COD concentrations were too high, and a cascade aeration system to increase the dissolved oxygen concentrations in the treated effluent prior to discharge. The systems are controlled by SCADA systems, with a PLC that receives level and pump status data and turns the reciprocation pumps on and off to automatically create the system cycling. The SCADA system was also used to help segregate influent flow with COD concentrations that could be directly discharged to the streams from diluted runoff that needed to be treated and concentrated runoff that needed to be treated. The system did include a final gravel bed cell that did not reciprocate as potential means for capturing biosolids from the reciprocating cells. Discharge of biosolids from the reciprocating cells (as represented by TSS) was kept low during winter by carefully

Airport Deicer Treatment System Summaries D-75 managing the growth of the bacterial films on the gravel and through summertime degradation of the dead biomass within the cells. Key Treatment System Sizing Parameters See Table 1 for sizing parameters. Treatment System Performance The data in Table 2 on the intended design performance of the system were derived from the 1998 Engineering Report associated with the Permit-to-Install Application to the Ohio EPA. The information in Table 3 on actual system performance was derived from daily data col- lected at the facility between 2000 and 2009. Component/Parameter Size/Capacity of Treatment Units Number of Treatment Units Total Capacity Stormwater storage capacity 15.8 million gallons 14.4 million gallons 2 30.2 million gallons Treatment unit volume 4.8 million gallons 3.6 million gallons 2 8.4 million gallons Treatment unit dimensions 6 acre x 7-ft D 3 acre x 7-ft D 2 1,180,000 ft3 887,000 ft3 Treatment facility footprint 6 acres 3 acres 2 9 acres Note: Not including conventional flow-through gravel bed. Table 1. Key system sizing parameters. Parameter Value Unit Design flow rates - Average - Maximum 250 1,000 Gallons per minute Design treatment load capacity 25,000 14,700* lbs COD/day lbs BOD5/day Design influent concentration - Range 0~3,000 0~1,700* mg COD/L mg BOD5/L Design effluent concentration (average) 270~690 160~410 mg COD/L mg BOD5/L Design treatment efficiency (average) 98% % influent COD load treated Loading is based on an average temperature of 41°F *Data based on conversion: [COD] = 1.7 [BOD5]. Table 2. Design basis for system performance. Parameter Value Lytle Creek System Value Indian Run System Unit Flow rates - Average - Maximum 439 1,967 413 2,680 Gallons per minute Actual COD treatment load rate - Average - Maximum 13,400 235,000 17,100 584,000 lbs/day Influent COD concentration - Average - Maximum 2,160 22,000 2,080 26,540 mg/L Effluent COD concentration - Average - Maximum 342 7,410 323 10,600 mg/L Treatment efficiency 87.7 85.5 % influent COD load treated Table 3. Actual system performance.

D-76 Guidance for Treatment of Airport Stormwater Containing Deicers Cost Assessment for the ILN Reciprocating Gravel Bed System See Table 4 for treatment system costs. Conclusions on Performance of ILN Reciprocating Gravel Bed System Influent Deicer Concentrations Influent deicer concentrations, although monitored regularly, are not a primary control parameter used in managing the loading to the system. If the overall mass loading is kept within the load capacity of the system, it can treat a wide variety of COD concentrations within the capabilities of the system pumps. Flow Rates The system was controlled initially to a set flow rate for concentrated deicer and diluted deicer conditions. Later the methodology was changed to a mass-load–based control system such that influent flows were turned on and off to meet the system loading targets. Based on the growth of airport operations (and deicer applied) and initial underestimation of the effects of prolonged cold weather, the design flow rates (treatment capacity) were insufficient for treatment needs, and additional storage had to be added. Treatment Load Rate The nominal design system COD treatment capacity was based on a water temperature of 41°F. It was known from standard biological treatment theory and pilot-study results that treatment rates would increase with higher water temperatures and decrease with lower temperatures. When perfor- mance data from 2001 through 2009 are averaged for the water temperature range of 41°F +/- 2°F, the COD treatment rates measured in the field compared well to the design treatment rates. At temperatures lower than 38°F, treatment rates were lower than anticipated in comparison to the results from the pilot studies. The most significant factor was not short-term decreases in treatment rates from cold temperatures, but the effects from prolonged air temperatures below 20°F (and the corresponding decreases in water temperature). It was also determined that at temperatures greater than 60°F, treatment rates also decreased to some degree because biofilm growth on the gravel began to clog the gravel beds, resulting in shorter hydraulic detention times. Effluent Concentrations When operating at influent COD loading levels suitable for the temperature, effluent COD concentrations of less than 200 mg/L could be obtained for water temperatures greater than 40°F. Cost Category Projected at Initial Implementation Actual Capital cost* $6M in 2000 $6.2M at construction, $0.6M added in 2004 for upgrades to control system and online monitoring Annual operating cost** - Utilities - Chemicals - Analysis - Material handling Total operating cost $40,000 $20,000 $5,000 0 $65,000 $44,000 $24,000 $4,000 0 $72,000 *Capital costs are for the two treatment systems combined, excluding collection, storage, and discharge support systems. **Operating costs are for Lytle Creek and Indian Run systems combined. They do not include labor costs, but equate to approximately two full-time operators. Maintenance costs, which vary, are not included. Major maintenance items included liner repair, weeding of gravel beds once per year, and pump preventative maintenance and repair. Table 4. Costs for the treatment system.

Airport Deicer Treatment System Summaries D-77 BOD concentrations of less than the detection limits of the BOD test could also be obtained. COD effluent concentrations rose when water temperatures were less than 40°F, especially when the system was overloaded. Treatment Efficiencies Over the course of most seasons, and over the 8 post–start-up years, the treatment efficiency (% of COD removed) was near 98%, very close to the design removal rates when tempera- tures were above 38°F. There were numerous instances where removal efficiency decreased for periods up to several weeks during the coldest portions of the winter, dropping average COD removal rates to near 85%. Most instances where treatment efficiency suffered were associated with the system being loaded beyond its treatment capacity for a given temperature. Through the course of operations, it was learned to decrease COD loadings in anticipation of extended stretches of cold weather. This allowed the system to recover more quickly when temperatures warmed. Cost The initial total cost of the deicer management system (collection, storage, treatment, dis- charge) was $11.5 million in 2000. $6.2 million of that total was associated with treatment. In 2004, $4 million in improvements were added, most of which was associated with additional storage. Direct additional costs for the treatment system were $600,000 for improvements to the treatment system controls. The increase in storage capacity was driven primarily by the decreased treatment rates at cold temperatures, which led to the need to hold more water during the middle portion of the deicing season. The operating costs were near design basis projections. The system required two full-time operators, as initially determined. Electrical costs were maintained at planned levels (~$40,000 per year), and use of the reciprocation method likely resulted in significantly lower electrical costs than would have been the case if blowers were used to supply air. Chemical use costs for nutrients were somewhat higher than projected. Lessons Learned for Potential Implementation of the Reciprocating Aerated Gravel Bed Technology at Other Airports Factors that have proved to be critical to the performance of the reciprocating aerated gravel bed treatment technology at ILN are: 1. Ability to control COD loading rate to the system, 2. Adjustment of the COD loading rate based on water temperature, 3. Location of nutrient supply to treatment system, and 4. Ability to reduce shorter reciprocation cycle times, which has proven to improve performance. The ILN treatment system performed as intended when the treatment system COD mass loading rates were within the system capacity. However, changes in operation at ILN, includ- ing the significant drop in operations and corresponding deicing operations, demonstrated the interdependency of treatment technology, deicing operations, deicing-affected stormwater collection, and storage. Conclusions from operation of the reciprocating aerated gravel bed system at ILN that can be used by other airports considering this technology include: 1. The reciprocating aerated gravel bed technology was the first use of its kind for a deicer man- agement system. Several years of experience reveal that that system can successfully treat large COD loads and volumes of runoff and produce, but loading of the COD and nutrients must be controlled such that a healthy biological population can be maintained.

D-78 Guidance for Treatment of Airport Stormwater Containing Deicers 2. At temperatures of less than 38°F, treatment rates were less than anticipated from the pilot studies. As the system operation evolved, it was learned that decreasing COD loading rates in anticipation of cold temperatures helped the system treatment rates rebound faster when temperatures warmed up. 3. Swings in air temperature had a somewhat larger negative effect on treatment rates than swings in water temperature, presumably because of the cold air drawn into the system in the reciprocation process. The effect was most pronounced when air temperatures were less than 20°F for prolonged periods. 4. A decrease in performance was also seen in the late spring in years when there was heavy deicer use because of the proliferation of bacterial colonies that thrived under warm temperatures and clogged air spaces between gravel cells, limiting the system hydraulic detention time. This biological growth can be managed by controlling COD loading rates at higher temperatures. 5. Effective tools for reducing the swings in treatment rates were installation of an online BOD monitor and implementation of a variable frequency drive on influent pumps in 2004, which allowed much better control and consistency of influent COD loading rates. When this system was implemented, the SCADA control system was programmed to maintain a constant COD loading rate with appropriate decreases in the COD loading rate with lower water temperatures. 6. The rates at which nutrients were added to the system had a significant effect on treatment rates, especially at the start of each season. Over time, the loading rates were adjusted to reduce the likelihood that nutrients were a limiting factor in performance. 7. Although it was never specifically quantified, it is likely that under certain conditions, the treatment was oxygen limited. 8. Over time, the gravel beds in the initial part of the treatment system became partially clogged with both inorganic sediment and biological solids. A challenge with this type of system is preventing the clogging from occurring by using pretreatment to remove inorganic solids and management of deicer loads to keep biological growth from being excessive. Documents and Information Review in Development of Airport Summary 1. Permit-to-Install Applications to Ohio EPA for Wilmington Air Park Deicer Management System, 1998 and 2004. 2. Daily system operating logs.

Airport Deicer Treatment System Summaries D-79 Airport Treatment Summary No. 10 Airport: London Heathrow International Airport—London, United Kingdom (LHR) Treatment Technology: Aerated Lagoon Aerated Gravel Beds Passive Facultative Treatment Years Operated: 2001 to 2012 (Currently Operational) Deicer Management System Description LHR is divided into four main catchments, each served by a separate balancing reservoir, designated as northwestern, southwestern, eastern, and southern. Deicing operations drain to either the eastern or southern catchment. The system discharges to surface waters. The Brit- ish Airports Authority (BAA) commissioned a reed bed treatment facility in 2001 at Mayfield Farm to treat deicing runoff from the southern catchment. Due to expansion of airfield opera- tions, the existing facility was upgraded in 2010 to provide a significant increase in treatment capacity. The 2010 treatment system at Mayfield Farm includes three major unit processes downstream of the main reservoir: the upgraded floating reed bed (a passive facultative technology), the balancing lagoon (an aerated lagoon), and the aerated gravel beds. During winter operations, as BOD meters detect elevated concentrations of BOD associated with deicing, stormwater flow is diverted to the main reservoir and is stored there before being pumped to the treatment system. Under normal operations, flow is pumped from the main reservoir so as to flow in series through each unit process until discharge. A schematic of the system is provided in Figure 1. Figure 2 and Figure 3 show photographs of the system. Deicer Treatment Technology Selection Considerations The reed bed treatment facility commissioned in 2001 at Mayfield Farm had a treatment capacity of 770 lbs of BOD5. BAA decided to upgrade the treatment capacity in 2010. The upgrade included the reconfiguration of existing unit processes and installation of new aeration equipment and nutrient feed system. Runoff From Southern Catchment – Process Water Main Reservoir KEY – Process Diversion Chamber Upgraded FRB Balancing Ponds Subsurface Reed Southern Balancing Reservoir Beds Effluent Figure 1. Mayfield Farm stormwater management system process flow diagram.

D-80 Guidance for Treatment of Airport Stormwater Containing Deicers The existing floating reed bed channels were transformed into aerated channels. The chan- nels were designed using aerated lagoon practices. The first part of each channel was designed as a complete mix lagoon. The remainder of the channel was designed as a partial mix lagoon. Floating reed bed racks were retained in the partial mix zones to improve the sedimentation of the bacterial solids generated in the complete mix zone. The balancing lagoon was added to provide process flexibility for the treatment train. The lagoon can be employed for either hydraulic equalization or as a middle process in the treatment train. It was designed as a partial mix aerated lagoon. The final reed bed was upgraded to a planted aerated gravel bed (also known as an intensified or aerated wetland). The upgrade included addition of aeration tubing and the reconfiguration of the flow path in the beds from horizontal (left to right) to vertical (top to bottom). The following considerations were factors in the selection of the upgrade design: 1. Ability to use existing infrastructure at Mayfield Farm. 2. Ability to quickly design and construct the system. Figure 2. Retrofit of aerated gravel beds (reed beds). Figure 3. Aeration in main reservoir.

Airport Deicer Treatment System Summaries D-81 3. Results from an on-site pilot test that demonstrated the capacity of the system. 4. Need to comply with green-zone requirements for the project location. Deicer Treatment Technology Description The original treatment technology employed at LHR was a reed bed system classified for the purposes of this guidebook as a passive facultative treatment system (See Fact Sheet 108). The upgrade of the floating reed bed and balancing lagoon used aerated lagoon technology (See Fact Sheet 103). The aerated gravel bed technology is described in Fact Sheet 102. Aeration is also provided in the main reservoir. Description of Support Systems The upgrade of the LHR treatment system included the addition of aeration equipment, a nutrient feed system, and related electrical and instrumentation work. The main storage reser- voir is equipped with floating aerators, which are used at the discretion of the operator. A nutrient feed system has also been included into the re-engineering of the system. The nutrients are added at various points in the process to support bacterial growth. By adding sup- plemental nitrogen, phosphorus, and other micronutrients at the influent, the aerobic bacteria can properly grow and degrade the hydrocarbons in the carbon-rich stormwater from deicing operations. The nutrient solution is prepared off-site and delivered to a chemical storage tank at Mayfield Farm. The feed system consists of a storage tank and four feed pumps. Each feed pump supplies nutrient solution to a dedicated dosing point. The primary feed points are to the main reservoir and the influent of the upgraded floating reed beds. Instrumentation for the system includes the collection and transfer of signals from blower panels, pumps, and online analytical equipment (online BOD meter, dissolved oxygen probe, phosphorus meter, and flow meter) to the existing SCADA system operated by BAA. The SCADA system is used to control the operation of motorized equipment (blowers, pumps, and valve actuators). Key System Sizing Parameters See Table 1 for sizing parameters. Treatment System Performance The data in Table 2 represent the intended design performance of the treatment system. Component/Parameter Size/Capacity of Treatment Units Number of Treatment Units Total Capacity Stormwater storage capacity 7.92 million gallons 1 7.92 million gallons Treatment unit volume Complete mix/partial mix channels Balancing lagoon Aerated gravel beds 4.0 million gallons 5.2 million gallons 1.8 million gallons 3 11 million gallons Treatment unit dimensions 529,000 ft³ 706,000 ft³ 247,000 ft³ 3 1,482,000 ft³ Treatment facility footprint Complete mix/partial mix channels Balancing lagoon Aerated gravel beds 2.5 acres 2.0 acres 5.1 acres 9.6 acres Table 1. Key system sizing parameters.

D-82 Guidance for Treatment of Airport Stormwater Containing Deicers The facility has numerous online BOD meters that are used to monitor real-time values of BOD within the system. These data are logged, along with related flow rates. The operation of the upgraded system began in February 2011. No monitoring data are avail- able at this time. Cost Assessment for the LHR Aerated Gravel Bed Treatment System Table 3 shows treatment system costs. Operational effort and cost consist primarily of management of pump, aeration, and nutrient feed systems. Biomass levels are monitored and managed as needed. Conclusions on Performance of LHR Aerated Gravel Bed System No performance data are currently available on which to make conclusions. Factors that are expected to be critical in the performance of the LHR aerated gravel bed system include: 1. The ability to provide adequate storage upstream of treatment, 2. Adequate nutrient dosing concentrations and dosing locations, and 3. The ability to properly operate aeration equipment. Lessons Learned for Potential Implementation of the LHR Treatment Technologies at Other Airports 1. The treatment system site should be located far enough from runways that bird strikes are reduced. Parameter Value Unit Design flow rates - Average - Maximum 634 1,270 Gallons per minute Design treatment load capacity 13,000* 7,700 7,700** lbs COD/day lbs BOD5/day lbs PG/day Design influent concentration 1,000 mg BOD5/L Design effluent concentration (average) 30 mg BOD5/L Design treatment efficiency (average) 98% % influent COD load treated *Data based on conversion: [COD] = 1.7 [BOD5]. **Data based on conversion: [PG] = [BOD5]. Table 2. Design basis for system performance. Cost Category Projected at Initial Implementation Actual Capital Cost 2001* 2011 $30M** in 2001 $4.5M*** in 2011 $27M** in 2001 Not available Annual Operating Cost*** - Utilities - Chemicals - Analysis - Material handling Total operating cost Not provided Not provided Not provided Not provided $250,000 Not provided Not provided Not provided Not provided Not provided *“Mayfield Farm Constructed Wetlands,” Constructing Excellence (2006). **Data based on conversion: $1.40 = £1.00 ***Data based on conversion: $1.60 = £1.00 Table 3. Costs for the treatment system.

Airport Deicer Treatment System Summaries D-83 2. The nutrient dosing points around the system should be flexible and accessible. 3. The ability to take tanker truck deliveries to the treatment site was a major design challenge. 4. Nutrient solution is prepared off-site and shipped to a storage unit on-site since the ability to take tanker trucks to the treatment site was a major design challenge. 5. Preliminary testing demonstrated that aeration and nutrient addition greatly improved per- formance in comparison to unaerated beds without nutrient addition. 6. Aeration lines were plowed into the existing beds, which greatly lowered the cost of the project. Document and Information Review in Development of Airport Summary 1. Naturally Wallace project fact sheet (http://naturallywallace.com/docs/NWC%20Mayfield% 20Farm%204-1-11%20F2.pdf). 2. 2011 WETPOL presentation: Glycol Treatment at London’s Heathrow Airport. BAA, ARM, and Naturally Wallace. 3. Naturally Wallace project files.

D-84 Guidance for Treatment of Airport Stormwater Containing Deicers Airport Treatment Summary No. 11 Airport: Oslo Airport, Gardermoen—Oslo, Norway (OSL) Treatment Technology: Moving Bed Biofilm Reactor Years Operated: 1998–2012 (Currently Operational) Deicer Management System Description The new OSL opened in 1998. Currently, OSL services 22 million passengers per year and has anywhere from 6,000 to 12,000 annual deicing operations. The deicer collection system was constructed as part of the new airport in 1998. Components of the ADS include three dedicated deicing pads and an ADF-contaminated stormwater collection/retention system, including four retention basins with a total storage capacity of approximately 17 million gallons. OSL has financed a moving bed bioreactor pretreatment unit based on the MBBR method used at the nearby municipal Gardermoen sewage treatment plant. Environmental regulations limit the acceptable concentration of COD in the groundwater to 15-mg COD/L on airport property and 0.5-mg COD/L off airport property during spring/snow melting periods. The regulatory limit for deicing fluid concentration in surface waters outside airport property is 0.5-mg/L glycol, formate, or acetate. All aircraft deicing at OSL is conducted on one of three remote deicing pads located next to runway entry points. Runoff from the pads, as well as from key taxiway/runway areas where deicing fluids generally drip from the aircraft, is collected in the ADF deicer management system. Figure 1 shows a process flow diagram of the system. Runoff contaminated with runway deicer from approximately 124 acres of apron area is col- lected in a separate system to keep it separated from stormwater contaminated with glycol. This system includes two retention basins with 20 million gallons of capacity. However, due to ongo- ing airport extension, apron areas have been doubled, and three new retention basins have been constructed; thus, a total storage capacity of nearly 40 million gallons will be available from the 2012–2013 deicing season on. Runoff from taxiways and runways is generally not collected but percolates into the soil along- side these areas. OSL monitors concentrations of COD in the groundwater to confirm natural attenuation of the deicing chemicals from these areas. Figure 1. OSL deicing-affected stormwater management system process flow diagram.

Airport Deicer Treatment System Summaries D-85 Stormwater runoff containing ADF collected from the deicing pads and portions of the run- ways is separated into three separate storage tanks at each deicing pad based on deicer usage and weather conditions. These tanks contain collected runoff at three different concentration levels (Table 1). High-concentration runoff is pumped to a treatment plant where a concentrate suit- able for trucking is produced. This concentrate is then transported to a chemical industrial facil- ity in Germany for distillation into a pure glycol product. The medium- and low-concentration runoff is pumped to the storage basins prior to usage or treatment at the Gardermoen sewage treatment plant (Table 1). Runoff containing runway deicer from the apron area is collected in two 9-million-gallon basins. Following collection, the contaminated stormwater is sent to Gardermoen sewage treat- ment plant for treatment in the ordinary MBBR train, mixed with sanitary wastewater. (COD varies from 100-mg to 400-mg COD/L.) The biomass in the MBBR has to be adapted to the runway deicer through a start-up procedure; thus initial capacity every new delivery period is low and then increases through buildup of specialized bacterial population. Deicer Treatment Technology Selection Considerations OSL is located on the largest unconfined aquifer in Norway. Consequently, strict regula- tions were introduced by the Norwegian pollution control authorities to minimize the envi- ronmental impact on the groundwater system. To mitigate impacts from deicing activities at the airport on the aquifer, the Norwegian pollution control authorities required OSL to not affect: 1. Groundwater balance, 2. Groundwater quality, 3. Natural erosion processes in the ravine system, or 4. Surrounding water resources. During the planning phase of OSL, it was determined that the local wastewater treatment plant did not have the capacity to treat the combined wastewaters from the local municipalities of Ullensaker and Nannestad in addition to that from OSL. Therefore, in 1994, the Norwegian pollution control authorities concluded that a new semi-regional wastewater treatment plant should be built. During the planning process of the new wastewater treatment plant, testing demonstrated that the glycol in the runoff could be used as an external carbon source for bio- logical denitrification as a substitute for the commonly used ethanol or methanol. Combining treatment of the wastewater and the deicer-contaminated stormwater into one treatment facility became a primary design focus. The Gardermoen treatment plant was then constructed to treat wastewater from the surrounding municipalities of Ullensaker and Nannestad as well as sanitary wastewater from OSL. Deicer-contaminated stormwater from OSL is treated in the winter and spring seasons. Concentration Collection and Treatment More than 2% glycol Recycled/reused in the glycol industry. Previously trucked to two different sewage treatment plants where it was used as a carbon source for the biological denitrification process (2004–2011). Between 0.2% and 2% glycol Sent to the Gardermoen sewage treatment plant as a carbon source for the denitrification process. Less than 0.2% glycol Sent to the Gardermoen sewage treatment plant for treatment as sewage. Source: Per Espen Jahren, Water Management Systems. Oslo Airport, Norway. Table 1. Treatment of aircraft deicing-affected stormwater at OSL.

D-86 Guidance for Treatment of Airport Stormwater Containing Deicers Deicer Treatment Technology Description The OSL treatment system is unique with respect to process concept. Stormwater is segregated by COD concentrations, and the medium-concentration fraction and the low-concentration fraction are treated in distinctly different parts of the biological treatment plant. The treatment reduces the high-COD of ADF deicer runoff as well as runway deicer runoff to a very low level. ADF-contaminated runoff containing less than 0.2% glycol is mixed with wastewater and sent through the pretreatment reactor prior to treatment in the ordinary MBBR train. The treatment plant includes an anoxic reactor for nitrogen removal, sludge removal (dissolved air floatation), and UV disinfection (summer only). ADF-contaminated runoff with glycol concentrations of between 0.2% and 2% is injected as a carbon source for the denitrification processes in the anoxic reactor of the MBBR train. A more detailed description of the MBBR technology can be found in Fact Sheet 107. Figure 2 shows a process diagram for the MBBR. The OSL deicer treatment system is unique compared to other airports because the biologi- cal treatment in the sanitary wastewater treatment plant is capable of treating large volumes of deicer-contaminated runoff from the airport. The mixing of stormwater and sanitary waste- waters is likely to have beneficial effects compared to treating stormwater alone because of the heat and nutrients supplied to the biological process from the sanitary wastewater, as well as creation of a more stable organic and hydraulic load and a more stable sludge volume and quality suitable for sludge dewatering and handling. Key Treatment System Sizing Parameters See Table 2 for system sizing parameters. Sewage Treatment Plant System Performance See Table 3 for designed system performance information. The ADF applications at OSL are integrated into the deicing management system. The ADF applied at OSL is a Type I proportional mix and Type II ADF, mono propylene glycol only. The aver- age ADF consumption is 290-lbs to 310-lbs glycol/aircraft (calculated as pure/100% glycol only). Figure 2. Gardermoen treatment plant, moving bed biofilm reactor process diagram.

Airport Deicer Treatment System Summaries D-87 The information in Table 4 on actual system performance was derived from facility data from the year 2000. Conclusions on Performance of OSL MBBR System Influent Deicer Concentrations The collected deicer-contaminated stormwater sent to the fixed film portion of the MBBR contains less than 2,000-mg PG/L, or approximately 3,400-mg COD/L. Higher-concentration stormwater runoff is used as a carbon source for the denitrification process in the treatment Component/Parameter Size/Capacity Number of Units Comments ADF runoff storage capacity, for recycling 3,500 m3 0.9 million gallons One tank at each deicing pad and two connected buffer tanks >2% glycol ADF runoff storage capacity, carbon source for denitrification 19,000 m3 5.0 million gallons One tank at each pad and two buffer basins <2%, >0.2 % glycol ADF runoff storage capacity, for delivery to sanitary sewage treatment 44,000 m3 11.6 million gallons One tank at each pad and two buffer basins <0.2% glycol Runway deicer runoff capacity 75,000 m3 (2012)/ 150,000 m3 (2013). 19.8 million gallons/ 39.6 million gallons Two basins (2012)/ five basins (2013) Originally 2 x 9.2 million gallons. Under construction due to expansion: 3 x 6.6 million gallons. ADF capacity in pretreatment unit Max 9,900 lbs COD/day, max 26,400 gal/hour ADF consumption in denitrification unit Approx. 280 tons glycol/year (470 tons COD/year) Runway deicer capacity Max 2,200 lbs COD/day, max 1.31 mgd Start-up capacity is approx. 220-lbs COD/day due to biomass adjustment to the deicer chemical. Source: Per Espen Jahren, Water Management. Oslo Airport Norway. Notes: Conversion factor: 1-kg glycol = 1.68-kg COD; U.S. gallon = 3.785 L. Table 2. Key system sizing parameters. Parameter Value Unit Design flow rates - Minimum - Average - Maximum 540 765 Gallons per minute Design treatment load capacity 12,000 7,100* 7,100** lbs COD/day lbs BOD5/day lbs PG/day Design influent concentration - Range 725 425* 425** mg COD/L mg BOD5/L mg PG/L Design effluent concentration (average) 33 10 0.5 mg COD/L mg BOD5/L mg PG/L Design treatment efficiency (average) 95% % influent COD load treated Source: Kruger Kaldnes, Case Study: Gardermoen Waste Water Treatment Plant. *Data based on conversion: [COD] = 1.7 [BOD5]. **Data based on conversion: [PG] = [BOD5]. Table 3. Design basis for system performance.

D-88 Guidance for Treatment of Airport Stormwater Containing Deicers plant. Prior to treatment in the MBBR, the deicing-affected stormwater is mixed with the influ- ent sanitary wastewater. Sanitary wastewater typically ranges between 250-mg and 800-mg COD/L.4 Mixing the sanitary wastewater with the deicing-affected stormwater typically reduces combined concentration before treatment by the MBBR and dampens the peaks. Mixing the sanitary wastewater with the deicing-affected stormwater also increases the temperature of the deicing-affected stormwater and adds biological nutrients. Flow Rate Data on the operation and performance of the OSL MBBR were not available. Treated Load Rate Data on the operation and performance of the OSL MBBR were not available. However, the presence of sanitary wastewater sources provides a steady baseline mass loading that likely sta- bilizes the biological population, potentially making for a more robust system in the face of the more fluctuating deicer load contribution. Effluent Concentrations Approximately 3 days a year, OSL experiences discharges to their surface waters that exceed the regulatory limits. These exceedances are not necessarily related to the deicing component. Treatment Efficiencies The MBBR has demonstrated a 96% removal efficiency for COD. Cost The OSL MBBR treatment system is used to treat municipal wastewater as well as deicing- affected stormwater. Therefore, the capital and operating costs for the treatment of stormwater runoff at the municipal MBBR treatment plant are less than for a separate stormwater treatment system with the same capability. Lessons Learned for Potential Implementation of the MBBR Technology at Other Airports Conclusions drawn from the operation of the MBBR at OSL that can be used by other airports considering implementing this technology include: Parameter Value Unit Flow rates Not available Gallons per minute Actual COD treatment load rate Not available lbs/day Actual BOD5 treatment load rate Not available lbs/day Influent COD concentration - Average 559 mg/L Influent BOD5 concentration - Average 330* mg/L Effluent COD concentration - Average 25 mg/L Effluent BOD5 concentration - Average 3.2 mg/L Treatment efficiency 96% % influent COD load treated Source: Van Haandel, A. C., and Van Der Lubbs, J. G. M. Handbook of Biological Wastewater Treatment: Design and Optimisation of Activated Sludge Systems. London: IWA Publishing, 2012. *Data based on conversion: [COD] = 1.7 [BOD5]. Table 4. Actual system performance. 4 Metcalf & Eddy. Wastewater Engineering: Treatment and Reuse. New York: McGraw Hill Publishing, 2003.

Airport Deicer Treatment System Summaries D-89 1. The MBBR technology is located in an enclosed facility, with the cold stormwater influent mixed into warm sanitary wastewater. This results in less treatment at cold temperatures compared to some other deicer treatment systems. 2. The MBBR technology typically has negligible effluent concentrations of BOD5 (3.2 mg/L) and COD (25 mg/L), although the system may exceed regulatory limits as often as 3 times a year. 3. The municipal wastewater MBBR system is operated year round, treating deicer-contaminated stormwater from OSL in winter and spring. The year-round operation allows a healthy bacteria population capable of treating the seasonal stormwater runoff. 4. Nutrient balance is provided by the municipal wastewater. 5. Some municipal wastewater treatment plants are in need of sources of carbon to facilitate the process of removing nitrogen from wastewater (the denitrification process). Airports may want to engage local wastewater treatment authorities to assess if there is a need. If so, it could be a lower-cost method than other alternative off-site destruction methods for disposing of the concentrate. The biggest obstacles to using collected deicer-contaminated stormwater as a carbon source for denitrification are: a. The means of transportation and transportation costs from the airport to the treatment plant, since it cannot be discharged to the sanitary sewer system. b. The means of storage for the deicer at the treatment plant. c. The means of metering the deicer into the denitrification process, given the potentially changing deicer concentrations. d. Matching the quantity of carbon needed for denitrification with the quantity of deicer that is available. This can be problematic since the availability of deicer is variable. e. Large-volume storage facilities have to be available in order to provide even delivery to the sewage treatment plants.

D-90 Guidance for Treatment of Airport Stormwater Containing Deicers Airport Treatment Summary No. 12 Airport: Portland International Airport—Portland, OR (PDX) Treatment Technology: Anaerobic Fluidized Bed Reactor, POTW Discharge Years Operated: 2011–2012 (currently operational) Deicer Management System Description In response to effluent limits for BOD5 in its NPDES permit, the Port of Portland constructed an airport-wide deicer management system at PDX that came into operation during the 2002–2003 winter season. This deicer management system included concentrate deicer runoff collection, storage (2 million gallons), and discharge to the sanitary sewer along with diluted deicer runoff collection, storage (13 million gallons), and metered discharge to the receiving water according to the limits of the NPDES permit. In 2005, the port began consideration of additional deicer system enhancements to improve the performance of its existing system based on limitations in the receiving water’s abil- ity to assimilate the discharges under all necessary conditions, in addition to unforeseen limitations in the BOD load that could be discharged to the local POTW. In 2011, construction was completed on the enhancements to the deicer system, which included additional diluted deicer runoff storage (13 million gallons), additional concentrated deicer runoff storage (3 million gallons), a new outfall to the Columbia River, and an on-site AFBR treatment system. The treated effluent from the AFBR can be discharged to either the sanitary sewer or to the Columbia River if BOD load limitations in the NPDES permit allow. Figure 1 shows a process flow diagram of the system. Figure 2 shows the treatment facility site. Deicer Treatment Technology Selection Considerations The limits in the NPDES permit at PDX are dependent on flow rate in the receiving water (Columbia Slough) and are therefore variable. There are periods during which the limits to the Columbia Slough are so restrictive that no discharge of BOD5 is allowed. The airport also has limitations on the daily load that they may discharge to the sanitary sewer. Lastly, the port may expand operations in the future as demand grows. These factors led the airport to decide that on- site treatment was necessary to effectively manage deicer runoff collected at the airport while main- taining compliance with the NPDES limits and the load limits for discharges to the sanitary sewer. Figure 1. PDX deicer management system flow diagram.

Airport Deicer Treatment System Summaries D-91 A wide variety of deicer treatment technologies was considered for PDX. An AFBR treatment facility was chosen because of its proven ability to treat deicer-affected runoff, its ability to remove a high percentage of influent BOD5, its ability to withstand the potential intermittent availability of deicer without needing to reseed the system, and because the system fit well into PDX’s existing airport-wide deicer management system. The AFBR system was also chosen due to its ability to consistently and predictably achieve desired effluent concentrations despite the great potential for variability of flows and concentrations entering the concentrate stor- age tanks. An RO treatment system was considered, and a pilot system was installed at PDX during design of the system enhancements, but an overabundance in silica in the stormwater prevented the RO facility from becoming a viable alternative. An aerobic treatment system was also considered but was discarded due to a lesser ability to handle the intermittent nature of PDX deicer discharges without reseeding or addition of supplementary and costly sources of BOD5 to keep the biology active. Deicer Treatment Technology Description AFBR See the AFBR treatment technology fact sheet (Fact Sheet 104) for a general description of the AFBR technology. The PDX AFBR system generally follows this description. Figure 3 shows the biological reactor units in the system. Figure 2. PDX on-site treatment facility site. Figure 3. Biological reactor units in the PDX AFBR system.

D-92 Guidance for Treatment of Airport Stormwater Containing Deicers Description of Support Systems The AFBR at PDX includes the following support systems for the treatment reactor-separator unit: storage (one 2-million-gallon tank and one 3-million-gallon tank for concentrated deicer runoff), influent pumping system, heat generation and exchange loop, chemical feed for nutrient addition and pH control, biogas handling, and biological solids removal and handling. Collected runoff water from the storage tanks is pumped to a small holding tank near the treatment facility and then pumped at a flow rate set by the system operators to achieve a constant BOD5 loading as influent BOD5 concentrations change. The cold influent water is heated first by passing it by warm effluent water in a heat exchanger and then by passing it by hot water from a boiler in a sec- ond heat exchanger. The hot water is obtained by heating potable water in a boiler using biogas captured from the reactor. The biogas is approximately 77% methane and 23% carbon dioxide and is used similarly to natural gas. For the PDX system, the heating system burns self-generated biogas, except for initial yearly start-up when natural gas is used. Any excess biogas is burned in a flare external to the building. The AFBR technology requires addition of a base chemical (sodium hydroxide) to keep pH in the reactors neutral, as well as addition of various chemical nutrients to support growth of the bacteria. Biological solids exiting the reactor-separator unit with the treated effluent are removed with a dissolved air flotation clarifier under certain condi- tions. The treated effluent from the reactors can be routed to bypass the dissolved air flotation clarifier, with the biological solids discharged to the sanitary sewer. Biological solids that are removed from the effluent are disposed of in a landfill. Key Treatment System Sizing Parameters See Table 1 for system sizing parameters. Treatment System Performance The 2011–2012 season was the start-up year for the PDX AFBR. During the year, the AFBR system was fed both from deicer-affected stormwater, off-spec deicer that could no longer be used, and purchased glycol. The feeding of purchased glycol was only for the start-up season. The goal was to test the capacity and capabilities of the system by slowly increasing the COD loading to the AFBR reactors over time, with the further goal of reaching the design COD loading and assessing whether the COD removal target of 98% could be reached. Through the process, the mechanical and control functions of the system were assessed. Design performance targets are presented in Table 2. Actual performance data are presented in Table 3. Figure 4 and Figure 5 graph actual performance data. Cost Assessment for the PDX AFBR Treatment System Engineering cost estimates indicated that the treatment facility capital cost, including the equipment, instrumentation, treatment building, and site/civil features external to the building, would be $9 million to $10 million. Component/Parameter Size/Capacity of Treatment Units Number of Treatment Units Total Capacity Treatment unit volume 29,000 gallons 2 58,000 gallons Treatment unit dimensions Reactors: 14-ft diameter 2 N/A Treatment facility footprint 0.28-acre building* 1 0.28-acre building *The 0.28-acre building footprint includes 0.06 acre storage and maintenance facility for GRVs. Table 1. Key system sizing parameters.

Airport Deicer Treatment System Summaries D-93 Parameter Value Unit Design flow rates - Minimum - Average - Maximum 5 35 200 Gallons per minute Design treatment load capacity 11,500 6,700* 6,700** lbs COD/day lbs BOD5/day lbs PG/day Design influent concentration - Range 1,800~34,000 1,050~20,000* 1,050~20,000** mg COD/L mg BOD5/L mg PG/L Design effluent concentration (average) <250 (after start-up) <150 <10 mg COD/L mg BOD5/L mg PG/L Design treatment efficiency (average) 98% % influent COD load treated *Data based on conversion: [COD] = 1.7 [BOD5]. **Data based on conversion: [PG] = [BOD5]. Table 2. Design basis for PDX AFBR system performance. Parameter Value Unit Flow rates - Minimum - Average - Maximum 11 32 90 Gallons per minute Actual COD treatment load rate - Average - Maximum 2,400 4,200 lbs/day Actual BOD5 treatment load rate - Average - Maximum 1,400* 2,400* lbs/day Actual PG treatment load rate - Average - Maximum 1,400** 2,400** lbs/day Influent COD concentration - Minimum - Average - Maximum 2,000 14,500 52,500 mg/L Influent BOD5 concentration - Minimum - Average - Maximum 1,200* 8,500* 30,800* mg/L Influent PG concentration - Minimum - Average - Maximum 1,200** 8,500** 30,800** mg/L Effluent COD concentration - Minimum - Average - Maximum 40 141 2,100 mg/L Treatment efficiency 98.66% % influent COD load treated *Data based on conversion: [COD] = 1.7 [BOD5]. **Data based on conversion: [PG] = [BOD5]. Table 3. Actual system performance.

D-94 Guidance for Treatment of Airport Stormwater Containing Deicers 0% 20% 40% 60% 80% 100% 120% 12/5 12/19 1/2 1/16 1/30 2/13 2/27 3/12 3/26 4/9 4/23 Pe rc en t R em ov al (% ) PERCENT LOAD REMOVAL BY DEICER TREATMENT SYSTEM Train 1 Train 2 Avg. % Removal at 100% Loading Train 1 = 99.1% Train 2 = 98.9 % Avg. % Removal at 60% Loading Train 1 = 99.0% Train 2 = 98.9% Figure 4. Actual system treatment efficiency performance. 20% Capacity 40% Capacity 60% Capacity 80% Capacity 100% Capacity 0 500 1,000 1,500 2,000 2,500 3,000 3,500 4,000 12/5 12/19 1/2 1/16 1/30 2/13 2/27 3/12 3/26 4/9 4/23 CO D Lo ad T re at ed (l bs C O D/ da y) COD LOAD TREATED BY DEICER TREATMENT SYSTEM Train 1 Train 2 100% Treatment Figure 5. Actual system treated load performance.

Airport Deicer Treatment System Summaries D-95 Conclusions on Performance of PDX AFBR System The conclusions presented here are based only on 2011–2012 data, which was the system’s start-up season. Influent Deicer Concentrations Influent COD concentrations were higher than would be expected in subsequent seasons due to the times when the system was fed from off-spec and new glycol sources. During the times when the system ran from deicer-affected stormwater, the COD concentrations were in the range of what would typically be expected. Flow Rate During the start-up phase, flow rates were not representative of flow rates expected in the future when runoff collected from the apron areas will be treated. Treated Load Rate The PDX AFBR reached its COD loading target of 7,700 lbs COD per day. Effluent Concentrations The PDX AFBR effluent concentrations averaged 131 mg/L based on measurements once the system reached the 50% loading mark. The average effluent concentrations for both reactors were similar. Treatment Efficiencies The COD removal efficiency was 99%, exceeding the design target of 98%. The system was able to maintain that removal efficiency at loading rates that ranged from 25% to 100% of design loadings. Lessons Learned for Potential Implementation of the PDX AFBR Technology at Other Airports The following factors have proven critical to effective and efficient performance in the PDX AFBR: 1. Proper seeding of bacteria. 2. Achieving target fluidization rates for the pumps. 3. Providing the necessary caustic feed to achieve pH targets. 4. Achieving target reactor temperatures. 5. Step increases in COD loading during the first year. Conclusions from operation of the AFBR at PDX that can be used by other airports consider- ing this technology include: 1. At initial start-up of the system, acquisition of the appropriate type of healthy bioseed is criti- cal. The bioseed must be obtained from a similar type of anaerobic operation. 2. Appropriate storage capacity and control of loading into the AFBR are important. 3. A well-planned and thorough commissioning of the system, including performance-based commissioning once bioseed is added, is critical to successful system implementation. This includes testing of the biogas handling system under field conditions.

D-96 Guidance for Treatment of Airport Stormwater Containing Deicers Documents and Information Review in Development of Airport Summary 1. Portland International Airport Deicing Facility Enhancement Project Schematic Design Report (2008). 2. RO Concentrate Treatment Alternatives Plan for PDX Deicing System Enhancements Sche- matic Design (2007). 3. PDX Airport Deicer Treatment System operational logs (2011–2012).

Airport Deicer Treatment System Summaries D-97 Airport Treatment Summary No. 13 Airport: Edmonton International Airport—Edmonton, Alberta (YEG) Treatment Technology: Passive Facultative Biological Treatment, Aerated Gravel Beds Years Operated: 2001–2012 (Currently Operational)5 Deicer Management System Description In the deicer management system at YEG, contaminated snow and ice is piled on the edge of pavement during the very cold winter months. During spring thaw, the contaminated snowmelt is diverted to a 90,000-m3 (~24 million gallons) pond for storage. The stored contaminated snowmelt is pumped into a treatment system prior to discharge to the adjacent creek. From 2001 through 2011, YEG treated deicer-affected stormwater in a 12-bed wetland-based passive facultative system. Due to the nature of snowmelt in Edmonton, the system functioned as a batch treatment process with initial influent BOD5 concentrations of approximately 600 mg/L when active treatment began in the spring. As each spring thaw continued, concentrations decreased notably. The system was upgraded in 2011 with the addition of an aerated gravel bed system to provide treatment of airfield runoff associated with deicing activity and meet Alberta Environment permit limits for discharge to a neighboring tributary to Whitemud Creek. Figure 1 shows the treatment system. The original horizontal flow wetland-based design suffered from performance issues primar- ily related to a lack of hydraulic capacity. The influent pump was improperly sized to handle the range of static head. Moreover, the horizontal flow configuration of the media bed reduced the hydraulic throughput due to the hydraulic resistance of the gravel. In 2011, the system was upgraded with a new set of influent pumps and reconfigured to a vertical flow configuration to increase the hydraulic capacity. Aeration and nutrient addition were also included based on the projected design load. Two of the original six treatment cells were upgraded to aerated gravel beds. The upgraded treatment system consists of two parallel trains. The first cell of each train is an aerated graved bed. The second cell is a surface flow wetland. The upgraded system is required 5 The YEG deicing-affected stormwater treatment system was upgraded from non-aerated gravel beds to aerated gravel beds in 2012. Figure 1. Photo of treatment system.

D-98 Guidance for Treatment of Airport Stormwater Containing Deicers to treat the full capacity of the storage pond within a 60-day time period once treatment starts in the spring, and to produce a high-quality effluent to protect the receiving stream. Deicer Treatment Technology Selection Considerations The primary considerations for the airport in selection of the upgraded treatment technology were low capital cost, a system that was compatible with existing systems, the ability to provide sufficient treatment load, and the ability to reliably meet effluent limits. YEG is located in a rural area with an abundance of land surrounding the airfield. A gun club pond and treatment system are located far away and present negligible bird strike hazards. Deicer Treatment Technology Description YEG uses a combination of passive facultative treatment (Fact Sheet 108) in the form of a horizontal subsurface wetland and aerated gravel beds (Fact Sheet 102). The flow distribution and aeration/collection piping are shown in Figure 2 and Figure 3, respectively. Figure 2. Vertical flow distribution system. Figure 3. Installation of drain and aeration lines.

Airport Deicer Treatment System Summaries D-99 Description of Support Systems Influent from the gun club pond is pumped into an aboveground splitter structure that divides flow between the trains. Influent pumps (two) have variable frequency drives. Water from the splitter structure flows by gravity to influent dosing, and nutrient addition lines lie atop the aerated gravel bed. Flow from the dosing line travels downward through the gravel to drains on the floor of the cells. A recirculation pump is installed in a sump prior to the effluent structure and is designed to provide water recirculation during seasonal start-up. Each recirculation pump is sized for 350 gpm. An irrigation propeller pump is used for the high flow, low head system. Effluent from the aerated gravel bed flows by gravity to the constructed surface flow wetlands (second cell). Influent is distributed along the leading edge of the system and picked up in a drain line running along the opposite side. The water level in this cell is to be maintained at 1-ft water depth. Key System Sizing Parameters See Table 1 for system sizing parameters. Treatment System Performance See Table 2 for designed system performance information. The upgraded system was started up in the spring of 2012. Deicer-affected stormwater enhanced with nutrients was gradually loaded into the system in a flow-through manner fol- lowing a 2-week acclimation period in which effluent was recirculated. The system was able to treat the contents of the storage pond in 60 days while meeting regulatory effluent limits. Figure 4 illustrates the decrease of COD over the sampling period. Figure 4 shows a near linear decrease in COD concentrations over time at approximately 7-mg COD/L per hour (168-mg COD/L per day). It is expected that after all biodegradable con- taminants are degraded, the COD values will level out to a practical floor representative of the Component/Parameter Size/Capacity of Treatment Units Number of Treatment Units Total Capacity Stormwater storage capacity 42 million gallons 1 42 million gallons Treatment unit volume 140,000 gallons 2 280,000 gal Treatment unit dimensions 141-ft L x 141-ft W x 3-ft 31/3-in. D 2 130,400 ft3 Treatment facility footprint 0.45 acre per train 2.5 acre total site 2 1 2.5 acre Table 1. Key system sizing parameters. Parameter Value Unit Design flow rates - Average - Maximum 275 733 Gallons per minute Design treatment load capacity 1,770* 1,040 1,040** lbs COD/day lbs BOD5/day lbs PG/day Design effluent concentration (average) 100 mg PG/L Design treatment efficiency (average) 98% % influent COD load treated *Data based on conversion: [COD] = 1.7 [BOD5]. **Data based on conversion: [PG] = [BOD5]. Table 2. Design basis for system performance.

D-100 Guidance for Treatment of Airport Stormwater Containing Deicers nonbiodegradable fraction of organics in the water. Figure 5 provides a visual confirmation of treatment after 24 hours of operation. Cost Assessment for Treatment System See Table 3 for treatment system costs. Conclusions on Performance of YEG Aerated Gravel Bed System Based on the initial start-up of the enhanced system in the spring of 2012, the addition of the aerated gravel beds met performance objectives. y = -6.868x + 620.25 R5 = 0.8647 0 200 400 600 800 0 10 20 30 40 50Ch em ic al O xy ge n De m an d (m g- CO D/ L) Hours After Initiation of Treatment Biological Commissioning – Environmental Impact Assessment COD Versus Time Figure 4. COD removal after initiation of treatment. Figure 5. Influent (left) and effluent (right) samples after 24 hours of operation.

Airport Deicer Treatment System Summaries D-101 Lessons Learned for Potential Implementation of the AFBR Technology at Other Airports The following factors are anticipated to be critical to effective and efficient performance in the YEG aerated gravel bed system: 1. The ability to maintain a loading rate below 0.051-lbs BOD/ft2/day. 2. The ability to provide upstream equalization from the beginning of the deicing season until spring thaw. 3. The ability to apply nutrients one time upstream of the treatment system was deemed the most cost-effective means to add nitrogen and phosphorus to the influent flow. 4. Presence of adequate material to operate effectively during extremely cold winters. 5. The ability to start up and shut down effectively since the system is uniquely designed to provide treatment of stormwater in the thaw period of the spring. Documents and Information Review in Development of Airport Summary 1. Wetland treatment upgrade–construction drawings (Associated Eng., et al.) 2. Design Brief, February 2011(Associated Eng. and Naturally Wallace) 3. Naturally Wallace project files. Cost Category Projected at Initial Implementation Actual Capital cost* - Initial system - Upgraded system $2,000,000 $3,000,000 $2,000,000 $3,000,000 Annual operating cost Not provided Not provided *Capital costs are for the treatment system only, including the building and basic building infrastructure. Costs do not include site-specific costs for collection, storage, and discharge. Table 3. Costs for the treatment system.

D-102 Guidance for Treatment of Airport Stormwater Containing Deicers Airport Treatment Summary No. 14 Airport: Halifax International Airport—Halifax, Nova Scotia (YHZ) Treatment Technology: Mechanical Vapor Recompression (MVR) Years Operated: 2004–2012 (Currently Operational) Deicer Management System Description YHZ has both a passive and active collection system in place for the capture of spent aircraft deicing fluid. All ADF that is applied at YHZ is EG-based. Deicing operations are conducted at both the terminal gate areas and the remote deicing pad. Active collection involves the use of GRVs on any area with deicing activity in order to maxi- mize the collection of high-concentrate fluids as soon as possible. Passive collection involves the use of a dedicated glycol collection drainage system for the deicing pad areas. These collection basins, piping, and pump stations allow conveyance of spent aircraft deicing fluid to on-site storage tanks adjacent to the deicing pads. The spent ADF cap- tured by this passive system is sent to two storage tanks (each with 65,000 gallons of capacity), which are connected to the deicing pad pumping systems. The two large storage tanks act as the interim storage until the spent ADF is transferred via tanker truck to the glycol processing facility at YHZ. All spent ADF that is collected is treated at an on-site recycling facility located at the Hali- fax International Airport. The glycol processing facility is owned and operated by a recycling subcontractor. This facility is on the airport, but it is located outside the airside secured area. Once transported to the processing facility, the collected ADF is segregated according to glycol concentration. The YHZ glycol recycling facility has two storage tanks with a combined volume of 950,000 U.S. gallons; these are used to store the low-concentration ethylene glycol (less than 10%). A third storage tank with a capacity of 66,000 U.S. gallons is designated for the collection of high-concentration ethylene glycol (more than 10%). The entire spent-ADF management system is operated by a subcontractor to ensure that unpermitted levels of glycol do not enter the sanitary sewer, via the airport wastewater flow, in order to comply with the Halifax Regional Municipality Wastewater Guidelines as directed by the Nova Scotia Department of Environ- ment’s 2004 permit approval. This coincides with the Canadian Environmental Protection Act, which requires that discharge of glycols into surface waters resulting from aircraft deicing and anti-icing activities at Canadian airports not exceed a concentration of 100 mg/L of ethylene glycol at the property lines. The recycling equipment includes three mechanical vapor recompression units, one dissolved air floatation filtration system, and one distillate aerator. All wastewater generated from the on- site treatment systems is discharged to an off-site wastewater treatment plant called the Aerotech Park Wastewater Treatment Facility. Deicer Treatment Technology Selection Considerations In 2002, the aviation industry in Canada established a “National Contract” to ensure that the airline community met the guidelines for the discharge of effluent generated from deicing opera- tions. The Halifax International Airport meets these terms by using a subcontractor to manage the drain management at the gate areas and the deicing pads to ensure all noncompliant fluids are collected and processed. The guideline for release to the environment is 100-mg/L ethylene glycol, as set under the Canadian Environmental Protection Act.

Airport Deicer Treatment System Summaries D-103 YHZ has deicing pads that consist of a pumping system that the airport owns and main- tains. This is used to pump all high-volume fluids from the deicing pads into aboveground storage tanks beside the pads. With Halifax being in a maritime climate, there are many weather events that require the collection of rainwater until the 100-mg/L release point is met. This typically could result in a large volume of dilute spent ADF being collected, but GRVs are also used to vacuum the glycol directly off the surface. The fluids recovered by the GRVs are normally higher in concentration levels. By proactively cleaning the pad surfaces with this equipment, the overall volume of stormwater that has to be collected is reduced, and compliance levels can be met more quickly. All the collected ADF, from the passive system on the deicing pads and from direct recovery on pads or gates with the recovery vehicles, is transferred to the glycol recycling facility at YHZ. The fluids from the passive system are transferred by tanker truck, while the glycol recovery vehicle offloads the fluid it collects directly into the appropriate tank located at the facility. The Halifax site recycling system was designed, installed, and implemented to meet the fol- lowing requirements: 1. Compliance with federal and provincial environmental regulations plus the existing munici- pal bylaws for wastewater discharge criteria. 2. On-site treatment that generates wastewater requires compliance with the following: limitations of 100-mg/L EG, 300-mg/L BOD5, 1000-mg/L COD, 15-mg/L oil and grease (mineral/synthetic), and 300-mg/L TSS. 3. Ability to meet an average production/removal rate of 400,000 U.S. gallons per month when spent-ADF volumes are present. Treatment technology was selected by the recycling contractor on the basis of being able to handle fluctuating glycol concentrations in spent ADF that occur with each weather-related deicing event. The analysis of factors included designing processing capability based on actual spent-ADF collection data from previous years. As part of the overall system, two independent processing trains were designed: one to recycle ethylene glycol of less than 10% concentration from the low-concentrate feed tanks, and another to process ethylene glycol of greater than 10% concentration from the high-concentrate feed tank. MVR was selected to treat both streams of concentrations because: 1. Alternative disposal options were limited and expensive (specifically, no off-site treatment facilities were in close proximity to the airport), 2. The MVR could handle anticipated fluctuations in glycol concentrations, and 3. The airport authority determined it was in its best interests to capture and recycle the glycol. Equipment was sized to accommodate the 400,000-gallon/month removal rate and so that no additional storage tanks would be required for storage and processing of the collected fluids. The glycol that is reclaimed from the system is sold, and the revenues generated are used to offset program costs to provide glycol management services. Deicer Treatment Technology Description The YHZ treatment system employs the MVR treatment process. A description of the MVR treatment technology can be found in Fact Sheet 106. Figure 1 shows the recycling facility. Description of Treatment Support Systems The MVR at YHZ includes the following support systems: two low-concentrate tanks capable of storing a combined 846,000 gallons, one high-concentrate tank capable of storing 66,000 gallons, a filtration system, a dissolved air flotation (DAF) unit, a chemical feed system, blowers

D-104 Guidance for Treatment of Airport Stormwater Containing Deicers with variable frequency drives, heat exchangers, a scrubber-absorber, electrical service, a control system, MVR maintenance, and solids disposal. Filtration systems installed to treat influent on the MVR units are an integral part of the overall recycling system. The DAF unit is an effective and integral part of the recycling process. Stainless-steel hot filter vessels with 1-micron filter bags are also used on each MVR prior to the feed entering the unit, which allows the influent to be filtered while it is hot in an effort to remove as much TSS as possible. Each type of filtration method increases influent throughput production by minimizing stoppages due to premature maintenance and cleaning requirements of the MVR heat exchangers. The DAF was designed as a support system to the MVR units to treat as much of the contaminants that make up TSS in the spent ADF as can be drawn out before the fluid is processed. Using the DAF increases production, and with less contamination of dirt in the MVR heat exchanger plates, the downtime for maintenance and cleaning is significantly reduced. The DAF adjusts the pH to a neutral level by reading and injecting caustic with a pumping system. With the fluid at neutral pH, a floccu- lent chemical is added. This fluid is then injected into the fluid-filled DAF unit along with air drawn in through the DAF pump. This mixture binds the contaminants (smaller than 1 micron) together to make larger particles that float with the air that was injected. These accumulated contaminants form a floating cake on the top of the fluids, which is skimmed off and disposed of. The fluids under the cake, now filtered by the DAF, are drawn in to the MVR units for processing. The concentrate is kept in tanks for interim storage until it can be trucked off the airport for sales into secondary markets. The control system includes many warning and emergency controls that, in the event of any mechanical failure or fluid overflow situation, shut off the units automatically. These were installed to minimize manpower requirements so that in most cases the facility can be run with one person. Heat exchanger plate changes are anticipated every 170,000 U.S. gallons on average. Downtime per shutdown is approximately 12 hours per machine for total maintenance. Solids from processing in YHZ filters are dried and sent to a landfill. Tank sludge at season end is disposed of at an off-site treatment plant. Key Treatment System Sizing Parameters The recycling contractor leases the airport land for the glycol processing facility, as well as the airside tanks at YHZ. The contractor supplied and installed the tanks, building, recovery trucks, and processing equipment for the Halifax site. The airport authority supplies the deicing pad Figure 1. YHZ glycol recycling facility.

Airport Deicer Treatment System Summaries D-105 and pumping systems and owns the drain blocking devices used in the spent-ADF collection infrastructure. See Table 1 for treatment sizing parameters. Treatment System Performance The processing/disposal rate per month will vary based on the glycol percentage. A processing performance of 400,000 gallons per month was designed based on two stages, with the concen- trator system running to achieve the performance indicated in Table 2 through Table 5. The information in Table 6 on actual system performance was derived from monthly average data collected at the facility between 2009 and 2012. Component/Parameter Size/Capacity of Treatment Units Number of Treatment Units Total Value Stormwater storage capacity Low-concentrate storage Airfield storage High-concentrate storage 423,000 gallons 66,000 gallons 66,000 gallons 2 2 1 846,000 gal 122,000 gal 66,000 gal Product storage volume 17,000 gallons 3 51,000 gal Treatment unit dimensions MVR MVR with scrubber 20-ft L x 6-ft W x 8’2” H 20-ft L x 6-ft W x 22 H 3 Treatment facility footprint MVR treatment building Treatment building and storage tanks 0.11 acre 4.94 acre 1 1 0.11 acre 4.94 acre Table 1. Key treatment system sizing parameters. Glycol Percentage Min. Removal Rate per Month (Two concentrators) Max. Removal Rate per Month (Three concentrators) <10% and lower 225,000 gallons 458,000 gallons Table 2. Stage 1 EG processing design basis (concentrator MVR systems). Glycol Percentage Min. Removal Rate per Month (One concentrator) Max. Removal Rate per Month (Two concentrators) >10% and higher 95,000 gallons 260,000 gallons Table 3. Stage 2 EG processing design basis (concentrator MVR systems). Parameter Value Unit Design flow rates - Minimum - Average - Maximum 2 Not available 4 Gallons per minute Design treatment load capacity 16,500* 9,700** 9,700 lbs COD/day lbs BOD5/day lbs PG/day Design influent concentration - Range 13,000~351,000* 5,000~135,000** 10,000~270,000 mg COD/L mg BOD5/L mg PG/L Design effluent concentration (average) <50~1000 Not available <50~1000 mg COD/L mg BOD5/L mg PG/L Target concentrate stream % EG 50 % Design treatment efficiency 94.1~99.7 % influent COD load treated *Data based on conversion: [EG] = 1.3 [COD]. **Data based on conversion: [EG] = 0.5 [BOD5]. Table 4. Design basis for MVR system performance.

D-106 Guidance for Treatment of Airport Stormwater Containing Deicers Parameters Single-Stage Production Two-Stage Production Stage 1 Stage 2 Influent flow rate range (gallons per hour) 150 to 200 170 to 230 130 to 170 Influent glycol concentration range (% glycol) 4 to 27 1 to 4 13 to 27 Influent temperature range (F or C) Ambient Ambient Ambient Number of effluent streams produced 2 streams– distillate and concentrate 2 streams– distillate and concentrate 2 streams– distillate and concentrate Distillate effluent flow rate range (gallons per hour) 60 to 184 136 to 219 52 to 126 Distillate effluent water quality (COD range in mg/L) <50 to 1,000 <50 to 1,000 <50 to 1,000 Distillate effluent water quality (mg/L COD) <50 to 1,000 <50 to 1,000 <50 to 1,000 Distillate effluent water quality (pH range) 3 to 8 3 to 8 3 to 8 Concentrate effluent flow rate range (gallons per hour) 12 to 120 8.5 to 61 33 to 102 Concentrate effluent concentration (% glycol range) 50 to 55 15 to 20 50 to 55 Heat source Electric-powered steam compression Control system PLC Energy consumption information 0.4 Kw per gal feed Estimate of waste to be produced Sludge and solids negligible and glycol in overheads less than 0.1% Footprint, dimensions, etc. Each MVR unit is 20’ (L) x 6’ (W) x 8’ 2” (H), with scrubber 13’ (H) or 22’ (H) Other support systems Feed pre heater heat exchanger, electric air compressor, cold and hot filter systems, piping for feed, distillate, concentrate, and storage tanks for feed, distillate, concentrate Notes: Specifications provided by Inland Technologies per design criteria for each MVR. YHZ has three MVR units installed. Table 5. Additional design basis for MVR system. Parameter Value Unit Flow rates - Minimum - Average - Maximum 0.11 5.9 8.6 Gallons per minute Actual EG treatment load rate - Average - Maximum 695 1,180 lbs/day Influent EG concentration - Minimum - Average - Maximum 27,000 52,000 105,000 mg/L Effluent COD concentration <100 mg/L mg/L Effluent BOD5 concentration Not available mg/L Effluent EG concentration* - Average - Maximum 27 70 mg/L Treatment efficiency 99.66 % influent EG load treated Values obtained from the monthly averages and totals in “Historical Data Assessment Based on Three Seasons of Data from 2009 Through 2012.” *Values obtained from monthly sampling provided in “Historical Data Assessment Based on Three Seasons of Data from 2009 Through 2012.” Table 6. Actual MVR system performance.

Airport Deicer Treatment System Summaries D-107 Historical Data Assessment Based on Three Seasons of Data from 2009 Through 20126 See Table 7 through Table 9 for actual MVR data at YHZ. Cost Assessment for YHZ MVR Treatment System The recycling contractor leases the airport land for the glycol processing facility as well as the airside tanks at YHZ. The contractor supplied and installed the tanks, building, recovery trucks, and processing equipment for the Halifax site. The airport authority supplies the deicing pad and pumping systems and owns the drain blocking devices used in the spent-ADF collection infrastructure. See Table 10 for treatment system costs. Conclusions on Performance of YHZ Treatment System Influent Deicer Concentrations and Flow Rate The MVR units at YHZ are configured to conduct two-stage processing. Based on data from 2009–2012, influent glycol concentrations of the low-concentration EG stream ranged from 3.9% to 6 All treatment data provided by YHZ operational logs. Stage 1 Nov Dec Jan Feb Mar Apr May Total or Avg for Season Influent volume processed (liters) 285,258 962,200 1,071,267 1,048,233 1,256,418 1,157,428 547,684 6,328,488 Average influent glycol concentration (% EG) 6.9% 3.7% 4.6% 5.1% 4.3% 3.1% 1.5% 4.2% Effluent volume of concentrate produced (liters) 83,328 266,124 334,766 332,045 349,691 270,055 106,638 1,742,647 Average effluent concentration of glycol produced (% EG ) 14.0% 14.0% 14.0% 14.0% 14.0% 14.0% 14.0% 14.0% Stage 2 Influent volume processed (liters) 0 294,078 289,742 356,237 380,658 235,500 212,223 1,768,438 Average influent glycol concentration (% EG) 0.0% 14.0% 14.0% 14.0% 14.0% 14.0% 14.0% 14.0% Volume of 100% EG in influent (liters) 0 41,171 40,564 49,873 53,292 32,970 29,711 247,581 Effluent volume of concentrate produced (liters) 0 60,535 75,403 89,617 101,056 52,072 50,935 429,618 Average effluent concentration of glycol produced (% EG ) 0.0% 53.0% 50.5% 52.0% 54.6% 53.0% 53.0% 52.7% Volume of 100% EG in glycol produced (liters) 0 32,084 38,079 46,601 55,177 27,598 26,996 226,533 Combined Discharges from Both Stages Total distillate discharged to sanitary (liters) 165,396 785,907 854,841 768,398 1,053,244 1,157,206 561,469 5,346,461 Average effluent concentration of distillate (mg/L EG) <100 <100 <100 <100 <100 <100 <100 <100 % ratio of glycol reclaimed from Stage 2 vs. infeed N/A 77.9% 93.9% 93.4% 103.5% 83.7% 90.9% 91.5% Table 7. Actual YHZ MVR data for 2009–2010 deicing season.

D-108 Guidance for Treatment of Airport Stormwater Containing Deicers 4.3%. This low-percentage feed is processed by two of the MVR units to make a concentrated stream with an average concentration of 14%. This stream is then comingled with any other spent ADF that is collected that is over 10% EG concentration so that the third MVR unit is configured to process the high-concentration EG feed, which is brought up to 50% to 52% EG. Based on feedback from the recycling operator, the MVR up-stages the concentration of spent EG to increase flow rates through the MVR units. In reference to the data, it is apparent that each concentrator running Stage 1 can process at least double the amount of influent when compared to a concentrator running Stage 2, higher-concentration glycol. This is very beneficial since this technique removes water from storage tanks more quickly than single-stage processing and keeps adequate storage for future storm events. Treated Load Rate The rate at which soluble or total COD is removed from the system is not a key measure of performance for the MVR system. Effluent Concentrations Each MVR at YHZ can be adjusted to produce a desired glycol concentration product. The MVR units produce two effluent streams, and the desired concentration set points in each effluent stream directly affect the performance of the concentrators. The operators have the ability to Stage 1 Nov Dec Jan Feb Mar Apr May Jun Jul Total or Avg for Season Influent volume processed (liters) 19,262 731,277 1,067,950 1,273,601 989,352 1,413,106 1,011,352 1,136,983 453,591 8,096,474 Average influent glycol concentration (% EG) 5.6% 5.5% 4.3% 4.3% 4.1% 3.2% 3.4% 3.4% 5.0% 4.3% Effluent volume of concentrate produced (liters) 5,517 213,741 292,599 314,743 315,608 335,442 260,667 327,040 133,409 2,198,766 Average effluent concentration of glycol produced (% EG ) 14.0% 14.0% 14.0% 14.0% 14.0% 14.0% 14.0% 14.0% 14.0% 14.0% Stage 2 Influent volume processed (liters) 0 165,444 402,440 184,862 387,560 250,735 316,829 293,245 245,730 2,246,845 Average influent glycol concentration (% EG) 0.0% 14.0% 14.0% 14.0% 14.0% 14.0% 14.0% 14.0% 14.0% 14.0% Volume of 100% EG in influent (liters) 0 23,162 56,342 25,881 54,258 35,103 44,356 41,054 34,402 314,558 Effluent volume of concentrate produced (liters) 0 38,749 100,184 39,166 103,328 73,916 85,043 72,142 69,717 582,245 Average effluent concentration of glycol produced (% EG ) 0.0% 54.2% 55.5% 52.7% 52.0% 52.0% 52.0% 52.7% 51.0% 52.8% Volume of 100% EG in glycol produced (liters) 0 21,002 55,602 20,640 53,731 38,436 44,222 38,019 35,556 307,208 Combined Discharges from Both Stages Total effluent distillate discharged to sanitary (liters) 0 476,219 720,682 830,623 715,328 1,132,068 909,635 1,065,889 525,778 6,376,222 Average effluent concentration of distillate (mg/L EG) <100 mg/L <100 mg/L <100 mg/L <100 mg/L <100 mg/L <100 mg/L <100 mg/L <100 mg/L <100 mg/L <100 mg/L % ratio of glycol reclaimed from Stage 2 vs. infeed N/A 90.7% 98.7% 79.8% 99.0% 109.5% 99.7% 92.6% 103.4% 97.7% Table 8. Actual YHZ MVR data for 2010–2011 deicing season.

Airport Deicer Treatment System Summaries D-109 Stage 1 November December January February March April Total or Avg for Season Influent volume processed (liters) 502,376 1,056,419 1,168,546 1,218,376 1,385,299 1,346,240 6,677,256 Average influent glycol concentration (% EG) 4.0% 2.9% 3.6% 3.9% 4.8% 4.4% 3.9% Effluent volume of concentrate produced (liters) 120,730 200,061 277,204 292,074 433,171 385,922 1,709,162 Average effluent concentration of glycol produced (% EG ) 14.0% 14.0% 14.0% 14.0% 14.0% 14.0% 14.0% Stage 2 Influent volume processed (liters) 0 219,335 294,294 251,748 512,606 440,452 1,718,435 Average influent glycol concentration (% EG) 0.0% 14.0% 14.0% 14.0% 14.0% 14.0% 14.0% Volume of 100% EG in influent (liters) 30,707 41,201 35,245 71,765 61,663 240,581 Effluent volume of concentrate produced (liters) 0 54,835 74,691 67,084 132,377 107,773 436,760 Average effluent concentration of glycol produced (% EG ) 0.0% 52.0% 53.0% 51.8% 52.8% 53.5% 52.8% Volume of 100% EG in glycol produced (liters) 0 28,514 39,586 34,750 69,895 57,659 230,404 Combined Discharges from Both Stages Total effluent distillate discharged to sanitary (liters) 261,080 989,616 1,104,459 862,233 1,383,081 1,423,335 6,023,804 Average effluent concentration of distillate (mg/L EG) <100 mg/L <100 mg/L <100 mg/L <100 mg/L <100 mg/L <100 mg/L <100 mg/L % ratio of glycol reclaimed from Stage 2 vs. infeed N/A 92.9% 96.1% 98.6% 97.4% 93.5% 95.8% Table 9. Actual YHZ MVR data for 2011–2012 deicing season. Cost Category Actual (Canadian $, 2003) Capital cost* Building and storage tanks Concentrators and support equipment Total capital cost $2M $1.15M $3.15M Annual operating cost** - Utilities - Processing supplies - Analysis - Repair and maintenance, labor, tank cleaning, other Total operating cost $90,000 $23,000 $5,000 $312,000 $ 430,000 *Capital costs are for the treatment system only and do not include glycol collection infrastructure costs. **Processing supplies include chemicals and filters. Table 10. Costs for the treatment system.

D-110 Guidance for Treatment of Airport Stormwater Containing Deicers adjust various parameters on each MVR via the PLC and can do so as conditions or influent characteristics change. In Stage 1 of the processing phase, the effluent glycol produced averaged 14% EG. This glycol concentration is continually monitored to balance the parameters on the machine to increase the processing rate of the low-concentration influent and to ensure that the effluent glycol produced is of a concentration that is ideal to feed Stage 2 of the processing. At Stage 2, the effluent glycol level is also crucial as the recycling contractor has a goal to produce a minimum of 50% EG. At this level and higher, the contractor is able to sell the glycol and gener- ate revenue to offset the expenses of the recycling operation. The second effluent stream produced from the MVR units is a distillate. This is the distilled water, which is continually monitored to ensure that glycol levels remain below 100-mg/L EG, and that BOD is below 300 mg/L to comply with the wastewater discharge permits. Based on the data, the MVR systems demonstrate that they are able to continually achieve distillate levels below the 100-mg/L requirement. BOD target concentrations are met through monitoring of COD concentra- tions and application of a site-specific correlation factor for COD to BOD. If the wastewater is not within the target concentration range, it can be sent through an aerator system prior to discharge. Effluent Concentrations The MVR has a concentrate and a distillate stream. However, the concentrate from the MVR is sold to vendors for EG reuse. The MVR distillate stream is sent back to the Aerotech Park Wastewater Treatment Facility for further treatment. Additionally, the distillate effluent EG con- centrations have consistently been demonstrated to be below 100 mg/L. Treatment Efficiencies Based on the data, 95% of the glycol that was fed through the MVR systems was reclaimed. The remaining balance of glycol was discharged through the effluent distillate stream to the POTW or for handling as solid waste at an off-site disposal facility. At a peak, the ratio of glycol produced from the MVR for reuse compared to the amount of glycol fed through the system reached over 97.7% during the 2010–2011 season. Based on operator feedback, this ratio is based on the data provided and is only from meter readings. Cost Before an on-site recycling facility was established at YHZ, all spent ADF was trucked to an off- site disposal facility in Debert, Nova Scotia. With the increase in volumes of ADF being applied at YHZ and the increase in volumes collected of spent ADF, trucking off-site became almost unsustainable. An average season at YHZ could generate 240 tanker trailer loads that would have to be trucked off-site. In addition, the cost was significant since each load would experience a 4-hour turnaround and unpredictable weather conditions during the winter that could halt transportation altogether. This affected the availability of on-site storage to support deicing operations. With an on-site recycling facility, fluid is transferred quickly, and manpower require- ments are reduced. The recycling contractor staff is used to conduct collection operations, recy- cling activities, and the management of wastewater discharges. The fluid is processed on-site, and adequate storage can be maintained for deicing operations. Lessons Learned for Potential Implementation of the MVR Technology at Other Airports The following factors have proven critical to the effective and efficient performance of the YHZ MVR system: 1. Variability in influent glycol concentrations and the ability to adjust the MVR systems to respond.

Airport Deicer Treatment System Summaries D-111 2. Quality of influent improved by filtration methods prior to treatment. 3. Desired effluent concentration of product produced affects influent processing rate. 4. Daily preventative maintenance is integrated into operations in order to optimize equipment performance. 5. Maintaining process variables such as temperature, flow rate, and pressures at consistent set points improves production rates. Although the distillate effluent concentrations are low, additional treatment of the low- concentration distillate is typically necessary. The MVR treatment system may commonly be installed at airports where there is an outlet for the effluent water produced such as a POTW or other type of system to treat low levels of COD and glycol. MVR units are more economical the greater the volume of ADF sprayed at the airport and, more importantly, the more glycol that can be captured at the airport for recycling. The greater the volume reclaimed, the larger the volume of product that can be sold to generate revenues to offset capital and operating expenses. Sale of the treated EG can reduce operational burdens and concerns associated with extensive trucking operations during winter weather events. Conclusions from operation of the MVR at YHZ that can be used by other airports consider- ing this technology include: 1. The MVR at YHZ requires additional treatment of the MVR distillate such as the use of a POTW or RO treatment system to be discharged to surface waters. 2. MVR heat exchangers require more maintenance and cleaning when dealing with ADF with higher concentrations of thickening agents, such as Type IV ADF. 3. The MVR technology can be effective for airports that consistently have variability in weather patterns and in influent concentrations. 4. The MVR can successfully conduct two-stage processing in an effort to efficiently remove large volumes of water in very diluted glycol concentration streams. 5. The MVR concentrators are modular, which means they can be installed in a relatively small footprint and can be adjusted to deal with varying influent concentrations or infrastructure needs. 6. If an airport generates a significant volume of spent ADF, then on-site recycling can be more cost-effective than transporting the fluid to an off-site facility. 7. Filtration systems are an integral part of the glycol recycling process with MVR technology. 8. The DAF system is a viable support technology to improve processing rates as well as other mechanical filtration methods to minimize equipment maintenance associated with heat exchanger plate fouling.

D-112 Guidance for Treatment of Airport Stormwater Containing Deicers Airport Treatment Summary No. 15 Airport: Zurich International Airport—Zurich, Switzerland (ZRH) Treatment Technologies: Passive Facultative Biological Treatment Distillation Treatment Years Operated: 2002–2012 (Currently Operational) Deicer Management System Description ZRH is Switzerland’s primary airport, transporting approximately 24 million passengers annually. The airport covers a total area of 800 ha (1,976 acres), including 250 ha (617 acres) of impervious area. Permit limits for stormwater discharges to the river Glatt require concentra- tions of less than 10-mg/L BOD5 and less than 20-mg/L dissolved organic carbon (DOC). As a result, much of the deicer-affected runoff from the airport has to be collected for treatment at the airport. In 2002, ZRH constructed a system for collecting, storing, and treating deicer-affected storm- water featuring passive, in-ground biological treatment. Testing of the system and establishment of operating parameters occurred over a 5-year period from 2002 to 2007. At present, two forms of passive facultative biological treatment technology, infiltration basins and spray irrigation-fed soil treatment, are used to treat low- and moderate-concentration fractions of runoff, respec- tively. The airport also uses distillation for treatment of high-concentration runoff to obtain a recyclable product. In the ZRH deicer management system, represented schematically in Figure 1, approximately 70% of the aircraft deicer is applied on two central deicing pads, with most of the remainder applied at the terminal aprons. Runoff containing spent deicing fluid is collected from the deic- ing pads, terminal apron, remote deicing areas, and several taxiways. All but 250 hectares of Figure 1. Zurich International Airport deicer management system.

Airport Deicer Treatment System Summaries D-113 airport surface area is currently collected. However, based on pressure from regulators, ZRH is planning to expand the area of runoff collected and treated. Currently, ZRH collects and treats 75% of the carbon contained in deicer-affected runoff. By the year 2015, this share is expected to be increased to 95%. Online TOC meters at various locations are used to measure the concentration of the col- lected runoff. The runoff is diverted to one of three locations based on concentration, as shown in Figure 1. The runoff is stored in underground reservoirs prior to treatment and disposal. The airport has 5.3 million liters of storage tanks available for runoff. Through over 5 years of pilot testing and monitoring, ZRH determined that a high degree of control of the quantities and timing of discharges to treatment are necessary to achieve the desired effluent quality. Deicer Treatment Technology Selection Considerations In the 1990s, the local Swiss canton (state) began pressuring the Zurich Airport Authority (ZAA) to meet cantonal concentration limits for discharges from ZRH to the river Glatt. To reduce impacts to the Glatt from deicing operations, a deicing task force, made up of the depart- ment of water protection and hydraulic engineering and Swissair, was created. The task force began evaluating treatment methods appropriate for reducing discharges from the airport to below the cantonal concentration limits. ZRH evaluated discharge to the local POTW (Werdholzli), in-situ soil treatment, aerated gravel beds, and aerobic membrane bioreactors as potential treatment technologies. The pas- sive biological methods included two alternatives: root (reed) bed wetland treatment and spray irrigation treatment. Pilot studies were conducted for root bed sewage, spray irrigation, aerated gravel beds, and the aerobic membrane bioreactor. All of the treatment methods demonstrated the ability to reduce concentrations from deicing operations at ZRH to below the cantonal concentration limits. However: 1. The Werdholzli POTW treatment capacity was determined to be inadequate to treat the flows or loads from ZRH. 2. A reed bed wetland-based treatment system was tested and had some success, but it was determined not to be a desirable long-term option because of the following: a. It had odor issues. b. It was an obstacle to aviation activities. c. It required too much space. d. It required too much maintenance. e. The effectiveness of treatment depended greatly on starting conditions and other factors difficult to control. f. A large initial capital investment would be required. g. Maintenance costs were high. The spray irrigation in-situ soil treatment technology tests demonstrated that the technology would require the lowest investment and lowest operational costs. Additionally, spray irrigation was selected because: 1. It is suitable for low concentrations, 2. ZRH has land available for irrigation, and 3. The ZRH climate is suitable for wintertime irrigation.

D-114 Guidance for Treatment of Airport Stormwater Containing Deicers The spray irrigation system was tested from 2002 to 2007 as described in the following and has been operational since. The testing was performed to establish the parameters for controlling influent flows to the irrigation system based on ambient conditions. ZRH also has a second passive facultative system described as infiltration basins (see Figure 2) that treat the lowest-concentration fraction of runoff (<50-mg/L TOC). Runoff for this system is supplied via piping rather than spray irrigation. In recent years, a distillation system was also added on-site to increase the concentration of runoff with greater than a 1% concentration. Deicer Treatment Technology Description The irrigation and infiltration fed treatment technologies below the ground surfaces at ZRH are classified as passive facultative biological treatment technologies in this guidebook. The sys- tems are considered passive because there is no active control over the biological treatment that occurs in the soil and media in the in-ground systems (e.g., there is no aeration, nutrient addition, or mixing). There is, however, significant active control over the timing and degree to which the treatment areas are loaded with deicer-affected stormwater. The systems are classified as facultative because, without active aeration, it is reasonably likely that the bacteria degrad- ing the deicer are a mix of bacteria types or bacteria that can function under both aerobic and anaerobic conditions. The focus of this summary is on the two biological technologies, although the basic parameters of the distillation system used to treat the bulk of the deicer loading are provided. General descriptions of the passive facultative technologies and distillation technolo- gies are provided in Fact Sheet 108 and Fact Sheet 105, respectively. The specific applications of the technologies used at ZRH are described in the following. 1. Infiltration Basins (for TOC < 50 mg/L) The lowest-concentration portion of the runoff (<50-mg carbon/L) at ZRH is treated with a passive facultative treatment technology described by the airport as infiltration basins. The basins include a vegetated 30-cm top layer of humus (degraded organic material) on top of a sand and stone gravel layer. A liner is located under the gravel layer to seal the treatment units and prevent contamination of groundwater. Treated water is collected with a perforated pipe and discharged to the river Glatt. Approximately 47% of the total surface runoff volume and 0.3% of the total spent deicer mass load are treated in the infiltration basins. 2. Spray Irrigation In-Situ Soil Treatment Technology (for TOC > 50 mg/L, <10,000 mg/L) The spray irrigation in-situ soil treatment at ZRH is a highly controlled system for manag- ing spraying of deicer-affected runoff to the soils based on ambient conditions and runoff characteristics. The irrigated areas cover approximately 21 hectares (51 acres). The irrigated areas are in the infield grass areas outside of the safety areas adjacent to the runways. The Figure 2. Surface of ZRH infiltration basin.

Airport Deicer Treatment System Summaries D-115 irrigation system (see Figure 3) requires a complex series of pumps, pipes, and approximately 700 pop-up sprinklers with heated heads to prevent the mechanism from freezing in winter. The irrigation pumps are fed from six reservoirs, which can hold a total of approximately 4,500 m3 (~1.2 million gallons). The irrigation system can process approximately 25 liters per second While irrigation could take place year round, due to relatively low storage capacity, irriga- tion is operated from October to April. The flow rates pumped through the irrigation system are controlled based on continuous measurement of several different parameters, including: • Influent and effluent TOC concentration, • Groundwater depth, • Precipitation, • Wind speed, • Air temperature, and • Soil temperature. The airport has developed ranges for these ambient conditions that are acceptable to achieve the desired effluent quality. The system has TOC (DOC) load targets on an hourly, daily, and total load basis specific to irrigation areas that are not to be exceeded. Vegetation is also monitored. The procedure for monitoring and controlling flows to the irrigation system is largely automated. The monitored area is divided into four quadrants. No irrigation can occur under the following conditions: • Rainfall of over 0.2 cm per hour. • Rainfall of over 1.5 cm per day. • Air temperature of less than –15°C. • Soil temperature of less than –2°C. • Wind greater than 6 m/s. • Groundwater less than 0.5 m below the surface. Pollutants are degraded biologically primarily in an aerobic zone in the top 20 cm (8 in.) of the soil. Based on testing, degradation is most complete at a depth of 80 cm (32 in.). Treated water from the irrigation system passes through perforated pipe drains that were originally installed at the airport for reducing the airport groundwater elevations. Therefore, unlike some other in-situ–based soil treatment systems, ZRH has the opportunity to moni- tor the treated concentrations. This monitoring led to the understanding that the loadings of deicer-affected stormwater to the soil needed to be controlled based on the factors shown previously. The monitoring and control system helps to reduce the exposure of the treatment system to stressful conditions. The treated water is discharged to the river Glatt. The typical detention time in the soils associated with the irrigation system is 7 days. Figure 3. ZRH irrigation system.

D-116 Guidance for Treatment of Airport Stormwater Containing Deicers The irrigation areas are located in the infield areas adjacent to runways and taxiways, although outside of the runway safety areas. Approximately 36% of the total surface runoff and 7% of the total spent deicer load are treated in the current irrigation system. ZRH is planning to collect an additional 77 hectares of area and treat it in a new irrigation area 25 hectares in size, which adds up to a total irrigation area of 45 hectares. 3. Distillation Distillation (see Figure 4) is used for high-concentration spent deicing fluid collected from two deicing pads. The high-concentration portion of the collected runoff is processed with an on-site distillation treatment plan paid for and run by a deicing chemical company. If the average concentration is below 5% glycol, the distillation process is not economically reasonable because of electricity costs. Collected concentrations sent to the distillation system range from 5% to 10%. The distillation process produces a concentrated and a diluted stream. The concentrated stream from the distillation process contains an average of 60% glycol and is transported off-site by the operator for reuse. The diluted stream from this process is mixed back into the runoff storage system for treatment by the irrigation system. Approximately 5% of the total surface runoff volume and 37% of the total spent deicing fluid load is treated in the distillation system. Key Sizing and Capacity Parameters Table 1 shows the sizing basis for the irrigation- and infiltration-based biological treatment systems. Figure 4. ZRH distillation system. yticapaC/eziS latoT retemaraP/tnenopmoC snollag noillim 523.1 yticapac egarots retawmrotS )serca 05( ah 12 tnirptoof metsys noitagirri tnerruC Planned expansion to irrigation system footprint 25 ha (60 acres) ah 7.2 tnirptoof snisab noitartlifnI ah 9.1 nisab noitartlifni ot noisnapxe dennalP Table 1. Key system sizing parameters.

Airport Deicer Treatment System Summaries D-117 Treatment System Performance The data in Table 2 reflect the intended design performance of the ZRH in-situ soil/irrigation treatment system. Table 3 shows actual values. Cost Assessment for the ZRH Irrigation Treatment System Table 4 shows treatment system costs. tinU eulaV retemaraP Design flow rates - Minimum - Average - Maximum 0 127* 1,500 Gallons per minute Design treatment load capacity 8,900 ** 5,200*** 5,200**** 2,500 lbs COD/day lbs BOD5/day lbs PG/day lbs TOC/day Design influent concentration - Range 178~35,500 100~21,000*** 100~21,000**** 50~10,000 mg COD/L mg BOD5/L mg PG/L mg TOC/L Design effluent concentration (average) <67***** <20 mg COD/L mg TOC/L Design treatment efficiency (average) Not provided % influent COD load treated *“Facts–Sheet Spray Irrigation System,” Unique. **[COD] = 3.55 [TOC] (theoretical PG stochiometric correlation for COD to TOC). ***Data based on conversion: [COD] = 1.7 [BOD5]. ****Data based on conversion: [PG] = [BOD5]. *****Assumed effluent concentration correlation [COD] = 7.25 + 2.99 [TOC] [Dubber, D. and Gray, N. (2010). “Replacement of chemical oxygen demand (COD) with total organic carbon (TOC) for monitoring wastewater treatment performance to minimize disposal of toxic analytical waste.” Journal of Environmental Science and Health, Part A, 45, 1595–1600]. Table 2. Design basis for system performance. tinU eulaV retemaraP Flow rates - Average 396 Gallons per minute Actual COD treatment load rate - Average - Maximum 1,280 4,460 lbs/day Effluent COD concentration - Average 20~32 mg/L Effluent TOC concentration - Average 4.4~8.3 mg/L Treatment efficiency 98.7% % influent COD load treated Table 3. Actual irrigation system performance. eulaV yrogetaC tsoC Capital cost* - Existing system - Planned expansion of irrigation system 25M CHF (Swiss Francs, approximately $31M in 2012) 35M CHF (~$43M) in 2012 Annual operating cost 1M CHF ($3.1M) in 2011 *Capital costs for the existing system are for the irrigation and infiltration basins’ treatment units, plus the costs for the collection, piping, storage, and distribution systems for supplying the stormwater to the treatment units. Table 4. Costs for the treatment system.

D-118 Guidance for Treatment of Airport Stormwater Containing Deicers Conclusions on Performance of ZRH Irrigation System Influent Deicer Concentrations Like many biological treatment systems, while influent TOC concentrations are measured, the flow into the treatment systems is essentially controlled based on TOC mass loading rate rather than concentrations. If collected concentrations are high, the flow rate to the treatment areas is reduced. Treated Load Rate Published data from 2005 indicate carbon inputs into the soil of 4,622-kg carbon (2001–2002) and 17,120-kg carbon (2003–2004). During the same periods, irrigation areas were 3.5 ha and 16.7 ha, respectively. This yields loading rates of 132-g carbon/m2 and 102-g carbon/m2. Effluent Concentrations The most extensive testing of the irrigation system was performed between 2002 and 2007. During 2003–2004 deicing season, only five out of 834 samples exceeded the effluent limit of 20-mg/L DOC. The average DOC concentration of the treated water in the irrigation system in that season ranged from 4.4 mg/L to 8.3 mg/L, which is 1 mg/L to 3 mg/L above the natural DOC level. Treatment Efficiencies Treatment in the irrigation-fed system occurs as the infiltrating water passes through bacteria located primarily in the top 80 cm (32 in.) of the soil. Overall, approximately 98% of the applied organic carbon is removed by the bacteria, with 90% of the removal in the top 20 cm (8 in.). Cost Although the ZRH irrigation and infiltration basin systems are considered passive from a treatment standpoint, the capital cost for the entire system is high because of the extensive amount of monitoring, storage, pumping, and piping that is needed. ZRH had one advantage in cost that not all airports will have: a ready-made pipe drainage system in the soil of the irrigated areas that was installed originally to drain groundwater. The operation of the system incurs costs for monitoring, power, and operations. Lessons Learned for Potential Implementation of the Irrigation Passive Biological Treatment Technology at Other Airports The following factors have proven critical to effective and efficient performance in the ZRH irrigation: 1. Measurements at ZRH suggest that most treatment (90%) occurs in the first 20 cm (8 in.) of soil. The upper layers also are typically composed of more natural organics from plant deg- radation that may be supplying nutrient-rich soils for the bacteria. 2. The degree of saturation of the soils with water from precipitation or groundwater is impor- tant. A saturated top layer is not conducive to treatment. 3. Wind speed is a factor in determining the feasibility in using the irrigation system at any given time because of a desire to avoid irrigation on roads and taxiways. 4. Water, soil, and air temperature are all factors in performance, and ZRH has determined the ranges in which effective performance can be achieved. System input is affected by the temperatures. Conclusions from operation of the irrigation-fed soil treatment technology at ZRH that can be used by other airports considering this technology include:

Airport Deicer Treatment System Summaries D-119 1. While there is no active control of the treatment elements such as oxygen supply and nutri- ent addition that is seen with other biological treatment systems, ZRH employs an extensive effort to control the timing of when the systems are fed with deicer-affected water, the mass loading rates, and the flow rates. The information used to control the influent flows is based on ongoing monitoring of ambient conditions, including real-time monitoring of multiple parameters. Therefore, while the treatment portion of the technology is passive, it would not meet performance criteria without a high active control of the loading of deicer-affected stormwater into the treatment areas. 2. ZRH spent 5 years performing extensive monitoring of the system performance and condi- tions that might affect performance. This resulted in the control system for the treatment system operation being based on field-collected data. Because of this extensive testing period, ZRH has developed a high degree of predictability for the treatment system performance. 3. Based on the measurements taken at ZRH, the passive biological treatment technologies are well suited for the higher-volume, lower-concentration fractions of the collected deicer-affected stormwater. At ZRH, a high percentage of runoff volume, but a relatively low percentage of the total spent deicer load, is treated in the passive biological treatment systems. 4. The hydraulic conductivity of silty soils would make getting sufficient detention time to get acceptable treatment more of a challenge. 5. Many passive biological treatment systems that use soil or media for treatment frequently have limited monitoring of influent pollutant concentrations and no means of measuring effluent concentrations. ZRH demonstrated that influent and effluent measurements are critical to achieving the desired treatment effectiveness. Documents and Information Review in Development of Assessment 1. Jungo, E, Schob, P., Disposal of De-Icing Effluents by Irrigation. 2. Jungo, E., Schob, P (2006), Disposal of Zurich Airport’s De-Icing Effluent by Irrigation, Water21. 3. Unique (2004), Treatment of De-Icing Sewage. 4. Unique (2005), Facts–Sheet Spray Irrigation System 5. Zurich Airport Annual Report (2011), Environmental Protection–Water and Wastewater.

Abbreviations and acronyms used without definitions in TRB publications: A4A Airlines for America AAAE American Association of Airport Executives AASHO American Association of State Highway Officials AASHTO American Association of State Highway and Transportation Officials ACI–NA Airports Council International–North America ACRP Airport Cooperative Research Program ADA Americans with Disabilities Act APTA American Public Transportation Association ASCE American Society of Civil Engineers ASME American Society of Mechanical Engineers ASTM American Society for Testing and Materials ATA American Trucking Associations CTAA Community Transportation Association of America CTBSSP Commercial Truck and Bus Safety Synthesis Program DHS Department of Homeland Security DOE Department of Energy EPA Environmental Protection Agency FAA Federal Aviation Administration FHWA Federal Highway Administration FMCSA Federal Motor Carrier Safety Administration FRA Federal Railroad Administration FTA Federal Transit Administration HMCRP Hazardous Materials Cooperative Research Program IEEE Institute of Electrical and Electronics Engineers ISTEA Intermodal Surface Transportation Efficiency Act of 1991 ITE Institute of Transportation Engineers MAP-21 Moving Ahead for Progress in the 21st Century Act (2012) NASA National Aeronautics and Space Administration NASAO National Association of State Aviation Officials NCFRP National Cooperative Freight Research Program NCHRP National Cooperative Highway Research Program NHTSA National Highway Traffic Safety Administration NTSB National Transportation Safety Board PHMSA Pipeline and Hazardous Materials Safety Administration RITA Research and Innovative Technology Administration SAE Society of Automotive Engineers SAFETEA-LU Safe, Accountable, Flexible, Efficient Transportation Equity Act: A Legacy for Users (2005) TCRP Transit Cooperative Research Program TEA-21 Transportation Equity Act for the 21st Century (1998) TRB Transportation Research Board TSA Transportation Security Administration U.S.DOT United States Department of Transportation