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.
21 This section presents five cases that illustrate the applica- tion of the guidance presented in Section 3 to different deic- ing management program situations requiring definition of a winter design storm event. 4.1 Baltimore/Washington International Thurgood Marshall Airport (BWI/Marshall) The following describes design criteria used for all of the deicing projects at BWI/Marshall. The latest of these proj- ects is to improve the collection system and pumping, add storage to meet peak needs, and provide redundant pumping capability. 4.1.1 Airport Overview/Project Definition BWI/Marshall is located approximately 10 miles south of Baltimore, MD. Owned, operated, and managed by the Maryland Department of Transportationâs Maryland Avia- tion Administration (MAA), BWI/Marshall is a medium- sized hub airport with approximately 22 million annual enplaned passengers and 276,457 annual operations (ACI- North America, 2010). January is the coldest month in Baltimore, with an aver- age temperature of 36.8°F. Snowfall occurs occasionally in the winter, with an annual average of 20.8 in. Freezing rain and sleet occur a few times each winter as warm air over- rides cold air at the low to middle levels of the atmosphere. When the wind blows from the east, cold air gets dammed against the mountains to the west, and the result is freezing rain or sleet. The average date of first frost in Baltimore is October 29, and the average last frost is April 11. The Bal- timore region experiences winter weather conditions that could require deicing runoff collection from November 1 through April 30. Project Definition MAA has worked closely with Maryland Department of the Environment and the EPA to develop a phased plan to reduce the amount of deicing runoff to the receiving waters surrounding BWI/Marshall. A consent order and decree was issued to MAA requiring a phased approach to provide pro- gressive improvement in water quality. One of the require- ments of the phased approach is that MAA must install deicing collection provisions at every newly constructed gate to capture gate deicing runoff. The projects included in this case study represent two separate phases of improvements. The most recent project included the rehabilitation of the terminal aprons at Piers C, D, and E with provisions to add eight additional deicing gates (four additional diversion vaults) at Piers C and D. In addition, infrastructure was also installed for three new diversion vaults at Piers D and E for future gate deicing capability and a future redundant lift station and bypass piping to provide backup to the Runway 15R lift station. The existing collection, pumping, and storage systems are undersized to handle the additional flow generated by these additional terminal deicing gates when they are brought online, so additional infrastructure is needed to handle the additional volume of collected fluid. The second project reviewed as part of this case study was the original construction of the Runway 15R and Runway 15L- 33R deicing facilities and the implementation of GRVs around the terminal apron areas. These were the original phases of the deicing fluid collection system, and they provide a significant period of data for the performance of the system with respect to the original winter storm design criteria that have been used for all phases of the development of the program. Existing (Pre-Project) System Components Deicing operations at BWI/Marshall typically occur on dedicated aircraft deicing pads, and gate deicing typically S e c t i o n 4 Case Studies
22 occurs at Piers A, B, and C. Gate defrosting may also occur at any gate. Trench drains around each gate deicing position collect deicing runoff, which then flows to one of 34 diver- sion vaults on the airside. The diversion vaults then discharge deicer-laden runoff to the deicing fluid collection piping and clean runoff to the storm sewer system. The collected deicing runoff is pumped through force mains to aboveground stor- age tanks with a total capacity of approximately 1.8 million gallons. Deicing runoff is discharged from the tanks to a sani- tary sewer for treatment at Baltimoreâs Patapsco Wastewater Treatment Plant. The permit for this discharge limits daily discharges to 10,000 gallons or 10,000 lbs of 5-day biochemi- cal oxygen demand (BOD5), whichever is most constraining. Performance of Existing (Pre-Project) Facilities The performance of the existing, pre-project facilities at BWI/Marshall was adequate for the number of deicing positions that were routinely used. The collection system, diversion vaults, force mains, and storage tanks appeared to provide adequate capacity for meeting the per-day limit of 10,000 gallons or 10,000 lbs BOD5. A system limitation, however, was that all of the pre-project deicing runoff was pumped through a single lift station near Runway 15R, with no allowance for redundancy if the lift station were to be taken out of service. If that were to occur, aircraft deicing operations would be severely limited by an inability to con- trol deicing runoff. 4.1.2 Application of Decision Support Process Risk Factors (Consequences of Failure) The primary risk factors for MAA at BWI/Marshall are non- compliance with the NPDES permit thresholds and exceed- ance of the collection system, pumping, and storage capacities (volume). The pre-project disposal conditions allowed for up to 10,000 gallons of glycol-water mixture or 10,000 lbs BOD5 to be discharged to the sanitary sewer system per day. The allowable discharge volume depends on the combined fluidâs glycol concentration and could be less than 10,000 gal- lons owing to BOD5 levels. With the addition of a new central deicing pad and more deicing gates, it was necessary to expand the existing collection system so that the system volume would not be exceeded during winter storm deicing opera- tions, when up to 30 aircraft could be departing per hour. Target Level of Service The target level of service at BWI/Marshall is to not exceed NPDES permit discharge limits and be capable of providing an aircraft flow rate of 30 departures per hour during a winter storm event with a snowfall rate of 1 in. per hour. Performance Goals of New Facilities The performance goals for the BWI/Marshall deicing runoff collection system, diversion vaults, pumping system, and stor- age tanks are to expand the deicing collection system to be able to accommodate additional gates and a centralized deicing pad for Runway 15R, thereby systematically reducing the impact of deicing runoff to receiving waters surrounding the airport. The goals further include designing a new system sized to accom- modate the new gate deicing and remote deicing locations and to provide redundancy for the single pump station. The performance goals include the need to contain runoff from a storm of memory to allow disposal at the predetermined rate of 10,000 gallons per day. This requires construction of deic- ing infrastructure adequate to divert and pump the additional runoff to the storage tanks so that the deicing runoff mixture to the local treatment plant can be metered. Assessing Existing Performance Relative to Target Levels The performance of the existing system is adequate with respect to the existing, pre-project deicing facilities; however, the system capacity will be inadequate to accommodate the runoff from additional deicing positions without additional diversion vaults, force mains, and storage tanks. Design Storm Features Critical to Performance Target The design storm feature critical to target performance is either the volume or the rate of winter precipitation. All components of the deicing runoff collection system are based on either a design storm of record or a defined hourly pre- cipitation rate. The collection piping and force mains were designed to accommodate the water equivalent of 1 in. per hour of snowfall. This snowfall rate was selected because it coincides with the rate at which the airfield snow removal equipment and air carrier ground support equipment can feasibly clear the airfield and maintain departure rates near 30 departures per hour. When the snowfall rate exceeds 1 in. per hour, the airlines typically go into a delay mode or begin canceling flights altogether, so this becomes the theoretical maximum rate at which deicing activities will occur. Exceed- ing these design storm assumptions could result in flooding or exceeding the storage capacity of the tanks, resulting in potential discharges to the storm sewer system and then to surface waters; however, the capacity of the collection pipes and pumps has never been exceeded.
23 Defining Duration of Design Condition (Multiday Storm or Isolated Event) The design storm for this project is actually different for different deicing collection system components. For diver- sion structures, pumps, and force mains, the design basis is an isolated storm event with a 1-in.-per-hour snowfall rate. This design element also includes a percentage reduction in assumed flows, which is applied because not all of the deicing locations will be contributing to the collection and pumping system at the same time. The airlines do not often deice at the gates, and deicing location is somewhat dependent on the selected departure runway. For overall deicing runoff storage volume requirements, the design condition takes into consideration cumulative effects from multiday or back-to-back events and is based on weather records and a storm of memory, although the record documents provided do not specify what design storm was actually used for that purpose. This storage design compo- nent is based on providing enough storage so that the daily rate of disposal (10,000 gallons) to the local treatment plant is not exceeded. The total storage volume capacity was derived over a number of years whereby MAA evaluated the system performance at the end of the deicing season to determine if storage capacity had been adequate. Additional tanks were added as new deicing gates or the Runway 15R deicing facil- ity came online or as MAA determined were necessary from evaluating the prior seasonâs performance. Frequency Analysis Are long-term records available? Long-term weather records are available for BWI/Marshall and were used by MAA in their initial storage tank sizing analysis; however, the storm used in the evaluation is not identified in the literature pro- vided for evaluation. The use of long-term weather records has been discontinued because MAA evaluates storage volume on the basis of the previous seasonâs actual performance. Is a robust risk analysis required? A robust risk analysis was not required for BWI/Marshall because the design solu- tion is dependent on runoff volume and only indirectly on load. Using this approach, MAA used the weather records to select a design storm of record for sizing the initial storage tank capacity but has since abandoned that approach and has based its storage requirements on actual system performance. Summary of the Decision Support Process Results The decision support process was applied to MAAâs collec- tion system improvements and is shown in Figure 4-1. The path taken is mapped by the arrows, and the decisions made are further explained in the following. Duration of design condition: Two different require- ments were identified corresponding to the different system components being designed: (1) multiday design storm for tank sizing, and (2) isolated design event (water equivalent of 1 in. of snowfall) for conveyance piping and pump sizing. These resulted in two different paths: Path Aâ(Storage Tank Sizing) ⢠Are long-term data records available? Decision = Yes ⢠Is a robust risk analysis needed? Decision = No. After initial sizing of storage tanks, BWI/Marshall changed its approach to evaluating storage adequacy at the end of each deicing season and has incrementally added volume as needed. Path A outcome: Storage tanks designed based on event or storm of memory. Path Bâ(Pump/Pipe Sizing) ⢠Is the design issue load or volume? Decision = Volume. At BWI/Marshall, the pipes/pump needed to be sized for the water equivalent of 1 in. of snowfall per hour, the maxi- mum rate of precipitation before the airfield and airlines start delaying/canceling flights. Path B outcome: Size pumps/piping in accordance with FAA guidelines for a maximum snowfall rate of 1 in. per hour. While not all of the suggested features of the process were followed in this example, the decision support process repre- sents BWI/Marshallâs identification of a winter design event based on the frequency of delivering the target level of ser- vice. Based on the selected storm and the addition of storage tanks as new facilities came online, MAA is able to provide adequate storage to allow a disposal rate of 10,000 gallons of fluid per day to the local treatment plant. 4.2 Port Columbus International Airport (CMH) This section describes a case study of the application of the decision support process to the design of a retrofit of drain- age, collection, and containment infrastructure at CMH to achieve necessary control of deicing runoff. 4.2.1 Airport Overview/Project Definition CMH is located approximately 8 miles from downtown Columbus, OH. The airport is owned and operated by the Columbus Regional Airport Authority (CRAA). CMH is a medium-sized hub airport with approximately 6.4 million annual enplaned passengers and 136,081 annual operations (ACI-North America, 2010).
24 Note: RON = remain overnight. Figure 4-1. Baltimore/Washington International Thurgood Marshall Airport.
25 At CMH, aircraft are deiced primarily on the terminal apron around Concourses A, B, and C and on the east remain- overnight apron. Storm water runoff from these deicing areas drains to Outfall 6, east of Sawyer Road and just south of the intersection of Sawyer Road and Bridgeway Road, eventually flowing to Big Walnut Creek. On rare occasions, aircraft are deiced at the Runway 10R hold pad located on Taxiway C at the approach end of Runway 10R. Storm water runoff from this area drains to Turkey Run on the south side of the airport. The Columbus region experiences winter weather condi- tions that could require deicing runoff collection from Octo- ber through April. Though Columbus rarely receives snow in October or April, October marks the beginning of freez- ing temperatures that require defrosting of aircraft that have been parked overnight on the terminal apron. The heaviest deicing activities occur from December through February. Project Definition The design component evaluated in this case study is a retrofit of the drainage, collection, and containment infra- structure that was in place prior to this project. Existing (Pre-Project) System Components Prior to this project, deicing runoff drained to the drainage infrastructure in the terminal apron and discharged to Outfalls 2 and 6, from which it flowed eventually to Big Walnut Creek. The deicing collection project includes the construction of dual-trench drains behind the aircraft gate positions at the outer edge of the terminal apron. These trench drains are connected by drainage pipes to diversion structures. Nor- mal wet-weather, nondeicing flows are diverted to the storm sewer outfalls around the airport. During deicing periods, a pump station transfers deicer-laden runoff to aboveground storage tanks, where it is held prior to being transferred to and disposed of at the local treatment plant. The design also includes snowmelt-collection areas where glycol-laden snow plowed from the terminal apron is held until it melts and eventually drains to the collection system and storage tanks. Performance Limits of Existing (Pre-Project) Facilities Under pre-project conditions, runoff from aircraft deicing was discharged directly to the storm drainage system. 4.2.2 Application of the Decision Support Process Risk Factors (Consequences of Failure) The primary risk factor associated with failing to adequately control the release of deicing fluid runoff to the surrounding streams is noncompliance with the airportâs NPDES permit and associated possible fines and enforcement actions. Other risks associated with exceeding the design storm collection capacity include damaging the diversion piping, damaging the pumps, and exceeding the storage tank capacity. The consequences of exceeding the design storm capacity include potential flooding of the terminal apron and overflows of contaminated deicing runoff to streams, with potential exceedances of NPDES and industrial permit limits. Target Level of Service The target level of service for CMH is defined by the Ohio Environmental Protection Agencyâs NPDES permit and City of Columbusâs industrial permit. The industrial discharge permit limits for CMH are a BOD5 maximum of 30,000 lb/ day for discharge to the POTW, with a monthly average of 12,000 lb/day. The NPDES permit limits for CMH outfalls related to deicing discharge are a daily maximum of 640 mg/L for propylene glycol for Outfall 2 and a daily maximum of 1,300 mg/L for propylene glycol for Outfall 6, with a monthly average not to exceed 71 mg/L and 950 mg/L for Outfalls 2 and 6, respectively. Maximum monthly 5-day carbonaceous biochemical oxygen demand (CBOD5) for Outfalls 2 and 6 are 200 mg/L and 1,300 mg/L, respectively. Furthermore, the NPDES permit stipulates the following regarding design storm overflows: The Airportâs containment system is designed based on col- lecting runoff equivalent to a 10-year winter storm event. Storm water flows above these design criteria will generate a bypass event resulting in the direct discharge of storm water to Big Wal- nut Creek. Computer modeling indicates the assimilative capac- ity of Big Walnut Creek during such a storm event is greater than the potential impact of aircraft deicing operations. Bypasses of this containment system according to the conditions of an approved Permit-to-Install are not violations of this permit as long as the maximum concentration limits for ammonia and glycol parameters are attained. (Ohio Environmental Protection Agency, 2011) Ammonia levels in Big Walnut Creek are no longer an issue since the airport has discontinued urea usage for pave- ment deicing. Performance Goals of New Facilities The goal behind installing the new deicing collection, stor- age, and disposal facilities at CMH was to dispose of glycol- laden runoff so that the NPDES permit discharge limits are achieved with only limited (and permitted) exceedances for storms that exceed the 10-year design storm, providing that maximum concentration limits for glycol parameters are not exceeded. The original objective of the design process was to
26 develop a collection and storage system with a deicer-runoff disposal rate to ensure that deicing tanks are empty after 24 hours of metered disposal to the local POTW or within 7 days for back-to-back events. Assessing Existing Performance Relative to Target Levels Pre-project deicing procedures did not include any pro- visions for collecting and disposing of deicer-laden runoff; therefore, the pre-project system did not meet the target levels. 4.2.3 Apply Decision Support Process Design Storm Features Critical to Performance Target Design storm features, including a combination of the total runoff generated by precipitation and deicing fluids sprayed, are critical to achieving the target of not exceeding the runoff equivalent to a 10-year winter design storm. Exceeding the 10-year storm could result in overflows to the storm water system with the potential of exceeding the maximum glycol parameters identified in the NPDES permit. Defining Duration of Design Condition (Multiday or Isolated Event) CRAA used two different approaches to defining design conditions relative to weather. The drainage infrastructure was sized on the basis of a single-day event, whereas a statis- tical analysis of weather records was used to size the storage tanks for deicing runoff. Storm drainage design at commercial airports is required to comply with FAA guidelines, which stipulate that surface drainage and collections systems must be designed for at least the 5-year design storm and a 10-year storm if surface pond- ing could affect aircraft traffic areas (FAA, 2006). The design storm for drainage system sizing at CMH was not specifically mentioned in the Deicing Study Final Report (Camp Dresser & McKee, 1998); however, it is specified in the July 2011 NPDES permit as a 10-year design storm. The deicing runoff collection, conveyance, and storage system was designed for a peak event, and it was anticipated (and permitted) to bypass the collection system when the precipitation event exceeded the design storm runoff volumes. The assumption was made that if flows were great enough to bypass the collection system, then glycol and ammonia concentrations in the runoff would be sufficiently low during such an event and would not exceed permit limits. Thus, the drainage infrastructure por- tion of this project followed the isolated design event path through the decision process. The need for an analysis of long-term weather records was identified to define the critical multiday storm that would be used for sizing the storage tanks. The airport used a model to determine the maximum required containment volume. Availability of Long-Term Data Records Forty-seven years of precipitation data for the months October through April (i.e., the winter period) were ana- lyzed for the purpose of characterizing a winter design storm. The average precipitation during the period from October through April is 19.4 in. Most deicing activities occur below 37°F, and the weather records analysis was limited to those days in the record where temperatures were below this level. Is a Robust Risk Analysis Needed? A robust risk analysis was required for the CMH project in order to size the deicing-runoff storage tanks. The tank sizing was based on the rate of discharge to the local POTW, so a robust analysis was required to determine the maximum daily storage volume. 4.2.4 Frequency Analysis Develop Time Series Water Budget from Historical Records Statistical programs were used to characterize storms that occurred during the winter period for the 47 years of avail- able data. The average storm duration for each month of the winter period was found to range between 5.5 and 12.8 hours, although occasional longer-duration events have occurred. The study showed that the greatest deicer usage coincided with the smallest precipitation events. However, occasional high-volume, high-intensity storms mandated that the storm water collection system be sized to handle large volumes and peak flows to minimize glycol-contaminated runoff during severe storm conditions. The water budget was therefore developed using the 10-year design storm frequency for siz- ing the drainage infrastructure. Develop Time Series Deicer Usage for Historical Weather Conditions Aircraft and airfield deicer usage has been tracked at CMH since 1992. However, at the time of the 1998 deicing study, only 7 yearsâ worth of data was available. More recently, the airport hired a consultant to assess the potential effect of a proposed deicing treatment system on the existing storage system. The consultant used a storm water model to simulate deicer system performance based on weather data from the
27 historical period of record. The 2006 flight schedule was used as the basis for simulating the runoff collected for 57 sea- sons of complete seasonal historical weather data (Gresham, Smith and Partners, 2009). Distribute Time Series Deicer Loads Among Fate Compartments It was assumed that 65% of the applied aircraft deicers would be captured in the storm sewer system and that this percentage would increase to 95% if the airport added GRVs for surface collection (optimistic, but conservative). The fate of the remaining fraction of applied deicers was not relevant to the design analysis. The basis for these assumptions was not described in any of the documents reviewed. Develop Time Series Water and Deicer Load Budget for Historical Weather Conditions The combined water and deicer load budget was used to support application of the STORM model, which allows for fate variables such as precipitation, fugitive glycol, applied glycol, and collected runoff. The generation, storage, and dis- charge of collected deicing runoff across the time series was modeled, and this provided the basis for sizing the storage. In simple terms, the model provided an estimate of daily storage volumes over the simulation period. These daily estimates were then analyzed to identify a storage volume that met the performance and target level of service criteria. The modeling analysis concluded that roughly 8 million gal- lons of aboveground storage tank capacity would be required to match the planned waste disposal rate. The design included snowmelt-collection areas that drain to the collection system and storage tanks. Evaluation of Results The goal of the deicing collection, storage, and disposal system is to maintain compliance with the industrial and NPDES permits without exceeding its total storage capacity. With the construction of 8 million gallons of storage with an additional 500,000 gallons in storage capacity in the pump wet well, the airport does not need to empty the tanks within 24 hours or 7 days for back-to-back events. To date, this storage volume has never been exceeded. One winter storm in recent years that started as ice and turned to rain nearly filled the storage tanks to capacity and depleted the airlinesâ supply of glycol; however, the tanks did not overtop and the system did not overflow. To date, the glycol collection, containment, and disposal system has adequately served its purpose and has successfully met the industrial and NPDES permit limits. Summary of the Decision Support Process Results The process followed by CRAA mapped closely to the deci- sion support tool. Were all features in the suggested process covered by the airport? The decision support process was applied to CRAAâs collection system improvements and is shown in Figure 4-2. The path taken is mapped by the arrows, and the decisions made are further explained in the following. Duration of design condition: Two different require- ments were identified corresponding to the different system components being designed: (1) multiday design storm for tank sizing, and (2) isolated design event for conveyance pip- ing and pump sizing. These resulted in two different paths. Path Aâ(Storage Tank Sizing) ⢠Are long-term data records available? Decision = Yes. Available data included 47 years of weather data, includ- ing hourly precipitation. ⢠Is a robust risk analysis needed? Decision = Yes. The tank sizing was based on the rate of discharge to the local POTW, so a robust analysis was required to determine the maximum daily storage volume requirement. ⢠Is deicer load or concentration an issue? Decision = Yes. â Develop time series deicer usage for historical weather conditions. CMH assumed captured deicing fluid vol- ume = 65% of applied fluid, or 95% when combined with GRV surface collection measures. â Distribute time series deicer loads among fate compart- ments. CMHâs consultant used storm-water modeling software that accounted for these variables. â Develop time series water and deicer load budget for historical weather recordsâthe water and deicer bud- gets were combined. â Analysis of exceedance frequency of deicer loads or concentrations. CMH looked at exceedance frequency of runoff/storage volumes. Path A outcome: Storage tanks designed based on mul- tiday storm based on frequency of delivery of target level of service (30,000 lb/day BOD5 maximum discharge to POTW). Path Bâ(Collection Infrastructure Sizing) ⢠Is the design issue load or volume? Decision = Volume. At CMH, the collection infrastructure needed to be sized for the ramp areas where deicing occurs in accordance with FAA AC 150/5320-5C. Path B outcome: Size collection infrastructure in accor- dance with FAA guidelines. As shown in Figure 4-2, the CMH project followed the process very closely. This project actually followed two paths
28 for determining the appropriate design storm, depending on the component of the deicing system being designed: (1) the multiday storm path was used for the deicing runoff storage tank design in order to capture multiday or back-to-back storm events, and (2) the isolated design event path was used to size the deicing runoff collection infrastructure. The use of the STORM model mirrored the development of a water and deicer load budget and how it was distributed to the various fate compartments. The result of the process was that a 10-year design storm was adequate for sizing the drainage infrastruc- ture and conveyance to deliver the performance and service level required, while a multiday storm based on 47 years of historical weather records was suitable for sizing deicing runoff storage tanks suitable for storing fluid at a rate that complied with the requirements of the airportâs NPDES permit. Did the airport include any features or considerations not covered in the suggested process? No. Figure 4-2. Port Columbus Regional Airport (CMH).
29 4.3.1 Airport Overview/Project Definition IAD is operated by the Metropolitan Washington Airports Authority (MWAA). IAD is situated on 11,830 acres in Chan- tilly, VA, approximately 26 miles from Washington, D.C. The IAD main terminal began operating in November 1962. IAD is a major hub for domestic and international air travel, serving 23.7 million passengers annually (2010). Figure 4-2. (Continued). 4.3 Washington Dulles International Airport (IAD) This section describes a case study of the application of the decision support process to the Washington Dulles Interna- tional Airportâs Biological Treatment of De-Icing Agents and Stormwater for Runway 1L-19R and Associated Taxiways project.
30 IAD experiences average annual temperatures of 55.3°F (1981â2000). December, January, and February are the cold- est months, averaging 36.6°F, 33.2°F, and 36.2°F, respectively. IAD averages 22.0 in. of snowfall each year (1981â2000), with the majority falling in January and February (7.3 and 7.6 in., respectively). Extreme monthly snowfalls occurred in Janu- ary 1996 and in February 2010, when 30.9 and 46.1 in. were recorded, respectively (NWS, 2011a). Project Definition MWAA invested in the infrastructure at IAD through a major construction program called D2, Dulles Development, which included construction of a fourth, 9,400-ft runway, referred to as Runway 1L-19R. A storm water management plan that considered environmental resources, storm water regulatory requirements, and long-term airport operational needs was developed for the new runway and associated taxi- ways, service roads, and future deicing pad. In order to meet Virginia Department of Environmental Quality water quality requirements, low-impact develop- ment (LID) concepts were incorporated into the storm water system plan. The principles of LID require that runoff be minimized by promoting infiltration and treatment on site or as near to the source as possible. This approach is increas- ingly advocated by regulators and reflected in changes in pro- fessional practice. In fact, the focus of the NPDES permit for IAD at the time of the design was on best management prac- tices (BMPs), monitoring, and communication with Fairfax County Water Authority. (Fairfax County Water Authority manages drinking water in the area.) The permit provided MWAA with the opportunity to lead by example, apply inno- vative treatment techniques, and exhibit a cooperative good- neighbor approach to storm water management. MWAA also wanted to be proactive in terms of storm water management on this project because it recognized that the next several years were going to bring unknown changes in the regula- tory environment. The airport was encouraged to implement biological treatment units (BTUs), as preferred by state and federal regulatory agencies, in order to comply with the per- mitâs BMP requirement (CH2M HILL, 2002). The storm water system plan featured BTUs, which are intended to treat storm water runoff not contained at cen- tralized deicing pads. The BTUs detain storm water for treat- ment through passive biological systems, which remove BOD and total phosphorus (TP) from fugitive deicing material. A total of five BTUs and their associated water quality swales and detention systems were planned; an additional five water quality swales were planned in areas not served by BTUs. The number of BTUs was driven by existing topography, stream channel locations, and the proposed airfield geometry. Treated storm water discharges to the nearest existing stream, usually Stallion Branch or its tributaries. Existing (Pre-Project) System Components There were no preexisting system components. Performance of Existing (Pre-Project) Facilities No system components existed for performance evaluation. 4.3.2 Application of the Decision Support Process Risk Factors (Consequences of Failure) The primary risk factor for MWAA at the time was non- compliance with its NPDES permit. Because the permit did not contain direct water quality effluent limitations, risk took the form of potential impacts to the airportâs relationship with local, state, and federal regulatory agencies and with the pub- lic rather than typical regulatory action, fines, and so forth, as one would expect with a water quality violation. Target Level of Service The NPDES permit was based on BMPs. The airport chose to implement BTUs in order to comply with the permitâs BMP requirement (CH2M HILL, 2002). The targeted level of service was for the system to meet performance goals in the average rainfall year. The selection of the average rainfall year was driven largely by the BTUsâ anticipated performance. A review of local dry-year and wet-year conditions concluded that BTUs designed for wet-year conditions would likely perform poorly during dry-year conditions. However, BTUs designed for average-year conditions were expected to per- form adequately during both dry and wet years. Performance Goals As mentioned earlier, performance limits were not included in the existing NPDES permit. However, it was anticipated that effluent limits might be specified during the next (2008) permitting cycle. The design team reviewed literature and case studies to determine appropriate performance goals for the selected treatment method, and it was decided to limit BOD concentration in storm water discharges to 100 mg/L or less. In the context of assumed storm water runoff water quality characteristics, this is equivalent to a concentration reduction of 91% for inflow concentrations of 1,140 mg/L and 80% for inflow concentrations of 500 mg/L. In addition, it was determined that the treatment system must reduce
31 total phosphorus concentrations by 50%. (The assumed water quality values were based on BOD concentrations mea- sured in storm water samples collected from various loca- tions at the airport.) Assessing Existing Performance Relative to Target Goals The proposed storm water management plan was for new development; no system components existed for performance evaluation. Design Storm Features Critical to Performance Target The primary design features critical to the performance target are the storm volume and the expected water quality characteristics, specifically BOD and TP concentrations of storm water runoff in the collection area. Define Duration of Design Condition (Multiday Storm or Isolated Event) The selected BTU treatment method requires that the col- lection and treatment capacity be sufficient for a variety of potential storm events. As a result, the design requires a con- tinuous simulation (multiday) approach. Availability of Long-Term Data Records Hourly meteorological data from Ronald Reagan Wash- ington National Airport (DCA) are collected and reported by the National Climate Data Center (NCDC). More than 50 years of detailed precipitation data were available. These data were compared to a shorter data set (5 years) collected at IAD. Appropriate adjustment factors were applied to gener- ate a long-term synthetic precipitation record for IAD. The synthetic weather data set was reviewed by the design team, and representative dry, average, and wet precipitation years were identified. Need for Robust Risk Analysis Statistical analysis of historical weather data was required because the design is based on the long-term average annual storm volume. This analysis qualifies as a form of frequency analysis, the frequency in this case being an average. Frequency Analysis The frequency analysis approach to identifying a design event differed in detail from that shown in the process flow chart, but the approach generally followed the process, as outlined in the following steps: Develop Time Series Water Budget 1. The precipitation data set at DCA was reviewed to iden- tify dry, average, and wet rainfall years. The year 1995 was selected as average on the basis of its comparison with the long-term annual average of approximately 45.7 in. Precipitation volumes at DCA were adjusted by a factor of 15% to estimate precipitation volumes at IAD. This is based on comparisons of available data that showed annu- al totals at IAD being between 9% and 17% greater than those at DCA. 2. Continuous flow simulation using the XP-SWMM soft- ware package was conducted to estimate surface runoff and system losses, including infiltration for dry, wet, and average rainfall years. 3. Runoff volumes to the BTUs, duration of storm events, and inter-event periods were determined from the contin- uous simulation output. The average rainfall year (1995) was selected as the basis of design. The annual average storm runoff volume was calculated by dividing the vol- ume of storm water runoff entering a given BTU during the year by the total number of storm events. Establish Deicer Concentrations in Storm Water Runoff 4. BOD in runoff to BTUs was assumed to be at antici- pated maximum average concentrations of 1,140 mg/L near future Deicing Pad B and 500 mg/L for the remain- ing collection system areas, which are farther from deic- ing operations. (The assumed water quality values were based on BOD concentrations measured in storm water samples collected from various locations at IAD.) The range of inflow concentrations is consistent with concen- trations treated by other systems receiving airport storm water with deicing agents, such as the Pearson Airport treatment system in Toronto and the Westover Air Force Base system. Apply Frequency Analysis Results to the Design and Sizing of Treatment Units 5. BTU sizing was based on water quality targets and accom- plished using empirical equations (Van der Tak et al., 2005). The independent (i.e., input) variables in the sizing equa- tion consisted of the oxygen demand (OD) for nitrification, BOD, chemical oxygen demand (COD), and the oxygen transfer rate (OTR). Observed ranges of oxygen supply per unit area [as taken from literature, including Kadlec (2000), Platzer (1999), Green et al. (1997), and Cooper and Green (1998)] were combined with an estimate of oxygen demand (OD) to determine the BTU bed area as: A = ( )OD OTR area in m2
32 Oxygen demand is, in turn, a function of water quality (NH4, N, and BOD) and flow rate. Oxygen demand is estimated as follows (Cooper and Green, 1998): OD NH N BOD oxygen demand in g O day2 = â( ) + ( )4 3 4. â â ( ) For example, for Runway 1L-19R System A, an OTR of 50 g O2/m2 îº day was assumed, based upon literature values. The OD was estimated as follows: 4 3 0 1 0 221 1 140 100 . . , i ig m g m m day g m g m 3 3 3 3 3 â( ) + â( ) =i 221 229 935m day g day3 , Dividing this OD by an OTR of 50 g O2/m2 îº day yields 4,599 m2, or 1.1 acres. 6. Once the BTU design capacity was determined, the sys- tem performance was evaluated given the 24-hour, SCS Type II synthetic storms for standard return periods of 1 to 100 years. The intensities and depths for these storm events were based on data provided in the Loudoun County Design Manual. Steps 1 through 3 developed the time series water budget. Step 4 established assumed deicer concentrations without needing to go through the process of estimating deicer usage or distributions, and steps 5 and 6 involved the analysis of frequency of runoff volumes to establish the design sizing. Summary of the Decision Support Process Results The storm water management design approach for the Runway 1L-19R and associated taxiways project at IAD maps to the decision support process as shown in Figure 4-3. The path taken is mapped using the arrows, and the decisions made are further explained: Define the nature of the project: Implement a biologi- cal treatment unit system to reduce discharged BOD to 100 mg/L and TP concentrations by half in accordance with Virginia Department of Environmental Quality NPDES permit requirements. Duration of design condition: Decision = Multiday design storm. A multiday storm was chosen because the collection and biological treatment system must work for a variety of storm events in order to comply with NPDES permit limits. ⢠Are long-term data records available? Decision = Yes. Fifty-seven years of meteorological data were available for analysis. ⢠Is a robust risk analysis needed? Decision = Yes. A robust risk analysis was needed because the design is based on the long-term average annual storm volume. This analysis qualifies as a form of frequency analysis, the frequency in this case being an average. A continuous flow simulation was performed using storm-water modeling software to determine runoff volume, event duration, and inter-event periods for the dry, average, and wet rainfall years. The average rainfall year was selected for design purposes, and average storm runoff volume was calculated for each BTU. â Is deicer load or concentration an issue? Decision = Yes. BOD in runoff diverted to biological treatment units was assumed to have a concentration of 1,140 mg/L near Deicing Pad B and 500 mg/L at remote facilities. â Develop time series deicer usage for historical weather conditions. Deicer usage at IAD was assumed, not cal- culated using historic records. â Distribute time series deicer loads among fate com- partments. MWAA made assumptions regarding how much of the applied deicer would reach each BTU (for example, maximum BOD concentration of 1,140 mg/L in runoff flowing to the BTU closest to Deicing Pad B). â Develop time series water and deicer load budget for historical weather records. MWAA used a storm-water modeling software for this purpose. â Analysis of exceedance frequency of deicer loads or con- centrations: This step was not specifically conducted in this example. Path outcome: The analysis conducted by MWAA was consistent with the decision support process outlined in this guidebook. In this case, the risk analysis included a simpli- fied frequency analysis that focused on average event runoff volumes during a representative precipitation year. The out- put of this analysis was combined with literature and applied empirical equations regarding pollutant removal to size the BTUs. While not all of the suggested features of the fre- quency analysis process were followed in this example, the structured approach of the decision support process leads to a winter design event based on the delivery of the target level of service. 4.4 Pittsburgh International Airport (PIT) This section describes a case study of the application of the decision support process to the design of a deicing storm water treatment plant at PIT. 4.4.1 Airport Overview/Project Definition PIT is approximately 17 miles west of Pittsburgh. Owned, operated, and managed by the Allegheny County Airport Authority (ACAA), PIT is a medium-sized hub airport with approximately 8.2 million annual enplaned passengers and 144,563 annual operations (ACI-North America, 2010).
33 Figure 4-3. Washington Dulles International Airport. (continued on next page)
34 The coldest month of the year in Pittsburgh is January, when the 24-hour average temperature is 27.5°F (-2.5°C), and subzero lows (below -18°C) can be expected on an aver- age of 3.9 nights per year. Annual snowfall is 40.3 in. (102 cm). Although Pittsburgh generally experiences moderate weather, a few extreme weather events occurred between 1990 and 2010. The blizzard of 1993 deposited over 23 in. (58 cm) of snow in under 24 hours, and the first North Amer- ican blizzard of 2010 deposited nearly 2 ft (60 cm) of snow in less than 24 hours. The Pittsburgh region experiences winter weather con- ditions that could require deicing runoff collection from Figure 4-3. (Continued).
35 October 1 through May 31. Though Pittsburgh rarely receives snow in October or May, October marks the beginning of freezing temperatures that require defrosting of aircraft that have been parked overnight on the terminal apron. Deicing operations at PIT typically occur on one of three centralized deicing pads: Deicing Pad Charlie (âCâ), con- structed in 1992, serves aircraft departures on Runways 10L- 28R and 14; Deicing Pad Echo (âEâ), constructed in 1992, serves aircraft departures on Runways 10C-28C and 10R- 28L; and Deicing Pad Sierra (âSâ), constructed in 2001, was built primarily to serve wide-body aircraft such as the Airbus A330 and Boeing 767 but was also sized to serve as a tempo- rary backup to either Deicing Pad C or E. Deicing Pad S now serves not only as the deicing pad for wide-body aircraft but also as a deicing location for aircraft with winglets. Project Definition ACAA was issued three administrative orders and a con- sent order and decree by the Pennsylvania Department of Environmental Protection (PaDEP) to cease the âunauthor- ized discharge of industrial wastes such as aircraft deicing fluid into waters of the Commonwealth of Pennsylvania.â ACAA is addressing these orders through a comprehensive storm water management program for the airport that is designed to (1) improve water quality so that these streams experience a decrease in nuisance bacterial growth (Spaeroti- lus) and odor observations, and (2) slow peak runoff rates to some extent, thereby potentially improving the morphologic condition of the streams (Camp Dresser & McKee, 2004). It should be noted that the receiving streams, including the East Fork of Enlow Run and the West Fork of McClarenâs Run, have a very small drainage area above the deicing oper- ations areas and offer minimal assimilative capacity (6 ft3/s and 4 ft3/s, respectively). ACAA has implemented available and practicable technol- ogies as well as known BMPs, including the construction of centralized deicing pads and the use of forced-air application technologies, collection structures, storage tanks, and surface collection measures (GRVs). The full employment of these technologies and practices has not and cannot sufficiently prevent deicing/anti-icing chemicals (DACs) from reach- ing and negatively affecting the receiving waters, including through point source and nonpoint source discharges. As a result, further control measures are required to reduce DACs in PIT discharges while maintaining the safe and efficient operation of the airport. Therefore, ACAA is proceeding with the design and installation of a deicing storm water treatment plant that uses a three-stage moving-bed biological reactor (MBBR) system. The design is based on collecting 90% of the average annual surface and subsurface runoff containing DACs in quantities sufficient to contribute to water quality impairment in Enlow and McClarenâs Runs during the deic- ing season (October 1 through May 31). This capture level will prevent the capacity of the deicing runoff collection and treatment system from being exceeded more than five times during the deicing season, on average. This capture level will also satisfy the water-quality-based requirements of the Pennsylvania Clean Streams Law. Existing (Pre-Project) System Components Current aircraft deicing practices at PIT consist of limited defrosting/deicing at the gates and deicing/anti-icing at the centralized deicing pads. The airlines, deicing pad operators, and spent-deicing-fluid-recycling/disposal contractor cur- rently use a variety of BMPs to capture deicer-laden runoff, including the following: ⢠Use of GRVs around the gate areas when defrosting activi- ties occur; ⢠Use of forced-air deicer application equipment at the deic- ing pads to reduce applied quantity; ⢠Collection of deicing runoff from the centralized deicing pads via subsurface and surface collection, as well as stor- age of runoff in low- and high-concentrate tanks; and ⢠Recycling of spent deicing fluid by a recycling contractor and treatment using a reverse osmosis process. Performance of Existing (Pre-Project) Facilities Pre-project BMPs resulted in the collection of approxi- mately 60% of the applied DACs, which correlated to approximately a 55% reduction in CBOD5 load. The annual sampling and monitoring program indicates that while the BMPs have drastically reduced the amount of DAC found in the receiving waters around PIT, nuisance bacterial growth in the streams has not been eliminated, and further reductions in DAC in airport discharges are needed. 4.4.2 Application of Decision Support Process Risk Factors (Consequences of Failure) The primary risk factor for the ACAA at PIT is noncom- pliance with its NPDES permit. A draft NPDES permit was issued in August 2010 with specific discharge limits imposed by PaDEP to meet water quality standards; however, the draft permit has yet to be finalized. The discharge pipe from the treatment plant will have specific limits, including for CBOD5 a maximum instantaneous value of 20 mg/L and a monthly average of 10 mg/L. The draft permit further stipu- lates for propylene glycol a maximum instantaneous value of 3.4 mg/L and a monthly average of 1.7 mg/L.
36 Target Level of Service The target level of service has been defined as not exceeding the capacity of the collection and treatment system more than five times per year. In addition, in-stream dissolved oxygen standards must be maintained 99% of the time, as required by PA Title 25 Chapter 96.3(c), and the draft NPDES permit thresholds must be met. Performance Goals of New Facilities The draft performance goals for the PIT treatment plant will be to achieve a total 93% CBOD5 load reduction to achieve less than 10 mg/L of BOD and less than 1.7 parts per million (ppm) of propylene glycol in the receiving waters of PIT. Water quality modeling studies show that a 93% CBOD5 load reduction will require collection and treatment of approximately 75% of the affected runoff. Assessing Existing Performance Relative to Target Levels The existing performance of the system is characterized by BOD5 concentrations on the order of 25 ppm to 30 ppm, with propylene glycol concentrations approaching 2,200 ppm. These levels do not meet the performance goals and lead to the conclusion that additional measures are necessary. Design Storm Features Critical to Performance Target The design approach for the MBBR treatment plant is based on continuous hydrologic simulation, yielding a per- cent load factor reduction similar to a wet-weather total maximum daily load (TMDL) or combined sewer overflow approach. This approach defines exceedance frequencies and recurrence intervals using runoff statistics, as opposed to tra- ditional design storm hydrology (e.g., a 2-year design storm), which assumes that rainfall and runoff recurrence intervals are identical. The analysis resulted in a design storm with an approximate 3- to 4-month 24-hour recurrence interval. As previously mentioned, the percent load factor reduction is based on a treatment system design capacity that is exceeded no more than five times, on average, during the deicing sea- son. This is in contrast to a point source industrial wastewater discharge, which typically uses the design storm approach. Defining Duration of Design Condition (Multiday Storm or Isolated Event) Continuous hydrologic simulation, coupled with deicing season receiving-water quality modeling of CBOD5 loading and the resultant in-stream dissolved oxygen concentration, was used to determine that a total CBOD5 load reduction of 93% would satisfy the water-quality-based requirements of the Pennsylvania Clean Streams Law. ACAA further deter- mined through modeling of the 2000 through 2006 deicing season flows and loads that the 2003â2004 deicing season and its use of DACs represented a conservative time period to be used for establishing the design capture volume. The 2003â2004 deicing season records are characterized by one of the highest uses of DACs at PIT since the start of their record keeping in the early 1990s. Also, there were signifi- cantly more aircraft departures in 2003â2004 than currently because of PITâs status then as a hub for US Airways, so oper- ations levels are believed to be very conservative. Thus, while a robust analysis of the storm water flow component of the design storm was performed, the deicing load component of the storm water runoff is actually a storm of memory, the 2003â2004 deicing season in this case. Another defining climatological factor was minimum tem- perature with respect to deicing collection operations. The efficiency of the MBBR treatment process will decline when influent temperatures are below 37°F, and ACAA is requesting PaDEP to take this fact into consideration when developing the limits in the new NPDES permit. Frequency Analysis The frequency analysis begins with developing a time series water budget, which is a tracking of runoff over a specified duration. For PIT, a knee-of-the-curve analysis determined the point at which additional storage and treatment would require significant additional investment for incrementally small environmental gains. The analysis concluded that treat- ment to achieve a 93% load reduction in CBOD5 was the opti- mal level of storage/treatment necessary and that it would meet dissolved oxygen standards 99% of the time, satisfying the water-quality-based requirements of the Pennsylvania Clean Streams Law. Because deicer load is not a factor, no further time series analysis was required in this case. The treatment system has adequate storage capacity and sufficient runoff volume to allow the treatment process to work efficiently regardless of deicing runoff load. At this point in the decision process, an analysis of the frequency of exceedance of the runoff or stor- age volume is the only remaining step in identifying the win- ter design storm. As mentioned earlier, ACAA determined the anticipated frequency of exceedance of dissolved oxy- gen standards to be four or five times a year, based on those times when the temperature during the treatment process dips below 37°F and/or the volume of the treatment system is exceeded.
37 Summary of the Decision Support Process Results The decision support process was applied to the ACAA MBBR treatment systemâs design approach, as shown in Fig- ure 4-4. The path taken is mapped using the arrows, and the decisions made are explained in the following. Duration of Design Condition: Multiday Design Storm for Storage Volume Sizing Flow path: ⢠Are long-term data records available? Decision = Yes. His- toric meteorological data, flow, historical deicing usage, and water quality data were available for the analysis. Figure 4-4. Pittsburgh International Airport. (continued on next page)
38 Figure 4-4. (Continued).
39 ⢠Is a robust risk analysis needed? Decision = Yes. A robust risk analysis was required to determine the design treat- ment plantâs hydraulic capacity. â Develop time series water budget: A knee-of-the-curve analysis was performed to determine the point at which additional storage and treatment would require sig- nificant additional investment for incrementally small environmental gains. The analysis further determined that a 93% reduction in CBOD5 would satisfy the water- quality-based requirements of the Pennsylvania Clean Streams Law. ⢠Is deicer load or concentration an issue? Decision = No. In PITâs case, calculating the design capture area and vol- ume were sufficient to develop the treatment plant model. However, the deicer usage from the 2003â2004 deicing season was used in the storm water model to be conserva- tive. The 2003â2004 deicing season represented a season with high deicer usage and correspondingly high aircraft departures prior to PIT losing its airport hub status. â Determine design capture area and perform design cap- ture volume calculations. Separate design capture area and design capture volume reports were performed for this purpose. â Analyze exceedance frequency of runoff or storage vol- umes. The exceedance frequency at PIT was the number of times the temperature during the treatment process fell below 37°F (four or five times per season). Flow-path decision outcome: For PIT, the design solu- tion was to identify a design event or multiday storm based on frequency of delivery of the target level of service (93% reduction on CBOD5 with an exceedance frequency of 4 to 5 times per year). While not all of the suggested features of the process were followed in this example, the decision support process led to an equivalent winter design event based on the frequency of delivery of the target level of service. 4.5 Portland International Airport (PDX) This section describes a case study of the application of the decision support process to PDXâs Deicing System Enhancement Project. 4.5.1 Airport Overview/Project Definition PDX is operated by the Port of Portland and provides both domestic and international passenger and cargo services. The airport has 14 passenger airlines that serve over 14 million trav- elers each year. The airport is also served by 11 all-air cargo carriers. PDX is the 34th largest passenger airport and the 24th largest cargo airport in the United States. In addition, PDX houses the 142nd Fighter Wing of the Oregon Air National Guard. Located in Portland, OR, PDX enjoys a moderate cli- mate with average annual temperatures of 53.6°F and aver- age deicer season (October through April) temperatures of 53.7°F. Portland typically sees 151 days of precipitation each year, but only 2.1 days per year with snow, ice, or hail exceed- ing 1.0 in. (NWS, 2011b). The PDX deicer management system that existed prior to the case study project collected, stored, and discharged deicer-laden runoff from Basins 2, 4, 6, and 7 to the Colum- bia Slough and to the City of Portlandâs Columbia Boulevard Wastewater Treatment Plant (CBWWTP). Discharges into the Columbia Slough are subject to the conditions of a NPDES permit issued by the Oregon Depart- ment of Environmental Quality (DEQ). Discharges to the CBWWTP are subject to the conditions in an industrial pre- treatment discharge permit with the City of Portland. Project Definition In 2003, the port undertook modifications to the then- existing deicing system in order to improve its performance and comply with the terms of its NPDES permit. However, the installed system was found to have limitations, and more significant infrastructure enhancement was determined to be needed. In 2006, the DEQ issued a Mutual Agreement and Order (MAO) to the port to bring the deicing system into compliance with environmental requirements. The port agreed to a larger project to enhance the existing system and committed to having it fully operational by April 2012. The targeted deicing system improvements include: ⢠Collection: Expansion of collection system to capture storm water runoff containing deicing materials from the western airfield. Approximate collection area to be con- tained is 2,100 acres (S. Aha, personal communication, Feb. 24, 2011). ⢠Storage: Increase storage capacities for concentrated and diluted runoff. Approximate storage capacities are 25.8 mil- lion gallons diluted and 5 million gallons concentrated (S. Aha, personal communication, Feb. 24, 2011). ⢠Treatment: Addition of an on-site anaerobic treatment facility for treating concentrated effluent prior to discharge to the Columbia River or the sanitary sewer. ⢠Discharge: Addition of a Columbia River outfall, which was completed in January 2010. Existing (Pre-Project) System Components In 2003, PDX completed construction of a $31 million deicing system designed to monitor, collect, treat, and discharge
40 deicing storm water runoff to the Columbia Slough and to the City of Portland sanitary system. The system includes the use of GRVs that capture concentrated runoff from aircraft deic- ing and anti-icing operations. All other runoff is collected by the storm water drainage system. Within the collection sys- tem, runoff is tested for BOD concentrations to determine if it is concentrated (1,000 mg/L or more) or diluted (800 mg/L or less). Concentrated runoff is diverted to an on-site stor- age tank, the contents of which are discharged to the sanitary sewer for treatment at CBWWTP. Diluted runoff is diverted to one of several dilute detention basins and eventually dis- charged to the Columbia Slough. Performance of Existing (Pre-Project) Facilities PDX exceeded its NPDES permit limits during each winter season from 2003 to 2006 (i.e., 2003â2004, 2004â2005, and 2005â2006). In general, the system was judged to be effective at collecting deicing runoff from runways, taxiways, and termi- nal areas. However, the collection system did not collect run- off from Basin 1, which serves the western half of the airfield. In addition, flows in the receiving waters (Columbia Slough) were much lower than anticipated, and as a result the allow- able discharge loads under the NPDES permit, which vary with flow in the slough, were often very low. These conditions contributed to the failure of the deicing system to consistently perform in compliance with the NPDES permit requirements. 4.5.2 Application of the Decision Support Process Risk Factors (Consequences of Failure) In conjunction with its Deicing System Enhancement Project, PDX obtained from DEQ a revised NPDES permit, which was approved in June 2009 (Port of Portland, 2010). The primary risk factor for PDX is noncompliance with this new NPDES permit. Target Level of Service The terms of the NPDES permit require that the system be able to handle all runoff from storm events up to the 5-year recurrence level. PDX may claim an upset if the system is insufficient for a given storm event, such that runoff bypasses the collection and treatment systems. However, the frequency of recurrence of upsets should not exceed once every 5 years, on average. Performance Goals Discharges to the Columbia Slough are to be limited to dilute runoff with BOD concentrations of 20 mg/L or less. Runoff with BOD concentrations in the range of 20 mg/L to â¤200 mg/L is either diverted to the dilute detention basins or discharged to the Columbia River via the dilute storage tanks. Runoff greater than 200 mg/L is diverted to the con- centrated storage tanks. PDX has an operational goal to col- lect and store concentrated runoff with BOD concentrations of 200 mg/L or greater and to then treat the runoff on site or deliver it to the CBWWTP for off-site treatment. In addition to regulatory compliance, the objectives of the PDX Deicing System Enhancements Project are as follows (CDM, 2008a): ⢠Reduced reliance on outside agencies for compliance; ⢠Efficient operation, cost-effective implementation, mini- mized ongoing maintenance costs, and a 20-year minimum project life; ⢠Increased confidence that the system will maintain com- pliance with current regulatory requirements and flexibil- ity to respond to potential changes such as effluent limit guidelines or potential changes in the Columbia Sloughâs TMDL allocation; ⢠Allowance for PDX growth, both in terms of BOD loading and changes in airport infrastructure; ⢠Use of proven technologies that complement the existing system infrastructure; and ⢠Meeting the milestones of the DEQâPort MAO compli- ance schedule. Finally, to accomplish these goals over the long-term, PDX considered future conditions as part of their analysis. A design year of 2022 was selected to represent airfield opera- tional levels. In addition, climate change impacts were incor- porated into the analysis. Assessing Existing Performance Relative to Target Goals The existing deicing system collects and treats runoff with BOD concentrations of 1,000 mg/L or greaterâa lower limit that is five times the new concentration collection goal (200 mg/L). The performance of the existing system was not sufficient to meet water quality limits specified in the 2004 PDX NPDES permit. Design Storm Features Critical to Performance Target The design approach for the PDX system expansion is based on providing system capacity sufficient to capture and treat all runoff with BOD concentrations above 200 mg/L, with a recurrence interval for upsets of not less than 5 years.
41 Defining Duration of Design Condition (Multiday Storm or Isolated Event) The permit requirement of a 5-year recurrence interval for system upsets required a multiday storm approach. Availability of Long-Term Data Records Hourly meteorological data are collected and reported by the NCDC at PDX. Over 50 yearsâ worth of detailed meteoro- logical data are available. Available parameters include ambi- ent air temperature, dew point temperature, and precipitation depth. These data were used by PDX to derive the precipita- tion type for each winter weather event, such as frost, ice, and snow. Historical, current, and projected facility and opera- tions data are available for PDX. Historical flight schedule data from 2003 to 2007 are available from the PDX noise department, while individual airlines have provided cur- rent (2007â2008) flight schedules. Future flight schedules through 2022 are available from the Terminal Area Forecast prepared for PDX by the FAA. Data regarding the number and type of aircraft used at PDX are available as provided by the airlines. Data regarding deicing practices, such as type (I or IV) and volume, are available as a function of weather conditions and aircraft type. Similarly, information related to the existing deicing runoff control system, including GRV operations, water quality thresholds for runoff diversion and discharge, and storage capacities, are also available. Hydrologic and water quality data are available to PDX for the Columbia Slough receiving waters. Continuous flow data are collected when discharge pumps are operating, and the data are available from 2003 to present. In addi- tion to historical data, slough discharge equations and flow assumptions are provided by the port (CDM, 2007a). The city has collected water quality data in the slough, including dissolved oxygen, temperature, and TMDL pollutant levels, for more than a decade (S. Aha, personal communication, Feb. 24, 2011). Data available for PDX are summarized as follows: ⢠Over 50 yearsâ worth of hourly meteorological data â Observed temperature and precipitation â Derived precipitation type (frost, ice, snow) ⢠Historical, current, and predicted future flight schedules (2003â2022) ⢠Current and future fleet mix ⢠Current deicing practices ⢠Current deicing controls ⢠Limited hydrologic and water quality data for the receiving water (Columbia Slough and Columbia River) Need for Robust Risk Analysis A robust risk analysis was needed to evaluate system design sizing to achieve the permit requirement of a 5-year recur- rence interval for system upsets. Frequency Analysis A simulation model was used to develop detailed repre- sentations of the existing and proposed PDX deicing man- agement systems and to evaluate sizing and other design considerations in terms of risk (i.e., frequency) of upsets. This analysis consisted of a long-term, continuous simu- lation using historic, current, and projected conditions. Historic meteorological data driving the analysis were con- strained to the most recent 22 years (i.e., 1985 to 2007) in order to account for climate change effects observed in the older data. Fate compartments were characterized based on a Transport Canada study on the fate and transport of deicer after it is applied to aircraft. The PDX analysis assumed that 90% of applied Type I and 23% of applied Type IV deicers are available for capture. Summary of the Decision Support Process Results The decision support process was applied to PDXâs enhance- ment project design approach, as shown in Figure 4-5. The path taken is mapped using the arrows, and the decisions made are explained in the following. Define the nature of the project: PDX collects and dis- charges deicing runoff to the Columbia Slough. In response to repeated exceedances of NPDES permit limits, PDX began designing and implementing deicing collection and treat- ment system enhancements in 2006 to achieve a BOD of less than 200 mg/L. Duration of the design condition: Decision = Multiday storm. A multiday storm evaluation was required to deter- mine the event that would result in a system upset no more frequent than once every 5 years. ⢠Are long-term records available? Decision = Yes. Fifty years of historical meteorological data were available; however, PDX limited the analysis to the last 22 years to account for climate change. ⢠Is a robust risk analysis needed? Decision = Yes. Compli- ance with the NPDES permit required a design that would result in a minimum recurrence interval of 5 years for exceedance of permit limits. â Develop time series water budget from historical weather records. PDX limited weather time series to 22 years (1985â2007) to account for climate change effects observed in the historical record.
42 Figure 4-5. Portland International Airport. (continued on next page)
43 Figure 4-5. (Continued).
44 ⢠Is deicer load or concentration an issue? Decision = Yes. Load or concentration is an issue at PDX and is driven by the risk factor of exceeding NPDES permit limits. â Develop time series deicer usage for historical weather conditions. In its analysis, PDX used simulated hourly deicer application on the basis of hourly weather, flight schedules, and aircraft wing span. â Distribute time series deicer loads among fate compart- ments. PDX assumed that 90% of applied Type I and 23% of applied Type IV deicer are available for capture based on a Transport Canada study. Additional fate compartments were then assigned on the basis of col- lection system characteristics. â Develop time series water and deicer load budget for historical weather records. PDX considered only the most recent 22 years of historic meteorological data to adjust for the effects of climate change. PDX then con- sidered future aircraft operations with a selected design year of 2022 to estimate the deicer load budget. â Analysis of exceedance frequency of deicer loads or concentrations: PDX determined the number of upsets over the 22-year simulation and sized the system to yield a 5-year upset-recurrence frequency. After exploring a number of design storm approaches (CDM, 2007b; CDM, 2008b), PDX targeted a system- performance-level approach to guide the design of the enhanced deicing system. The analysis conducted by PDX was consistent with the decision support process outlined in this guidebook. However, rather than selecting a specific design storm, PDX defined an acceptable system performance level based on a selected recur- rence frequency of 5 years. Slight modifications were made to the decision process to allow for future conditions, including operations and climate change. While not all of the suggested features of the process were followed in this example, the deci- sion support process worked in leading to a winter design event based on the frequency of delivery of the target level of service.
