Defining and Measuring Aircraft Delay and Airport Capacity Thresholds (2014)

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.

38 Airport capacity is a measure of the maximum number of aircraft operations that can be accommodated on the air- port or by an airport component within a given period of time. The context of this report is airside or runway capacity. This chapter discusses definitions of airport capacity, some of the challenges and various threshold metrics used to define capacity, how the capacity values can be used in airport plan- ning, and guidance on the use of capacity metrics for differ- ent audiences. 3.1 How Capacity Is Calculated ACRP Report 79: Evaluating Airfield Capacity discusses how to analyze capacity including the factors that influ- ence capacity, and readers are directed to that report for specific information on airport capacity evaluations. This section will briefly summarize common approaches and the strengths and weaknesses of each approach. Capac- ity calculations range from low-fidelity estimates based on simple assumptions about the runway layout through quick spreadsheet models to complex and high-fidelity simulation results. 3.1.1 Historical/Actual Observations Airports and the FAA keep records of actual runway operations. Using the actual throughput counts of an exist- ing airfield to estimate capacity may greatly underestimate capacity. The actual throughput is certainly achievable, since it occurred in the real world. However, without any additional knowledge of the delays or saturation condition, there is no way of knowing if more operations could have occurred had more flights wanted to land/depart. So if the demand (e.g., current or historical flight schedule) is lower than the capac- ity, then the observed throughput is not a good indication of capacity. For example, as shown in Figure 3-1, the airports displayed have actual throughput much lower than their theoretical capacity. Each dot shown in the lower left of the Figure 3-1 graphs represents an actual hour’s operations con- sisting of a certain number of arrivals and departures, while the blue lines represent the capacity estimated by analytical methods. Simply estimating capacity from the actual opera- tions counts would underestimate true achievable capacity. Rather than using actual throughputs as indications of capac- ity, planners usually determine capacity through more rigor- ous approaches. 3.1.2 Basic Models The FAA’s advisory circular AC 150/5060-5, Airport Capac- ity and Delay, Change 2, can be used for long-range planning purposes for simple capacity calculations of hourly capacity and ASV. • Hourly capacity is defined as the maximum number of aircraft operations that can take place on a runway sys- tem with a specific runway use configuration in a 1-hour period. • ASV “is a reasonable estimate of an airport’s annual capac- ity. It accounts for differences in runway use, aircraft mix, weather conditions, etc., that would be encountered over a year’s time.” The FAA’s Airfield Capacity Model (ACM) has automated the methodology described in AC 150/5060-5 to analyti- cally calculate the maximum throughput of a runway sys- tem. Capacity is computed by determining the minimum time between successive arrivals and inverting this time to find the maximum number of arrivals per hour. The maxi- mum number of departures that can be inserted between the arrivals is then calculated to give the arrival-priority capac- ity. If a specific ratio of arrivals to departures is specified, the C H a P T E R 3 Metrics to Define Airport Capacity

39 departure-priority capacity is calculated. The ACM produces hourly capacity at varying arrival-departure ratios. Hence, an airport runway layout has one capacity or maximum throughput when arrivals are given priority, another capac- ity when departures are given priority, and other capacities at other mixes, such as 50% arrivals and 50% departures. AC 150/5060-5 and ACM rely on specific runway layout configu- rations and the ATC rules in place at that time. Strengths These approaches provide initial capacity estimates when more exact information is not necessary and also for smaller airports or at airports where the demand is much lower than capacity. Also, they provide comparable estimates when com- paring alternatives or airports. These approaches may over- estimate capacity in comparison to what can reasonably be achieved during actual operations. Weaknesses The base capacity curves were developed using older FAA rules and procedures (the AC was published in 1983 and last updated in 1995) and are not easily adaptable as new tech- nology is developed and implemented at airports or as new aircraft types require changes in wake turbulence separations, etc. The methods tend to overestimate capacity because FAA has modified aircraft separation rules and reporting. Analysts are challenged to apply the method when the airport layout does not exactly match one of the 19 runway configurations in the AC. Although these approaches are still commonly used, the methodology is quite outdated. Note: ACRP Report 79 includes a new Prototype Airfield Capacity Spreadsheet Model that allows the analyst to update certain characteristics and ATC procedures that were rigid (hard-coded) in the ACM. The tool calculates hourly capacity and ASV. The model displays capacity charts similar to those shown in Figure 3-1. 3.1.3 Simulation Models Simulation models for evaluating airfield and airspace facilities and air traffic procedures, such as SIMMOD and TAAM, allow for the evaluation of almost any airfield lay- out, procedures, aircraft characteristics, etc. The analyst can input (and generally override any defaults) to represent run- way rolls, taxi paths, runway crossings/priorities, queuing rules, air traffic procedures (separations between aircraft on the same runway, different runways, airspace), runway usage, aircraft performance, touch-and-go operations and many more situations. In addition to the runways, most of the simulation tools also can analyze operations through the air- space, taxiways, and gate/ramp areas. Although one can use generic flights representing a specific fleet mix and demand level, analysts often use a detailed flight schedule of current or forecast traffic levels. Most of the simulation tools report throughput or flow for specific units of time, but do not necessarily output a value for capacity. Analysts may input more demand than they estimate can be accommodated, then report the maximum throughput as the capacity. Alternatively, most simulation Source: Airport Capacity Benchmark Report 2004, FAA Figure 3-1. Airport capacity exceeding actual throughput.

40 models output delay estimates (described in Section 2.2.2), so a threshold value for “acceptable delay” is pre-determined, then when that level of delay occurs in the model, the corre- sponding demand level is reported as the capacity. Strengths Almost any runway layout and operation can be modeled, including new procedures, technologies, and aircraft models. An analyst can choose what delay values to use for determin- ing capacity. The sophisticated predictive capabilities of these models allows for extensive and reliable what-if analysis to support large capital expenditures. Weaknesses This method of determining capacity by using a threshold of acceptable delay is commonly used, but is highly depen- dent on flight schedule patterns. Yet, flight schedule changes should not result in different capacity estimates. Delay thresh- olds cannot be applied across the board to all airports—larger hub airports may tolerate higher levels of delay while smaller spoke airports may use a lower delay threshold. Also, passen- gers using airports that are close to metropolitan areas may tolerate higher delay thresholds than at farther-out airports. Note: MITRE is expected to release a new runway Simulator (rS) capacity model in 2013. This simula- tion tool captures the dynamics of a complex airport, without the labor-intensive setup of a SIMMOD or TAAM model. rS randomly generates flights accord- ing to a specified mix to estimate hourly capacity as a Pareto frontier of arrival-departure throughput (the charts shown in Figure 3-1 were generated by rS). It does not calculate delays. The model uses a definition of capacity that is described as “average maximum sustainable throughput,” where • “Throughput” is a rate of aircraft operations per unit time. • “Maximum” refers to the demand levels; the airport is experiencing demand at or greater than its capacity. • “Sustainable” accounts for variability in perfor- mance and the separations between aircraft to account for the normal variability due to pilot, controller, and environmental variations. • “Average” refers to the output being averaged over many hours of operations. 3.1.4 Reporting Throughput Planners generally do not use the absolute maximum capac- ity as the primary value to measure capacity. Maximum capac- ity can be achieved only with a perfect mix of aircraft, arrivals/ departures, and the minimum separation between aircraft. Given the human actions and reactions involved in the com- munication and coordination between ATC and pilots, as well as aircraft performance, there is much variance in real-world separations. Maximum capacity may also be called saturation capacity. While it is possible to achieve this number of opera- tions when conditions are just right, it is not a capacity that can be maintained or sustained for several hours. A more realistic measure may be called sustainable capac- ity. This is a measure of the hourly capacity that can be realis- tically achieved for several consecutive hours. Due to queuing phenomena of stochastic arrivals, practical capacity is gener- ally 10%-20% lower than maximum capacity. This measure of capacity is designed to take into account the effects of ATC workload. Although desirable to operate at maximum capac- ity, studies have demonstrated that this level of performance cannot be maintained for more than 1 or 2 hours at a time. 3.2 Timeframe of Capacity Measure The basic definition of capacity refers to the number of aircraft that can land/depart on a runway system during a specified time period. Capacity is typically only an issue when demand is higher than capacity. Thus, when discuss- ing capacity, there often is a focus on peak periods, which are expressed in terms of hourly or annual operations, but other timeframes also can be used. The strengths and weakness of each category and the applicability to certain stakeholder groups also will be addressed. 3.2.1 Hourly Hourly increments are generally short enough timeframes to account for the capacity effects of fleet mix, runway depen- dencies, arrival/departure mix, and variances in aircraft sep- arations while the system is still experiencing a continuous demand. This also coordinates well with the peak-hour incre- ments often used in airport traffic demand forecasting. When measuring hourly capacity, most analysts recog- nize that the capacity varies depending on whether the hour consists mostly of arrivals or departures, or a more balanced arrival/departure mix. At slot-controlled airports, airlines may schedule a block of flights near the end or the beginning of the hour for marketing purposes. Thus, merely assigning slots to a 60-minute span can still result in operational delays due to airline scheduling practices. In these cases, the slots should be allocated to smaller time periods, perhaps in 10-minute blocks.

41 the ratios of certain hours per year under certain conditions) is not be useful because it would typically either overstate or understate the actual capacity on any given day. 3.2.2 Daily Daily capacity is infrequently used or reported. However, daily capacity may provide a means of reporting weighted capacity based on the typical wind/weather conditions dur- ing certain scenarios. One of the challenges with daily capac- ity is that generally there are fluctuations such that a demand is not continuous; in other words, there are gaps or lulls in the demand for several minutes or even hours. Even though an airport may be open to operations 24 hours/day, gener- ally there are several hours (perhaps overnight) when there is very little traffic. If an analyst simply multiplies hourly capacity by 24 hours in a day or even by some typical oper- ating time period (say, 6 a.m.–10 p.m.), capacity is likely to be overstated. Strengths Daily capacity can be useful if an airport has a very typi- cal wind/weather pattern such that certain runway(s) and weather conditions are used for mornings, then it switches to different wind/weather configurations throughout the day. Then the hourly capacity in each configuration could be summed to report a daily capacity, as long as that capacity does not include hours in which there is no traffic at the air- port. Daily capacity also can be helpful for airports that have extreme traffic variations due to seasonality or special events to estimate the maximum operations achievable during their high traffic demand days. Some ATC facilities may report throughput in 10-minute or 15-minute increments, however the actual traffic demand can be heavily weighted to arrivals or heavy departures dur- ing such a small time increment. To look at time periods longer than an hour, the arrival/ departure mix would be more balanced; in other words, the longer the time period being considered, the more likely there will be an even number of arrivals and departures. However, there may be brief lulls in the flight demand over several hours such that the demand is not continuous. Strengths Describing capacity in operations per hour is very effec- tive for comparing the capacity under different wind/weather conditions, estimating the capacity change of a new technol- ogy, air traffic rule, or procedure, and ensuring that neces- sary capacity is maintained during construction phases. As depicted in Figure 3-2, it also is quite easy to compare the hourly operations in a flight schedule to hourly capacity. For airports that have highly peaked schedules throughout the day, hourly capacity is helpful to demonstrate to officials why periodic delay/congestion issues occur even though there are other periods with little activity. Weaknesses As described, hourly capacities at a single airport can vary greatly depending on the percentage of arrivals vs. depar- tures and the wind/weather conditions. It is unrealistic to list a single hourly capacity value for an airport, because it would have to be qualified as the hourly capacity under certain con- ditions. A weighted average of hourly capacities (based on Figure 3-2. Forecast demand vs. hourly capacity.

42 a metric of capacity, especially if the landside facilities are more of a constraining factor than are the airside facilities. This would be described as annual enplanements or mil- lion annual passengers (MAP). Or, for smaller increments, the airport could use passengers per hour or departure seats per hour. Translating airside or runway capacity in terms of operations to passengers merely requires one to multiply the number of operations by the average seats or passengers per aircraft. This metric provides the ability to combine number of passengers and number of flights. If the fleet mix changes, the airport may choose to report the number of operations to supplement their information on capacity. Strengths All areas of the airport—airside, landside, terminal processors—can be reported with the same metric to ensure the airport development provides a balanced capacity. Mea- suring capacity in number of passengers may more appro- priately account for fleet mix changes. A large commercial airport provided the example shown in Table 3-1, which dis- plays capacities of various airport components in the com- mon metric of MAP. Weaknesses This requires the airfield capacity to be translated into an annualized MAP value. If the fleet mix (or average seats per aircraft) changes at an airport, the annual passenger metric also may change. 3.3 Capacity Coverage “Curves” (Graphs) Many people outside the airport community do not recog- nize that capacity is not a single number at any given airport, but is really a range of capacity rates. Even for a given airport/ layout, capacity varies in different wind/weather conditions, and varies according to the ratio or mix of arrivals and depar- tures. Capacity coverage curves show the percentage of time that different throughputs can be achieved. In the industry, we sometimes add to the confusion in that we may speak of it as a number. For example, while the FAA’s Airport Capacity Benchmark Report 2004 provides ranges of capacity in different weather conditions, it only reports capacity for the most commonly used runway configuration for each weather condition. 3.3.1 Wind/Weather Conditions When the cloud ceiling and/or visibility are reduced such that IFR applies, there often is a reduced capacity from VFR Weaknesses Daily capacity is not a straightforward output from any of the analytic approaches. For airports that have highly peaked schedules, the daily operations count or flight schedule may be well below daily capacity and it would mask the inability of the airport to accommodate the peaked demand during certain peak hours/times of day. 3.2.3 Annual ASV can be estimated from the FAA’s AC 150/5060-5, ACM, and some spreadsheet tools. Although simulation analyses do not use annual demand or output annual capacity, the daily flight schedule analyzed often represents a daily demand (or perhaps average day, peak month [ADPM]) of a particular annual operations forecast. Thus the simulation analysis of a particular daily flight demand is assumed to be an analysis of a certain annual demand. Annual capacity must account for the various wind/weather conditions throughout the year and the percent of time (or percent of operations) in each condition so as to not over- estimate or underestimate. Annual capacity is most often stated in terms of aircraft operations or flights as the unit of measure. If the aircraft fleet mix changes, the annual operations capacity, as well as the airport’s ability to accommodate that number of annual passengers, may have changed. Strengths Measures of annual capacity are useful for initial high-level estimates of airport capabilities and for comparisons between alternatives, especially in early planning stages. For small air- ports, ASV may be the only metric needed. Weaknesses When calculating ASV, an analyst must make some assumptions as to the number of hours in a day when peak operations will occur and the number of “equivalent” days in a year. Small changes in these types of values result in quite different annual capacity figures. Many in the industry these days believe that, in general, annual capacity measures are not that useful. Annual calculations mask the impact of peak or hub-based scheduling and, for most large airports, have little merit when looking at detailed planning. 3.2.4 Passenger Capacity as a Metric To use a common metric for all airport functional areas, some commercial service airports prefer to use passengers as

43 this table, capacity at several airports is reduced by 5-10% in marginal weather conditions, as compared to optimal condi- tions. Then when in IFR rules, the capacity is only 60-80% of optimum capacity. The reduction in capacity during marginal or IFR may be due to such things as the inability to land on certain runways in IFR conditions (e.g., converging or intersecting runways) or the extra separation required between arrivals on one run- way and departures on a closely spaced parallel runway. Although the capacity values reported in the Benchmark report were for only the primary runway usage configuration in each weather condition, frequently multiple configura- tions also are used, depending on the winds. Table 3-3 gives an example of hourly capacities for a simple airport with just two primary wind conditions in three weather conditions. Displaying this type of capacity information is overwhelming even for planners familiar with the situation. One type of chart to provide this information visually is a capacity coverage curve (CCC), Figure 3-3, which shows capacity by percentage of time for which that configuration conditions, and even certain runway combinations may not be able to be used. Many airports have different capaci- ties in VFR, IFR, and interim conditions that may be called marginal VFR. For example, the FAA’s Airport Capacity Benchmark Report 2004 provided a summary table of hourly capacities at 35 U.S. airports, assuming an equal number of arrivals and departures. Ranges of capacity values were provided for each airport to represent the hourly capacity in the primary con- figuration for three weather conditions: 1. Optimum: Visual approaches possible when the ceiling and visibility are above the minima for visual approaches. 2. Marginal: Ceiling/visibility does not allow visual approaches, but is better than instrument conditions. 3. IFR: Radar separation required due to low ceiling and visibility. The first few airports with their hourly capacity ranges from the FAA’s 2004 report are shown in Table 3-2. As seen in 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51+ (JFB) Forecasted Year To Achieve MAP 2008 2009 2010 2011 2012 2013 2014 2016 2018 2019 2020 Passengers Security Checkpoints Ticketing (Queue Area) On-Airport RAC Entering and Exiting Weaves Terminal Ramps Curbs (Enplane, Deplane, Comm.) Baggage Claim Baggage Handling Systems (BHS) Gates Terminal Roads Ticketing Counter Positions Parking Airfield Automated People Movers (APM) 44 44 50 44 45 45 45 45 Capacity (Million Annual Passengers) Terminal Operational Systems 2007 2015 2017 Current MAP Existing Capacity Enhanced Capacity Table 3-1. Capacity of all airport systems (in MAP). Source: Airport Capacity Benchmark Report 2004, FAA Table 3-2. Sample capacity benchmarks (hourly operations).

44 tions used throughout the day. For example, when the cloud ceiling lowers for a few hours during the day, an airport may lose the ability to use certain runways, which reduces capacity. 3.4 Delay Thresholds Used to Define Capacity Planning reports and studies from a variety of U.S. airports were reviewed by the research team. Studies came from airports with fewer than 2,000 monthly operations to large commercial service airports with over 1,000 daily operations. The reports were examined to determine what types of capacity analyses were conducted and the method or threshold to evaluate if additional capacity was needed. is in use in a year. The same sample airport’s balanced mix capacity is depicted in Figure 3-4 as a CCC with a series of blocks representing the year with the hourly capacity and percent of time that configuration is used. With the CCCs, it is possible to quickly spot the weather conditions that are problematic at an airport and the percent of time that the capacity drops and by how much. A more complex CCC for the 1990 configuration of Chi- cago O’Hare International Airport is shown in Figure 3-4. Although this still does not cover all of the capacity possibili- ties of the airport (e.g., it does not display arrival priority or departure priority values), it does better represent the vari- ances in the capacity as wind/weather changes throughout the day and from day to day. Many airports have winds that shift throughout each day, with several different configura- Figure 3-3. Sample CCC. Wind/Weather Configuration Percent of Year Hourly Capacity – Arrival Priority Hourly Capacity – Departure Priority Hourly Capacity – Balanced Mix Primary wind – VMC 50 105 85 102 Secondary wind – VMC 20 76 65 76 Primary wind – marginal 10 90 82 90 Secondary wind – marginal 8 66 60 66 Primary wind – IMC 10 80 64 74 Secondary wind - IMC 2 30 30 30 Table 3-3. Example of hourly capacities for an airport in each wind/weather configuration.

45 utes. Some have used a different average delay value for arriv- als than departures; for example, 3 minutes average delay for arrivals and 6 minutes average delay for departures. The thought is that additional flying time in the air (delayed arriv- als) is more costly than taxiing/queuing time on the ground for departures. When the simulated average delays per flight reach those thresholds or delay values, the analyst identifies that additional capacity will be required. There tends to be general agreement that when average annual delays per operation exceed 10 minutes, an airport is experiencing severe capacity constraints; and at 20 minutes delay, there is a risk of gridlock. The delay thresholds that identify capacity or trigger the planning of additional capacity vary because the needs are dif- ferent at each airport. The triggers vary according to the lay- out of the airport, taxiway infrastructure, regional airspace serving multiple airports, as well as the typical flight schedule. There is also some indication that the proximity of the airport to the local population influences how much delay is tolerated at that airport. Passengers at close-in airports may tolerate higher airport delays than at further-out airports, indicating that they consider total travel time in their overall view. For initial/strategic planning and master planning, many airports under 200,000 annual operations have used the ana- lytical approach in AC 150/5060-5 to calculate ASV, then sim- ply applied the FAA’s guidance of starting to plan for capacity improvements (e.g., new runway) when the airport traffic demand reaches 60% ASV. However, the forecast growth trend also needs to be considered, as well. If the forecast growth is quite slow in that it will be another 40 years before reaching 80% ASV, then there is no compelling need to plan now for additional capacity. For master planning, most commercial service airports with 150,000 or more annual operations have evaluated aver- age aircraft delays using either analytical methods (such as in AC 150/5060-5) or computer simulation analyses. As stated previously, many of the simulation tools output delays rather than capacity, thus the criteria is often stated in average delay minutes per flight operation. As the planning progresses into Benefits Cost Analysis and Environmental Planning, simula- tion analyses are more common to quantify the delay savings. However, the average delay threshold used to calculate airport capacity or planning criteria for new infrastructure varies from an annual average of 3–4 minutes to 10–15 min- Source: Volpe/Boeing/LMI/FTA/FAA Report, “A Preliminary Design Process for Airspace Systems” Figure 3-4. Runway CCC—Chicago O’Hare.

46 higher capacity, environmental concerns put constraints on the airport that limit the throughput. • Economics/business case vs. delay thresholds—Airport capacity projects should be based on economics and should be done if the improvement is cost beneficial regardless of the level of delay. Projects are based on business case analy- sis, whether for taxiway improvements or future runways. When delay savings exceed costs to build and operate, the project is justified. • Capacity is not a single metric—Some reports describe capacity as a single measure, usually using either hourly or annual throughput. Although this becomes a simple way to compare airports or the capacity of a specific airport during different weather conditions, it oversimplifies the issues and ignores capacity variances that occur in actual operations. Metrics for capacity need to take into account minute-to-minute, day-to-day and/or season-to-season variance in capacity. • Just-in-time capacity—Many analyses, from the FAA’s AC 150/5060-5 to some recent estimates by major airlines, show that airports need additional capacity when demand steadily reaches 80% of the capacity or acceptance rates. As depicted in Figure 3-5, the arrival on-time performance significantly drops off when the number of flights exceeds 80% of the stated capacity (FAA’s airport acceptance rate [AAR]). Most major capacity projects take several years to plan, design, approve (through environmental processes), and construct. Often, it is difficult to get users or the com- munity to support capacity enhancement projects until it is too late. Typically, U.S. airports have a just-in-time delivery of capacity. Projects are not planned while traf- fic is down because the projects are not justified then. But when traffic demand has returned, it is difficult to expedite the planning/development of new projects. If an airline operates a peaked flight schedule (such as con- necting complexes) at an airport, high delays during those peaked times greatly affect the airline’s ability to meet sched- uled times and maintain schedule integrity. So, this type of traffic demand might tolerate a lower overall average delay to ensure that the delays during those peak times do not negatively impact on-time performance. However, the flight schedule pattern should not influence the capacity. In other words, the capacity is established according to the layout and infrastructure and, while the throughput may change according to the procedures and technology, the flight schedule or traffic demand does not change the capacity. 3.5 Challenges in Capacity Measurements Although many aspects of airside capacity are well devel- oped, there remain some challenges and issues in airport capacity measurements and evaluations, including the following: • Airport geometry, ramp capacity, and proximity of gates/ ramp to runways—Runway capacity and gate capacity are fairly straightforward calculations; taxiway capacity and ramp capacity are not so well defined. One can observe that certain taxiway infrastructure systems provide lower capacity (e.g., a single taxiway parallel to a runway instead of dual parallel taxiways, small amount of space between the ramp and the departure queue), but it is difficult to put a value to this. • Environmental effects on capacity—Capacity or through- put is sometimes constrained by environmental concerns (e.g., noise, emissions), just like air traffic procedures. Although the infrastructure or layout exists to achieve a Source: TransSolutions, based on ASPM analysis Figure 3-5. Airport arrival performance as function of acceptance rates.