lable at ScienceDirect

Journal of Air Transport Management 42 (2015) 1e9

Contents lists avai

Journal of Air Transport Management

journal homepage: www.elsevier .com/locate/ ja ir t raman

A Monte-Carlo approach to estimating the effects of selected airportcapacity options in London

Daniel Irvine, Lucy C.S. Budd, David E. Pitfield*

Transport Studies Group, School of Civil & Building Engineering, Loughborough University, Loughborough, LE11 3TU, UK

a r t i c l e i n f o

Article history:Available online 11 July 2014

Keywords:Airport capacityMonte Carlo simulationLondonUK

* Corresponding author.E-mail address: D.E.Pitfield@lboro.ac.uk (D.E. Pitfie

http://dx.doi.org/10.1016/j.jairtraman.2014.06.0050969-6997/© 2014 Elsevier Ltd. All rights reserved.

a b s t r a c t

The issue of future airport capacity in London is currently the subject of much political debate in the UK.Although realistic estimates of the effects of capacity enhancement may be desirable, such estimates aredifficult. Through the use of Monte Carlo simulation, this paper quantifies and compares the relativecapacity enhancements that may be afforded by the construction of a new hub airport in the ThamesEstuary, additional runways at Heathrow, Gatwick and Stansted and changes to operating practices atHeathrow. The simulations show that a new hub airport would be the most effective way to increasecapacity, although the reported financial and environmental costs of such a development indicate acomparatively poor rate of return. Proposed new runways at Heathrow, Gatwick and Stansted and theremoval of runway alternation at Heathrow provide more modest increases in capacity.

© 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Debates surrounding the provision of future airport capacity inLondon and the South East have had a long and controversialpedigree with the issue polarised between those who claim con-nectivity is vital for economic growth and those who believe thatairport expansion creates an unacceptable social and environ-mental burden. At the time of writing, a UKGovernment-appointedAirports Commission, chaired by Sir Howard Davies, has publisheda shortlist of possible solutions that they are currently evaluating. Itcould be argued that, given the political sensitivity of airport ca-pacity in the South East of England, realistic estimates of the effectsof all the proposed capacity enhancements are desirable. Thesecapacities should be comparable and be based on hypotheses thatreflect real or near-real conditions and the method of estimatingcapacities should reflect current traffic conditions. Nobody couldargue with that ideal. However, achieving such estimates is a timeconsuming task and one which is hampered by a number of factors.For example, if traffic is based on schedules, it would not reflect theairports at full capacity and, in reality, aircraft rarely arrive anddepart in the order they are scheduled and, for certain airports,schedules do not include charter traffic. In addition, realistically

ld).

simulating the decisions of air traffic controllers who combinearriving and departing traffic is difficult.

As an indicator of the relative contributions to capacity thatrepresent the maximum usage of the facilities,1 Monte-Carlosimulation can provide comparable estimates based on minimumseparations.2 These estimates are not realistic as operations atairport are not conducted so as to gain maximum usage, irre-spective of arrivals and departures, but they indicate a correctrelative estimate of the contributions to capacity as they are allestimated on exactly the same basis. It is unlikely that the relativeimpacts would be altered if estimates were more refined.

This paper employs Monte Carlo simulation to provide estima-tions of the relative effect on airport capacity that five proposals,most of which are still under consideration by the Commission.These proposals are: the development of a new hub airport in theThames Estuary (as advocated by the Mayor of London); additionalor extended runways at Heathrow (either to the north or the westof the existing airport); a second runway at Gatwick; a secondrunway at Stansted; and the adoption of mixed mode operations atHeathrow.

1 There is no necessary Correspondence between these estimates and declaredcapacities.

2 It could be argued that due to technological advances that these will lessen infuture so increasing capacity estimates.

Table 1Traffic mix and aircraft probabilities for Heathrow.

Aircraft category Percentage ofmovements

Probability Cumulativeprobability

Light (L) 0% 0 0Small b (Sb) 0.2% 0.002 0.002Small c (Sc) 2% 0.02 0.022Lower Medium (LM) 60.4% 0.604 0.626Upper Medium (UM) 2.3% 0.023 0.649Heavy c (Hc) 16.15% 0.1615 0.8105Heavy d (Hd) 16.15% 0.1615 0.972Super (J) 2.8% 0.028 1

Source: Heathrow Airport (2010a)

Table 4Example of departure simulations at Heathrow (using arrivals data presented inTable 3).

Observation Random A/ccategory

Separation Runwayoccupation

Enterrunway

Exitrunway

1 0.313926 LM n/a 0.83 0.83 1.662 0.247403 Hc 1.33 0.83 6.22 7.053 0.570501 LM 2 0.83 8.22 9.054 0.453592 LM 1.33 0.835 0.631647 UM 1.33 0.836 0.471327 LM 1.33 0.83

D. Irvine et al. / Journal of Air Transport Management 42 (2015) 1e92

2. Method

Monte Carlo simulation will provide the basis of the method forthis paper. It has been used before in simulating runway capacity,most notably by Pitfield and Jerrard (1999), and has proven to beequally as suitable as precision modelling tools such as Simmod forsimulating the effect of runway configuration changes. In this casethe main aim is a comparison of the broad changes in capacity ofvarious proposals to give maximum potential usage and thereforeprecision modelling is not necessary.

The first step in the simulation is to define the variables that willbe randomised, in this case the traffic mix at an airport. Inboundand outbound aircraft must be separated according to size to enablepotentially dangerous wake vortices to dissipate. Larger aircraftcreate more wake turbulence than smaller aircraft and conse-quently each aircraft type is assigned to awake turbulence categoryto ensure safe separation is maintained. The traffic mix at an airportthus directly influences its capacity. At London Heathrow, forexample, the majority of aircraft are Lower Medium or Heavy air-frames (Table 1).

Each aircraft in the simulation is assigned a random numberwhich is married to the cumulative probability column in Table 1 toassign a wake turbulence category. This process is then repeated tosimulate peak hour flows. A separation time for each aircraft isdefined according to international metrics (Table 2). It is noted thatthe peak hour traffic mix may vary slightly from the overall trafficmix (e.g. cargo airlines may use larger aircraft and avoid the peak

Table 2Separation times for arriving aircraft (in minutes).

Following aircraft

PrecedingAircraft

Super (J) HeavyD (Hd) HeavyC (Hc) UppeSuper (J) 1.78 2.34 2.67 3.07Heavy (H) 1.78 1.56 1.78 2.19Upper Medium (UM) 1.11 0.97 1.11 1.31Lower medium (LM) 1.11 0.97 1.11 1.09Small (S) 1.11 0.97 1.11 1.09

Light (L) 1.11 0.97 1.11 1.0

Source: Based on separation distance data from NATS (2010)

Table 3Example of arrival simulations at Heathrow.

Aircraft Random Aircraft category Separation (mins) Runtim

1 0.666653 Hd 0 0.832 0.568444 LM 2.22 0.673 0.43141 LM 1.11 0.674 0.307736 LM 1.11 0.675 0.597959 Hc 1.11 0.676 0.661426 Hc 1.11 0.83

hour), however with the data that is available, and the “peak hour”being a theoretical rather than actual time period, the overall trafficmixes are broadly appropriate.

All the aircraft in the simulation are then also assigned a runwayoccupation time (see Section 3.4). In this exercise, arrivals aresimulated before departures as inbound aircraft have priority.

The simulation starts at 0 min and it is assumed the first arrivalcrosses the runway threshold at this time. The next aircraft isassumed to be following the first at minimum separation. Thus ifthe required separation between the aircraft is 1.11 min, the nextaircraft is 1.11 min away from crossing runway threshold when thefirst aircraft is at the threshold. Table 3 illustrates simulated aircraftarrivals and shows how separation distances determine the time inthe simulation at which they enter the runway. The runway occu-pancy times for arriving aircraft are required to calculate de-partures. The process continues until the simulation reaches 60minat which point the arrivals are stopped and counted.

Simulating departures similarly requires aircraft category, sep-aration time and runway occupancy time to be calculated. Suitablegaps must then be found in the sequence of arriving traffic inwhichto slot in departures. A departure requires both the minimumseparation time from the preceding departure and a gap to the nextarrival that is greater than its own runway occupation time(Table 4).

This process is repeated for 60 min. The simulation is thenstopped and the departures are counted. Any aircraft that hasbegun its take-off roll but which has not left the runway withinthese 60 min is not counted. The arrivals and departures are then

r Medium (UM) Lower Medium (LM) SmallC (Sc) SmallD (Sd) Light (L)3.11 3.21 3.65 4.532.22 2.75 3.13 3.961.78 1.83 2.09 3.401.11 1.37 1.57 2.831.11 1.37 1.57 2.26

9 1.11 1.15 1.30 1.42

way occupatione (mins)

Enter runway(mins)

Exit runway(mins)

Unoccupiedduration (mins)

0 0.83 1.392.22 2.89 0.443.33 4.00 0.444.44 5.11 0.445.55 6.22 0.957.17 8.00 1.39

Table 5Separation times for aircraft wake turbulence categories, in minutes.

Following aircraft

Super (J) HeavyD (Hd) HeavyC (Hc) Upper Medium (UM) Lower Medium (LM) SmallC (Sc) SmallD (Sd) Light (L)

Precedingaircraft

Super (J) 1.78 2.34 2.67 3.07 3.11 3.21 3.65 4.53Heavy (H) 1.78 1.56 1.78 2.19 2.22 2.75 3.13 3.96Upper Medium (UM) 1.11 0.97 1.11 1.31 1.78 1.83 2.09 3.40Lower medium (LM) 1.11 0.97 1.11 1.09 1.11 1.37 1.57 2.83Small (S) 1.11 0.97 1.11 1.09 1.11 1.37 1.57 2.26Light (L) 1.11 0.97 1.11 1.09 1.11 1.15 1.30 1.42

D. Irvine et al. / Journal of Air Transport Management 42 (2015) 1e9 3

added together to give an overall figure of theoretical maximumpeak hour capacity. The only exceptions to this are the simulationsat Heathrow under segregated mode operations (in which onerunway is used for arrivals the other for departures) and in caseswhere runway layout dictates that runway crossings are simulated.Each simulation is repeated 10 times (the pre-determined point atwhich simulations started reaching a steady state) and the averagenumber of movements defines that configuration's capacity.

It should be noted that giving total priority to the flow ofarriving aircraft does not entirely reflect the nature of air trafficmanagement and may lead to some “arrival-heavy” results. In re-ality, arriving aircraft would be expected to hold in stacks and thenumber of arrival slots may be limited, to allow the departingaircraft to use the runway. This paper is focused on analysingmaximum capacities, and not in precisely replicating real worldairport operations.

3. Data

The simulations employed four main variables.

3.1. Aircraft separation minima

All aircraft create wake turbulence which can pose a danger toother air traffic. The International Civil Aviation Organisation(ICAO) assign commercial aircraft to wake categories according tothe strength of the vortices they produce and publish minimumseparation distance guidelines. These guidelines have beenexpanded by the UK's Civil Aviation Authority (CAA) to six cate-gories, from the “Light” aircraft such as a Jetstream 41, up to the“Super” category for the Airbus A380. These categories have beenassigned separation distances for both arrivals and departures.However, as these simulations need to work in time, rather thandistance, aircraft speeds need to be estimated too.

3.2. Aircraft speeds

Aircraft speeds are needed to convert arrival separation dis-tances into time. However, this task is not straightforward. An air-craft's final approach speed depends on a number of factorsincluding weight, meteorological conditions and local operatingrequirements. In order to convert wake turbulence categories into

Table 6Traffic as percentages of each wake vortex category.

Category Heathrow Gatwick Stansted

Super (J) 0.7% 0% 0%Heavy (H) 34.4% 9.35% 4.3%Upper Medium (UM) 2.3% 4% 0.6%Lower Medium (LM) 60.4% 75.5% 91.9%Small (S) 2.2% 11% 2.6%Light (L) 0% 0.15% 0.6%

average approach speeds, a number of assumptions needed to bemade. These included identifying the approach speeds of specificaircraft within a wake turbulence category and then calculating anaverage speed. The speeds used are the manufacturer's designated“approach speed”. Two of the categories provided some differencesin approach speeds (for instance small jets and large turboprops inthe “Small” category) and these categories were therefore split intotwo distinct categories. The categories and their respectiveapproach speeds can be seen in the table below. Table 5

3.3. Traffic mix

Information on the average traffic mix of aircraft currently usingHeathrow, Gatwick and Stansted airports was obtained fromHeathrow Airport (2010a) and Airport Coordination Ltd (2010).Each aircraft type was cross-referenced with an aircraft database toassign it to a wake turbulence category. These were then collated toproduce probabilities of each category at each airport (Table 6).

A slight alteration was made to the Super category to replicatethe expected increase in A380 aircraft at Heathrow in the comingyears (Heathrow Airport, 2012).

3.4. Runway occupancy time

The time an aircraft occupies the runway depends on numerousvariables including size, weight, acceleration/deceleration and in-dividual flightcrew. For this investigation, the runway occupancytimes derived from ICAO rules and used by Pitfield and Jerrard(1999) were adopted.

4. Simulating current capacity at London airports

Current peak hour capacity simulations were performed forLondon's three busiest passenger airports: Heathrow, Gatwick andStansted.3

4.1. Heathrow

Heathrow airport is surrounded by densely populated urbanareas and has two independent parallel runways, 27L/09R and 27R/09L. These runways have historically been used in segregatedmodeto give local residents some relief from aircraft noise. Owing toprevailing wind directions and the frequency of their use, the “27”runways are used in these simulations. Both runways feature high-speed taxiway turnoffs which facilitate rapid access and egress. Oursimulation assumes that the arrival runway is 27L and that de-partures can use both runways. Owing to a lack of available slotsand the number of long-haul flights at the airport, Heathrow'straffic mix is dominated by ‘Heavy’ aircraft.

3 London Luton was also simulated to provide an all London airport total. Itincluded the assumption that new taxiways are developed.

Table 7Heathrow “current situation” simulations.

Simulation 27Larrivals

27Ldepartures

27Rarrivals

27Rdepartures

Total

1 42 13 0 40 952 40 15 0 40 953 44 12 0 38 944 41 15 0 40 965 41 14 0 40 956 42 12 0 41 957 41 14 0 40 958 42 11 0 42 959 38 18 0 40 9610 41 13 0 41 95Average 41.2 13.7 0 40.2 95.1

Table 9Stansted “Current situation” simulations.

Simulation Arrivals Departures Total

1 54 0 542 53 2 553 51 4 554 52 2 545 50 5 556 50 5 557 51 4 558 52 3 559 50 4 5410 51 3 54Average 51.4 3.2 54.6

D. Irvine et al. / Journal of Air Transport Management 42 (2015) 1e94

Table 7 shows that Heathrow's current peak hour capacity isaround 95 movements. This is close to the official figure of 96(Heathrow Airport, 2010b) and shows that the assumptions in thesimulation are appropriate and accurate.

4.2. Gatwick

Gatwick Airport is the busiest single runway airport in theWorld. Although the facility has two runways, they are too closetogether to be operated simultaneously and usually only the longer26L/08R is active. Our simulation assumes a single-runway opera-tion. Owing to the dominance of B737 and A320 family aircraft thatare used by the low cost and charter operators, 75% of aircraft atGatwick are in the Lower Mediumwake category. The results of thepeak hour simulations at Gatwick appear in Table 8. They areslightly higher than the official declared hourly capacity of 53 (ACL,2012a).

4.3. Stansted

Stansted has a single runway and a single terminal. Like Gat-wick, Stansted's traffic mix is dominated by Lower Medium cate-gory aircraft on account of easyJet and Ryanair's use of A319 andB737 aircraft respectively. The simulations in Table 9 indicate thatStansted can accommodate 54 aircraft per hour, four more than theofficial declared capacity (ACL, 2012b).

4.4. Current combined capacity of Heathrow, Gatwick and Stansted

Table 10 shows the combined capacity of the three main Londonairports that were simulated in this section, and Luton, which wasalso simulated using the same methodology. The estimated totalpeak hour capacity is just under 263 movements. This is around 31movements (11.7%) higher than the sum of the airports' official

Table 8Gatwick “Current situation” simulations.

Simulation Arrivals Departures Total

1 48 12 602 49 6 553 48 10 584 47 10 575 46 11 576 48 11 597 46 10 568 47 9 569 47 9 5610 44 14 58Average 47 10.2 57.2

declared capacities and mean that the simulations yield results4e5% higher than the declared capacity.

5. Monte-Carlo Simulations of proposed capacityenhancements

This section reports on the findings of capacity simulations forfive proposed capacity enhancements in order to compare andcontrast the capacity effects of the individual options against thecurrent baseline capacity figures. The results of the Monte Carlosimulations into the effects of a new hub airport in the ThamesEstuary will be presented first. This will be followed by simulationsinto the effects of new runways at Heathrow, Gatwick and Stanstedand the abolition of segregated mode operations at Heathrow.

5.1. New hub airport

A number of different proposals and locations for a new four-runway airport to the east of London in the Thames Estuary havebeen proposed by construction companies and developers. Thesedesigns typically feature four parallel runways and a central ter-minal area andwould result in the closure of Heathrow (BBC, 2012).The Airports Commission has not “shortlisted” this idea, but iscurrently investigating the idea further, with an eye to potentiallyadding it to the shortlist before the final report is published. For thesimulations it is assumed that the new facility will have a similartraffic mix to Heathrow, albeit with a slightly higher proportion ofsmaller aircraft owing to the availability of more slots (see Table 11).Actual traffic mixes at airports in the future would be extremelydifficult to forecast, but may vary considerably from those used inthese simulations. The runways will not be segregated by aircrafttype, as the proposals are at an early stage and no such plan hasbeen suggested. However it should be noted that this has been usedat some airports with a large number of runways, and may result ina small increase in peak hour capacity.

Table 12 shows the output of the simulations for the Thamesestuary airport. The maximum capacity of the airport is estimatedat around 217 movements per hour, 119 more than at Heathrow.

Table 10Total hourly capacity for London airports in current situation.

Airport Arrivals Departures Totalmovements

Airport'sdeclaredcapacity

Difference

Heathrow 41.2 53.9 95.1 96 �0.9Gatwick 47.0 10.2 57.2 53 þ4.2Stansted 51.4 3.2 54.6 50 þ4.6All London

Airportsa191.6 71 262.6 232 þ30.6

a Includes Luton.

Table 11Traffic mix at Heathrow and (assumed) traffic mix at a new hub airport.

Aircraft category Percentage atHeathrow

Percentage at newhub airport

Super (J) 2.8% 2.8%Heavy d (Hd) 16.15% 16.15%Heavy (Hc) 16.15% 16.15%Upper Medium (UM) 2.3% 2.3%Lower Medium (LM) 60.4% 50.4%Small c (Sc) 2.0% 7.0%Small b (Sb) 0.2% 5.2%Light (L) 0% 0%

Table 14Comparison between Heathrow Third Runway and Current Situation simulations.

Currentsituationcapacity

Heathrowthird runwaycapacity

Difference Percentagechange

Heathrow 95.1 149.9 54.8 57.6%All London

airports262.6 317.4 54.8 20.9%

D. Irvine et al. / Journal of Air Transport Management 42 (2015) 1e9 5

The simulations show that the construction of a new hub airportwill have a dramatic effect, potentially increasing London's airportcapacity by 46%. Although the construction of a newhub airport hasbeen advocated by politicians, developers and some sections of theaviation community, concerns about cost, environmental impactand airspace conflict with Amsterdam Schiphol means that alter-native proposals to expand Heathrow are also being considered.

5.1.1. Heathrow 3rd runwayOne initial option for enhancing Heathrow's capacity involved

constructing a third runway to the north-east of the existingairfield. In plans outlined by the Airports Commission, the runwaywould be built far enough to the north to allow for the simulta-neous operation of all three runways. This might, however, obligesome aircraft to cross the active centre runway when transitingbetween the northern-most runway and the central terminal area.In the simulation, we assume all aircraft using the new thirdrunway will use the proposed new sixth terminal (therebyremoving the need to cross any active runways) and that the newrunway would handle a similar traffic mix. The existing segregatedmode operationwill be retained, although it is assumed arrivals canland on the new runway at any time.

Table 12Simulations for Thames Estuary airport.

Simulation R1 arrivals R1 Dep's R2 arrivals R2 Dep's

1 39 16 40 152 37 15 40 133 42 11 41 144 38 15 40 135 37 18 40 146 38 18 39 157 37 18 38 148 41 12 45 109 39 16 38 1710 38 18 41 13Average 38.6 15.7 40.2 13.8

Table 13Simulations for Heathrow with third runway.

Simulation 27L arrivals 27L departures 27R arrivals 27R de

1 42 13 0 402 40 15 0 403 44 12 0 384 41 15 0 405 41 14 0 406 42 12 0 417 41 14 0 408 42 11 0 429 38 18 0 4010 41 13 0 41Average 41.2 13.7 0 40.2

Table 13 shows the simulations for the two current runways andalso the simulations for the third runway. The new total peak hourcapacity is just under 150 movements. Compared to the currentsituation, this option represents an increase in total capacity of57.6% and an overall increase in the capacity of London airports of20.9% (Table 14).

5.1.2. iiHeathrow runway extensionAn alternative being considered by the Airports Commission, is

to extend Heathrow's runway 27R/09L to roughly double its length.This will allow arrivals to use the first half of the runway and de-partures the second half, without effecting each other's operations.This is fairly straight forward to simulate, by effectively splittingthis extended runway into two runways; one solely for arrivals andone solely for departures. The results of this can be seen below inTable 15.

As expected, the arrivals are roughly the same, but the numberof departures has been allowed to triple without the restrictions ofthe arrival flow. Although in reality the number of arrivals anddepartures would be balanced in both scenarios, it is reasonable toassume the increase in total movements is representative of theincrease in both arrivals and departures.

This proposal provides a 15.65% increase on the current Londoncapacity as shown in Table 16, however that is over 5 percentagepoints lower than the plan for a new runway. The main reason for

R3 arrivals R3 Dep's R4 arrivals R4 Dep's Total

41 14 39 14 21840 15 41 15 21638 14 40 13 21339 18 42 13 21837 16 39 16 21738 16 40 13 21739 15 43 10 21441 13 40 16 21838 16 41 12 21738 17 41 13 21938.9 15.4 40.6 13.5 216.7

partures 3rd Runway arrivals 3rd Runway departures Total

44 11 15041 14 15040 14 14842 13 15141 13 14941 15 15137 19 15145 9 14944 11 15143 11 14941.8 13 149.9

Table 15Simulation results for Heathrow Runway Extension.

Simulation Extendedrunwayarrivals

Extendedrunwaydepartures

Old runwayarrivals

Old runwaydepartures

Total

1 39 40 42 13 1342 40 39 40 15 1343 42 40 44 12 1384 42 40 41 15 1385 42 41 41 14 1386 41 40 42 12 1357 43 40 41 14 1388 41 40 42 11 1349 43 39 38 18 13810 41 39 41 13 134Average 41.4 39.8 41.2 13.7 136.1

Table 17Output of simulations for Heathrow westward expansion.

Simulation Runway#1“Wide-Body”arrivals

Runway#2departures

Runway#3departures

Runway#4“Narrow-Body”arrivals

Total (allrunways)

1 34 38 40 51 1632 36 40 41 48 1653 34 39 40 52 1654 37 40 38 50 1655 37 38 41 49 1656 36 40 39 50 1657 36 38 41 50 1658 35 37 40 50 1629 36 38 41 51 16610 36 39 40 48 163Average 35.7 38.7 40.1 49.9 164.4

D. Irvine et al. / Journal of Air Transport Management 42 (2015) 1e96

this is that the extended runway still only allows one aircraft toarrive and one to depart at a time.

5.1.3. Heathrow westward expansionAnother option that has been proposed is to expand Heathrow

westwards by building two new parallel runways to the west of theexisting pair along the same alignment (Leunig, 2012). The mainsuggested advantages of this scheme are that it would permit asignificant increase in capacity, would not require the destructionof local villages and would reduce the number of people adverselyaffected by aircraft noise. This is because the approach path wouldbe moved a few kilometres to the west meaning aircraft should besignificantly higher (and quieter) to the east of the airport. Thesimulations show that this layout can accommodate around 164peak hour movements (Table 17).

Table 18 shows how the outputs of the simulations comparewith the two proposals already simulated; the plan to build a thirdnorthern runway at Heathrow and the plan to construct a new hubairport in the Thames Estuary.

The westward expansion proposal increases the total runwaycapacity of London by a little over 26%. However, this comparesunfavourably with the replacement of Heathrowwith a new hub inthe Thames Estuary which would provide nearly double the ca-pacity, even though the number of runways is exactly the same.Indeed, the westward expansion only provides around a 5%improvement on the capacity a third runway at Heathrow wouldprovide. The reason for this is simply that the runways in thewestward expansion plan can only accommodate either arriving ordeparting traffic, and not both types, as the other proposals do.

5.2. 2nd Runway at Gatwick

Although a local agreement prevents the construction of a sec-ond runway at Gatwick before 2019, a second runway has beenproposed as a way of alleviating some of the existing capacityconstraints at London airports. Under plans outlined in the AirportsCommission's Interim Report, the new runway would be located tothe south of the existing one, with a new terminal located between

Table 16Comparison between Heathrow runway extension and current situationsimulations.

Currentsituationcapacity

Heathrowrunway extensioncapacity

Difference Percentagechange

Heathrow 95.1 136.1 41.0 43.11%All London

Airports262.6 303.6 41.0 15.65%

the two. Table 19 shows that the new runway roughly doubles theairport's capacity. Interestingly, it also suggests that Gatwick withtwo runways would be able to handle 9 extra movements per hourthan Heathrow's two runways (with no runway use restrictions)and nearly 5 more per hour than Stansted with two runways (seeSection 4.4).

Table 20 shows that building a second runway at Gatwick almostdoubles the airport's capacity and increases the capacity of Londonairports by 22%.

5.3. 2nd Runway at Stansted

Stansted's location in rural Essex means that it is theoreticallythe easiest to expand, although it has been rejected from the Air-ports Commission's shortlist. The airport's 2006 interim masterplan (Stansted Airport, 2006), indicated that a new runway could bebuilt to the south east of the existing runway and it is this config-uration which we use for simulation. It is assumed that the newrunway has identical dimensions and can support the same trafficmix as the existing runway. Table 21 shows the results of thesimulations for Stansted with the second runway.

Table 22 shows how the Stansted second runway plans wouldchange the capacity of Stansted and the London airports as a whole.This shows that the increase in capacity is very similar to that of aHeathrow third runway, which is potentially significant sinceexpanding Stansted may be cheaper and more politically accept-able than developing Heathrow.

5.4. Optimising Heathrow

The final optionwe consider here is the proposal to permanentlyadopt mixed mode operations at Heathrow. The advantage of thisproposal is that it would not require any expensive or disruptiveconstruction but it would impose an additional noise burden onairport residents.

The current runway alternation agreement “provides for onerunway to be used by landing aircraft from 06:00 until 15:00 andthe other runway to be used from 15:00 until after the last depar-ture” (Heathrow Airport, 2013). What this means is that, while onerunway can be freely used for arrivals and departures, the otherrunway can be used for departures only. While in an off-peak flow,having one runway designated primarily for arrivals and one fordepartures may be fine, during peak periods commercial andoperational imperatives dictate that the runways should be used totheir maximum capacity. At present, therefore, Heathrow is notoperating at its maximum capacity. The simulations that examinethe effect of removing runway segregation adopt the same as-sumptions as for the current situation simulations. The simulations

Table 18Comparing the results of the westward expansion simulations with similar proposals.

Expansion proposal London “current situation”capacity

Capacity with expansionproposal

Difference Percentage change Estimated cost

Heathrow Third Runway 262.6 317.2 þ54.6 þ20.79% £10bnThames Estuary Airport (replace LHR) 262.6 384 þ121.4 þ46.23% £50bnHeathrow Westward Expansion 262.6 332 þ69.4 þ26.43% £10bn

Table 19Simulations for Gatwick Airport with the new runway.

Simulation 1st Runway arrivals 1st Runway departures 2nd Runway arrivals 2nd Runwaydepartures

Total movements Average runwaycrossing delay

1 48 12 49 9 118 1.45 min2 49 6 46 9 110 0.36 min3 48 10 47 11 116 0.70 min4 47 10 47 9 113 0.46 min5 46 11 49 9 115 0.64 min6 48 11 50 7 116 0.74 min7 46 10 44 12 112 0.40 min8 47 9 49 9 114 0.54 min9 47 9 43 14 113 0.50 min10 44 14 47 9 114 0.57 minAverage 47 10.2 47.1 9.8 114.1 0.64 min

D. Irvine et al. / Journal of Air Transport Management 42 (2015) 1e9 7

for runway 27L will be kept as the variables are unchanged. Thesimulations for runway 27R will be run as if the runway alternationagreement did not exist.

Table 23 details the outputs of these simulations. The unre-stricted runway can handle around 54 movements per hour (40

Table 20Comparing the capacity of Gatwick Airport. New runway simulations versus thecurrent situation.

Current situationcapacity

Gatwick new runwaycapacity

Difference Change

Gatwick 57.2 114.1 56.9 99.5%All London

Airports262.6 319.5 56.9 21.7%

Table 21Simulations for Stansted second runway.

Simulation 1st Runwayarrivals

1st Runwaydepartures

1st Runwayarrivals

2nd Runwaydepartures

Total

1 54 0 51 4 1092 53 2 51 3 1093 51 4 50 5 1104 52 2 49 7 1105 50 5 51 5 1116 50 5 51 5 1117 51 4 51 5 1118 52 3 50 6 1119 50 4 48 7 10910 51 3 52 2 108Average 51.4 3.2 50.4 4.9 109.9

Table 22Comparison between Stansted expansion capacity and current situation capacity.

Currentsituationcapacity

Capacity withStansted expansion

Difference Percentagechange

Stansted 54.6 109.9 55.3 101.3%All London

Airports262.6 317.9 55.3 21.1%

arrivals, 14 departures). The simulations with the runway alterna-tion agreement in the current situation section showed just under40 departures were possible. Therefore removing the runwayalternation agreement would raise Heathrow's maximum capacityby around 14 movements per hour.

Table 24 shows the capacity of London and Heathrow with andwithout the runway alternation agreement. It can be seen thatremoval of this agreement would increase Heathrow's capacity by15% and London overall by over 5%.

Significantly, this proposal affords a 14.6% capacity increase atthe airport at minimal cost.

6. Discussion

The results of existing airport capacity simulations were pre-sented in Section 3 and the results of simulations into five proposedcapacity options were presented in Section 4. This section

Table 23Simulations for Heathrow without runway alternation restrictions.

Simulation 27L arrivals 27L departures 27R arrivals 27R departures Total

1 42 13 39 15 1092 40 15 43 12 1103 44 12 41 14 1114 41 15 40 14 1105 41 14 39 15 1096 42 12 39 14 1077 41 14 38 16 1098 42 11 42 12 1079 38 18 40 15 11110 41 13 37 16 107Average 41.2 13.7 39.8 14.3 109

Table 24Comparison of Heathrow capacity with and without runway alternation.

Currentsituationcapacity

Capacity withoutrunway alternation

Difference Change

Heathrow 95.1 109.0 13.9 14.6%London total 262.6 276.5 13.9 5.29%

Table 25The proposals compared.

Proposal Total capacity Change in capacity Year capacity constraints reach 2013 levels Estimated cost

Current Situation/Do nothing 262.6 0% 0 0Heathrow third runway 317.4 þ20.9% 2021 £10bnHeathrow Runway Extension 303.6 þ15.7% 2018 £10bnStansted second runway 317.9 þ21.1% 2021 £2.7bnGatwick second runway 319.5 þ21.7% 2021 £5bnThames Estuary airport 479.3 þ82.5% 2050þ £50bnThames Estuary airport without LHR 384.2 þ46.3% 2033 £50bnLHR Westward Expansion 332.0 þ26.43% 2023 £10bnNo runway alternation at LHR 276.5 þ5.3% 2015 N/A

D. Irvine et al. / Journal of Air Transport Management 42 (2015) 1e98

compares the relative capacity benefits of these proposals with thecurrent situation to assess their relative merits and limitations.(This cost/benefit analysis is not the focus of this paper, and wouldcertainly benefit from a more thorough investigation in a futurepaper, which could use the capacity figures produced in this paper)

Table 25 shows the results of all the proposals in this investi-gation. “N/A” has been given under costs where there are not anyinfrastructure costs, but other costs may exist. It is important toemphasise that these cost estimations are derived from publicallyavailable sources and are indicative only. The “year capacity con-straints reach 2013 levels” column uses UK Department for Trans-port estimates to see at which point demand reaches the sameproportion of supply as in 2013.

All proposals to add a new runway increase the capacity byaround 20%. All the proposals that do not involve the constructionof a new runway are comparatively ineffective at increasing runwaycapacity. Increasing the length of Heathrow's runway only providesaround three quarters of the benefits of a new runway, so unlessthere are any significant benefits unique to this alternative, it doesnot appear as attractive as a new runway. The proposal to removethe runway alternation agreement at Heathrow increases total ca-pacity by just over 5%. This proposal should be easier to implementthan any of the proposals requiring high levels of construction.There is potentially a problem that extra terminal capacity wouldbe needed and therefore this may not be a “free” proposal. Thereare also further issues, mainly concerning the effect on localresidents.

The Thames Estuary proposal is by far the most effective atincreasing capacity. However it also has the obvious disadvantagesof being the most expensive. If the government believes thatincreasing capacity is of the highest priority regardless of cost, thenthis could be an option. The decision about whether or not to keepHeathrow should a new hub airport open will arguably not bedown to capacity. It is more likely to be either a question ofcompromise (removing Heathrow to increase the proposalsacceptability) or airspace congestion as a result of the proximity ofthe two airports. Overall, this proposal does provide an effectivelong-term solution to the capacity problem, but considering otherfactors, this option is perhaps not as attractive as some of thesmaller expansion proposals.

The proposals to construct a second runway at Stansted andGatwick both increase capacity but there is not a lot to choosebetween them in terms of capacity increases. If financial estimatesare taken into account, then Stansted appears to offer the best valueoption. Stansted also has plenty of land around the airport toaccommodate any expansion. The research findings suggest that allthree airports have strong cases for a new runway. Gatwick offersthe highest increase in capacity, Stansted arguably provides thebest value for money and Heathrow is reportedly the preferredchoice among certain business and aviation groups.

Constructing all three runways would cost in the region of£20bn but increase the capacity in the South East by around 62%.

This would offer a significant saving over the £50bn cost of a newhub airport whilst maintaining much of the capacity increase.

Expanding Heathrow to the west, as per Leunig's proposal,would give a capacity increase of around 26% (only 5% more thanthe Heathrow third runway proposal). However, more researchneeds to be carried out to assess the true costs of such a project,both financially and socially.

7. Conclusion

Through the use of Monte Carlo simulation this paper hasexamined five proposals that have been advanced to address theissue of future airport capacity in London and the South East ofEngland. While the exercise has provided comparable figures forcurrent and potential future capacity at individual airports andacross the three main London airports it does not advocate anyparticular development as it is not based on realistic estimates ofabsolute capacity but on simulations of relative capacity at thepeak. Indeed, we are acutely aware that any future decision aboutUK airport capacity needs to consider not only the practicalities ofairport operations and optimal airfield configurations and locationsbut also the myriad socio-economic and environmental implica-tions of any development at both local and global scales.

However, with all assumptions considered, this paper hasclearly shown how the most popular proposals compare againsteach other in terms of peak hour capacity. It shows that a newunrestricted runway generally translates to around a 21% increasein peak hour movement capacity for the London area, althoughthere are some small variations depending on where it is placed,and the consequent aircraft types that would use this new runway.This paper has also shown that removing the runway alternationagreement that is currently in place at Heathrow could releasearound 5% of extra runway capacity for London comparativelysimply, although in the wider context London's capacity issues areunlikely to be solved without the building of new additionalrunway infrastructure. The increase in capacity figures producedthrough this research provides an easily comparable basis fromwhich the relative merits of these projects can be examined ingreater detail.

References

Airport Coordination Ltd, 2010. LGW W10 Start of Season Report. Available at:http://www.acl-uk.org/UserFiles/File/LGW%20W10%20START%20OF%20SEASON%20REPORT_1.pdf (accessed 03.03.13.).

Airport Coordination Ltd, 2012a. LGW S12 Start of Season Report. Available at:http://www.acl-uk.org/UserFiles/File/LGW%20S12%20Start%20of%20season%20report_V6.pdf (accessed 11.03.13.).

Airport Coordination Ltd, 2012b. STN S12 Start of Season Report. Available at: http://www.acl-uk.org/UserFiles/File/STN_S12_Start_of_season_report2.pdf (accessed16.03.13.).

BBC, 2012. New Thames Estuary Airport Proposal Unveiled. Available at: http://www.bbc.co.uk/news/uk-england-19577410 (accessed 04.04.13.).

D. Irvine et al. / Journal of Air Transport Management 42 (2015) 1e9 9

Heathrow Airport, 2010a. BAA Airports Traffic Statistics 2010 Annual Review.Available at: http://www.heathrowairport.com/static/Heathrow/Downloads/PDF/Traffic_Statistics-2010.pdf (accessed 03.03.13.).

Heathrow Airport, 2010b. Patterns of Traffic 2010. Available at: http://www.heathrowairport.com/static/Heathrow/Downloads/PDF/Patterns_of_Traffic-2010.pdf (accessed 10.03.13.).

Heathrow Airport, 2013. Runway Alternation. Available at: http://www.heathrowairport.com/noise/what-we-do-about-it/measures-already-in-place/runway-use/runway-alternation (accessed 08.03.13.).

Leunig, T., 2012. Bigger and Quieter: the right answer for aviation. Policy Exchangeavailable at:[. http://www.policyexchange.org.uk/images/publications/bigger%20and%20quieter.pdf (last accessed 20.08.13.).

Pitfield, D.E., Jerrard, E.A., 1999. Monte Carlo comes to Rome: a note on the esti-mation of unconstrained runway capacity at Rome Fiumucino Airport. J. AirTransp. Manag. 5 (4), 185e192.

Stansted Airport (BAA), 2006. Interim Masterplan. http://www.stanstedairport.com/static/Stansted/Downloads/PDF/STN_interim_masterplan.pdf (accessed01.04.13.).