Benchmarking Airport Airside Performance: FRA vs. EWR

Transcription

1 Benchmarking Airport Airside Performance: FRA vs. EWR Amedeo Odoni, Thomas Morisset, MIT Alexander Zock, Wilhelm Drotleff, ECAD ATM Research & Development Seminar Berlin, June 14, 2011

2 Compare airside performance characterisjcs: European vs. US Airports Part of FAA EUROCONTROL inijajve. MIT effort funded by FAA through NEXTOR Stage 1: 34 top airports in Europe vs. 34 top in US Stage 2: Deep dive at selected airports (FRA vs. EWR) Impacts of: (i) use of VMC procedures, weather permi^ng, at US airports (ii) limited or, mostly, non- use of declared capacity limits at US airports on the performance of US vs. European airports w.r.t. capacity, delays and reliability of flight schedules 2

3 Based on combinajon of data analysis (primarily from 2007) and modeling (throughputs, and delays) Databases: US: ASPM (34 busiest airports, complete data) EUR: CODA (34 busiest airports, parjal data collected from airlines on a voluntary basis) Germany: DFS, Fraport, LH, German Airport Coordinator (GAC) Models: MACAD, MIT DELAYS, MIT AND (Airport Network Delays) [not in this talk] 3

4 FRA vs. EWR: Main Characteris>cs US & GERMAN AIP 2007 Layout FRA & EWR Compara>ve overview: FRA EWR Number of pax 54,2 mln 36,4 mln No. of movements pax/ movement No. Of runways 3 3 Cargo volume (t) 2,2 mln. 0,9 mln % intern. pax 85% 29% Carrier Lulhansa ConJnental ACI World Traffic Report 2007 Both airports are considered congested Number of runways and configuration are similar Number of movements are comparable Both airports serve as hubs for a network airline 4

5 FRA vs. EWR - airport vicinity EWR and its surrounding airports FRA and its surrounding airports EWR is embedded in a metroplex environment wherein the airspace is shared with six other airports two of which are major international airports within 25 nm of EWR FRA has almost no constraints deriving from other airports in its neighborhood 5

6 In US: Opera>ng Policies (somewhat simplified) In Visual Meteorological CondiJons VMC) use Visual Flight Rules (VFR); this may increase throughput significantly compared to use of Instrument Flight Rules (IFR) which are used in Instrument Meteorological CondiJons (IMC) With very few excepjons, no caps on number of movements that can be scheduled Almost everywhere else (including Europe): Operate (officially) under IFR all the Jme Set the declared capacity (=number of slots) at or below the IFR capacity of the airport (but LHR, LGW, FRA, MUC, other?) 6

7 Some Consequences US airports operate with VFR for about 80% of the Jme on average As a result, they increase their average capacity (weighted by weather condijons) by about 25% compared to their IFR capacity Given extensive use of VFR, US airlines typically schedule airport operajons with reference to the VMC capacity of airports Even when caps on schedules have been set by the FAA, these caps are close to the VFR capacijes of the respecjve airports In Europe, the number of scheduled movements typically does not exceed the IFR capacity of the airport 7

8 EWR vs FRA: Scheduling Limita>ons Newark (since 2008) Frankfurt (Summers 2010 and 2011) Hourly*: total; arrs; deps 30-min: 43 total; 23 arrs; 25 deps Hourly: 81 scheduled movements + allowance for 2 unscheduled movts Slots specified within 30-minute time windows (6:00 6:29, 6:30 6:59, etc.) 10-min: 16 total; 9 arrs; 9 deps * Hourly limits vary by time of day depending on arrival departure mix, widebody operations Slots specified within 5-minute time windows EWR schedule limitations administered by FAA Slot Coordination Office; initiated 5/15/2008, extended to expiration date of 10/29/2011 FRA schedule limitations administered by German Airport Coordinator (GAC); 15 full-time staff, coordinates 16 airports with total of about 2 million annual movements; no expiration dates for schedule limitations

9 FRA Average daily schedule by month (2007) Daily demand profile movements / hour time January February March April Mai June July August September October November December Evenly distributed demand profile from 07:00 to 21:00 Hourly demand peaks at 84-movement hourly slot limit 6/3/11 FAA - FRA airport operajon analysis 9

10 FRA Max throughput es>mates (95%- ile) (movements per 15 minutes) Hourly capacity: movements 10

11 FRA Average daily schedule by month (2007) Daily demand profile movements / hour Scheduling limit VMC Capacity IMC Capacity time January February March April Mai June July August September October November December Evenly distributed demand profile from 07:00 to 21:00 Hourly demand peaks at 84-movement hourly slot limit 6/3/11 FAA - FRA airport operajon analysis 11

12 EWR Average daily schedule by month (2007) Daily demand profile Unevenly distributed demand profile with extended afternoon/ evening peak Average hourly demand peaks at about 90 movements 6/3/11 FAA - FRA airport operajon analysis 12

13 EWR Max throughput es>mates (95%- ile) (movements per 15 minutes) Hourly capacity: movements 13

14 EWR Average daily schedule by month (2007) Daily demand profile VMC Capacity IMC Capacity Unevenly distributed demand profile with extended afternoon/ evening peak Average hourly demand peaks at about 90 movements 6/3/11 FAA - FRA airport operajon analysis 14

15 Due to: (i) intensive scheduling (close or above VFR capacity) (ii) variability of the operajng capacity with weather, delays (relajve to schedule) at EWR are high, in general, strongly variable with weather Increasing over the course of the day and very high in IMC and during the peak evening hours 15

16 FRA - all runway configura>ons: schedule and delays Scheduled vs. actual movements vs. delays all weather (2007) movements/hour Minutes hour clock ARR delay min DEP and TAXI out delay min Scheduled movements actual movements 0 ARR delays and DEP delays up to +300min (including punctual movements with a delay of 0) are taken into calculation. Average calculation includes hours before 5am. Those are simply not displayed. Taxi out delays are not in accordance to the standard definition of Eurocontrols PRU, due to lack of data. 6/3/11 FAA - FRA airport operajon analysis 16

17 EWR - all runway configura>ons: schedule and delays Scheduled vs. actual movements vs. delays all weather (2007) movements/hour Minutes hour clock sched ARR delay min DEP and TAXI out delay min Series3 actual movements movts 0 ARR delays and DEP delays up to +300min (including punctual movements with a delay of 0) are taken into calculation. Average calculation includes hours before 5am. Those are simply not displayed. 17

18 Average Delays (min, vs. schedule)

19 Reliability/Predictability Due to (i) variability in performance with weather and (ii) airline scheduling pracjces, schedule reliability and predictability is much lower at US airports than at European airports Reliability of schedules in US declines sharply over the course of a day, parjcularly in the presence of poor weather 19

20 FRA - delays do not worsen over the course of the day; roughly constant schedule reliability ARR delay at FRA 3.5 average delay standard devia>on Frequency (%) am- 9am 0,9 min 24,9 min 12pm- 1pm 4,1 min 30,6 min 4pm- 5pm 1,7 min 28,4 min 5pm- 6pm 2,3 min 35,5 min ARR delay minutes 6/3/11 FAA - FRA airport operajon analysis 20

21 EWR - delays worsen during the course of the day; schedule reliability declines ARR delay at EWR 5.0% average delay standard devia>on 4.0% 8am- 9am min 29.0 min 12pm- 1pm 8.1 min 38.1 min frequency 3.0% 2.0% 4pm- 5pm 30.1 min 55.0 min 1.0% 0.0% arrival delays (1-minute intervals) 6/3/11 FAA - FRA airport operajon analysis 21

22 JFK delays worsen during the course of the day; schedule reliability declines ARR delay at EWR 5.0% average delay standard devia>on 4.0% 8am- 9am 7.6 min 46.6 min 12pm- 1pm 10.8 min 49.3 min frequency 3.0% 2.0% 4pm- 5pm 23.2 min 56.9 min 8pm- 9pm 34.7 min 65.4 min 1.0% 0.0% arrival delays (1-minute intervals) 6/3/11 FAA - FRA airport operajon analysis 22

23 Delay under IFR is much greater than delay under VFR Delay increases steadily during the day 23

24 24

25 But! since 2007 Roughly 6-8% reduction in demand at JFK and EWR between 2007 and 2010 Roughly 35% reduction in departure delay relative to schedule

26

27

28 DOT Inspector General s Report RecommendaJon 1: (10/28/10) We recommend that FAA re- examine flight caps at JFK, LGA, and EWR, basing the caps on more realisjc airport operajng condijons, air carrier scheduling pracjces, and a goal towards reducing delays to an acceptable rate. 28

29 The Bofom Line Scheduling limits or market- based demand management schemes are desirable at some airports to avoid over- scheduling and extreme condijons But the approach currently ujlized to set scheduling limits and declared capacijes lacks sophisjcajon and may yield very conservajve results, wasjng valuable potenjal capacity and excessively constraining access to some major airports Need a bever approach that considers: the interests of diverse stakeholders full range of tradeoffs between throughput and delay 29

30 The se^ng of declared capacijes (or of target levels of demand for market- based schemes) involves finding an opjmal tradeoff: throughput vs. delays and predictability/reliability [An issue which, quite surprisingly, has not been addressed in sufficient depth to date!] 30