F065-B © 2006 The MITRE Corporation. All rights reserved. System Performance and Convective Weather Kenneth Lamon, Ph.D. 15 March 2006 ©2006. All rights reserved. This is the copyright work of The MITRE Corporation and was produced for the U.S. Government under Contract Number DTFA01-01-C and is subject to Federal Aviation Administration Acquisition Management System Clause , Rights in Data-General, Alt. III and Alt. IV (Oct., 1996). No other use other than that granted to the U.S. Government, or to those acting on behalf of the U.S. Government, under that Clause is authorized without the express written permission of The MITRE Corporation. For further information, please contact The MITRE Corporation, Contracts Office, 7515 Colshire Drive, McLean, VA 22102, (703)

F065-B © 2006 The MITRE Corporation. All rights reserved. 2 June, July and August are the Peak of the Convective Weather Season Source: NOAA, FSL; measures the average area of thunderstorms over the US at 19Z, 21Z, and 23Z, weekly average

F065-B © 2006 The MITRE Corporation. All rights reserved. 3 Delays versus Scheduled Arrivals Delays = a*Weather*(Schedule- 16,500)

F065-B © 2006 The MITRE Corporation. All rights reserved. 4 Minutes of Delay versus Scheduled Arrivals Minutes = 100,000 + b*Weather*(Schedule- 16,500)

F065-B © 2006 The MITRE Corporation. All rights reserved. 5 An Airport-specific Convective Weather Metric Take an airport’s maximum storm intensity each hour Multiply by the number of scheduled arrivals that hour Sum over all hours of the day Sum over 45 airports (Divide by total number of scheduled arrivals)

F065-B © 2006 The MITRE Corporation. All rights reserved. Show Movies Here

F065-B © 2006 The MITRE Corporation. All rights reserved. 7 ATL Delays Versus Weather Score

F065-B © 2006 The MITRE Corporation. All rights reserved. 8

F065-B © 2006 The MITRE Corporation. All rights reserved ABQ ATL BNA BOS BWI CLE CLT CVG DCA DEN DFW DTW EWR FLL HOU IAD IAH IND JFK LAS LGA MCI MCO MDW MEM MIA MSP MSY ORD PBI PDX PHL PHX PIT RDU SEA SLC STL TEB TPA YYYYMM Monthly Convective Weather Score by Major Airport

F065-B © 2006 The MITRE Corporation. All rights reserved. 10 Number of Hours Having Thunderstorm Intensity>3 BROWN indicates locations where the frequency of thunderstorms increased from 2004 to 2005

F065-B © 2006 The MITRE Corporation. All rights reserved. 11 SEA PDX TPA TEB STL SLC RDU PIT PHX PHL PBI ORD MSY MSP MIA MEM MDW MCO MCI LGA LAS JFK IND IAH IAD HOU FLL EWR DFW DEN DCA CVG CLT CLE BWI BOS BNA ATL ABQ - 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 -5,00010,00015,00020,000 Weather Score Delays (>1 Hour) DTW Delay and Convective Weather Score: June  August 2004

F065-B © 2006 The MITRE Corporation. All rights reserved. 12 R-Squared Improves (a little) if Airports are Grouped

F065-B © 2006 The MITRE Corporation. All rights reserved. 13 Delays Versus Weather Score: 45 Airports R 2 = 0.67 R 2 = ,000 2,000 3,000 4,000 5,000 6,000 -1,0002,0003,0004,0005,0006,0007,0008,000 Weather*Schedule Delays (>1 Hour)

F065-B © 2006 The MITRE Corporation. All rights reserved. 14 Regression Results 2003 and 2004 Delays = *Weather*(Schedule- 15,429) + 298*Year2004 R-squared = 0.65 Minutes = 106, *Weather*(Schedule- 16,870) + 24,454*Year2004 R-squared = 0.67

F065-B © 2006 The MITRE Corporation. All rights reserved. 15 Airborne Holds Versus Weather Score 2003 and Airports

F065-B © 2006 The MITRE Corporation. All rights reserved. 16 Diversions Versus Weather Score 2003 and 2004 R 2 = 0.50 R 2 = Weather*Schedule Diversions Airports

F065-B © 2006 The MITRE Corporation. All rights reserved. 17 Cancellations Versus Weather Score 2003 and Airports

F065-B © 2006 The MITRE Corporation. All rights reserved. 18 Delays Versus Weather Score 2003, 2004, and Airports

F065-B © 2006 The MITRE Corporation. All rights reserved. 19 Airborne Holding Up in Summer 2005 Year There were 28 additional holds per day in summer 2005 compared to Airports

F065-B © 2006 The MITRE Corporation. All rights reserved. 20 Airborne Holding Up in Summer 2005 *June-August Year Holding was up sharply at ATL in summer 2005

F065-B © 2006 The MITRE Corporation. All rights reserved. 21 Schedule Delays Lasting More Than One Hour Up 16% Compared to , , , , , , , , , , , , Arrivals and Departures More than One Hour Late Relative to Schedule Source: ASPM, 45 Airports

F065-B © 2006 The MITRE Corporation. All rights reserved. 22 Delays Due to GDPs Up in 2004 and 2005 Source: OPSNET, includes delays attributed to en route centers thousands GDP GS Departure OPSNET Delays by Type, 45 Airports, June-August

F065-B © 2006 The MITRE Corporation. All rights reserved thousands ATL, ORD, PHL, EWR, LGA, IAD, IAH, JFK 37 other airports Proportion of OPSNET Delays at the 8 Most Delayed Airports Has Grown Includes delays attributed to ARTCCs. Delays up 7% overall compared to 2000 Up 23% compared to 2000 Down 39% compared to 2000 OPSNET Delays, 45 Airports, June-August

F065-B © 2006 The MITRE Corporation. All rights reserved millions ATL, ORD, PHL, EWR, LGA, IAD, IAH, JFK 37 other airports Operations, on Average, Up 9% at the 8 Most Delayed Hubs, and Down 9% Elsewhere Down 9% compared to 2000 Up 9% compared to 2000 Operations down 5% overall at OPSNET 45 airports compared to 2000 Operations June-August (OPSNET)

F065-B © 2006 The MITRE Corporation. All rights reserved. 25 ATL Simulated Minutes of Arrival Delay Using Arrival Capacity from 8/7/ ,000 20,000 30,000 40,000 50,000 60, Scheduled Arrivals Arrival Delay Minutes (Simulated) Aug , 2003 Aug 10, 2003 Aug 9, 2003 Schedule Dates 2003 Schedule Dates 2004 Aug , 2004 Aug 8, 2004 Aug 7, 2004 Aug , 2005 Aug 7, 2005 Schedule Dates 2005

F065-B © 2006 The MITRE Corporation. All rights reserved. 26 Summary Introduced a convective weather metric defined as the product of an airport’s maximum hourly thunderstorm intensity and the number of flights scheduled to land that hour Metric has several advantages –Can be aggregated and disaggregated –Small, computationally simple dataset –Yields many insights about NAS performance

F065-B © 2006 The MITRE Corporation. All rights reserved. 27 Other Findings Factoring out differences in weather and the total number of scheduled ops, hour-long delays increased substantially from 2003 to 2004 and from 2004 to 2005 –Increases caused by greater concentration of traffic at already delayed airports Airborne holding and diversions were unchanged from 2003 to 2004, however, holding increased from 2004 to 2005

F065-B © 2006 The MITRE Corporation. All rights reserved. 28 And Lastly Similar approaches for measuring system performance based on delays, cancellations, diversions, holding, etc. ought to yield results similar to those presented here.

F065-B © 2006 The MITRE Corporation. All rights reserved. Appendix

F065-B © 2006 The MITRE Corporation. All rights reserved. 30 Difference (Hours) Source: National Convective Weather Detection Number of Hours Having Thunderstorm Intensity>3 (June  August 2005 Minus June  August 2004)

F065-B © 2006 The MITRE Corporation. All rights reserved. 31 OPSNET Delays by Type, 45 Airports, June-August OPSNET Delays Up 7% Compared to , ,312 94,087 92, , , thousands GS/GDP Arrival Departure Source: OPSNET, includes delays attributed to en route centers

F065-B © 2006 The MITRE Corporation. All rights reserved. 32 Average Number of Seats per Departure has Decreased -8% -19% -15% -16% -1% -17% -10% ATLEWRIADIAHLGAORDPHL Average Seats per Departure Jul-00 Jul-05 Average number of seats per scheduled departure down in July 2005 compared to July 2000 Source: OAG

F065-B © 2006 The MITRE Corporation. All rights reserved. 33

F065-B © 2006 The MITRE Corporation. All rights reserved. 34

F065-B © 2006 The MITRE Corporation. All rights reserved. 35 Delays from GDPs: June  August 2005 In summer 2005 ATL accounted for 33% of all delays from GDPs The average duration of this type of delay was 74 minutes Source: OPSNET

F065-B © 2006 The MITRE Corporation. All rights reserved. 36 PHL Simulated Minutes of Arrival Delay Using Arrival Capacity from 7/13/05 - 5,000 10,000 15,000 20,000 25,000 30,000 35,000 40,000 45,000 50,000 55, Scheduled Arrivals Arrival Delay Minutes (Simulated) Jul , 2003 Jul 12, 2004 Schedule Dates 2003 Jul , 2004 Jul 11, 2004 Jul 10, 2004 Schedule Dates 2004 Jul , 2005 Jul 10, 2005 May 30, 2005 Schedule Dates 2005

F065-B © 2006 The MITRE Corporation. All rights reserved. 37 PHL: Actual Minutes of Arrival Delay, June  August - 5,000 10,000 15,000 20,000 25,000 30,000 35,000 40,000 45, Scheduled Arrivals Minutes of Arrival Delay (Actual) Actual Delay 2003 Actual Delay 2004 Actual Delay 2005 Source: ASPM, Each point represents one day, one outlier day excluded from 2003 and 2004

F065-B © 2006 The MITRE Corporation. All rights reserved. 38 Ordered by busiest airport, includes delays attributed to en route centers OPSNET Delays for 22 Most Delayed Airports: June  August

F065-B © 2006 The MITRE Corporation. All rights reserved. 39 ATL: Actual Minutes of Arrival Delay, June  August - 10,000 20,000 30,000 40,000 50,000 60,000 70,000 80,000 90, Scheduled Arrivals Minutes of Arrival Delay (Actual) Source: ASPM, Each point represents one day, two outlier days excluded from 2003 and 2004 Actual Delay 2003 Actual Delay 2004 Actual Delay 2005

F065-B © 2006 The MITRE Corporation. All rights reserved. 40 To Improve Airport Model Adding a variable representing amount of system- wide delay raises r-squared to 0.66 for ATL For ORD two-variable model has r-squared of 0.71

F065-B © 2006 The MITRE Corporation. All rights reserved. 41 Three-Letter Airport Abbreviations ATL - Atlanta International MDW - Chicago Midway BOS - Boston/Logan International MEM - Memphis International BWI - Baltimore-Washington International MIA - Miami International CLE - Cleveland-Hopkins International MSP - Minneapolis St Paul International CVG - Covington/Cincinnati International MSY - New Orleans International DCA - Reagan National OAK - Metropolitan Oakland International DEN - Denver International ORD - Chicago O'Hare International DFW - Dallas-Ft Worth International PBI - Palm Beach International DTW - Detroit Metro Wayne Co PDX - Portland International EWR - Newark International PHL - Philadelphia International IAD - Washington Dulles International PHX - Phoenix Sky Harbor International IAH - Houston/G Bush Intercontinental PIT - Pittsburgh International IND - Indianapolis International RDU - Raleigh-Durham International JFK - John F Kennedy International SEA - Seattle Tacoma International LAS - Las Vegas/Mc Carran International SFO - San Francisco International LAX - Los Angeles International SJC - San Jose International LGA - La Guardia SLC - Salt Lake City International - STL - Lambert-St Louis International TEB - Teterboro ABQ - Albuquerque International Sunport BNA - Nashville International CLTCharlotte/Douglas Internationa MCIKansas City International MCO - Orlando International FLL- Fort Lauderdale/Hollywood International HOU - William P. Hobby Airport SAN - San Diego International TPA - Tampa International

F065-B © 2006 The MITRE Corporation. All rights reserved. 42 Glossary ASPMAviation System Performance Metrics FSLForecast Systems Laboratory GDPGround Delay Program NASNational Airspace System NOAANational Oceanic & Atmospheric Administration OAGOfficial Airline Guide OPSNETOperations Network USUnited States ZZulu Time