AD-AI05 112 WILSON-MILL ASSOCIATES INC WASIIINGTON DC F/B 1/5 ICOMMU TER AIRLINE FOI4ECASTS.(U) MAY 81 H MEUVILLE. C STARRY, 6 8ERNSTEIN DOT-FA79WAX-138 UNCLASSIFIED FAA-APO-81-7 NL m ME ummii omhmhmmMENm Iliii 11111 AA- yz5-,7 - k .14 ,- Aro, N fitt COMMUTER AIRLINE FORECASTS FINAL REPORT May 1981 Hazel Medville Claire Starry Gerald Bernstein Prepared For: U.S. DEPARTMENT OF TRANSPORTATION FEDERAL AVIATION ADMINISTRATION Office, of Aviation Policy and Plans Washington, D.C. 20591 Io7_ _ _~'' Technical keport Documentation Page 1. Report No. 2. Government Accession No. 3. Recipient's Catalog No. d. Titl anid Subtitle S twe caurwter Airline Forecasts ..... .. t. __18. Performing Orgo . Report No. 1/z? Hazel/i~edville, Claire, Starry an Gerald/iBernstein i - 9; PN t itlniml-, 1 4 i clu fa - - e'oW---ss r - 10. Work Unit No. (TRAIS) Wilson Hill Associates, Inc. 1025 Vermont Avenue, N.W. .. %shington, D.C. 20005 D-FA79W--138 -- ~ ype Tof oJ~' Cove,.red 12. Sponsoring Agency Name and Address Forecast 1981-1992, Federal Aviation Adninistration , l. ..- , / Office of Aviation Policy and Plans . - . ............ _" 800 Independence Avenue, S.W. 14. Sponsoring Agency Code Washington, D.C. 20591 FAA 15. Supplementary Notes I Abstract This publication presents forecasts of cammuter air carrier activity and describes the models designed for forecasting Contenninous United States, Puerto Rico and the Virgin Islands, Hawaii, and individual airport activity. These forecasts take into account the recent dynanic growth of the commter industry and the effects of the changed operating and marketing enviromient created by the Airline Deregulation Act of 1978. A separate forecast is provided for cammter cargo activity. Lastly, ndeling approaches were evaluated to forecast activity at individual airports which the FAA expects to utilize in preparing its annual Terminal Area Forecast. Descriptions of the models are provided in the pendices. The models developed under this contract represents an initial effort by FAA to describe and Forecast a major sector of Aviation. It is anticipated that there will be further refinvvents to the models as the comuter industry evolves and additional data becomes available. 17. Key Words 18. Distribution Statement Camuter Air Carrier, Caunuter Cargo, Document is available to the public Passenger Enplanements, Revenue through the National Technical Passenger Miles, Operations, Seat Infonnation Service, Springfield, Miles Virginia 22151 19. Security Cleseif. (of this report) 20. Security Clesiif. (of thil page) 21. No. of Pages 22. Price Unclassified Unclassified 109 Form DOT F 1700.7 (8-72) Reproductlio of Completed page outhorlized 01,'J Preface The contents of this report reflect neither a position or an official policy of the Department of Transportation. This document is disseminated to the public in the interest of information exchange. The United States Government assumes no liability for the contents of this document or use thereof. The work culminating in this report, Commuter Airlines Activity Forecasts, was performed under Department of Transportation contract number DOT-FA79WAI-138 with Wilson Hill Associates Inc., Washington, D.C. as the prime contractor and SRI Ijiternational, Menlo Park, CA as the subcontractor. The contract deliverables also included documentation of the data sources and files used in the development of the model which was delivered to FAA in November 1980 and a description of the computerized models for passenger commuter service delivered in May 1981. Interim reports, that were prepared during the development of the base data and the models for presentation at briefings, have been updated and incorporated where appropriate into this document. Accession For NTis rRA&I DTIC TAB U!,-ri.!r;:nced E Ju tiiC ")t i0 By Distribution/ Availability Codes :Avail cnd/or Dist I Special hi i i im i i I I Ii TABLE OF CONTENTS SECTION PAGE PREFACE .................................................. i ACKNOWLEDGE14ENTS .......................................... vi INTRODUCTION ............................................. viii 9 CHAPTER 1: PASSENGER MODELS ................................ 1 Background ..................................... 1 U.S. Conterminous States ...................... 4 Hawaii and Puerto Rico/Virgin Islands ....... 12 CHAPTER 2: CARGO MODELS .................................. 20 CHAPTER 3: LOCAL MODEL EVALUATION ....................... 25 CHAPTER 4: HISTORIC DATA COLLECTED FOR MODEL DEVELOPMENT ......................... 34 APPENDIX A - Description and Specification of Econometric Models ............................ A-i - Conterminous US - Hawaii and Puerto Rico/Virgin Islands - Cargo APPENDIX B - Quarterly Data and Forecasts ................ B-i - Definitions - Conterminous US Consensus Growth High Growth - Hawaii - Puerto Rico and the Virgin Islands APPENDIX C - Local Model Evaluation Addendum ............. C-i - Variable Definitions - Sample City Pair, Market Forecast Model Ii LIST OF FIGURES PAGE FIGURE 1. CERTIFIED AIR ROUTE CARRIERS INCLUDED IN THE COMMUTER FORECAST ......................... 3 FIGURE 2. DEFINITIONS AND SOURCES OF VARIABLES USED IN CONTERMINOUS US MODEL ........................ 7 FIGURE A-1. DEFINITIONS OF VARIABLES ........................ A-2 •i. II iii 1 I. LIST OF TABLES PAGE TABLE 1. HISTORICAL AND FORECAST ANNUAL AVERAGE RATES OF GROWTH FOR CONTERMINOUS US MODEL INDEPENDENT VARIABLES .............................. 6 TABLE 2. HISTORICAL AND FORECAST ANNUAL AVERAGE RATES OF GROWTH FOR SELECTED CONTERMINOUS US COMMUTER AIRLINE ACTIVITY VARIABLES ....... 8 TABLE 3. ACTUAL AND FORECAST VALUES OF CONTERMINOUS US COMMUTER AIRLINE ACTIVITY, CONSENSUS SCENARIO ..................................... 10 TABLE 4. ACTUAL AND FORECAST VALUES OF CONTERMINOUS US COMMUTER AIRLINE ACTIVITY, HIGH GROWTH SCENARIO ...................................... 11 TABLE 5. HISTORICAL AND FORECAST ANNUAL AVERAGE RATES OF GROWTH FOR HAWAII AND PUERTO RICO/ VIRGIN ISLAND MODELS INDEPENDENT VARIABLES... 13 TABLE 6. HISTORICAL AND FORECAST ANNUAL AVERAGE RATES OF GROWTH FOR SELECTED COMMUTER AIRLINE ACTIVITY VARIABLES: HAWAII AND PUERTO RICO/VIRGIN ISLANDS ................... 15 TABLE 7. ACTUAL AND FORECAST VALUES OF COMMUTER AIRLINE ACTIVITY FOR HAWAII (IN MILLIONS) .... 16 TABLE 8. ACTUAL AND FORECAST VALUES OF COMMUTER AIRLINE ACTIVITY FOR PUERTO RICO/VIRGIN ISLANDS IN MILLIONS .............................. 17 TABLE 8-A. ESTIMATED EFFECT OF NEW COMMUTER CARRIER ON COMMUTER AIRLINE ACTIVITY FOR HAWAII (IN MILLIONS) .... . .........................19 TABLE 9. CARGO CARRIED BY COMMUTER AIRLINES ............. 22 TABLE 10. HISTORICAL AND 'FORECAST ANNUAL AVERAGE GROWTH RATES FOR POUNDS OF CARGO CARRIED ..... 23 TABLE 11. COMPARISON OF LOCAL MODEL RESULTS .............. 26 TABLE 12. LOCAL GROWTH RATE MODEL ...................... 28 TABLE 13. LOCAL TO NATIONAL GROWTH MODEL .............. 30 TABLE 14. MARKET FORECAST MODEL ........................... 32 iv - -J -7 1 ' LIST OF TABLES (CONT'D.) PAGE TABLE 15. OAG SCHEDULED OPERATIONS AND OPERATIONS REPORTED BY FAA TOWERS .......................... 37 TABLE 16. BOARDING AND LOAD FACTORS ...................... 38 TABLE A-i. ECONOMETRIC ESTIMATES FOR REDUCED FORM EQUATIONS CONTERMINOUS US MODEL ................ A-5 TABLE A-2. ECONOMETRIC ESTIMATES FOR REDUCED FORM EQUATION: HAWAII AND PUERTO RICO/ VIRGIN ISLANDS .............................. A-8 TABLE A-3. ECONOMETRIC ESTIMATES FOR CARGO EQUATIONS... A-10 F2 V¢ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ 1 ACKNOWLrDGENTS The Commuter Airlines Activity Forecasts were prepared for the Planning Analysis Division of the FAA Office of Aviation Policy and Plans by Hazel Medville of Wilson Hill Associates, Inc. and by Claire Starry and Gerald Bernstein of SRI International. Regina VanDuzee of the Aviation Activities Division, represented 4 FAA as the technical officer on the project and contributed much from her knowledge of the commuter airline industry. As part of the task force to produce reasonable long-term forecasts of this dynamic industry, a committee selected from airlines, manufacturers, trade associations, and local and Federal government reviewed and commented on the forecasting effort throughout the study. Individuals serving on the committee included: Bar Harbor Airlines - Allyn J. Caruso and Jeff Jenner Baltimore Washington International Airport - Jim Truby Civil Aeronautics Board - Robin Caldwell, Bruce Goldberg, Paul Gavel, and Mary Vavrina Commuter Airline Associ~tion of America - Steve Smith and Alan R. Stephen The de Havilland Aircraft of Canada, Ltd- Joseph Gude and Arthur F. Toplis Ransome Airlines - Larry Crawford San Jose Municipal Airport - Raul Regalado Scheduled Skyways, Inc. - Raymond A. Young III S.M.B. Stage Line - Robert Grammer Transportation Systems Center - Robert N. Tap James Hines of Wilson Hill Associates and Marika Garskis of SRI International contributed to this report by collecting and com- puterizing the data base and forecast models and by evaluating vi the data to assure contiruty and accuracy. Gene Mercer of the FAA Office of Aviation Policy and Plans, Forecasting Branch provided guidance and support throughout the forecasting process. Hazel Madville Project Manager Wilson Hill Associates vii = - INTRODUCTION This publication presents forecasts of commuter air carrier activity and describes the models designed for forecasting Conterminous United States, Puerto Rico and the Virgin Islands, Hawaii, and individual airport activity. As the forecasts were developed, the forecasting team relied heavily on advice of members of the Forecast Committee whose knowledge of the real world problems of operating a commuter airline, handling expanding commuter traffic at airports and p,.ducing aircraft for commuter airlines was an invaluable aid. These forecasts take into account the recent dynamic growth of the commuter industry and the effects of the changed operating and marketing environment created by the Airline Deregulation Act of 1978. Computerized models were used to prepare forecasts of passenger enplanements and operations through 1992 for the 48 conterminous states, Hawaii and Puerto Rico/Virgin Islands. A forecast for Alaska is not included because the structure of the commuter industry there differs markedly from that of the other states. A separate forecast is provided for commuter cargo activity which depends heavily on information and insights obtained from a number of major commuter cargo carriers. Lastly, wodeling approaches were evaluated to forecast activity at individual airports which the FAA expects to utilize in preparing its annual Terminal Area Forecast. Descriptions of the models are provided in the Appendices. If further information is desired concerning the structure and use of these models, interested persons can contact the Office of Aviation Policy and Plans, Forecasting Branch, Federal Aviation Administration. Viii 4L CHAPTER I. PASSENGER MODELS BACKGROUND Commuter airlines constitute a growing sector of the United States commercial passenger aviation industry. During a time when certificated route air carriers have shown modest annual growth, the commuter air carriers have shown a 47.7 percent gain in the number of passengers enplaned within the conterminous States (Continental United States) and a 16.2 percent gain in scheduled operations. This is impacting both the national air system and groundside handling of passengers, cargo, and aircraft. Service by these smaller aircraft at hub airports means more operations by a mixed fleet of aircraft which is adding to air traffic controllers workload. Due to the flexibility of these aircraft, service to non-FAA controlled airports has also increased, and with the increase, requests for sophisticated landing and takeoff systems to insure safe, timely, and non-weather dependent flights to cities and towns serviced by commuter airlines have also increased in number. At airports gate and terminal space shortages have occurred as carriers provide additional service with commuter aircraft and more frequently scheduled flights. Commuter airlines were first required to register with the Civil Aeronautics Board (CAB) in 1969 and operated under CAB Economic Regulation Part 298. These airlines were defined as "operators which perform, pursuant to published schedules, five round trips per week between two or more points or carry mail." They were subject to limited regulatory and reporting requirements, were allowed free access to all markets and had no route protection. Under Part 298, the maximum size of commuter aircraft was first set at 12,500 pounds maximum takeoff weight and nineteen passen- ger capacity. This was later increased to a 7,500-pound payload with a maximum of 30 seats. The Airline Deregulation Act of 1978 permitted larger aircraft with the CAB regulation finally setting maximum capacity at 60 passengers. 1 Passage of the Deregulation Act accelerated the industry's growth because it (1) facilitated withdrawal of large certifica- ted carriers from uneconomic short-haul routes, their commuter replacements offered frequent well-timed flights generating ad- ditional traffic, (2) permitted use of larger aircraft more attractive to the public and offering greater capacity, (3) made commuters eligible for the Equipment Loan Guarantee Program which aids them in the purchase of larger, improved aircraft, and (4) encouraged more joint fares and interline agreements with major carriers. Before 1979, only two commuters, Air New England and Air Mid- west, had become certificated carriers. The Deregulation Act made the certification process easier and less costly and the number of certificated commuters grew. These airlines either applied directly for certification or for certification on dor- mant routes no longer operated by the large carriers. A list of commuter carriers now wholly or partially certificated (dual authority) with their first Form 41 reporting dates is shown in Figure 1. One of the first tasks of the working group (with the advice of the Forecast Committee) was to decide which carriers should be included in the forecast in order to properly describe the industry. The final decision was to include both part 298 and certificated passenger carriers utilizing aircraft seating 60 passengers or less and providing regularly scheduled service in accordance with published schedules. Cargo carriers included were those operating aircraft with a maximum payload of 18,000 pounds. These carriers typically have stage lengths of under 250 miles and serve as feeder lines to the trunk carriers with 11 approximately 70 percent I of the passengers interlining in 1979. Commuter Airline Association of America, Annual Report, November 1980, p. 6. 2 UIII a i CARRIER NAME ist Report on Form 41 Air Midwest 11/76 Air New England 1/751 Air Wisconsin 7/79 Altair Airlines 1/79 Cochise 1/79 Golden West 2/79 Mississippi Valley 6/79 New Haven (New Air) 5/79 Sky West Aviation 7/79 Southeast 7/79 Swift Aire 1/79 Apollo 5/79 Big Sky 6/79 I Empire 1/80 Imperial 12/79 1. 1st data month used in this study FIGURE 1. CERTIFIED AIR ROUTE CARRIERS INCLUDED IN THE COMMUTER FORECAST 3 Past modeling efforts have produced models in which the growth 2 of commuter activity was linked to air carrier activity. Suffi- cient data have been collected during this project to permit the use of more sophisticated and improved modeling techniques. Forecast models of the commuter airline activity were developed for the conterminous United States (48 contiguous states and the District of Columbia), for the State of Hawaii, and for the U.S. Carribean areas, Puerto Rico and U.S. Virgin Islands. These latter two areas were considered significantly different from conterminous US to warrant separate analysis. A technical dis- cussion of the models is provided in Appendix A. CONTERMINOUS UNITED STATES As indicated earlier, commuter activity has been increasing much faster than trunk and regional airline activity. Although this growth has not shown signs of tapering off, new market opportun- ities and the growth of per capita income will probably slow down, so that the growth in commuter activity will remain strong, but will not be as rapid as in the late 1970s. Market saturation, which will likely occur sometime after the forecast period, will eventually limit the rate of increase in commuter airline activity. The generation of new origin destination pairs significantly in- fluences the rate of growth in commuter enplanements and opera- tions. The number of origin destination pairs increased about 15 percent per year during the late 1970s, and although many new market opportunities are expected to emerge in the 1980s, this growth rate is forecast to gradually taper off to an average of 8 percent to 9 percent per year during the 12-year forecast period, 1981 to 1992 as shown in Table 1. 2 See Systems Analysis and Research Corporation (SARC), "Forecasts of Commuter Airlines Activity," Report No. FAA-AVP-77-28 (July 1977) and "Update of Commuter Forecast Model," Final Report (May 1978). I 4 n I I I I . Another important factor is general economic growth, as represented by constant dollar gross national product (GNP). As a rule air carrier traffic has grown faster than GNP, and commuter activity is no exception. Forecasts of constant dollar GNP growth in the 1980s were obtained from Wharton Econometric Forecasting Associates, Inc., and formed the basis for the scenario summarized in Table 2. Another scenario, based on lower GNP and origin-destination pair growth, was also examined. Other independent variables included in the model are average seating capacity, fuel prices, automobile operating costs, average stage length, average distance per enplanement, and a variable to capture the effects of deregulation. The historical and average forecast growth rates for these variables are provided in Table 1; actual quarterly data and quarterly fore- casts are shown in Appendix B. The model forecasts seat miles, revenue passenger miles, opera- tions, enplanements, and an inflation adjusted index of passenger fares charged. Definitions of these variables, as well as the independent variables used, are listed in Figure 2. The model results indicate continued rapid growth in commuter activity, although at rates below those experienced in the late 1970s (see Table 2). The ability of commuters to find new market niches will decline as market saturation occurs, but it appears complete market saturation will not occur during the forecast period. Changes in the regulatory environment or re- duced assistance for essential air service may inhibit growth later in the 1980s. By the end of the 1980s and early 1990s, the rate of increase will be about half the rate recorded during the late 1970s, which started from a very small base. Actual enplanements are expected to reach at least 25 million by 1987 and more than 37 million in 1992. I, 5\ - II TABLE 1. HISTORICAL AND FORECAST ANNUAL AVERAGE RATES OF GROWTH FOR CONTERMINOUS US MODEL INDEPENDENT VARIABLES PERCENT GROWTH ,I FORECASTED 1981-1992 ACTUAL HIGH GROWTH LOW GROWTJ 1975-1979 SCENARIO SCENARIO Gross National Product (1972 dollars) 4.5 3.9 3.3 Origin-Destination City 15.0 9.0 8.0 Pairs (O-D) Average Seating Capacity 1.8 4.4 4.4 Fuel, in 1972 Cents Per Gallon 8.0 2.5 2.5 Index of Automobile Operating Costs Relative to General Inflation 3.1 0.1 0.1 Average Stage Length n.a3 3.0 3.0 Average Distance Per Enplanement 2.8 2.1 2.1 1. GNP growth based on Wharton Econometrics Forecasting Associates, Inc; forecast dated March 1980. 2. GNP growth based on a more consecutive growth than the Wharton forecast prepared by Claire Starry, SRI International. 3. Average stage length between 1975 and 1979 was 106.7 miles. The lowest value was for 1978 second quarter (96.1 miles) and the highest values was for 1979 first quarter (115.2 miles). i 6 DEPENDENT VARIABLES DEFINITIONS/SOURCES Passenger Enplanements From Civil Aeronautics Board (CAB) Part 298 data and selected Form 41 data. Revenue Passenger miles Calculated by multiplying enplaned passen- gers by distance flown. Data from CAB Part 298 and selected Form 41 data. Operations OAS scheduled flights multiplied by 2 to compute takeoffs and landing operations. Seat Miles Calculated using scheduled flights by aircraft type from the Official Airline Guide (OAG) computer tapes multiplied by average segment (distance), also OAG data, and then by number of seats normally installed in the aircraft type. Fare Index Consumer Price Index on all airline faces (no separate one for commuters is avail- able) divided by the index of general Lnfla- tion. Data from Department of Labor, "Con- sumer Price Index," and Department of Com- merce, "Survey of Current Business," var- ious issues. INDEPENDENT VARIABLES Origin-Destination Pairs Sum of the number segments provided by each airline, from CAB Part 298 data. Gross National ?roduct in constant 1972 dollars, from Department of Commerce, "Survey of Current Susiness." Auto Operating Cost Index Consumer Price Index for owner onerated transportation, adjusted for general infla- tion. Data from Department of Labor, "Con- sumer Price Index." Fuel Costs Price for fuel paid by commuters, approxi- mately equal to retail price per gallon of aviation gasoline (taxes not included) as reported in Department of Energy, "Monthly Energy Review," various issues. Average Seating Capacity Computed from mix of aircraft reported in OAG computer tapes ieighted by scheduled flights. Average Distance Flown Mean trip length in miles per enplaned oas- senger, from CAB ?art 298 data and selected Form 41 data. Average Stage Length Mean stage .enqth per operation, from OAG .omputer tapes. Seasonal Factors 3ecause quarterly data were used, sei 3onal adjustment factors were included in the model. The variable is equal to a I or 0. Deregulation A ':aciable to capture the effects of deregulation, before second quarter 1978 the variable is a zero and from that date on it is equal to 1. FIGURE 2. DEFINITIONS AND SOURCES OF VARIABLES USED IN CONTERMINOUS UNITED STATES MODEL 7 TABLE 2. HISTORICAL AND FORECAST ANNUAL AVERAGE RATES OF GROWTH FOR SELECTED CONTERMINOUS US COMMUTER AIRLINE ACTIVITY VARIABLES PERCENT GROWTH ACTUAL FORECAST 1975-1979 1981-1985 1986-1992 1981-1992 CONSENSUS SCENARIO Seat Miles 19.1 10.6 10.1 10.3 Revenue Passenger Miles 23.8 14.0 10.5 11.9 Operations 16.2 6.6 5.2 5.7 Passenger Enplanements 47.7 11.6 8.4 9.6 Fare Index nil 0.7 1.6 1.2 HIGH GROWTH SCENARIO Available Seat Miles 19.1 11.2 10.6 10.8 Revenue Passenger Miles 23.8 15.3 10.9 12.7 Operations 16.2 7.5 5.6 6.1 Passenger Enplanements 47.7 12.8 8.8 10.4 Fare Index nil 1.3 2.2 1.8 8 I Even with a less optimistic scenario, the forecast rate of growth for commuter activity is higher than for the large certi- ficated carriers. Under the consensus scenario, there are fewer new market opportunities available to the commuters, and slower growth in GNP limits the increase in air travel. The inflation adjusted index of fares charged by commuters is expected to increase gradually. During the mid to late 1970s, fares increased at about the same rate as inflation. With higher fuel prices and expanded service areas, the average fare is forecast to increase about one to two percentage points above inflation. Forecasts for the two scenarios are given in Tables 3 and 4. In the consensus growth scenario, revenue passenger miles and seat miles increased by about 330 percent between 1979 and 1992, while enplanements increase by about 235 percent and operations by about 150 percent. Longer distances flown by passengers and larger aircraft account for the differences. In the high growth scenario, seat miles and revenue passenger miles increase by over 350 percent between 1979 and 1992, enplanements by about 270 percent and operations by 170 percent. The average load factor increases from 51 percent in the late 1970s to 55 percent in the consensus growth and 57 percent in the high growth scenario by 1992. The fares charged are higher under the high growth scenario, reflecting the ability of the carriers to raise fares when demand is strong. 9 ~~- - ~ - - - ~ - _ _ _ _ Table 3 ACTUAL AND FORECAST VALUES OF CONTERMINOUS US COMMUTER AIRLINE ACTIVITY, CONSENSUS SCENARIO SEAT PASSENGER PASSENGER COMMUTER MILES MILES ENPLANEMENTS OPERATIONS YEAR (in millions) (in millions) (in millions) (in millions) Actual: 1975 1415.050 616.09 5.26 2.036 1976 1482.779 674.90 5.83 2.214 1977 1668.369 841.32 7.05 2.531 1978 1913.835 1031.09 8.55 2.812 1979 2843.252 1447.06 11.04 3.677 Forecast: 1981 3863.7 1803.25 13.50 4.843 1982 4328.9 2101.44 15.36 5.228 1983 4778.0 2396.07 17.17 5.559 1984 5260.6 2713.66 19.01 5.903 1985 5788.2 3050.45 20.93 6.244 1986 6359.4 3421.30 22.92 6.615 1987 7013.6 3817.38 25.03 6.985 1988 7748.4 4239.24 27.26 7.354 1989 8562.5 4694.30 29.61 7.734 1990 9430.0 5171.75 32.02 8.114 1991 10382.6 5676.16 34.50 8.497 1992 11382.6 6217.55 37.11 8.890 10 i -" Table 4 ACTUAL AND FORECAST VALUES OF CONTERMINOUS US COMMUTER AIRLINE ACTIVITY, HIGH GROWTH SCENARIO SEAT PASSENGER PASSENGER COMMUTER MILES MILES ENPLANEMENTS OPERATIONS YEAR (in millions) (in millions) (in millions) (in millions) Actual: 1975 1415.05 616.09 5.26 2.036 1976 1482.78 674.90 5.83 2.214 1977 1668.37 841.32 7.05 2.531 1978 1913.84 1031.09 8.55 2.812 1979 2843.25 1447.06 11.04 3.677 Forecast: 1981 3884.8 1819.97 13.63 1982 4341.4 2123.98 15.53 1983 4811.6 2445.67 17.53 1984 5348.6 2812.89 19.70 1985 5943.8 3213.89 22.05 1986 6570.0 3651.49 24.46 1987 7283.9 4117.10 26.99 1988 8078.7 4596.32 29.55 1989 8957.8 5106.49 32.21 1990 9894.4 5635.58 34.88 1991 10906.6 6191.64 37.63 1992 12010.3 6788.63 40.52 11 |7 HAWAII AND PUERTO RICO/VIRGIN ISLANDS MODELS The models developed for Hawaii and Puerto Rico/Virgin Islands are structured differently than the national model. Reasons for this difference include the structure of the 48-state market versus those of Hawaii and Puerto Rico/Virgin Islands and the lack of data available for these latter two areas. Over 70 per- cent of commuter passengers in Hawaii are state residents, and most of them are on business. Most tourists use the larger inter-island carriers, Aloha and Hawaiian Airlines. Because level land for airports is very scarce in Hawaii, the possibili- ties for the construction of new airports is limited. Growth will come primarily from increased business related travel and from local and tourist recreation or personal business related. trips. Commuter aircraft used in Hawaii are small, about 4.5 seats per aircraft. However, the establishment of a new carrier operating 60-seat aircraft will raise the average substantially to about 24 in 1982. Table 5 gives the historic and forecast growth rates of the independent variables used in the models. Disposable income (in constant dollars) is an important factor influencing enplane- ments in Hawaii. In general, disposable income will be growing faster than it did in the late 1970s, but will not increase as fast as for the United States as a whole. Average distances flown in Hawaii are short--about 85 to 90 miles. Because no new airports are being planned before 1992, it is likely that this variable will not change significantly in the forecast period. Puerto Rico/Virgin Islands commuter activity experienced little growth during the period from 1975 to 1979. Tourism, which pro- vides an important part of the region's commuter passengers, leveled off, and the construction of a new airport reduced the 12 {{{| | -I | •I TABLE 5. HISTORICAL AND FORECAST ANNUAL AVERAGE RATES OF GROWTH FOR HAWAII AND PUERTO RICO/VIRGIN ISLAND MODELS INDEPENDENT VARIABLES PERCENT GROWTH ACTUAL FORECAST 1975-1979 1981-1992 Hawaii Disposable Personal Income (1972 dollars) 2.2 2.5 Average Stage Length 2.0 0.6 Average Distance per Passenger 2.0 0.6 Average Size of Aircraft -0.9 2.0 Puerto Rico/Virgin Islands Personal Income (1972 dollars) 4.8 3.5 Average Stage Length 0.2 nil Average Distance per Passenger 0.2 nil Average Size of Aircraft -1.0 nil *A new commuter carrier has started operations in Hawaii utilizing 60 passenger aircrafts. This should affect the average size of aircraft and the forecasts of operations, revenue passenger miles, and enplanements. 13 9--I l I | II i need for some short-haul traffic. As with Hawaii, the average distance flown in Puerto Rico/Virgin Islands is short, about 65 to 70 miles, but the average size of aircraft is about 15 seats. There does not appear to be'any potential for adding new, longer --haul markets. Based on the assumed changes in the independent variables described above, forecasts of commuter activity in the two areas were developed. Summary growth rates are provided in Table 6. Seat miles in Hawaii are forecast to increase about 30 percent between 1981 and 1985 while passenger miles are forecast to increase 33 percent for the same period. This sudden spurt in traffic is the result of the start-up of a new commuter carrier operating 60-passenger YS-lls and offering fares substantially under the certificated carrier rate. After the carrier becomes fully established, growth is expected to level off during the last part of the decade. This pattern is a reversal of the one recorded during the late 1970s, when passenger miles increased twice as fast as seat miles. Enplanements will be growing somewhat slower than in the late 1970s, but still at 8 to 11 percent per year. Operations, which increased dramatically in the late 1970s as short trips to less traveled parts of Hawaii became prevalent, will increase rapidly as the new carrier adds aircraft and departures, but will stabilize after 1988. The stagnant condition of commuter acitvity in Puerto Rico/Virgin Islands is forecasted to continue. Seat miles actually declined in the late 1970s, but are expected to inc- rease between 1.5 percent and 2.0 percent per year. Passenger miles are forecast to grow slightly faster, as are enplanements. Operations should increase at a slightly slower rate than seat miles because of the addition of larger aircraft to the fleet in 1980. There does not appear to be reason for average segment length to change. 14 TABLE 6. HISTORICAL AND FORECAST ANNUAL AVERAGE RATES OF GROWTH FOR SELECTED COMMUTER AIRLINE ACTIVITY VARIABLES: HAWAII AND PUERTO RICO/VIRGIN ISLANDS PERCENT GROWTH ACTUAL FORECAST 1975-1979 1981-1985 1986-1992 1981-1992 Hawaii Seat Miles 9.8 30.5 4.2 13.8 Passenger Miles 19.6 33.6 3.9 14.7 Passenger Enplanements 18.0 33.7 3.7 14.6 Operations 23.5 16.4 6.6 10.4 Puerto Rico/Virgin Islands Seat Miles -0.2 1.6 1.7 1.7 Passenger miles 3.7 1.9 2.1 2.0 Passenger Enpilanements 2.7 1.9 2.1 2.0 Operations 2.7 1.6 1.7 1.7 TABLE 7. ACTUAL AND FORECAST VALUES OF COMMUTER AIRLINE ACTIVITY FOR HAWAII (IN MILLIONS) SEAT PASSENGER PASSENGER COMMUTER YEAR MILES MILES EN PLAN EMENTS OPERATIONS Actual 1975 30.62 14.60 0.178 0.095 1976 36.13 18.59 0.220 0.113 1977 39.55 20.50 0.245 0.136 1978 44.21 28.23 0.349 0.200 1979 44.52 29.90 0.345 0.221 Forecast 1981 128.63 74.45 0.864 0.253 1982 246.20 148.25 1.726 0.326 1983 290.06 180.68 2.103 0.373 1984 329.36 206.71 2.404 0.418 1985 374.35 236.97 2.757 0.464 1986 414.80 267.39 3.108 0.510 1987 425.29 294.60 3.422 0.544 1988 477.20 316.55 3.673 0.596 1989 489.02 322.53 3.736 0.633 1990 503.81 326.79 3.778 0.670 1991 517.22 331.24 3.822 0.709 1992 531.39 335.84 3.867 0.750 16 TABLE 8. ACTUAL AND FORECAST VALUES OF COMMUTER AIRLINE ACTIVITY FOR PUERTO RICO/VIRGIN ISLANDS (IN MILLIONS) SEAT PASSENGER PASSENGER COMMUTER YEAR MILES MILES ENPLANEMENTS OPERATIONS Actual 1975 168.35 82.36 1.284 0.254 1976 162.69 78.45 1.200 0.244 1977 157.64 78.88 1.185 0.242 1978 146.54 83.57 1.239 0.266 1979 166.80 95.26 1.428 0.283 Forecast 1981 165.18 90.56 1.372 0.275 1982 167.70 92.22 1.397 0.279 1983 170.32 93.95 1.424 0.284 1984 173.02 95.74 1.451 0.288 1985 175.82 97.59 1.479 0.293 1986 178.81 99.50 1.508 0.298 1987 181.71 101.48 1.538 0.301 1988 184.81 103.53 1.569 0.308 1989 188.04 105.66 1.601 0.313 1990 191.36 107.86 1.634 0.319 1991 194.80 110.13 1.669 0.325 1992 198.37 112.49 1.704 0.330 17 Actual forecasts are presented in Table 7 for Hawaii and Table 8 for Puerto Rico/Virgin Islands; the quarterly forecasts are shown in Appendix B. Forecasts of average fares were not developed because data were not available. Total percentage in- creases in commuter activity in Hawaii are similar to those forecast for the 48 states. Puerto Rico/Virgin Island commuter activity is forecast to grow under 20 percent over the 12-year forecast period. Adjustments were made to the model forecast for Hawaii to show the expected effect of Mid-Pacific Airlines entering the commuter airline market with 60-seat YS-ll's in spring of 1981. The airline is expected to attract passengers from the trunk carriers. By 1988 their options on eleven aircraft should be fulfilled and the growth in passengers should level off. Table 8-A reflects our assumptions of Mid-Pacific Airlines' effect on commuter airline activity in Hawaii. 1 18 __ _ _ _ _ _ _ _ _ _ _ _ _ __ _ _ _ _ _ _ _ _ _ _ _ _ _ I& TABLE 8A. ESTIMATED EFFECT OF NEW COMMUTER CARRIER ON COMMUTER AIRLINE ACTIVITY FOR HAWAII (IN MILLIONS)* SEAT PASSENGER PASSENGER COMMUTER YEAR MILES MILES ENPLANEMENTS OPERATIONS Actual 1975 30.62 14.60 0.178 0.095 1976 36.13 18.59 0.220 0.113 1977 39.55 20.50 0.245 0.136 1978 44.21 28.23 0.349 0.200 1979 44.52 29.90 0.345 0.221 Forecast 1981 138.00 89.73 1.035 .279 1982 257.18 164.38 1.905 .337 1983 309.76 198.87 2.304 .382 J 1984 343.89 227.01 2.627 .427 1985 390.83 259.39 3.000 .473 1986 433.35 292.04 3.376 .520 1987 446.04 321.60 3.714 .561 1988 500.29 345.93 3.989 .612 1989 514.53 354.41 4.077 .656 1990 532.48 361.30 4.145 .700 1991 548.70 368.44 4.216 .746 1992 565.84 375.82 4.288 .794 * MID PACIFIC AIRLINES ENTERED THE HAWAIIAN COMMUTER MARKET WITH YS-11, 60 passenger aircraft in spring 1981. 19 - 4 CHAPTER 2: CARGO MODEL All-cargo commuter activity has also experienced record growth during the past several years. The use of overnight small pack- age delivery by commercial and manufacturing establishments has been augmented by the growth in the service and financial sec- tors, the more widespread use of high technology products, and more effective inventory control. Commuter airlines have re- sponded to this market opportunity by increasing their common carrier and contract operations. Future opportunities for commuters are good. The growth areas of the economy are those that tend to support the use of air freight and overnight deliveries, but the poor facilities available to all-cargo commuters at many airports may limit growth. A lack of sufficient facilities has reduced these operations at many airports, while other airports that have provided adequate space have attracted considerable growth in all-cargo operations. A second growth-limiting factor is the greater use of trucking operations for many short haul market pairs. Deregulation of air and truck services allows for more flexibility in intermodal operations that may result in the demise of air service for some of the routes easily and efficiently served by trucks. Routes where trucks are able to provide overnight service are most likely to show declines. Other routes, longer and more heavily used, should continue to experience substantial growth. Aircraft used by all-cargo carriers will increase in size, at rates higher than for the passenger carriers. In a survey of some of the major commuter cargo airlines, many reported planned retirement of their aircraft in the 3000-pound payload range with replacement by larger aircraft with payload capacities of up to 18000 pounds or even larger jet aircraft. The use of 20 larger aircraft implies that operations will not increase, or increase slowly, even though cargo carried is forecast to increase substantially. Forecasts developed for pounds of cargo carried by all commuter airlines are shown in Table 10. Because of data limitations, it was not feasible to identify cargo carried by all-cargo carriers, excluding Federal Express, as was the specification for this study. For the years 1975 to 1977 these data were collected and are given in Table 9. For the three years available, pounds of cargo carried by all-cargo airlines excluding Federal Express accounted for about 52 percent to 53 percent of the total for all cargo carriers. All-cargo operators indicated that this percentage share should remain at or exceed the historic level because of the limited cargo carrying capacity of most passenger commuter operations. Growth in cargo carried by air freight forwarders, however, should exceed the growth in the other all-cargo commuter airlines. In general, the percent rate of increase in cargo carried by the all-cargo carriers (excluding major air freight forwarders) can be assumed to approximate that for cargo carried by all commuters. A definite break in the upward trend of cargo pounds carried by commuters occurred in 1978. Pounds carried increased by 48 per- cent between 1977 and 1978, significantly in excess of the 30 percent per year average between 1970 and 1977 and 12 percent per year between 1978 and 1980. The large increase in 1979 is generally attributed to the impact of deregulation. The major impacts of deregulation are fairly well completed, and large jumps are not expected in the future. Growth in cargo carried is forecast to increase about 6 percent to 8 percent per year to 1985 and then from about 4 percent to 7 percent per year between 1986 and 1992. (See Table 10.) Three 21 L | {Jim - --:-- TABLE 9 CARGO CARRIED BY COMMUTER AIRLINES (in thousands of pounds) Predicted GNP-High GNP-Low Actual Growth Growth Trend 1970 43,527 15,191 15,191 14,976 1971 51,203 46,006 46,006 48,476 1972 74,573 99,180 99,180 81,976 1973 92,963 157,043 157,043 115,475 1974 138,279 151,181 151,181 148,975 1975 169,203 (89,304)* 139,374 139,374 182,475 1976 216,811 (116,203)* 195,060 195,060 215,974 1977 271,242 (137,722)* 254,765 254,765 249,474 1978 401,638 433,730 433,730 425,380 1979 475,000 472,333 472,330 458,879 1980 500,000 470,575 470,575 492,379 1981 482,968 469,066 525,879 1982 511,690 486,066 559,378 1983 535,807 528,438 592,878 1984 592,162 567,208 626,378 1985 651,868 607,319 659,878 1986 704,623 649,021 693,377 1987 789,784 693,653 726,877 1988 816,915 740,295 760,377 1989 876,453 788,696 793,876 1990 938,168 838,687 827,376 1991 1,002,140 889,767 860,876 1992 1,068,630 942,941 894,370 Source: Historical data: Commuter Airline Association of America "1980 Annual Report, Commuter Airline Industry" Washington, D.C. (November 1980); forecasts generated by SRI International. Cargo carried by all-cargo operators and excluding Federal Express. 22 4 _ _ _ _ _ _ _ | || TABLE 10 HISTORICAL AND FORECAST ANNUAL AVERAGE GROWTH RATES FOR POUNDS OF CARGO CARRIED BY COMMUTERS (percent per year) Pounds of Cargo Actual Forecast Carried 1970-1977 1978-1980 1981-1985 1986-1992 High growth GNP 29.9 11.6 7.8 7.2 Low growth GNP 29.9 11.6 6.7 6.4 Trend line 29.9 11.6 5.8 4.3 23 &i I separate forecasts were developed; the first two relate cargo carried to GNP (in constant 1972 dollars) and the third is a trend extrapolation. The high and low GNP scenarios discussed in Chapter 1 were used to generate the cargo pound forecasts given in this section. The trend extrapolation equation pro- vided a better fit between predicted and actual historic data, and also resulted in the lowest growth scenario. The more con- servative forecasts, therefore, may be the more realistic ones. As indicated earlier, all-cargo operators are moving toward larger turboprop and even jet aircraft. The growth in size of aircraft in this segement of the industry will probably exceed that for the passenger oriented airlines. Average annual growth of 4 percent to 6 percent, with more rapid increases early in the forecast period, indicates that all-cargo operations will not grow significantly. This result is somewhat misleading, because growth will be strong in many of the busier airports, but will drop off significantly on routes where trucks can easily and effectively replace aircraft. Also, because the amount of cargo handled will increase significantly, about doubling in size over the forecast period, the storage and handling space at airports required by all-cargo commuters will likewise need to increase. 24 p -~ ~ .-- ~ - _ ____ ___ - ~- -~- CHAPTER 3: LOCAL MODELS EVALUATION Commuter enplanements at an airport are influenced by a variety of transportation and demographic factors unique to each commu- nity. Availability of alternate forms of travel--either certi- ficated air service or the automobile--will influence a travel- er's decision to take a commuter flight, as well as timing and routing of flights. Perceived levels of comfort and safety will also have an effect. The location of a major hub airport rela- tive to the originating community will influence mode choice, as will the economic activity within that community. To reduce the evaluation of demand at over 600 locations served by commuter carriers to manageable proportions, a sample approach was adopted in which six cities of varying size and service were selected and a five-year (1975-1979) history of their commuter service and traffic was examined in detail. These cities were: Boston, Massachusetts (a large hub); Indianapolis, Indiana (a medium hub); Spokane, Washington (a medium/small hub); Abilene, Texas, and Lake Tahoe, California (towered nonhubs); and Altoona, Pennsylvania (nontowered, nonhub). Three types of models were evaluated; these included a local growth rate model, a model relating local to national growth, and a market (origin-destination) forecast model. Examples of the results from each approach are described separately; the conclusions are summarized in Table 11. Two of the approaches evaluated yielded fair to good statistical and intuitive descriptions of commuter activity of a city. The overall re- sults, demonstrate that local commuter activity is dependent on an exceptionally diverse set of causal factors that will require 25 Table 11 COMPARISON OF LOCAL MODEL RESULTS Statistical Type Model Validity Advantages Disadvantages Growth rate Poor Ease of application to any Poor statistical results location Oversimplified variables readily available Fails to reflect adequately or easy to forecast many causal relations Computationally simple Applicable to previously served locations only Local to Mixed, some Ease of application to any Requires national commuter national good location and certificated carrier growth estimates Variables are generally enplanement models as available or fore- prerequisite castable Applicable to previously served locations only I Market Fair to Replicates specific city- Contains location specific forecast good pair characteristics dummy variables * Applicable to previously . Has to be incorporated in served and new locations a larger program to be * Independent of any other used nationally forecast 26 g I further effort to identify and quantify before a definitive model can be developed. None of the models reported herein is considered complete at this stage. An increase in the number of sample cities would provide a broader base for testing alternative model formulations by size or hub-status. Growth Rate Model This technique measures the annual percentage increase in com- muter enplanements at a community as a function of changes in community and air service characteristics. The variables used, their definition, and the results are indicated in Table 12. The rationale for this approach is that increased population, increased commuter frequency and changes in certificated carrier frequency and enplanements will affect commuter enplanements. The results in the table are typical of those obtained from this approach; introducing dummy variables to identify unexplained jumps or drops in traffic between quarters of a year improved results slightly. Tahoe and Altoona were eliminated from the data base due to data problems. The signs of the coefficients are mixed; frequency effects are positive (as expected) with the year previous change exerting a strong influence. Population changes would be expected to show positive correlation, not negative. Certificated carrier frequency and enplanements would also, in most of these cities, be expected to show a positive influence. In Abilene, the relationship might be negative due to the effect of Texas International on Chaparral. Reestimating the model without Abilene did not change results significantly. None of the dependent variables is significant without a 90% confidence interval. The strength of this approach is its apparent simplicity and ease of application in that all dependent variables are readily 27j a MEMO Table 12 LOCAL GROWTH RATE MODEL Dependent variable: annual enplanements (TRAF) Independent variables: Abbreviation Description POP Population of locality in specified year FRQl Number of commuter departures in the specified year FRQ2 Number of commuter departures the year previous CFRQ Number of certificated carrier departures in the specified year CEN? Number of certificated carrier enplanements in the specified year How measured: The percent change (positive or negative) in all variables is measured from the preceding year. These are indicated by the prefix "D" before the abbreviation. Results: Variable Estimate t-Ratio Intercept 0.38 1.35 DPOP -8.90 -0.43 DFRQ1 0.15 0.28 DFRQ2 0.47 0.74 DCFRQ -0.03 -0.02 DCENP -0.03 -0.02 2 N-14 R -0.37 28 } I II I|4 obtainable or capable of being estimated. It is, however, the least accurate of the model approaches evaluated. Local To National Growth Model The second forecast technique evaluates local changes in enplanements or operations in relation to national changes. Such a technique provides a local forecast by developing unique local growth factors that are derived from national growth. The relation between local growth and national growth in enplane- ments and operations was explored. Table 13 summarizes the results. Both operations and enplane- ment growth forecasts were evaluated. Those for enplanement growth yielded better results. Two enplanements (TRAF) forecasts are presented. The first of these, using all independent variables, obtains inconclusive re- sults. Population change acts as a strong predictor, but the signs of disposable income and frequency are contrary to expec- tation. The expected sign of the certificated frequency coefficient is unclear as a positive value would be anticipated where feeder effects dominate (e.g., Boston and Spokane), but negative where competition is heavy (Tahoe, Abilene and Indiana- polis). Attempting to group the cities by these criteria did achieve some improvements. The forecast model for the smaller cities (the second entry in Table 13) shows an acceptable predictor. The two variables--disposable income and certificated departure frequency--can be seen to act as good predictors of commuter enplanement growth. Additional effort at categorizing cities and developing a "family" of predictors is indicated. Forecasting local enplanement growth in relation to national enplanement increases has potential. 29 *1 - Table L3 LOCAL TO NATIONAL GROWTH MODEL Dependent Variable(s): annual enplanements (TRAF) or operations (OPNS) at each locality. Independent Variables: Abbreviation Description POP Population of each locality in the specific year DISINC Per capita disposable income FRQ Number of commuter departures annually (NTAKEOFFS is the equivalent national variable) CFRQ Number of certificated carrier departures performed annually at each locality (ACOPS is the equivalent national variable). LDI Local Dummy Variable for Altoona LD2 Local Dummy Variable for Lake Tahoe LD3 Local Dummy Variable for Abilene LD4 Local Dummy Variable for Spokane LD5 Local Dummy Variable for Indianapolis How measured: Each variable is expressed as the ratio between that variables percent change in the locality from the previous year to the equivalent national variables change between the same years. RC prefixes the variables to indicate the ratio of change. Results: RCTRAF (all cities) Variable Estimate t-Ratio Intercept 9.05 0.84 RCPOP 14.07 4.88 RCDISINC -3.82 -0.53 RCFRQ -0.56 -0.69 RCCFRQ -3.01 -10.12 N-24 R2m0.90 RCTRAF (Altoona, Lake Tahoe, Abilene and Indianapolis) Intercept -10.21 -0.40 RCDISINC 21.21 1.57 RCCFRQ -3.14 -5.61 2 N-16 R -0.77 30 ----.--. ,.I. . .. . . ____________________________-_____ Market Forecast Model The third technique evaluated forecasts of the actual traffic volume for any origin-destination pair of cities as a function of their demographic characteristics, their distance, the avail- ability of markets to alternate hub airports, and the level of air service provided. Information was compiled for markets be- tween the six sample cities and other cities to which they are connected by commuter service. Minimum criteria were established to select markets for the data base: o Service in the market had to be provided for at least two full, consecutive years by at least one carrier; o Service was not started during any of the sample years; o No unusual jump or drop in traffic (other than seasonal variations) occurs during any quarter of a year; and o A minimum of 1,000 annual passengers should travel in the market. All Lake Tahoe markets were eliminated due to a very unstable commuter history. The results are summarized in Table 14. The city-pair traffic volumes are summarized in Appendix C. The results are good statistically. Only the sign of ALT 2-- which identifies better hub access in another market than the one modeled--is contrary to expectation. GDIST appears to be a good replacement for distance, avoiding the problem that traffic would increase with distance up to approximately 120 miles (positive correlation), then would decrease (negative correla- tion). 31 Table 14 MARKET FORECAST MODEL Dependent variable: Annual traffic volume (TRAF) in a directional market (between a specified origin-destination pair of cities). Independent variables: Abbreviation Description POPO SMSA or county population (if not SMSA) in which the nonhub or smaller hub airport ("origin") of a city-pair is located by year. POPD SMSA population in which the hub or larger hub airport ("destination") is located by year. GDIST A statistical value derived from the distance between cities. See Appendix C FRQ The number of annual commuter departures from origin to destination. CFRQ The number of annual certificated carrier departures from origin to destination. FRQPOPO, CFRQPOPO The number of annual commuter and certificated departures divided by the annual origin population. ALTI, ALT2 Dummy variables reflecting air access from the origin to hub airports other than the destination. See Appendix C . How measured: Each variable is expressed as its numeric value for each year. Results: TRAF Variable Estimate t-Ratio Intercept -37,857.20 -5.71 POPD 9.55 10.94 GDIST 213,898.10 6.25 ALTI -1,476.17 -0.67 ALT2 19,418.34 4.94 FRQPOPO 231.65 2.48 CFRQPOPO -283.44 -1.82 2 N-58 R .0.82 32 ir e.Il The results of this model suggest that a "family" of similar models could be developed by segmenting market types into dis- trict categories (e.g., large hub to large hub, nonhub to large hub, etc.). For the FAA to utilize this type of forecasting model, a comprehensive computer program would be needed in which all O-D volumes would be calculated, reversed to provide two-way traffic, then assembled by origin city to provide a forecast for each locality. 3 33 . .... _ _ _ __ i -- ~ . -- - - - -- - - . i CHAPTER 4: HISTORIC DATA COLLECTED FOR MODEL DEVELOPMENT Data collected for evaluation during the model development ef- fort was extracted from the following sources: o Civil Aeronautics Board Commuter Air Carrier Statistics - Online 0 & D (Part 298) data, 1975-1979 as extracted on tape by the National Archives and Record Service in July 1980 o Official Airlines Guide (OAG) data for March, June, September and December 1975- 1979 as edited by FAA and stored on the OAG nMiles" data tapes o Air Traffic Activity system data for FAA towered airports, summed to quarters, 1975- 1979 o Aircraft descriptions keyed from data in tables printed in the OAG 1977-1979, from Jane's, and miscellaneous sources o Civil Aeronautics Board, aircraft inventories, published annually in June, for 1975 through 1979 o The Computer Company (TCC) data based on the CAB form 41 data for the small certificated carriers considered as commuters in this study (See Figure 2). o Base year and forecast variables for GNP, fare, other trans fuel, personal income, disposable personal income. The OAG data, CAB Part 298 data, and Air Traffic Activity data were accumulated by geographic entity, year and quarter, and class of service. The geographic entities used for accumulating the base data were: 34__ _ _ _ _ _ _ _ US: The 48 conterminous states and the District of Columbia HI: Hawaii AK: Alaska VI: US Virgin Islands PR: Puerto Rico IN: International (CAB only) The US Virgin Islands and Puerto Rico were combined for forecasting purposes because of their proximity and similarities in their economies. Due to the nature of the airline industry in Alaska, it was felt that a separate study would be necessary to successfully forecast growth. The International category for the CAB data was accumulated to cross-reference the figures with the published past years' total traffic data. Individual airlines may update data after the publication of the detailed data in the CAB Commuter Air Carrier Traffic Statistics. Flights and trip segments that originated or terminated in the ( same geographic entity were assigned to that area. The number of times a week that an OAG flight was scheduled was multiplied by 13 (the number of weeks in a quarter) and then by 2 to obtain scheduled operations. When a flight originated in one geographic entity and terminated in another, then the number of flights per week was only multiplied by the number of weeks in the quarter. CAB data was assigned based on the originating airport of the trip segment. All the small certificated carrier data was included in the US figures. The CAB data was separated into two classes of service: cargo only and passenger service. The airline records for each trip segment were evaluated to determine if any passengers had been carried. Segments without passengers were coded as cargo-only. 35 The OAG data contains codes to indicate cargo only flights and three classes of air carriers: certificated, intrastate, and air taxi. The OAG air taxi class and the selected certificated carriers providing commuter service were grouped as commuters for this study. Some of the Allegheny commuter flights may have been missed because of inconsistencies in the OAG data record coding of these flights. The Air Traffic Activity system operations are reported for FAA towered airports only. In this system commuters are counted as either air taxi or air carriers depending on their aircraft size and flight numbers. The air taxi operations include some com- muter flights, carriers carrying only mail, and air taxi ser- vices. Table 15 lists the OAG scheduled flights and the FAA Air Traffic Activity figures for comparison. For operations historic data the OAG scheduled. Available seats were based on the OAG scheduled flights and the minimum seating capacity of the aircraft as published in the OAG or found in other sources. Many of the OAG aircraft codes in 1975 and 1976 were not listed in the OAG aircraft codes tables. Judgmental determinations were used in equating the codes to aircraft codes used in Jane's or to other codes on the OAG list. Seat miles as computed from the OAG scheduled flights and pas- senger miles from the CAB part 298 and TCC form 41 data were used to calculate load factors and appear in Table 16. These load factors were within the range suggested as profitable by commuter operators. The load factors tend to be lower than a those of the large certificated carriers because of the greater number of multiple segment flights and the need to have seats available for passengers that board down line. The smaller aircraft must also be operated at lower load factors to reduce the need to turn away customers because of flights filled to 100% capacity. 36 ..- ~ -- TABLE 15 OAG SCHEDULE) OPERATIONS AND OPERATIONS REPORTED BY FAA TOWERS 03S GEOENT TYPE YR ATAUPS OAGSCH RATIO I AK AC 75 91609 234546 2.56029 2 AK AC 76 90114 221036 2.46284 3 AK AC 77 80568 247598 3.07316 4 AK AC 78 87989 168084 2.13759 5 AK AC 79 102710 255125 2.48394 6 4K CO 75 120376 38558 0.31899 7 AK Cc 76 122869 92612 0.75375 8 AK CO 77 153153 81640 0.53304 9 AK CO 78 142726 61224 0.56909 10 AK CO 79 148Z71 106386 0.71798 11 N1 AC 75 200659 175487 0.87455 12 HI AC 76 201346 j76904 0.87861 13 HI AC 77 215054 1C7135 0.87018 14 mI AC 78 229748 214162 0.93216 15 HI AC 79 240591 243334 I.C1140 16 HI CO 75 95090 16Z214 1.70590 17 mI CO 76 112991 194792 1.72396 18 HI CO 77 135674 223470 1.64711 19 HI cO 78 199762 z49673 1.24988 20 HI CO 79 220537 23675a 1.03262 21 PR AC 75 53280 46436 0.87155 22 PR AC 76 50761 46202 0.91019 23 PR AC 77 48968 43147 0.86113 24 PR AC 78 47038 46670 0.9112 25 PR AC 79 481356 44109 0.91596 26 PR CO 75 163222 ^08221 1.23778 27 PR CO 76 158916 -C222& 1.27255 28 PR CO 77 147010 189345 1.28780 29 PR CO 78 144451 170924 1.18327 30 PR CO 79. 146206 181506 1.24144 31 us AC 75 8364491 9563691 1.08113 32 us AC 76 91490429 9810138 1.07226 33 us AC 77 9522670 I0C73375 1.05783 34 Us AC 78 9779700 10843898 1.10882 35 us AC 79 9863742 10964369 1.11158 36 us CO 75 2282535 2291172 1.00378 37 us CO 76 2440667 2634619 1.07947 38 us Cc 77 2830763 2976519 1.03324 39 us CO 78 3247563 2982408 0.91835 40 us CO 79 3823937 3950713 i.03315 41 VI AC 75 12779 14332 1.12309 42 VI AC 76 12487 14690 1.17642 43 VI AC 77 10274 14534 1.41464 44 VI AC 78 11775 25870 2.19703 45 VI AC 79 12641 17992 1.42351 46 VI Co 75 35684 122096 1.42494 47 VI CO 76 85083 117546 1.38155 48 VI CO 77 95345 128427 1.34697 49 VI CO 78 121494 141167 !.16193 5 VI CO 79 136476 151372 1.10915 37 I TABLE 16 BOARDING AND LOAD FACTORS Available Revenue Seat Passenger Load , Boarding Year Miles * Miles * Factor Operations Enplanements Factor 1975 1415.05 616.09 .44 2.04 5.26 5.17 1976 1482.78 674.90 .46 2.21 5.83 5.27 1977 1668.37 841.32 .50 2.53 7.05 5.57 1978 1913.83 1031.09 .54 2.81 8.54 6.08 1979 2843.25 1447.06 .51 3.68 11.04 6.00 * in millions Source: CAB and TCC: revenue passenger miles and enplanements OAG: available seat miles, operations 38 Local model data was extracted for specific airports by origin and destination city pairs from the OAG data files, the CAB part 298 data file, and the A-AIMS (American Airlines Information Management System) data base. The data was accumulated by air- line, class of service, and non-directional market for model de- velopment. Quarterly data used by the Conterminous United States, Hawaii, and Virgin Islands/Puerto Rico models and the quarterly forecasts are listed in Appendix B. 39 I..- . . APPENDIX A DESCRIPTION AND SPECIFICATIONS OF ECONOMETRIC MODELS ... . . - ron .......... 1 I I I I I I I l I. , ... CONTERMINOUS US MODEL The amount of commuter airline activity in the United States is determined by the factors influencing both the supply of these services and the demand for them. The model developed in this study is based on simultaneous estimation of the supply and de- mand equations and is designed to produce forecasts of commuter airline passenger operations (OPS), passenger enplanements (PE), revenue passenger miles (RPM), and seat miles (SM). The demand for commuter airline services or passenger enplane- ments is considered to be a function of the price of these ser- vices (FARE); constant dollar gross national product (GNP72); the price of competitive modes of transportation, primarily automobiles (AUTOPR); the number of origin-destination pairs served by commuters (O-D); the average distance flown by passen- gers (ADF); and a series of dummy variables representing sea- sonal factors (Dl, D2, and D3) and the impacts of deregulation (DREG). This equation is specified in a general form in la. De- finitions of variables are provided in Figure A-1. PE = f (FARE, GNP72, AUTOPR, O-D, ADF (la) a DI,D2,D3, DREG) Revenue passenger miles are estimated by multiplying passenger enplanements by average distance flown, as specified in equation lb. RPM = PE x SEGDIST (lb) The supply of commuter operations is generally considered to be a function of the fare received, fuel costs, other costs, the size and number of aircraft, and the same dummy variables repre- senting seasonal variations and deregulation. The general form of the supply equation is given below. OPS a ga (FARE, FUELPR, OTHERPR, SIZE, FLEET Dl,D2,D3, DREG) (2a) A-1 ___ ___ ___ ___ ___ __ ___ ___ ___ ___ ___ __ VARIABLE DEFINrTION PE Passenger enplanements per quarter -- Civil Aero- nautics Board (CAB), Part 298, data on enplaned passengers and selected Form 41 data. RPM Revenue oassenger miles per quarter -- CAB Part 298 data and selected Form 41 data developed by multiplying the sume of enplaned passengers times the average distance flown by passengers. 0-0 Sum of the number of trip segments provided by each airline -- CAB Part 298 data. OPS Scheduled operations per quarter -- derived from the OAG data tapes fields for scheduled days multiplied by weeks in a quarter multiplied by 2 (take off and landing operations). SIZE Average number of seats per operation -- data cal- culated from Office Airline Guide (OAG) computer rapes. SM Seat miles flown -- OAG derived data on flights reported times by aircraft type times seats for standard aircraft configuration times average stage length. ADF Mean trip Length in miles per enplaned passenger -- CAB Part 298 data and selected Form 41 data. AVESTAGE Mean stage length per operation -- OAG derived data. FUELPR Price of Eule to commuter airlines -- data from the Department of Energy, "Monthly Energy Review,' on retail price per gallon of aviation gasoline fuel (taxes not included). rice adjusted Eor inflation by using GNP deflator. FARE Index of ai:line fares -- data from Department of Labor, "Consumer Price index" and represent an index of faces charged by all domestic airlines. No index was available for commuter airlines only. index ad- justed for general inflation using GNP deflator. GNP 72 Gross National Product, in constant 1972 dollars -- data from Department of Commerce, "Survey of Current Business." AOTOPR Index of the price of user-operated transporta- tion -- data from Department of Labor, "Consumer Price Index," adjusted for general inflation by GNP deflator. Dl, D2, 03 Seasonal dummy variables for first, second and third quarters, respecti.ely. ,alue - I or 0.) DREG Deregulation dummy 7ariable, zero through 1378, first quarter and one thereafter. It is assumed some of the effects of deregulation started oefore the 1978 act took effect. (value - I or 0.) QUAAT!wRLY DATA COULD NOT BE OBTAINED FOR THE FOLLOWING VARIABLES: OTIERPR Index of non-fuel costs of operating commuter ai:- craft. The few obeervations obtained and Aviation Data Services, Inc. statistics for general aviation aircraft operating coats, indicates that this variable appeared to follows general inflation. FLEET Total fleet seats available, or capital stock. Annual estimates indicate relatively smooth growth in this rartaOle. FIGURE A-1. DEFINITIONS OF VARIABLES A-2 Seat miles are calculated by multiplying the number of scheduled flights (OAG) for an aircraft type by average seating capacity of the aircraft by average stage length. SM = AVESTAGE x SIZE x OPS/2 (2b) Equilibrium conditions are met when supply and demand for these services are in balance, or that the airlines are achieving their desired load factors. Equation 3 specifies these condi- t ions PE = h (OPS) (3) a where ha indicates the desired average number of passengers per operation. Because of nonavailability of data, not all of the variables could be included in the model. No time series data on the costs of nonfuel inputs to providing commuter services (i.e., equip- ment, services, and personnel) were available and estimates of fleet size could be determined only on an annual basis. The few estimates of nonfuel costs of operating commuter airlines indi- cated that these costs tended to follow general inflation and, therefore, after they were adjusted for inflation, the impact of these costs on explaining changes in activity would be slight. A linear form of equations la, 2a, and 3 was estimated, as shown in equations 4 through 6. This specification provided the best overall results. PE - a1 0 + all FARE + a1 2 GNP + a 1 3 AUTOPR + a1 4 O-D + a15 Dl + a 1 6 D2 + a 1 7 D3 + a 1 8 DREG + ual (4) OPS - a2 0 + a 2 1 FARE * a 2 2 FUELPR + a2 3 SIZE + a 2 4 Dl + a 2 5 D2 + a 2 6 D3 + a 2 7 DREG + Ua 2 (5) A-3 PE = a 3 1 OPS + Ua3 (6) Several techniques can be used to econometrically estimate the above models. A reduced form procedure was selected becasue of the limited number of observations and the poor quality of the data on fares charged by commuter airlines. By using the price index for all airline fares, we are assuming that commuter air- line fares follow the average pattern of all fares, which tends to be dominated by trunk and regional carriers. With reduced form procedures, the fare variable does not affect the fore- casting equations for the other dependent variables. Table A-1 presents the results of the econometric estimation for the national level model. The results for all three equations are fairly good. Most of the non-dummy independent variables have significant t-statis- tics and the signs of the coefficients generally follow the ex- pected patterns. GNP positively affects commuter enplanements and operations, but decreases fares; the cost of automobile transportation has a negative effect, indicating that as ground travel becomes more expensive, people tend to travel less. The fuel price index variable has a positive coefficient, because commuter airlines become more competitive in short haul markets when fuel prices are high. Average passenger distance has a slight positive effect on operations and essentially no effect on enplanements. This variable does have a significantly negative impact on fares (per mile), as would be expected. The number of origin-destination pairs served is important in generating passenger enplanements, but also adds to the average fare charged. Size of aircraft is not significant in any of the equations. The impact of deregulation over and above that incorporated into the other independent variables is fairly small. Seasonal factors vary with equation. A-4 .I I TABLE A-1. ECONOMETRIC ESTIMATES FOR REDUCED FROM EQUATIONS CONTERMINOUS US MODEL PASSENGER INDEX OF OPERATIONS ENPLANEMENTS FARE R2 .9889 .9958 .8684 Constant -1.5338 -2.861 173.08 (-2.4) (-1.9) (4.4) GNP72 0.0006 0.0018 -0.02 (2.2) (3.0) (-1.1) AUTOPR -0.0005 0.0018 -0.02 (-0.2) (-3.4) (1.7) FUELPR 0.0096 0.065 0.34 (1.9) (5.4) (1.1) ADF 0.0096 -0.001 -0.91 (2.6) (-0.1) (-3.9) O-D 0.0001 0.001 0.02 (0.5) (3.3) (2.5) SIZE -0.0067 0.057 -1.08 (-0.2) (0.8) (-0.6) Dl -0.0186 -0.085 2.30 (-1.0) (-0.8) (2.0) D2 0.004 0.064 -1.88 (0.2) (1.4) (-1.6) D3 -0.0088 0.218 -1.50 (-0.4) (4.3) (-1.1) DREG 0.0070 -0.082 -2.63 (0.2) (-0.9) (-1.1) t-statistics given in parenthesis A-5 - -|| | | | - - HAWAII AND PUERTO RICO /VIRGIN ISLANDS MODELS As indicated in the main body of the report, commuter activity in Hawaii and Puerto Rico/Virgin Islands is influenced by fac- tors that are different from those that affect commuter airlines in the 48 states. The models developed to forecast commuter activity in these two areas reflect this difference and are structurally simpler because of less available data, as speci- fied in equations 7 and 8. PE = b10 + b1 1 Income + b1 2 D1 + b 13 D2 + b14 D3 + v1 (7) OPS = b2 0 + b2 1 PE + b2 3 SIZE + b2 4 Dl + b2 5 D2 + b2 6 D3 + v2 (8) (Income denotes constant dollar disposable personal income for Hawaii and constant dollar personal income for Puerto Rico/Virgin Islands and vI and v2 represent random error terms.) (Personal income data are used in place of GNP because the lat- ter data are generally not available on a state level.) The model was specified in the above fashion for Hawaii because, after discussions with local airlines, it was determined that most commuter passengers are business people. Therefore, enplanements are specified as a function of the income generated in the state. Operations are assumed to be a function of enplanements and the average size of aircraft. Because quar- terly data are used, seasonal adjustment is accomplished through three dummy variables. Ordinarily, a price variable would be included in equations 7 and 8, but we were unable to find time series data on fares charged in Hawaii. The same basic model structure is used for Puerto Rico and the Virgin Islands. Tourism, however, is a more important source of A-6 - , ~ ' I ______ __II_ _I___ ___I_. .. . commuter passengers in this area, but personal income is still used as the independent variable for enplanements because tour- ism expenditures help to generate personal income. Results from model estimation for reduced form equation are pro- vided in Table A-2. The Hawaii model gave better results than the Puerto Rico/Virgin Islands model. The seasonal adjustment and size of aircraft variables have the insignificant t-statis- tics, but there are very significant relationships between enplanements and income and operations and enplanements. A-7 2_ = Table A-2 ECONOMETRIC ESTIMATES FOR REDUCED FORM EQUATION: HAWAII AND PUERTO RICO/VIRGIN ISLANDS Hawaii Puerto Rico/Virgin Islands Enplanements Operations Enelanements Operations R2 .7993 .8991 .3904 .7693 Constant -0.456 0.0 0.143 0.006 (-6.6) (0.0) (1.3) (0.2) Income 0.125 0.008 (7.6) (1.5) Enplanements 0.660 0.166 (10.9) (5.8) Size -0.0012 0.001 (-0.2) (0.3) D1 0.0016 0.001 0.052 -0.004 (0.3) (0.3) (0.0) (-1.4) D2 0.0029 0.003 0.003 -0.004 (0.5) (-0.8) (0.9) (-1.5) D3 0.0043 -0.0004 0.024 -0.005 (0.7) (-0.1) (0.3) (-1.9) * Disposable personal income, 1972 dollars for Hawaii and personal income, 1972 dollars, for Puerto Rico t-statistics given in parenthesis A-8 Cargo Model Because of data limitations, an al-cargo commuter airline model could not be developed. Instead, data on pounds of cargo carried by all commuters were used to estimate the overall growth in this variable. The resulting growth rate was used to approximate the rate of increase in pounds of cargo carried in all-cargo commuter operations. All-cargo operations (excluding Federal Express) accounted for about 52% to 53% of cargo carried by all commuters between 1975 and 1977 and, assuming this share remains relatively constant, the amount of cargo carried by the all-cargo operators can be calculated from the model forecasts. Two econometric equations were specified, as shown below: POUNDS = C1 0 + C11 GNP72+ C1 2DREG + w, (9a) and POUNDS = C 20 + C 21YEAR + C 22DREG + w 2 (9b). POUNDS denotes thousands of pounds of cargo carried by all com- muters1 , YEAR is a trend variable where 1970-70, 1972-72, and so forth, and w1 and w2 are random error terms. GNP72 and DREG are defined in Table A-1. The results from estimation are given in Table A-3. Both equa- tions yielded good results; however, the second specification, given in equation 9b, gave a better fit. Both equations were used to generate the alternative forecasts provided in Chapter 2 of this report. 1Commuter Airline Association of America (CAAA), "1980 Annual Report, Commuter Airline Industry", Washington, D.C. (November 1980). A-9 ,, I I I i I I I I I- TABLE A-3 ECONOMETRIC ESTIMATES FOR CARGO EQUATIONS R 2 Constant GNP 72 Year DREG Pounds .9676 -893,870 837 124,367 (-5.3) (6.1) (3.1) Pounds .9894 -2,330,000 33,500 142,406 (-10.8) (11.4) (6.8) t-statistics given in parenthesis. A-10 t &POWNIX B QIJAR'INL! DATA AND FPORECASTS DEFIITIONS QUARTERLY FORECASTS AND BASE DATA - DEFINITIONS ADF the average distance flown by a commuter airline passenger ATOPS commuter operations AVEDIST the average stage length AVESEATS the average available seats per operation DEREG dummy variable used to take into account changes in growth since deregulation (0 before 2nd quarter 1978, 1 thereafter) DUMlJ dummy variables used to forecast seasonal fluctuations DUM2( in the data (DUMI=l if first quarter data; DUM2=I DUM3D if second quarter data; and DUM3=1 if third quarter data; DUMI, DUM2, and DUM3=0 for fourth quarter data) FARE a fare index for all air transportation (1972=100) FUELAD the cost per gallon of aviation fuel for commuter airlines in 72 constant dollars GNP72 the GNP in 1972 constant dollars NTAKEOFF the number of commuter airline pairs weighted by airlines (CAB Part 298 data only) OBS a record number generated by the SAS procedure that sequences the data for printing OTC the consumer Price Index for owner operated transportation PATOPS forecasted commuter operations PFARE forecasted fare index PRPTDIST forecasted revenue passenger PRPTI forecasted enplanements PSEATMI forecasted available seat miles QTR the year and quarter of the data or forecast RPTDIST revenue passenger miles RPTI enplaned passengers SEATMI available seat miles B-1 i - CONTERMINOUS UNITED STATES ____ ___ ____ ___ ___ ____ ___ __ Nd . nN 4 .7 PN .0 0.0 an ey do W ry11 O. N N 0 - Q dAW a r%4 Nf% @. n in N W a : Wob 0 P%2 P C, C!d.7 0:.. . . .- .- ea. LU~~ ~~ "o@ N~eLN .4~ 242t p 1 ue $ANini NN@0,O.a I bn . f.0 W.NAen 4 ---- a -a - fta-- - a --- am a aa a aa-oaa- Ma a-a a a a m -C 0 N a ,4E ' N P . N do' - W cv CL a Cl No N- a W ... if a do en . &nIn AD a kA 10 P.O i P, N o N d . Ie. M wl -' W W. N C i-Is ",a 4p w 4. 0 oenN@F- o4A &M. ae7isen.7do.In Ina... aNWddo 0~ ~~~~ Ns . oai - r.NNaP W .79 w9 W 0 . W, VN~ od - . 0 po " 4 ~ ~ ~ ~ ~ ~ C C4 - --- - - - N 94 44 M 1%o 0-1 .4 ,& l do l 001fn 01 wt W W C6 00w ~ -CI 4p 0. p04paais 414 3h 2 . 4IN a 4 . a a P- .- dNNNNNN 0 a 0 a- - B-2 01 La Wj 4c - -j U. U . 4 4c -C 0 4.J UjW 4b47 cac oI r0 r m -A a..,' A lW rP ;r4 C6 I PCP C, CD0 4 D r-W0 S,, a 10 4!4 Li &n' r-w ngw- .a - 0 Z' -CI n 4cL Y0 4' 00 do- co - e a aa -0d 4j ,2 -C . CL 1 4 0 0 '- 2 L.fl'JJ - j 4 4c 14z . 0. Zgi LUL LA -a U. LL - 22 00 0 =m x L6. I4O .4 -C 0 0 a.= n F .0 r 7.C 00 09 -C -U 20( -L.. ao"N nWr_ 09 ~ A 0 3 : m I% ,m" z ; 0 C C, .1 -L4% UA ' 0P C 1, 0, ~ .0 4 N X a, LU4 41 so- 641 cya@ 'm 0 l qM ***a*m-C = * 4* 6n U.. 4W 4n t .r o r 4. 4C ' n IN % 4D. -j Ca p4 10 N CL -w0. 1 q4ma U. - -- ------ - 0u C6 U *4 -i p-0- im 40V1B- 0 6 029 LU W. u U .wW4 9 a m0 -' =U .- 3.aZ - 'A1 W U. U 4 - W4 UJ E- Z . a V 4-0 0 :- L. -J . U. . - C3 4W w~ 4 0 4900 cc0. . .. . 4CJ T%0 N - CDN~'.--- a. CP*. 45 *0 .aI 0.- : ..n InN'.4 U~~. 0r 10 0Y r dc~lo. -. 0 -. M ~ 42 4D4 04 ~~.-0 4b 4.2~ 0~~4 4 .'NJC!d1Nr 404 -j L64uj . 0 v -n L6. 0 w - P I, au 7 vi.1 a. P- A ,W-4i 4- P- W, %0 1a W rC N40c -C '7 40 , 0 4, C, 41 C6 I.. P. 2@D II 1 - Cas a *D 0a Q a La NIO 10 -. aeouU *t4 w u-w t.t o *oou N P. a Z6 agg w e.e ., LLU 024 Z -c W U4J La -L .c ~ ~ ~ ~ 0 Q0. 0a aPNC n0i% I- =,a CD 0C Z wt ' 7 41 CDd 4 001 4nC C 42 - Wn goo44 ' .0 -Aa .NI : 40 .4 09. .N 40w l ' Nr ,4 .. 04 CDL4- W W~0 U4 wlp % Q - III 0K @0 NU. do u .0 -4 4 ac .9 w~ @C"IN 0 B- 1~ -a '0 m0 u. ,l Wr * j -C .z vi- u Ln uii -L LU .00, ~ ~ 0 0 aN- "0 I- -- z J 0. (12 0 C 0 M ;1 .4 LU 0 4 L6 0.01 0- C l I U6%0 Nc 1 0. -x c 0 , 0 N 4. Na&n 'o -. IL ?1 0 00 " 'r. - 'mc .4 r a-a. Inrr Ir *u In do. In -a aa aa N %n- -CCD 0 L64 *rI n@ a a , 'nh .r 1.1 0 -0~ LULU kW P- ie LU N4 0-04 oC 4 ~C4A 0 0 .6 0 w ..-. 4. of U W ZCOOO L e6 e.. .4 WI N W In 1 n' - J A ,Z -C4' - CL ac c U a x x 0 w -C;- P-t0Z U.'0 0 2 U - Cf== cw U.1 r n )- U. -d- UU -90 L6 .q z CD CD -V a 04 G SL cw ~ ~ 0 0 . Aa , n U. c Cl J 0- A ~ ZZ U"' e4 43. 44 0l. 1* 01 "1. L a Z r s Z 0 -aj 4 .. 40 "0 - c0 a 0, aa aaaa 4-,, w L LUw-rm aa dlot% .0 W,2 P I= r4.U. -n w vw "PU~un, -N.,- N-F a a a-aI.- ,a a .ua. LL.. flU~a U1 0. 2. a . . mg W 2w ).&i 40 10- U4 L zn z L LU A- 4c - Cb ac. .. - c.. 0001; *g 0- AZZ u 0 K 44 0 L6- $Ac. - r(j- r 3 to Aai 40 W, 41~a~ K 4-- 10 C2 41C ~ UQ ~ . IA.. -d aV'- a U - o ItI L" L 40 W L 4n -I I LA. 96 ILg F,4 In0 41 N U N -r doJ F, ' 1 %4 W 0 44 .. j 0-.? w mN 4 @~@in cc -C LuE I "4 n4 -C I- Cd 0-N 0 ~-4NrE. Im -C 40 V @iQOs An L6 0. ag at a 4Q U W U. 44 1S 0 ocEn Enf U~ 4 l~n eL..04 in LLU4 LU LU - -. U . - LA -c - LLLL - CD0 0 0 . ae 0=F l 0u Cm0 - U. 0 1 4In .00 N0 00 8 a 09 0.g 0c m oCn I Ifi a O " * f U-"nI& -1 a- CF I0 F 4' r n or 1 .d A ~ B-11P. X 6W U L-4U W .4. 6u 6n .0 n n o NP aF%0 . 0C'JU 00; -. r% W1£lrd wa' :0 '7r @' N %C . a "9 ca a@ aO-@@.o@ -0 a a boO@- *a@@-- - - NNNN. ,. " Nf a ca .0 0. ,-4 4 ou . ou0 a 01.tas m Da0 uj C&0 fn0 %aF 0V% r a. Q , . , ' 2.L YN n.. n- LI. '0 -01 6006m 0 wlP NaNwla 0(pC or - :0 a o q ooo 1404 O. .' .. 6 . . .s. . ... -n a ,4 nm rla0C 604 0D 0' ' N' ' 0 mNNNCDN QD. .U0 'O Ir CY0' 05 , z? :I0.. "II 6 a @0 fN- -~~: NN #-I f N. m 42 %IIra @ o6 NA POI.@ 1%~ W NNOOnW CD0I U 0 --- ~ 4.- NNN NN~i S'~ 0 N N~t 60~1 N..~CL L ~ ~ ~ ~ ~ ~ ~ ~ ~ C 4 4-.4f 0 S I~ 1 .01. ' -n a. m.7Q1U5W60.C -O ,NNNNNC a. , Y. N 'r 'r N 1%@6060, 7@ .4 'r 2 N waW 00I d . 0 @ 7 N 610 40 k0m -4 -. in ra a - 0 o5%no a a. '7 Qfs ca 07. ., 0 m o00 N 01 01 . . J N N N N V . . . .. . 400 - L- -N- - - - -- - - - - - u 0.00 4. CF a. a0.0S7. Ol c .a-a a, 0. , 0.7. 7sa, , Q Q50. o , go N N@a Wdo 0 I'61 .. , a WD 9 d I '060@60 N NOM P, "It".. w lrww 0Ifl @ = ulM~ N w N'7 a a-0 0 w a id 2 I0 0 r% 00c 4 * 0 ~ ~ ~ W N0@0 0000 0000 000-Cin LU N@ 00 0' .7 '0 .. o~~~~~~a 4 1. 6611%1."5 @A0 u.I41: I CYN W0t .0 0 .0.0. *N %0In U0 % 060 .' ' @0 4p a-4NW rW *f .a I %4 p4 Y t NNN" 1% 0~ 1%51 1%.0 NS~B-121 LS0. 4WD0e 0- 00' 004 NW4. . . . 4 .4 -. .4 a.00 a4 F Y 4% 0 74- 0. V 4404U,' In 0 7 m% a. IF Q 4 0 a.r 00' a- - -4 N "A aIV 0....0 MIN.74 U, An a0 a04 .4 .4 P, m% r, 1%1% 40 .000 0' 04 0' 47 0=4 0 0 - N CJN Nj(\ I WlI 47.. .J* N , U, 60 40 a' "!? 14 00.4 ' 400N4!00 ! l'0'. 4! 1 w -40w0 a N 'r0 to 40. - .74 1%7. 410 N - N.4 A- - ulf nP of N% .74. 404704 40 N m447%..0%4 40 14( %A .. 40 N .71% 00.01 400. N 0'40 L~~flWv4Z7 400744I%1NV4'1404,'.0I~4400b. c2 C0.040..'r 1%%a 'o V400 .4- -- -- ----- -- - - NN N N N N N N N N N ON N N wmN.440.7"Ca CF ,VI00'V %Dl'4. ) Im P%1%F,061% 47U,0'V u,.4a'4N ON a LM a N - P. 7r.414 '4. 40 N.4n4n 0 W) PI .71% 470' 4, 40000 0 U, -n PM r 60.0 41.40 ^, 1J -0 U 0 N p - %'10 P4.1% 41 0 C 44...4444. .4...... - - 44N... 4N CJN N N CN NNNN N a VI0 pV1.7.N .0 Ng. N .00 N a000 0.442a l l N 'ON .440 0' el 0..1 Nn N. w74 0 Nd NA:A" -0.I4 r I N40VI1%N1 0 in .7..0 N 10in471U0 .4 " VI'44U,40VIa,. 0' 01 N47 .a U,4 .0 N.7 .74 0' -14--U.40 040'.7 ' %a.4744 N .7' .1% oU'p a,1 *. In . N4 a. . ~uj 1-0 CL 0 -.- -. - -- -- - - -4... -4444... .. .-- - --- N N Nl N N Nj N N N N 40 Na0.70 4 N" "0o 40 ol .0..4 i ,' W ag No W,404. 1 0 4 ' VI a N.0. U,'r a7. U, 'a co ON 'r 'a ~ 0,VIA D . . . . . . . . . . . . . .. . . I .7 ,V . . , 04 . . .. 7 . ' 4 N I4 0. 7, 0. .. . . O I . . . . 01 0 C~~~~~ .U . .U .. .444444444.. .44444..- - - - - ..... .44 4 . .. IfEV u.4 i0&A0.4 u 00 u0 4 00.4 'a.4000 4 .- 00. Q 00 0: ;4 0-.0002000.4000. (a 4l %a 4p 40. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0 U 1%.0..I.'4%004n."40''10 '740V 00N01OI, 0I4. % ---------------------------------- %%400.000''0''00 O--~--N--N----- 4p 4r.4 N 4. .444'.4 n0 % 7 & w r n& r ntr- NNNNm NfNaN.N. . 4p'4a41404000.4NNI 740 40'4% 40 N VI.7'7U . 47. u1a% <*'0VII7 ,a w -^ a0 Q~zw~-do .7'N00'0'.4U I 1 # , 'ee'4400"4.7VIro . CU 401% OO'40 N.1 Go0 0'0.t do140 O ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ . g ... 4.40,4'VO00N4'.N4.4,'00411%%%%%4'..-.NI,' ~~~,11: Z14WO.J 440004NI.. 411 999999 .4V" "77,4%0 g ^ 0-.NVI.7 ", In .00'n m .7U,'4F NNNIVVIIVVIIVVIIVVIIV.----------------.7,U400UU,,UU,,'--------. 4!0 !4 ! 4 a4m aa . o o!4, bc 2a. a .am1.m ( .1n. * N. la4" maQ I nmi LL.CW.4 04 4040 'a 40 %%111%%111%040040 40 a-, *0.0040440400404p'0'"0'. 0' 40 000 run.,00 00 00 0 000C0000 0000 0000 000 coo NVV.VV.VV..V.VVI4II4IIIIIIIIIIIIVVVV.VVV"V..VVV.V .444...444...44....4..44.444...44... . . . . . . .413.. 000 0 U 00 00 000 00 00 00 000 000 00 00Woo000 IL~~~I U. -Ca -. o 4c -4u ad CLP' I.-@ OOC% U@@..M@@ -- 00 42 kn W iW f w 13-4 wOU U. ..... 2c U. z L. c go... . 04 .B41 00010 11ai 40 a1 1 z-0 0 ca 1 1C nw . . N a a@ aa 10 I* ; . @ . a' 0. - - - - - - --- - ---- - ---- - NeNO rN l r$w. I6nU s 6A. rr, odogdo 4 am CL 0 ~ 0n~~f0 40 r 0 a P, @P a- a- c N 0 N NCD 'a 0.0 0F~~ 0 O6 nn 'a -t .0Cn-in - ~ ~ON 47 CF @00. &1Q.a-a 4 3 7 o a, da6 a o a -- a-- cm- a- - -- - --- - -- -- -- -- - -- - - - - - zp nNN0"F n. A4.0a NP r L 440 -Nr lN n 4,gk 4. . . . . . . . 'n 4Da 0.N 4 rN0 a aN(100a ,0U l17a -e -a1 A ui t * 2 r a. 4040 1 .0naiD 44-a L&J .0 N7.@ cC : W N 0. 60 l 0 ., Go N :a -o inr.. n r 61% nr e, 0 co-b1 O O~~~~ In N0~ 0 r C f N cm 0.I ,f.a,% 4C W,' ,u n% nCa . ,' aa -6 A lot 0 a ,^ ac wl w~~ N t' It . 47. 400 W a- 0 6. . . . . . . . . . . . N1 N B-15yWC - % ~i40 a r P"NI40@ A. - --- -In -- - LU .4FNIAfn V40If., P .%I 0. c N N cyN AcmNNN N N Y, 4 C km .0 14 N cp . a. N - 1% 'A.. I . - .4 F" &A. a . ;; IA VI n IA in 4.A V LU .7IA*10-1 CN 61 4 FI . In P .4r. ... 4 r FVI.NIA-N e4,r w. = ~ ~ F 0. -- Fn -- - -- - O 0 4.-VI0%Ip. ~0N - , ' 0 w In 'A In a wI. I* r a I- 40 00M~N ad .. . . .. . .N N... CL N 'V0 IAr.4 40 N0 do 40I 40 4 IA@ 4w ac CA. - .'~@040?VoVtJAA . . . . . . . .4@'. . .P. ... I. ag o AJ C9 .0 ObV 1,1 , rI as4 o4 0d p.0 . 4,@ r ,q -a a .4 B- 1 a 4040 aa at 40a= 0014 L~CL N s W1 o N0'C l 6% a- p .'7Jr..-.J r Ln- ~-C -ft t- 40 .W1 w a w a in0in40ftCa.90.404a4aaaaa 0 a ft n M -D 49 42 -40 40 aa 40 40ft a a 40 a a .4Cl r 40 20. 0'ini4 W 0 n 'CW1 40-o r 4..a 0.nw (~.1 6n .7.9. 0173. a, 0. .5F Qi(Y.'a3 F, w A C, 0 N 4. m n inft-. 41C N 'o "I0447.04 ww wN00"CD0 V, e C! .''7i a-.. w N in ". Q''a a'aNNdo^ ' M n n47 a (Na'a'a 0. N ini,. S no- 00 - -- Pl; aaI c . W 2I 0~ o inN w4'~ c. A w00%cj - NN N V. N~ N k 0 mN o0 1 ol a CK 06 r" .a'J..7 C; ThP** a aN 1..fS -~~~~~ ~ ~ ~ - ------- n . F r i 0 mNW rm a 'a -,n 1 -r -0l-i 4 w00t w-O U t N N - 0r~ .4I - oco wC,.-N. "'a 40 aN N NN . Go'. - Wa.m0i % ji~~~ ~ ~ ~ -- -- - -4 Mt nUi ,-I 4Qc. - Nin CF u'% cp 0 cp% Q c, a u%,acw o &n u.7 a.* n-inh~ m aO% 0.. -C % J.mNN N ' P -- 4 4 - - - N. Nw a. 0 FnJ N N N4 42 m 'A 44 C.0 .0 W, a .~-. 0- 700i ~ 10LAI . i nIn r.4 00sfa l rU I1. -ja40 61 - 0 ! - ina 740Min40 Q .4 Q a 0.. 44 w m 92% 40 0 4 " 14m m N o r-f4p,. 4! 45 0 O40N i N, C N! r.c ' n4 - v 7 SM N i I, ~ i i ' i.40 - ~ ~ ~ ~ rlw 000n7 ~ 0-..... .4rinO I; a4 ,1 ,1 W ;aI l 0 N m57.i 0 in ft in a r..N.a- 007 40d n%*m o i-n a 4 a% ft pm N 0. 40 a 0- 1 SnQ -1 t - 0m 00 40 In-s z0 'iP.?' 'i-4,P.~ 0 Sn6 wa' 7i.7n w CD40 nN 0 in~ ~ ~ ~ ~ ~ ~~~~% mO?4,'@.0~C.-n,0 a A w 'a........... ~~~~~U~a4 g0e 0 ' .rn w n ,.' a o aa ------ ------------ f1p = W9. -.. 01 .- owl N=-N N N 0... 75040,00.040 'A 'a0 P.' - ". in ,~-5.i P.. In on m- 40m m~4 05o '7 47. 03 .0 .\ .' .5 . . I..~~~ ~ ~ ~ a'fS.4t 'r0i m ,m mmmf.- :loz l "NW I N,?N N,?Nr~,?-..f Pi,?~ NF,?.N - -? N,',,P.000..,040 4 B- 17 -'~~~~~~~~~ -i r.0 IA '0r a N NA .-I ioK N, -n N4 n.80.' CF- r-. W) - AI 400 A''I I j C6 t 4 P-12 A - 6M-W-F - N I A a a UM 0 IA 0 n'0a'0 1 01 r- 'I-$ ad f, a. N 0. IA 0 0 0(.? A N 0 NU 0 % WIN 7- W0 lo 1%0 N . in' 0'?a IA -11 'i A0 In In hn 0 W 1 NC-6IA a7' -a '00 111 '0. Q A - n - I 'l "~I- , C4 0 - .7'? * l 00 I00 '0 W .0 N a 0 a 041 'a IA Q 1 '7 0 so Nr 0 a, a0'- A'. A.01%- 0. %n I IA 0' '0 1AN 0A N P- - M '0 02 A&A0'-. 7A0 N 40' r% N7 WA N , 0 M ' 0 0 IA 01 N A C, 'n-'. .0 'r0'%n N7% 40 N 0. C2, '0- NI -a 0p% co0O 'n.'n 00.IAIA. 40- IA 1 0. wt .r In U, N %A 0. N r, 42 MA .0 wl a.00 I .7 N IV0. ' = A a 0I 01% = . J." I 71 -%I N 0.0 ~ a Cp 9% N0 4r, I'i pot7 '0AIA. N % 1% w V0 '0 .0 N do I 1 a - '0 l N, 0 0.0 a In N0.-. rl IA 00N40 A' AI A IA IA r%; 0.01 a, 'n a '0 N -1 N 0 NA % . % '00 Q in NA 0%I C6 U. -C MW - - - --- ---- - --- - -------- - -- --- N' 6A 9 0 m a -. 4. 7 A co % I Ar U N c p r n o a n - 4 7? ' n . . . 0 0 n0 0=' V o ' I ia ' 40 I r ' 0 I fe n ' v ' r -.n C FN- IA 'I '0 0. A ' n.'.'.. U, -.. 4 1 00 '00. N V N0 a. 43. '0%?NI A0%'000.(P . , .0W, a. -C b.0 a 0-' %0A -A 0 0. -A - -7 -, y N.. ON "0' ('1 N0' NA 1% N% ( AI..1 A 7 ''N kA' - 4M, 'N P" pn -n IA' -U 1 N oIA p - .7 %nU -1 0D N 09%Pn4 aA 0 . N c0 N 0P - :040cr N U, N C60C -- r C.UC -j roAA0@I'0A00-~-47%0%I ?I70-0-0Ial' 06nA9% NNIA'001 00A'0'01%o- WI'$ In0r n n n n l l 41' *a * D 'PF*4,0 0 0 o 0 p'0,C"C- F'0 b .. . . . . . . . . .. . . . . . . . . . . .. 0 %. .1% % 0 0 0 0 0 0 0 .000000 .... ... %f - 4---- . . . . . . . . . . . . . . . . . . . . . . U6. 4. . . . . . . . . . . . . . . . . . . . . . .4 a ama 04 4 D40 4 L & DC C 0lba 4 09 a a1M CM m ------ 4---.- - - - -- 4- - - -CD-.0-- - - - - - -- CD 4-4m CD C2 a. .43 00000000000N0N00 In In 00 a 0. .0 u0 - 0 0. a.00&M 00.a- -UQ. 1m0 rr - oV a09 r% Qr v ra0a1 n0 4 ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ l 0 Z- =00 0 00 0 0 0 0 -0. 0 .I 00.0 0- 0 0..0 0. fn fn W% W, to, I "I N''0r ''% -''r r '00r n i l A n0 n %n % 41.4 010.4 .0 a .7 01Inq nVI lQF,0, F I . 00- MU .rVNj , W fl AIIAAIIA el 40 0O .00 1%%99%0 0'0 V Q 0 0 V Q Q w w 0 0 Q NNIa, I aa. :J 9~UnUI-f .- -- 49 ..-. 10.~.NN NNC.IN C, NNNN ,10C C 0 C 0 C 1 40M. AIAIIAIAIIAIA?' - - -7' -'-?'- - 21I - 2 -W1 - - - AIAI'0.:2'0- '0-0- .0- 00W0' O~~~~~~~~~~~~~~~~ 'I ''1 ''0- N I rI I a1 '004 I r.U.1 es0 O N IIA7'A N'0 %'0 C- 1 - - - - - - - - - - - - - - - - - - - - - - Im- No '0 %0 I O00% N^It N w .7 : i 1 NN zNPMO0 N 1- lIA- N P.4 Ir O-IA-4OOIA ci W .7'0'0I W'0'0s rrr U,00 0r. .400 04.I N N A7u..04.700 U0 3 ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 7 CF-~UO .M a,0-N I.7A0s %00-I'I 1 0.04I'U'1 '0'.I~', % (YN. U ~ c oo % ; o m w % r V alII I IA A A A A AI I . . ' .7 ?' ? =7? AI U I AIA AIaAAI '0 0' '0 0' '0 01 1% %1 rw r sroautiii ~ '0 0'0%N %. 0 A . 0. 0IA'I 'I 'I 'I 07 . 0. '0.B0U '?'18%-' 0U 7 L. LU .... . L4I..... 0 L 4..-0a..% CLOf9 . . . . . OOEN~C a.a qcl, z 9L F.- . . . . B-1 '. -- coe "ac-T-lk -- IL p f - 0 o 4 N0. 0 t " l rI r I rk A n In0- v C JQ.==a .0 cD w 40 .4.In a, N .4 0. a c~ n0 7 a. In a- 0o4.NI 4.0 o ,w .4r L6az QI 4 0or ;0 ;r 1 i 1'0 n , ~~~~ r~ aoN 0 0 N0' _j . N r Nc oI Q 1" 0(k0 ~ I 4- n c4u r , -CJ W 4 CF ? a j r oj * N. A- D I ,< rI I rW - 4a-N = r USA M6 n . a. ot F1 ca r yc r' r; 4 l oaar . If om 04 4, C- m* 102- cb 2c 0c o ., ;c nc ;1 a-N0 o . r h n0 k 6 D6 -1 I3 r0 0 Q- 2C P 0 'AlI 4,T ., z lr. 0 ul13c lG 111 a 3I lc n 4 =r . r Q rN- m 0c a r%) 6 )cj " , w - . 4"0w kA 0. ., N 0 ..... "a Q a= 1 o ... %r a w ro C-b .4 N 40 F ou 4I l o N a oPIa 1 N, CD 0 006nIF00-07040070 0000%nn00 a* .0 Q 0. ,.. "j ~. 1, .A Q. -1 --1 wl 101 "1 "1 'r ' 'r 'r .20. C'J 0 -4 0 - - . 4 1. 0 Ir Q 6n N4 "' I0~ U C6 4O a0. 9 r nFl0 r D-4 I c r ad 5 4,U0 40 0U,. cDclc m0 c o 0. 'aF t0 n l'r ,i 1' wJ N N I 4 0 0'C 6 v- C lo0 N . N ~ ~ ~ ~ ~ ~ ~ e N 7-f% ~ . N N~-E . - U, u..0.0 4.4 .1 e .w -4040,GO0V0 0 G %n 17 a- 40 c2 0D in N. 40so If% 4 d. U, d .0 do do N r 41., kn20 10 l,-- - --- - -; 1- N U,%m . @0 'a Q w 40: wl 0 40@'@7,1"C 0 el cr- t W -a. Fn % i P og K4 N - mc.' cl GP In -. NNaNN n JNNC oo - %A47 ri N04N 04 ir -1"N( 9 w=cfia 04 7 a.V ,' 0V y o" =I 0 w .0 A L6. 400 0 N , I wl@ N F 430@ 4 0 " :;, .4. o o B- 21 -. mo -- :- ---- 6-f 40 UNN a-r NjZ"%na, o kn .4 42'. 'n f'aIri : 2 a sn ISr' 'a Nr % a n'aO' LL. N"NN" .aa Ij N-::,a N LNO~I U Ni Ln 'Ge.n a- 4a .' U ~ ~ ~ ' 't.?a@' %Ao@LIN 29 c.Ji .- a LL.o tnr o c 0b N ulC 40 . .. .~ . . 06a 'a'e I 1% P, le %n.4 0 a P -N 47 a B-22 iL.Q n lna- . . . . . . -1 e lL% .4, up . .404 N 47 U)C Ya- 0 I d'n wt r a 0@a In. . .. r ."hflrfr .4 . a, Co @aC Da ca aI'D a a@.@@4 c Q .4 L6 C j. O km ,? .4 i, CD O f~G .. . . . . . .. 4 E . . .4 W% -Ir-a C r o 1.4 I rs p N do ~r.4~ In o q o -, = 40 I . m In P .A. . I-, CP Ok "I' D 0 W Ufl~o 01 CN sr A "N-14 -AL6 N in "~ w Inke L6 .11 U. .4 C6 U.' N cy W; N 0 -0 I )... N N F .a.0q. . .4. . . . a. I P, 40 EUAfAI I InI cn IN In 'aa cc rC. , c. a ti . N' n n'.6 a p C ; In Jr 0- n 'p, un 0,0e.Q -c C3,~.. a - - - - - -@'- - - - C. a .f -. 0V0 . 1.C0N.0N~'' 0.7w .1 1r.4 t 4 . -!! C00 --- I C~i 1% N C I c2. v O'In - 00 00' -r 2,n ww ,Oa G -C -r %A'0NO'NN'0In in Q, J,' -~l a0--.I - n'0'I 0 NN N ey0. - 0-N- - -0. -* .n0 'nn .'. 0 .N 7 . . . .70 . . .0' . . . . . . . . . . . . . . . . . . *~~~~~C 1o "1ooaoa oaa .0 . 4cl' - :aao.0aooaoa-Ca2 ( O 0. o C* 0a 0" t .4 . m Q Q00.. 00 00 .000 0.0V,-N0-000 r 0rrrrI'rrr I rI rwI r I rI I IV 1 ' n0I .Y7 I .7 .77777777........ r.7.7.7.7.7.7.7.7;.7.7.7.7.r7.7.r7... r 7.7.n 7 n n n n ,, v-% In u -I I n .1 N V4e D 4 1ulI In.~ 7'0 ' ONn0I7n r. O O OO a oa~o O 000000Z30000o0000. M Na.7.00.7-.7-1777InN-In0 0 I 0a o o o o o o oo o o o o 0 0 0 0 -0 -. N0 a Q-0 0 -0 0 do-0 '0NNI.0 0 00a0000003N0'a0000000z00007 000 ' ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ' '00n0f'I.'.77-I.''Ij. N0000 N 1-0000 'O00 0 - 43 N77 7 7 7 7 7 7 . . . . . . 7 * * 7 7 7. . . . . . . . . . 7 7 7 7 7 7 7 I n I n I n I n I n I n . 7 I n I .n0N0~7e77 N''7 70',.e.0r o oa oo o oo oo ooB-24oo oo o a. f 0I ca 0c CDD~ r. ,4. In0 o a N N NC4NN NN4 Nep. N c%6N AN N 0.. " o4o G.a. 4 4---n l -n -- =-- .. .c e . . . .a . .- . . .. . .G. F. .'a40'a*pmtor fowwM n p 00..lI 6 . 0c ;0000080a0000. 4r . U4- . . Z- 0 . . . . . . . . . . . . . . . . 29 x .6 &aS D4 c 5 c ad 4 3 .8. 0 p 40.- o a a ID "j N e I ID B-2 ak -1 LA O.- 0. 0- :'CI .4 10 "1 VI.~ 0 C- .C . ' . . . . . . . . P.. .. .7 . .. -; .* 0 . . . . . . . . . . . . . .C C C a a C = oaa ta42C do =r cOa OjaIaw fZaI e r I s W a a'r' 00C %a- If 4). COO- r. -J, 04 . . ..- .. . .. .. .a .. .. . . . .aa s a ~ a - a - a - a - e - ~ - O . -o-o-o..a.a . . aaoa@@0@@-a@@aaa0 00'0 Cea o a.a CM. a aa .0 C *. a . . .a . .aa-. c.. o0.a .. ea. 00 4 PI .AL - . '0. -0. af N a . N~ a 0,. 'a. . aaaaaaaaaeaaaa .aa . . . . . . . . . . e . . . . . . . . e a e m a m. .a. -C 0 a a a a emaaD a10 m;ea 1a a 'a eo C 0,am@A~A~mefUSAfm~UkUllaa M 0.@2 10N U9d' t..q000 @ @----------- 4 Q. 1 a, 0.0000....000....0000....000 .--.-.-.-.-.-.-.--.-.-.-.-.-.-.--.-.-.-.-.---- . ' . . . .. .. a -- 2 pn I0n 'D6 1 r s j n- a N 3~ a0 - - ca C.J .J N . Q N6N M C N 17 Un 120 o P, c C', " Go CDw oc .I0 * ' J "AA IAAI, Ln a- P, P. .?S 0 V, 42.O rl: c Du *0 000 Pco oc c c oo o c coo co co oc a-w a 0 - ~~0 - c . CZ, 0, '0 V 4 ,0 ,4 PM I . . .- J . .C ,0 0* 01 01 IN6 -14 eecla l 4 DDo 0a4 bf D4a= mz - - - - - - - - - - ~N~, 0 00 a C 4 aoC C oM4 co 0 0 00 0 00 CC 4* o aCDCco=cao= Lm - 'Tmi- II10 46 . r d -D 43 aC 1 ! I.J -11! ----- 2 11 NN . . . .. i 0 0 . . . . .. . . . . 00 4r 0 O . 0 ' 7,' oj U, - -- - - --- S-3. - -0 NO. Ln NInna a-?DWIAL~'AIAIAI 4 m 00 Na0 'a e 0 a N 00 0 C "i Ny- 0 Ir - 'n N Cj N- 10 4p -0- ', 40 =ZN, cv IA f- a33 -P, 'a0 s o 40~ .0 .S . .- .. . .~ .S .S . .e .. t .A .0.3 . . . . . . . . 0 4 1 .0 OS a- .J0 IA .2 0N eA -4 ca 000 ri - Ir ,?I A.t.N A. % IA~ ~~~ ~~ .. ~ . . . . . . . . . . . .3s 0 S NI ,I , P .OS... .?I..3.s.S....I.?I... . NI ?l. . CO 4 4040o D 4 c 4 4 g 4 c A c C C doA:: ya yP y C U" ~~ ~B 27linw u', In P r Z" & -r ?-t In M.r-= P 0 rwr"Z% HU -6 w aOP wl rWW .W.P.O6 cLU w4'o ~ a..i F C- fu 0. so a JCso "1-..0 p N, " to D N s c*7 cl, 0 0 a w. 4 y -r I -O -s0 v- -s-ac'- at 46 4 13-2 PUERTO RICO AND THE VIRGIN ISLANDS !J~m IN 1140. a-~' 'M 0.0.0%n 'Q 0. CD 00 . C, 0 0 , Aj . . . . . . . .f.0.J W3 00Qn11 DIn. 1 0 7-Ma 4N 000 00 00 -0 D .0 01 r -- r O OOrCaO2 r 0. C -o rr d t..0 ".0' .4 NO W 00 -. ? r'-. 6A. 40 CD.l Z0r Go" Cl1 00 Calfi VNa o D ni C-4 In ac i" ' Ca - 10 .' . . 6n 4 C - .' . . . . . . .? . .I . .~ .0. , I.. C3 C)- ; C ac 4 2C 3 o DC3 2c D 0 aC mC 00c 3 C0 o0 DNC 0a ' 1 l r(i aIw N - P6 1."-Q " 010n - 4 '- O4 . . .0. . . . . . .' . . . . .0 0. . . . . . . . . . . . . . . . . . . . 10I 40 .4 2 -4 C4d 0 .N X "i 000000000000400u 000C000c000040C A 2 CD 00C20000D00m00050000I=CD000000a000 - ~ ~ . 60. a 00 0 00 0 0 00 00Q0 0 00 0 A< c3 0?0.CJ.T 'Z ~ ~ -4 cc z. In n~ .7 4= Z ACD< 1 =C -M - =6n '-' CD ' 'J = -Z 9 - =1 = C1=z= =I o% rrQ Cj 2 p % " == a1= Q CDQ C200=( 4 0 ZQ D=_ _m 61C.: c?' . " No4'4 .40 .4. ~ 0 @0 An n n a 0-9*C as C do 4 co CD 10 Cca aff 4p 40 9; cb Ij -m oj cafo ob a a aCe.4 a dbda a ao 4a C a* 9 a0 a 0 I" U. In 4c B-3 AO-AL05 112 WILSONHL SSOCIATES INC WASH4INGTON OC F/6 1/5 COMMUT H AIRL NE FOHECASTS.lUL NAY 81 H NEUVILLE. C STARRY. 0 BERNSTEIN DOT-FAT9WAI-138 UNCLASSIFIED FAA-APO-81-7 NL 22 fllllffff 2 IIIl -2 m~.. r IN..nd 4 f.'4ag 3 do*P-.4 40 d N m Ar - .7 N.7.04 00@ "N ON 'aN c8 .0 wV J. . 0 'IO i0~7~. .- U. 'r ,r W II W)~ N1 *.ri ,' I@ 0 COSrA 4i . U.7 w w' I. 7 N. -r N WA.? 00.0i.400a 40 0 ag 0 42 .70 7 aO 9I40L .4 7 c .t0P .. f'... o. NA.tr YN .9 ca ifl~~~ mL Nf ca -0 'n-JrII 1r , S.:0% d 4 4 b -~~~ .'' 000 P00 00 0 00 0 00 0 00 0 00 0 tv4 -N- 0 NU 47OO -10 4 a ,wa . a ol W .0'IL ra .. 0 i 0L-.C0 &A 0 ON. - '0. n-NmW Ana. 0r6 wt G. CI .r.k P IN-0-2 4L mJ N 0000 .1- 7 07 r p %c J 0 in 9 0 W, rim00 N07 N o ;9 04% a. :0- U, r' L - l - 0 0 ca A 0'0 . 7'I0 N Nwlo - M "n mL A ' V 0 .* CF4 F. 0. Ca ~ ~ ~ O I' 4NO' .. , 0IgOC0 a"40. - . . . . . . . . N. . . . . . . . . . . . . . . . in 41 -,W M 0.w .. 000..0Z 0- @N A.9 'a 0%MJ% U,, CPAAV.p-; am .O00N0'@0'@0' P.0. i 0. 0 a0 mm 00A AOa d.o in do w- * .4 0.oaa m 1ee =Qee = C-400 0.I 0r: C 0 " O Mw N00 'r N 4.- N oAr= :1 - l b mONV9NONNONOO'I0a,**00'Fz~w 1. Ir% r r t n u ZB 3 - LA 0004 NN9% 0600'0'0sum0 C% Wa%%N c.I In a 40 NdorZ V a. NN e NC %a .. n Ui 4. .4.. 0 F,% 0.J .U, .4 %nl C. ~ aW 11I ,I - N P. . 1-1 . P." wl go Ir %r .4oe.J 10 .4C,% - a.V oD cb C 440:u '7~ O 4 C 0c ! C f. V . L I or tr- CN 0 L64 InU I -r 40 40d o4 .V 0," %~eo~00.0 wm inW~. %.4 441 .4 B- 32 I W. I APPENDIX C LOCAL MODEL EVALUATION ADDENDUM I:I .. .. i i I iai Two of the variables employed in the market forecast model require special comment. These include the distance measurement and the dummy assigned as a measure of competing markets to alternative hubs. Distance Previous studies of commuter activity have employed distance as a predictor of community "isolation" and the demand generated for commuter service. But the relationship is not clear; Figure 14 indicates the problem. The steps correspond to the annual domestic traffic volumes (left scale) by mileage group in 1979.* As can be seen, ridership increases in markets up to 100 miles, then decreases. The relationship between the mileage in a market and commuter traffic is slightly more complex than can be measured by distance directly. Superimposed on the step-function is a gamma distribution; a stat- istical approximation to the step function. On the right scale are values of the gamma distribution corresponding to mileage intervals. The ga distribution achieves its maximum value at approximately 100 miles, and is less for other distances. For the purpose of modeling commuter traffic behavior, the values of the gamma function are interpreted as "likelihood" estimates. That is, commuter carriers have the same "likelihood" of attracting passengers in 35 mile marekts as they do in 160 mile markets (a "likelihood" estimate of 0.095). In either of these markets there is less "likelihood" to attract passengers as there is in 100 mile markets where the estimate is approximately 0.184. The gamm distribution value has been used in lieu of diatance. A positive correlation is expected between traffic and this distribution (GDIST). This was demonstrated in the market forecast model. Table 7, Coiter Air Carrier Traffic Statistics, 1979. CAB. C-1 I!I 2,000 0.190 1,500 0.143 0 * 4J 0 S--- - .0095 INV 0 25 50 75 100 125 150 175 200 225 250 275 300 Mileage Figure 1 Distribution of commuter passengers by mileage, 1979, and ga distribution approximation. " Source: CAB data and SRI analysis C-2 i n u n n n 1 1 I . .. .. .. . .. . . . Alternatives The extent to which a passenger has a choice of alternative hubs at which to connect will reduce the traffic in any single market. For example a passenger in Grand Rapids, Minnesota has commuter access to only one hub--Minneapolis/St. Paul. A Fresno, California traveler, in contrast, has access to a choice of hubs, including San Francisco, Los Angeles and Sacramento. We expect the presence of alternate hubs to reduce traffic from what it might be if there were no alternative. Two dummy variables are introduced to reflect this competition, ALT 1 and ALT 2. Both are assigned a value of zero if no competition appears to exist for the route being examined. If the originating city has comparable access to an equivalent hub or slightly less convenient " access to a better hub than is provided on the route being modeled, then ALT 1 is assigned a value of 1 for that route. If hub access is better in another market than on the one modeled, ALT 2 is assigned a value - I (ALT I - I, also). We anticipate the sign of these dummy variables to be negative as traffic will decrease as ALT I and ALT 2 increase from 0 to 1. C-3 ____ ___ ___ ___ ___ ____ ___ ___ ___ ___ 1 i I4 SAMPLE CITY-PAIRS, MARKET FORECAST MODEL, YEAR AND TRAFFIC Origin Destination Year Traffic Abilene Austin 1978 2,578 o " 1979 3,319 Abilene Houston 1977 4,070 " 1978 5,489 " 1979 7,577 Altoona Pittsburgh 1975 17,568 " " 1976 20,822 " " 1977 23,219 " " 1978 24,811 " " 1979 25,679 Walla Walla Spokane 1975 3,653 " " 1977 3,762 " I 1978 4,004 " " 1979 4,849 Lewiston Spokane 1977 1,927 it " 1978 2,402 " " 1979 4,511 Pasco/Richland Spokane 1975 2,788 " " 1977 2,868 It I 1978 4,767 1979 5,938 Pullan Spokane 1975 4,498 " " 1976 3,587 " 1977 4,733 1978 4,926 " "1979 6,771 Yakima Spokane 1975 4,575 i it 1977 5,123 " " 1978 5,469 " " 1979 5,995 Bloomington Indianapolis 1975 3,303 " " 1976 4,376 " " 1977 6,229 i t 1978 6,840 " 1979 6,706 (continued) C-4 " ' I I I I I I • II II II I I II Destination Year Traffic Terre Haute Indianapolis 1975 7,097 " i 1976 7,490 i i 1978 9,607 " " 1979 10,594 Muncie Indianpolis 1975 1,924 1976 1,595 1977 3,440 1978 6,828 Augusta Boston 1977 28,059 " 1978 27,537 1979 26,930 Hyannis Boston 1977 39,467 1978 42,838 1979 39,368 Lebanon Boston 1977 35,267 " 1978 47,775 " 1979 40,064 Montpelier Boston 1977 11,767 " " 1978 8,596 " " 1979 5,147 Portland Boston 1977 25,544 " 1978 22,913 1979 15,366 Source: CAB data C-5 APPENDIX D AIRCRAFT INVENTORY AND OAG SCHEDULED OPERATIONS BY AIRCRAFT CODES 10 In In f. -Y . n 46 - N FaF.4 t aW y C . N4D %a M4 0s 4 A.-I4ZWP. . . . . . . N .4 . . . . . . . .4 . - . . .4 . . 40 P" .4- 0 A fU A59. SAM L.SZS N ^ . ft 0 tV 61 a ~ 0 W 0 ft M " 0 N M ft b . CD N NY V4 0 in -r A.4 4WSP 96 40) w 0 -C4 * a a ND b so In 0 f . . a . . In . 4* NY L CY cy I., h.. do"I r. a. X, wnu -g . n. . . . . . in U.4. 44 KZ -I 12 10 0- w a. W -5 5 =4 = c 4 w. .4 N a a m a a a. a. .4 z 2 a inVSM bI qn a ag 0 * 4 j a 0 . a OF - -Uy .r 'a- .16 S0 in a a, .4 .4 , Ul In a 4 44 On I& Fa- - D-1 U. . n n .4 N -W4 N I, in cy . a N N - In 4 ~ ~ ~ ~ ~ ~ P Cy4W1 U a * a .. N0 N 4 a 4pU in 4. * p a a a .N .4 . 4 . .. .. .... .. . . . . in a UO* NN 4in 4. . .4. ~z r4 a a M 0D a V. 0 40 N4 M * a a r. n 4 . a aN 4 ~ ~ F inJ0.n P.4 La in 0Q.4 4 nr-n . . . . . 4. . . N . . . . . . . . . . . . . . in 4V in 0 N . =~I r4.I ~ O..d4WN4 a a a e 0 4 a 4 N U . a . a. W~A .0pnU.4 a... 04 in Isn9 of 0.t4 . D- a~- -M a 'A0IA As 4 NZ - ty IAUCO.I Nz 44 IIV 40 to~4 N a -D a0 m- t 0 IA'00.IAV%40~ . .. . . . . 4 . . 0 . *N In 0 N N A 4 0 a. 40 61 : 4 4F W. AI IIn 21. Ni 40 I" IAI - N - - N N zw 0 I= 44IA 41I f - C . . . .. . . . . 4 4W 40 *U 40 O S O > -C- - . C C - A IA 4A 0- 40 w N 4 -a -C -0 of a0 ad 0 4 4 I* 0 4 0 4 AL~r - 0 - - - - - I -C - l C C C - 4 IA U6 UU -D-3 |0 Irj . . . . . . . . . . . . . . . D 0 0 a NO 4 0 1 LI- l M IaNI -P 4 NIO O * i~ l +4 N • ,' f 4 1 N O NI e, IV ob -. NN 4 46 9 Z MA. V'I IP 4 00 N + * W * a * I * 0 * * W a .C a a wp4o6npm . 4 . . . . . . . . F . .i . . . a.-§4ZW~nW -d - 0 4 au .4N4 . L-,. J'q ZMIIeD P N I.O O : * pO 'e * * 'D * 4: +4 * O O 0 * a e-. WCIA 0d rd in aJ IF uJ t t In in4 in a l a i.n w . Z M - 0 0 04 4' - N * - 0 , 0 4 0 4 * 0 W D40 ma a a 4 - 4- I. ... 4ZW 4 N 0 0 0 0 0 *N *0 N * *0 An W w W W x 0 a aaI - a.-dSao A 0 a le ft 0 : 0 0 D- aa -~~ ~ P% aA I . . 0. L.J4ZWNU. . .~ . . A . . . . . . . . . . . . . . .* 4 - % 0 0 D b In au *I e-u 4 a 0 N0 I-m .OZ P- "W 0 6 0 . . . . . . . . . . A . N . A . . . . . . . . . 0 0 L...OZWPin n a *P 4 A 0 Ps v A 0 'A * w 0 a 0 -a 0 A4 n 4c in Z In o a. n p = Nt IN NY .0 P 0 a 4 O 4 . . 4 0 . . .. . In M a a a pn Nn on 'AN 's 0 Ps ~ (lWPsf.....................a N . . * AV - N 4c -61 6A 14 16.6 4 slo a 4 46 161 O n4. : 0" w . ob 0 D- .4. .fU ~ s . . . . 0 NY 4~ - CF- Ov=4ZLO * 40 . . . . . . . fy 6n N6 02 0 . -C LaN6 4 D nA I-- 4p U~~C .410~lN ... C6 qs -C X L0 P-0* * , a a aY a 0 0 0 ou= . . . . . . . . N . N . 0 C6-C U0 in cy a 4 0 aN In o 4 -0 4p o * at NN atN in I 16. * m 0. . . ... N aa z vi 0 4O 0 p l- a, ox tog 15 m X. No N~~P " n0 in 4 d - ~ .~D-6 DAT I M I DTIC