1y ago
448.18 KB
20 Pages
Last View : 1m ago
Last Download : n/a
Upload by : Elisha Lemon


Appendix D.1:Transportation—Ridership ModelingA. INTRODUCTIONThis Appendix describes the ridership forecasting methodology used by New York City Transit forthis SDEIS. The primary model is the Transit Demand Forecasting Model (TDFM) developed byNYCT.1 A second model, the MTA Regional Transit Forecasting Model (RTFM), provided forecastsof commuter rail and other suburban transit trips that included a NYCT subway transfer and modeshifts between transit and non-transit modes. The TDFM provided transit ridership forecasts fornumerous alternatives and sub-alternatives proposed by the Second Avenue Subway Study. Transitridership forecasts include projections of the number of transit trips by origin and destination and bysubway and bus route. The main components of the methodology include the following: A computer model containing a detailed representation of New York City’s subway and busroutes within the context of a geographic information system (GIS) A census tract based zone system for locating the origin and destination of trips A base year (2000) AM peak hour origin and destination trip table Future year (2020) AM peak hour origin and destination trip tables for projecting usage ofexisting and proposed routes using the subway and bus route computer model.This modeling work has been closely coordinated with regional transportation modeling at the NewYork Metropolitan Transportation Planning Council (NYMTC), the federally designatedmetropolitan planning organization (MPO) and the Metropolitan Transportation Authority (MTA).The regional models developed at both of these organizations use NYCT’s subway and bus networkdatabases.B. MODELS USEDTRANSIT DEMAND FORECASTING MODEL (TDFM)The TDFM is NYCT’s in-house model of all subway and bus routes in New York City. This modelhas been developed and enhanced over the past eight years using TransCAD, a software package thatcombines transportation demand modeling with a Geographic Information System (GIS). TransCADis a travel demand forecasting package that is used by many highway and transit agencies, as well asMetropolitan Planning Organizations throughout the country (including the NYMTC, theMetropolitan Planning Organization for the New York metropolitan area). The TDFM has been usedextensively at NYCT for short-range planning [e.g., Lenox Avenue Line and Williamsburg BridgeReconstruction projects) and long-range planning (e.g., East River Crossings Study and the MajorInvestment Study (MIS) and Draft Environmental Impact Statement (DEIS) prepared for theManhattan East Side Transit Alternatives (MESA) Study in 1999]. It has also been integrated into1Modeling Transit Demand in the Big Apple from a Transit Agency Perspective, James J. Barry, Jeffrey Erlitz,Robert Newhouser and QiFeng Zeng; presented at the Seventh Transportation Research Board Conference on theApplication of Transportation Planning Methods, March 1999.D.1-1

Second Avenue Subway SDEIStwo TransCAD based regional models, the MTA’s RTFM and NYMTC’s Best Practice Model(BPM), which includes a full highway demand modeling capability in addition to transit modeling.The GIS capability in TransCAD has enabled NYCT to include the spatial detail that is needed toaccurately represent the interaction between transit travel demand and the location of bus stops andsubway stations. For example, the model includes a walk “network” that allows it to predict trips thatmight use the street grid to access alternative bus and subway services, rather than moresimplistically assuming that all trips are made from the center of an origin zone in a straight line to asingle subway or bus stop. The model also represents the characteristics of individual transit routes(e.g., the Lexington Avenueandservice) separately, rather than the characteristics of a wholetransit line or service (e.g., the Lexington Avenue express service). Thus, the model can predictresults for each route separately.The TDFM estimates how people use the transit system in New York City during the morning (AM)peak hour. The AM peak hour was chosen because the travel demand is highest during this hour, andtransit facilities and services are sized to accommodate this level of demand. (Typically, trips aremade in a more concentrated time period during the morning rush; during the evening, trips arespread over a longer period.) Travel demand is represented in the model by transit “trip tables,”which list trips made by subway, NYCT buses, and other local and express buses in New York City.The trip tables contain estimates of the origins and destinations of patrons using the transit system, byanalysis zones, e.g. census tracts or split census tracts. The origin and destination trip tables wereconstructed from the journey-to-work data provided in the 1990 U.S. Census of Population andHousing, Census Transportation Planning Package (CTPP). These are the latest Census data of thisinformation currently available. The trip tables were also updated using recent ridership counts,information on origins and destinations derived from MetroCard data, and a prediction of futuregrowth in population, labor force, and employment by census tract in Manhattan prepared by AKRF,Inc. A more detail discussion on the prediction of future growth is in Chapter 6 of this SDEIS,“Social and Economic Conditions”.The TDFM estimates route choices using information on transit service (routes, travel time,frequency, transfer connections) to predict which route patrons will choose. It is used to produceforecasts of riders by subway or bus route. These forecasts are suitable for general planning analysis,such as comparing the No Build Alternative and Second Avenue Subway alignment options in thisdocument. The model outputs from the TDFM provide information on the volumes (i.e., transitriders) on various subway and bus routes, including passengers boarding and exiting trains andtransferring between subway routes at various stations, and summary statistics on passenger hoursand miles traveled by transit. These data were used to evaluate and compare the changes in ridershipand service operations between the No Build Alternative and alternative alignments of the SecondAvenue Subway. Potential improvements and decline in service were compared using such measuresas station leave-load volumes1 and volume-to-capacity (v/c) ratios.2REGIONAL TRANSIT FORECASTING MODEL (RTFM)The RTFM is a model of regional travel in the New York metropolitan area, including NYCTsubway and bus riders; commuters using Metro-North Railroad, Long Island Rail Road (LIRR), andNew Jersey Transit; automobile travelers; and people using other travel modes, including taxi,1Leave-load volume is the estimated number of passengers on-board a subway train as it leaves a station, basedupon field observations or TDF model outputs.2Volume-to-capacity ratio is the ratio of passenger demand to capacity of scheduled subway or bus servicesduring the analysis period (e.g. AM peak hour).D.1-2

Appendix D.1: Transportation—Ridership Modelingbicycle, and walk. The model divides such trips into three types: home-based work (i.e., a trip towork from home), home-based other (i.e., a trip from home to another destination), and non-homebased (i.e., a trip that begins away from home). The RTFM (and its earlier versions) has been usedfor major transit studies in the region including the LIRR East Side Access Project, the MTA’sLower Manhattan Access Study, and Metro-North’s Penn Station Access Study.The RTFM network database includes NYCT’s subway and bus network from the NYCT TDFM aswell as the commuter rail system and is used by NYMTC for the MTA region portion of the BPMtransit network. For this SDEIS, the RTFM was used to provide forecasts of commuter rail ridersusing the subway and estimates of changes in mode usage (e.g., from auto to transit). The RTFMprovided station-by-station estimates of commuter rail riders who would also use the LexingtonAvenue and Second Avenue Lines, as well as commuters who would shift to transit from othermodes.C. APPLICATION OF TDFM FOR SECOND AVENUE SUBWAYThe steps involved in applying the model for the Second Avenue Subway Project are describedbelow and illustrated in Figure D.1-1. The transit modeling process entailed four basic steps: 1)development of the zone system and transit network for the entire city (including coding andvalidation); 2) the development of existing and future travel demand (origin and destination tripmatrices); 3) inclusion of the RTFM forecasts of commuter rail to subway transfers and mode shareshifts; and 4) assignment of future trips to the future networks.TDFM PROCESSThe ridership forecasting process must account for both the supply of transit service and the demandfor the transit service. A change in either one can change the forecast of ridership on a proposedroute. Figures D.1-2, D.1-3, and D.1-4 contain charts of the process used in this model with transitservice shown on the left side of each chart and transit demand shown on the right side. The currentprocess was updated from 1995 base year networks and trip tables used in the MESA Study’sMIS/DEIS. The base year (2000) process includes coding scheduled subway and bus service into themodel and developing an origin and destination trip table to represent transit demand. This verydetailed process must ensure that the travel choices faced by travelers are carefully simulated,including transferring from one subway route to another or between a subway route and a bus route.Similarly, sources of information on transit demand must be carefully selected and combined toensure that an accurate origin and destination table of transit trips is developed for use in the model.The calibration and validation process, shown at the center of Figure D.1-2, is used to accuratelyestimate the volume of passengers on each subway and bus route by adjusting parameters in the tripassignment model, updating the trip table and verifying coded service characteristics.Figure D.1-3 shows the development of the future year (2020) No Build transit service and trip table.To ensure that the service and demand are interacting in a reasonable way, a checking and validationprocess is used. Major transit service improvements in the No Build alternative include the free busto-subway transfers that started July 1, 1997, and the opening of the 63rd Street connection with theQueens Boulevard Line in 2001. Additional transit trips from the commuter rail and mode choicemodel components of the RTFM are added to the trip table to produce the final 2020 No Build triptable.With the No Build network completed and validated, the coding of build alternatives proceeds byadding new services and modifying existing services to accommodate use of existing lines by theSecond Ave Subway service. The final build alternative is passed on to the RTFM Mode ChoiceModel, which estimates new trips attracted to transit. Figure D.1-4 outlines this process and showsthe final product to be a set of alternative specific trip tables that include the new trips attributed toD.1-3

Second Avenue Subway SDEIStransit. Each of these trip tables is used as input to the transit trip assignment model. The modelproduces the final ridership forecasts for each alternative.DEVELOPMENT OF ZONE SYSTEM AND TRANSPORTATION NETWORKSDEVELOPMENT OF A ZONE SYSTEMThe TDFM evaluates a travel network made of different travel modes. For this evaluation, a zonesystem was created to represent the origins and destination of trips made in the network. These zonesare used to build trip tables or origin and destination matrices that show the number of current (orfuture) trips that people typically make between each zone. Census tracts were selected as the zonesfor the study so that 1990 (and in the future, when data are available, 2000) census journey-to-workdata could be used.In this model, New York City is divided into 2,288 zones (see Table D.1-1). Each zone has a centroidthat represents an approximate location in the zone where trips begin or end. Using over 2,000 zonesreduces the size of the zones so that any errors due to this approximation are minimized. To providegreater detail close to the alternatives’ routes, a number of the census tracts that straddled Lexingtonand Second Avenues were split into smaller units (census block groups).Table D.1-1Zones by BoroughBoroughZonesNew tal2,288

Appendix D.1: Transportation—Ridership ModelingFigure D.1-5Manhattan East Side Zone System with Subway RoutesROUTE CODING AND NETWORKS DEVELOPMENTNYCT has revised and enhanced the TDFM model to take advantage of computer speed and memoryimprovements and upgrades in the software package, TransCAD. Transit ridership forecasting for thelong and short list of alternatives in the 1999 MESA MIS/DEIS was developed using the traditionaltrip assignment modeling approach with the following enhancements: A detailed NYC transit network with all NYCT subway and bus routes and all NYCDOTfranchised bus routes, plus an extensive walk network covering most of Manhattan. All transitroutes and related transfer passageway are re-coded into the “route system” format. A detailed zone system with 2,288 internal zones comprised of individual census tracts or censustracts split along block group lines. Two external zones were developed to account for MetroNorth users transferring to subway routes within or in the vicinity of Grand Central Terminal and125th Street and Park Avenue.D.1-5

Second Avenue Subway SDEIS A Stochastic User Equilibrium Assignment procedure that accounts for crowding constraints aswell as the multiple transit route choices available in many areas of New York City.Transit Route SystemsForecasting transit ridership changes resulting from major service changes or additions requires amodel of New York City’s large and complex transit network. It is important that this model not onlybe accurate and verifiable but that the updating process be efficient and the addition of proposed newroutes be manageable. To achieve these objectives NYCT switched its route coding to the new “routesystem” feature in TransCAD. A route system is a collection of routes stored together in a map layer.Each route is composed of a series of line features that are stored in another layer in the map.Bus routes are composed of line features from a layer containing all the streets in New York City.Subway routes are composed of line features from a layer containing all the subway, surface andelevated rights-of-way in the New York City subway system and the Staten Island Rapid TransitSystem. More than one route can operate over a single line segment, whether it is a street or a railstructure. Adding a bus route to existing streets or adding a subway route to existing lines is arelatively simple operation with this route system feature. When a route uses a new street or subwayline, the street or subway line must be added to the line layer first.CREATION OF AM PEAK HOUR TRANSIT NETWORKSTransit networks were created for existing conditions in the base year (2000) and for the No Buildscenario and each Build alternative in the future analysis year (2020), based on route-by-route serviceplans.Transit NetworkThe transit network is a computer-based model of the transit travel choices available to passengersduring the AM peak hour. It was designed with sufficient detail to represent accurately the numerousroute choices available for travel to areas, such as Manhattan south of 60th Street and downtownBrooklyn. The components of the transit network include the following: All NYCT subway and Staten Island Railway (SIR) routes with AM peak hour servicefrequencies, travel times, and passenger-carrying capacities. A walk network with walking times and capacities for passageways between subway routes. All NYCT and New York City Department of Transportation (NYCDOT)-franchised local andexpress bus routes with AM peak hour service frequencies, travel times and passenger-carryingcapacities. NYCDOT Staten Island Ferry route with AM peak hour service frequencies, travel times, andpassenger-carrying capacities. Access and egress links for connecting zones with subway stations and/or bus stops or with streetintersections in the walk network areas. Fares and transfer charges in accordance with the MTA’s current fare policy.For future conditions, the model’s transit network was adjusted to reflect a future service plan for allroutes expected to be operating in the analysis year. The No Build Alternative’s service plan includesthe recent completion of the 63rd Street Tunnel Connector in Queens and the future restoration of fullsubway service on the Manhattan Bridge. The No Build Alternative and the Second Avenue Subwaybuild alternatives also include the LIRR East Side Access Project, which will bring LIRR trains intoGrand Central Terminal.D.1-6

Appendix D.1: Transportation—Ridership ModelingTransit Network SpecificationIn order for TransCAD to assign trips to routes, a network file must be created and configured todepict the key features of New York City’s transit system such as express/local subway servicesand numerous transfer passageways between subway lines. In creating a transit network onemust specify the source of all attributes, i.e., the route layer, the stop layer, and the line layer.The route layer has descriptive information such as route name, mode, headway, passengercarrying capacity and fare amount. The line layer contains length, transit in-vehicle time, andwalk time. For in-vehicle time, NYCT uses the travel time table look up option. The line layerincludes all streets (see Figure D.1-6) in New York City, but in order to provide reasonableconnectivity several link types were specified. These link types can represent either walkinglinks or different types of station access or transfer links. Together these links are called nontransit links. The stop layer has subway stations and bus stop information. Listed below aredefinitions of the various non-transit links.A transit network combines all the routes and associated model components described aboveinto a model of the transit system as passengers would view it when choosing their “path” fromorigin to destination. Transit networks are used to solve “best path” problems and for performingtransit trip assignments. Transit trip assignment (with capacity constraint) is a mathematicalprocedure that takes information from the transit network, computes the “best path” for all thetrips in the trip table and checks the results for links with total passengers in excess of thespecified passenger carrying capacity. Links over capacity are given a penalty value that makesthe link less attractive for the subsequent iteration of the trip assignment.Figure D.1-6Infrastructure Layer (Streets) VWW39E4XI WROCKEFE9AVEW52SDISONWWDRELLPLENTRYRP

Second Avenue Subway SDEISA walk network consists of street links included in the model to allow trips to go between zonesand bus stops or subway stations. These streets can also be used to transfer between bus orsubway routes if there is no better connection. Most streets in Manhattan and DowntownBrooklyn are included plus all other streets used by at least one bus route (Figure D.1-7). Thiswalk network allows the model to account for the typical New York City practice of someonewalking past the nearest subway station to acc

Appendix D.1: Transportation—Ridership Modeling D.1-3 bicycle, and walk. The model divides such trips into three types: home-based work (i.e., a trip to work from home), home-based other (i.e., a trip from home to another destination), and non-home based (i.e., a trip that begins away from