APPENDIX:  METHODS AND SOURCES

_______________________________________________________

Acknowledgements:  Staff of transit systems and other local, regional and state agencies has provided data and source material for this report.  Many have carefully reviewed drafts of the report and offered many helpful suggestions to improve its accuracy, clarity and usefulness.  NVTC appreciates those efforts and cautions that any errors should not be attributed to those organizations.  Notification of errors or questions about any aspect of this report should be directed to NVTC (nvtc@nvtdc.org).

TRANSIT SERVICES

Rail Services

            All data provided directly by individual transit operators.  All passenger trips figures refer to one-way unlinked passenger trips.  Passenger miles and vehicle miles figures for Metrobus and Metrorail in Figure 2 were calculated based on Northern Virginia ridership as a percentage of total system ridership.  Operating expenses for Northern Virginia Metrorail and Metrobus were estimated by adding the operating subsidy to estimated passenger revenues.   Average trip lengths and on-time performance for Metrobus and Metrorail in Figure 2 are system-wide figures. Ridership in Figure 7 and Figure 9 calculated from Monthly weekday/Saturday/Sunday average ridership reports.  To obtain an average daily or average monthly passenger trips figure for each of the Metrorail stations in Northern Virginia, total daily boardings for each station were doubled, with 7.93% subtracted for intra-Virginia trips.  The 7.93% figure is a WMATA estimate.  Annual ridership numbers for each Metrorail station in Figure 8 were calculated from average daily figures in Figure 7, using the annual numbers of weekdays, Saturdays and Sundays as a multiplier.  Annual Northern Virginia Metrorail ridership in Figure 10 is the sum of the Annual Weekday/Saturday/Sunday ridership for the individual stations listed in Figure 8.  VRE average daily and annual passenger trips for individual stations (Figure 11) calculated by multiplying FY2002 ridership (NTD) by the percentage of total ridership for each station (calculated using VRE data).  Annual and Monthly passenger trip figures for VRE (Figures 12 and 13) provided by VRE.

BUS SERVICES

   All data provided directly by individual transit operators.  All passenger trips figures refer to one-way unlinked passenger trips.  Monthly average weekday/Saturday/Sunday passenger trips for Northern Virginia routes are provided by WMATA.  The monthly averages were used to calculate the average daily ridership in Figure 15 and the annual weekday/Saturday/Sunday ridership in Figure 16.  Monthly Northern Virginia Metrobus ridership in Figure 17 calculated from Monthly average weekday/Saturday/Sunday ridership by route figures provided by WMATA.  Annual Northern Virginia Metrobus ridership in Figure 18 is the sum of the weekday/Saturday/Sunday ridership for the individual routes listed in Figure 16.

HOV FACILITIES AND USE

   Disclaimer:  The mode share and HOV performance data presented in this report are based on three research studies conducted by the National Capital Region Transportation Planning Board of the Metropolitan Washington Council of Governments (COG/TPB); namely, the 1999 Metro Core Cordon Count, the 2001 Beltway Cordon Count, and the 1999 Performance of Regional High-Occupancy Vehicle Facilities on Freeways in the Washington Region. Each of the studies is in draft form and has not been presented to the TPB but has been approved by COG staff for use in planning analyses.

 Mode Share

      Mode shares for the four major HOV corridors in Northern Virginia were calculated from 2001 Beltway Cordon Count and 1999 Metro Core Cordon Count data compiled by the COG/TPB. 

COG/TPB Beltway and Metro Core Cordon Data Collection Methodology

     Data collection occurred for five hours in both the morning and afternoon, the data represented here include counts only for the peak periods of maximum demand on the region’s transportation facilities (6:30 to 9:30 A.M. and 3:30 to 6:30 P.M.). It should be noted that this also does not correspond exactly to the HOV-restricted periods for these corridors. For the Beltway Cordon Count, the inner loop of the Beltway defined the cordon, and counts were taken at points just inside this boundary. Inbound and outbound traffic was counted at each of the 61 counting stations located along the cordon line on a Tuesday, Wednesday, or Thursday in the spring of 2001. The Metro Core Cordon line encompasses the central employment area of Washington, D.C. and Arlington, Virginia. Inbound and outbound traffic was counted once at each site on a Tuesday, Wednesday or Thursday in the spring of 1999. For both cordon counts, all vehicles crossing the cordon were classified by vehicle type, and in the case of automobiles, further classified by number of occupants (one to seven persons). Pickup trucks, panel trucks, sport utility vehicles (SUVs), and vans (with the exception of 15-passenger vanpool vehicles) were counted as automobiles if they had two axles and exactly four wheels.  

The COG/TPB defines a vanpool as a stretch-van with privacy windows and an overhang of at least three feet from the rear wheel to the rear bumper or a van without privacy windows having eight or more occupants. Vanpools crossing the cordon lines were counted and a load factor of 12 persons per vanpool was applied. This load factor was derived from a mail-back survey that the COG/TPB conducted in 1989.   

The COG/TPB included all Metrorail, Metrobus, and other public transit services crossing the cordon lines in the mode share analysis. The ridership data provided by the transit operators were assigned to the appropriate counting station. Metrorail and VRE passenger volumes were assigned to the traffic counting station closest to the point at which the lines cross the cordon line. Commuter bus data, including routes, headways, and average passenger loads, were obtained through telephone interviews with bus company operators. From these data load factors were developed for the I-95/I-395 and I-66 corridors. The number of “in-service” commuter buses that cross the cordon is then multiplied by the load factor for that corridor to come up with the total commuter bus ridership. Commuter bus ridership across the cordons was distributed by counting station and time period. 

NVTC Mode Share Analysis

    Figures 51 and 52 summarize the counting stations and transit lines used in NVTC’s mode share analysis for each of the four corridors with HOV facilities; namely I-95/I395, Route 1, I-66, and VA 267. The Blue and Yellow Metrorail lines were assigned to the Route 1 Corridor, as the Metrorail lines are geographically closer to this corridor. This had the effect of increasing the transit mode share on Route 1 and lowering the transit share on the I-95/I-395 corridor. One could also assume that persons taking these Metrorail lines might otherwise drive on I-95/I-395 and thus show a higher transit mode share on the I-95/I-395 corridor and lower transit mode share on Route 1.  NVTC’s mode share analysis for the I-66 corridor includes travel on I-66 and the Metrorail orange line. It does not include parallel travel on US Routes 50 or 29. 

Data Limitations of Mode Share Analysis

    The purpose of the COG/TPB Cordon Counts is to measure the radial flow of persons and vehicles to and from the denser parts of the region, at the Beltway, and at the core of the region. NVTC has used this data to estimate mode shares for corridors with HOV facilities. This traffic count is only a “snap shot” of current peak period travel patterns crossing the Capital Beltway. The data for the counts were collected only for a single weekday A.M. and P.M. peak period. Although COG/TPB staff attempted to avoid unusual travel situations as much as possible (inclement weather, traffic accidents, legal holidays, school vacations, periods of increased tourist activity, etc.) peak period travel can vary significantly from day to day, and thus individual site data should be interpreted cautiously. Furthermore, caution is advised in extrapolating the mode share figures to the entire length of a corridor as the analysis is based data gathered at a set of points along the Beltway and Metro Core cordon lines only.

Figure  51:    2001 Beltway Cordon Count Corridors with Associated Counting Stations and Transit

Routes Used in NVTC’s Mode Share Analysis

Figure 52:  1999 Metro Core Cordon Count Corridors

with Associated Counting Stations and Transit

Routes Used in NVTC’s Mode Share Analysis

 Performance of Regional HOV Facilities

      The COG/TPB Report entitled 1999 Performance of Regional High-Occupancy Vehicle Facilities on Freeways in the Washington Region is the source for NVTC’s data on HOV travel time, person carrying capacity, and compliance. The underlying data were collected on a Tuesday, Wednesday, or Thursday in the fall of 1999, from 5 A.M. until 10 A.M. in the inbound peak-flow direction and from 3:00 P.M. until 8:00 P.M. in the outbound peak-flow direction along HOV corridors, including I-95/395, I-66, and VA 267. 

Travel Time

   Travel time data were collected using the “floating car” methodology—the car conducting the travel time floats with traffic, passing the same number of vehicles that it is passed by during the trip. An HOV and a non-HOV run were always paired within a corridor, departing the starting location at the same time, so that the two modes could be compared. Morning runs were scheduled to start at the outer limits of each HOV corridor, between 6:45 A.M. and 7:50 A.M. in the in bound direction. Afternoon runs started from the inner limits of each corridor, between 4:45 P.M. and 5:20 P.M. To compare the travel times of HOV and non-HOV routes, a set of travel time runs was conducted in each corridor during peak commute periods. In each corridor, five travel time runs were conducted on different days during both the morning and evening peak commute periods to obtain a mean travel time for each corridor. Personnel making the travel time runs had stopwatches and a list of designated points along the route, so that they could record the time at which they passed each point on the list. 

Occupancy Counts

     Occupancy counts were collected at a series of locations along each HOV corridor. Temporary personnel were hired and trained to classify vehicles according to the standard COG/TPB counting methodology. Generally one person was assigned per lane. Using a laptop computer running a simple BASIC denominator program each vehicle and the number of passengers observed was tallied as it crossed the counting station. The numerator of the calculation is the number of auto passengers, including vanpoolers, observed passing a count station. The denominator is the number of autos, auto-like vehicles and vanpools counted. Trucks and transit passengers are not included in the calculation. 

Vehicle occupancy counts are performed for automobiles, station wagons, vans, pickup trucks, sport utility vehicles and panel trucks, and any other motor vehicle with exactly two axles and exactly four wheels. The number of persons observed in each passing vehicle is counted and recorded, for vehicle occupancies from one to seven. Vehicles with seven or more occupants are classified as having seven occupants.

Stretch vans with privacy windows and an overhang of at least three feet from the rear wheel to the rear bumper or vans without privacy windows with eight or more occupants are counted as vanpools and multiplied by 12 (the average number of passengers per vanpool according to the COG/TPB 1989 mailback survey of vanpool operators). Occupancies for vans with less than eight passengers are tallied for the number of persons observed in a manner identical with automobiles and other vehicles with four wheels (see above). 

Vehicles with exactly two axles and six tires, and all motor vehicles with three axles or more that are not classified as buses are counted as trucks. Occupancy counts are not performed for trucks. 

Motorcycles are motorized vehicles with two or three wheels. All motorcycles are presumed to carry one person, even if more than one individual is observed riding astride the cycle or in a sidecar. 

Pickup trucks, panel trucks, SUVs, and vans with exactly four wheels are considered autos. 

Person Carrying Capacity

    Person carrying capacity was calculated from occupancy counts and includes all vehicles (autos, vanpools, motorcycles, buses, and trucks). All buses are counted in the vehicle classification procedure. Buses are classified as either ‘transit bus” or “other bus.” Transit buses include Metrobus, Fairfax Connector and Alexandria DASH. Other buses include all commuter buses, such as those operated by PRTC, school buses, and airport, inter-city and charter coaches. Transit and commuter bus patronage data were obtained from providers of bus and rail service in each corridor. 

Data limitations of HOV Performance Analysis

    The data used in the occupancy counts and person carrying capacity calculations are based on one-day counts. Although COG/TPB staff attempted to avoid unusual travel situations as much as possible (inclement weather, traffic accidents, legal holidays, school vacations, periods of increased tourist activity, etc.) peak period travel can vary significantly from day to day. Furthermore, caution is advised in extrapolating the mode share figures to the entire length of a corridor as the analysis is based data gathered at a set of points along the Beltway cordon line only. Travel on parallel roadways may be reduced due the effect of HOV restrictions that may tend to “draw” vehicle and person trips meeting the HOV requirement off of parallel roads.