FHWA → Planning → TMIP → Publications → Other Reports → Commercial Vehicles Transportation

Accounting for Commercial Vehicles in Urban Transportation Models

Magnitude and Distribution

3.0 Data Sources

This section describes the data sources evaluated to provide information on the spatial and temporal distribution of commercial vehicles in urban areas. There are five general types of data reviewed for this study: commercial vehicle surveys, vehicle registration data, vehicle count data, category-specific data sources, and data from individual contacts.

3.1 Commercial Vehicle Surveys

Vehicle Inventory and Use Survey

The 1997 Vehicle Inventory and Use Survey (VIUS) is a probability sample of private and commercial trucks registered (or licensed) in the United States as of July 1, 1997. This survey excludes vehicles owned by Federal, state, or local governments; ambulances; buses; motor homes; farm tractors; unpowered trailer units; and trucks reported to have been sold, junked, or wrecked by the respondents prior to July 1, 1996. A sample of about 131,000 trucks was surveyed to measure the characteristics of nearly 75 million trucks registered in the United States.

Many states allow pickups, small vans, and sport utility vehicles to be registered as either cars or commercial vehicles. Therefore, during the development of the VIUS sampling frame, passenger car registration files were searched and appropriate vehicles were included. Some vehicles, such as 'off-highway' trucks used exclusively on private property, do not have to be registered. These vehicles were not included in the sampling frame.

The following information is available from VIUS for each vehicle:

• Number of miles driven during 1997;
• Number of miles driven since the vehicle was manufactured;
• Weighted annual miles;
• How the vehicle was most often operated (business use, personal transportation, for-hire, daily rental, or mixed);
• If usage is mixed, the percentages of mixed use, business use, and personal use; and
• The principal product carried by the vehicle.

Table 3.1 shows the number of vehicles in the VIUS database by body type and vehicle size. Table 3.2 gives the number of vehicles located within metropolitan statistical areas (MSA) in the eight states considered for this study.

Table 3.1: Number of Vehicles by Body Type and Vehicle Size

Body Type	Number of Vehicles by Vehicle Size: Total	Number of Vehicles by Vehicle Size: Light	Number of Vehicles by Vehicle Size: Medium	Number of Vehicles by Vehicle Size: Light-Heavy	Number of Vehicles by Vehicle Size: Heavy-Heavy
Pickup	36,191,818	36,009,449	182,369	0	0
Panel or Van	5,572,678	5,547,280	25,396	2	0
Multi-stop or Step Van	560,420	313,216	222,705	19,372	5,128
Platform with Added Devices	308,176	58,156	84,959	67,953	97,109
Low Boy or Depressed Center	111,054	4,401	6,329	8,892	91,432
Basic Platform	1,176,066	409,246	290,540	154,914	321,365
Livestock Truck	39,069	3,661	11,153	5,725	18,530
Insulated Non-refrigerated Van	34,520	2,079	2,199	3,698	26,544
Insulated Refrigerated Van	233,977	8,613	23,807	19,070	182,487
Drop-frame Van	54,858	3,834	8,586	8,996	33,442
Open-top Van	20,781	1,527	1,690	1,580	15,984
Basic Enclosed Van	1,008,959	98,205	134,562	135,173	641,019
Beverage	70,233	2,403	8,017	15,284	44,529
Public Utility	151,950	44,441	43,599	31,434	32,475
Winch or Crane	55,017	6,157	12,167	9,209	27,485
Wrecker	111,899	38,925	56,898	9,005	7,071
Pole or Logging	55,705	1,312	2,625	5,713	46,055
Auto Transport	20,103	2,182	4,779	924	12,218
Service Truck	168,620	97,658	51,926	12,062	6,973
Yard Tractor	10,798	478	2,384	505	7,431
Sport Utility	13,762,470	13,739,880	22,591	0	0
Station Wagon	1,770,676	1,765,985	4,691	0	0
Minivan	9,837,926	9,828,651	9,275	0	0
Oilfield Truck	26,106	3,453	2,787	3,035	16,831
Grain Body	299,078	13,197	46,231	59,631	180,019
Garbage Hauler	91,633	2,129	8,506	6,921	74,078
Dump Truck	670,821	83,654	129,067	95,876	362,224
Tank Truck (Liquids or Gases)	249,382	6,273	29,320	45,538	168,250
Tank Truck (Dry Bulk)	39,724	649	2,190	4,003	32,882
Concrete Mixer	73,092	201	362	1,963	70,566
Other	22,642	2,616	3,819	2,787	13,421
TOTAL	72,800,252	68,099,912	1,435,528	729,263	2,535,549

Source: Vehicle Inventory and Use Survey (1997).

Table 3.2: Number of Vehicles within MSAs in Selected States

State	Number of Vehicles
California	8,087,382
Colorado	1,032,943
Florida	2,870,581
Georgia	1,333,548
Michigan	1,980,215
North Carolina	1,124,455
Oregon	816,205
Texas	3,206,313

Source: Vehicle Inventory and Use Survey (1997).

The VIUS data set was modified for use in this project so that average daily vehicle miles traveled (VMT) and average daily VMT per vehicle could be estimated. Vehicles whose home bases were outside MSAs or had more than 50 percent of their miles driven more than 50 miles away from their home bases were excluded. After trimming the dataset, it was decided to exclude observations that listed the following as their major use:

• Daily rental;
• Not in use;
• For hire transportation; and
• One-way rental.

The daily rental categories were excluded because they have been captured separately elsewhere. "For hire transportation" and "not in use" were not included because of the difficulty in categorizing them.

Table 3.3 shows the daily VMT for six categories available in VIUS data. While VIUS data can be reported either for an entire state or for all MSAs in a state, data cannot be reported separately for a specific city or urban area. As a result, VIUS data for the 12 urban areas used in this project (see Table 1.1) cannot be reported separately. However, for this study VMT per vehicle data have been calculated using VIUS data and used with other data for estimating the total VMT by category. Table 3.4 shows daily VMT per vehicle.

Table 3.3: Daily Vehicle Miles Traveled in MSAs by Commercial Vehicle Category

State	Business and Personal Services	Construction Transport	Public Safety	Public Utilities	Trades and Services	Urban Freight Distribution
California	191,184,016	5,770,580	331,941	905,218	574,938	2,905,847
Colorado	24,330,116	702,068	49,231	138,096	157,911	397,717
Florida	75,437,337	2,549,107	89,078	640,260	383,978	1,315,266
Georgia	34,710,330	1,196,564	174,992	265,348	99,270	550,991
Michigan	48,700,595	1,608,217	56,396	278,389	44,522	739,032
North Carolina	25,927,200	1,703,922	103,516	191,591	181,445	754,688
Oregon	18,975,350	428,885	27,210	34,656	38,807	364,341
Texas	91,799,636	2,279,053	119,408	296,610	345,972	1,275,083
National	1,174,389,225	41,163,792	2,311,391	7,377,819	4,534,727	19,583,562

Source: Vehicle Inventory and Use Survey (1997).

Table 3.4: Average Daily VMT per Vehicle in MSAs by CS Commercial Vehicle Category

State	Business and Personal Services	Construction Transport	Public Safety	Public Utilities	Trades and Services	Urban Freight Distribution
California	41.3	45.7	52.6	60.0	34.3	74.5
Colorado	38.6	57.2	44.3	56.4	35.3	47.2
Florida	45.9	62.2	49.4	68.6	57.6	66.6
Georgia	43.5	49.2	70.2	64.3	62.6	61.7
Michigan	44.0	44.1	56.3	58.4	42.8	57.3
North Carolina	41.1	51.1	45.4	50.7	83.8	52.6
Oregon	38.3	38.9	68.1	38.9	54.8	63.3
Texas	47.4	62.2	47.0	68.7	84.3	60.3
National	40.7	46.0	46.7	58.4	51.8	53.4

Source: Vehicle Inventory and Use Survey (1997).

VIUS reports annual VMT. The daily VMT was calculated based on the number of days in a year that vehicles in a specific category operate. The number of days used for estimating daily VMT was developed by Cambridge Systematics based on average number of days per year that each category was open for business. These estimates are shown in Table 3.5.

Table 3.5: Number of Days in a Year Used for VMT Calculations

Commercial Vehicle Categories	Number of Days in Year
Urban Freight Distribution, Warehouse Deliveries	306
Construction Transport	260
Public Safety	365
Public Utilities	260
Trades and Services	260
Business and Personal Services	306

Source: Cambridge Systematics, Inc.

Atlanta Area Commercial Vehicle Survey

The Atlanta Area Commercial Vehicle Survey was conducted by NuStats International for the Atlanta Regional Commission (ARC) in the spring of 1996.¹ The primary objective of the survey was to provide insight into truck movements in the Atlanta region. Specifically, the goals of the study were to determine the number of trips per truck and the average truck trip length, and to develop a truck trip table that would provide critical information for the regional travel demand model.

The Atlanta Area Commercial Vehicle survey was conducted in two phases. First a "recruitment interview" was performed to identify suitable businesses that were willing to participate in the survey. Firms were randomly selected from a 1993 commercial vehicle listing from the Georgia Department of Environmental Regulation. Participating businesses were assigned a 24-hour period (the travel day). All trips made using the selected vehicle(s) were recorded for the travel day. If the business maintained detailed vehicle manifest information, the travel data could generally be obtained from the manifest.

The survey sample was expanded based on the fleet size of the survey firm. Table 3.6 lists the vehicle groupings and the expansion factor for each group.

Table 3.6: Atlanta Expansion Factors

Fleet Size	Universe	Sample Size	Factors
1	10,808	35	308.8
2	15,560	44	353.6
3-5	19,580	108	181.6
6-10	14,060	208	67.6
11-20	10,950	153	71.6
21-50	12,280	164	74.9
51+	32,840	31	1,059.4
TOTAL	116,078	743	N/A

Source: Atlanta Area Commercial Vehicle Survey.

Commercial Vehicle Category Groupings

Table 3.7 illustrates how the various commercial vehicle types are defined in the survey. The top portion of the table lists the descending order of precedence in which the vehicle types are defined. For example, if a vehicle meets the criteria to be defined as both Package Delivery and Business and Personal Services, the vehicle is classified as Package Delivery.

Table 3.7: Atlanta Commercial Vehicle Categories

Code	Vehicle Type	Selection
1	School Bus	Land Use = Educational and Purpose = Delivery or pick-up
6	Package Delivery	Cargo = Printed matter and Purpose = Delivery or pick-up
8	Construct Transport	Cargo = Lumber or wood or transport equipment
10	Utilities	Cargo = Electrical and Land Use = Office/government or Purpose = Maintenance
12	Out Sales	Purpose = Driver need or Land Use = Office/government
7	Urban Freight	Purpose = Any and Land Use = Any and Cargo = Farm products; or Crude petro/natural gas; or Food; or Apparel; or Furniture fixtures; or Chemicals; or Rubber/plastic; or Leather; or Clay, concrete, glass or stone; or Primary metal products; or Fabricated metal; or Machinery; or Electrical; or Transport equipment; or Instrument: cameras/optical, watches; or Miscellaneous manufacturing products; or Waste, scrap; or Miscellaneous freight; or Containers; or Miscellaneous; or Empty

Summary of Survey Results

The results from the survey are summarized in Table 3.8. These data may differ from those presented in the survey report due differences in the vehicle type groupings. The "urban freight" category is the largest category in this survey (62 percent of the total), and "business and personal services" contains a large percentage as well (23 percent of the total). The longest average trip length is for urban freight vehicles, and the shortest average trip length is for school buses.

Table 3.8: Atlanta Expanded Survey Data

Vehicle Type	Vehicles	Average Daily Trips	Total Daily Vehicle Miles Traveled	Trips per Vehicle	Average Daily Miles Traveled
School Bus	2,212	2,414	40,177	1.09	18.17
Package/Product/Mail	4,681	12,644	155,215	2.70	33.16
Urban Freight	66,239	280,589	4,901,560	4.24	74.00
Construction Transport	8,267	31,596	481,804	3.82	58.28
Utility Vehicles	1,420	3,835	58,043	2.70	40.88
Business and Personal Services	24,463	44,721	660,730	1.83	27.01
TOTAL	107,282	373,385	6,297,528	3.48	58.70

Source: Atlanta Area Commercial Vehicle Survey.

Denver Commercial Vehicle Survey

The Denver Regional Council of Governments (DRCOG), in partnership with the Regional Transportation District, the Colorado Department of Transportation, and the Regional Air Quality Council, initiated the Regional Travel Behavior Inventory (TBI) in 1996. The TBI was undertaken to provide a snapshot of travel patterns and characteristics of travelers in the Denver region and to collect the data needed to develop and "freshen" traditional travel models, while providing for the possible development of new modeling techniques. The Denver Commercial Vehicle Survey was one of four surveys conducted as part of TBI.

The Denver survey was designed as a two-stage survey - a business and vehicle survey and a vehicle travel survey. Numerous businesses were surveyed to verify or correct business characteristics listed for the business and to determine the number and types of commercial vehicles garaged at these businesses. The list of businesses included all businesses listed within the Denver area in 1996 (90,558 entries) and was obtained from DRCOG. This first stage of the survey was completed prior to the selection of any vehicles for the second stage.

Commercial vehicles selected for the second stage survey were selected from vehicles listed in the first stage survey based on designated sampling procedures. The sampling procedure allowed a single business to have multiple vehicles included in the travel survey. A travel diary was collected for the selected vehicles.

It is important to note that the Denver commercial vehicle survey specifically excluded auto and truck rental businesses and police, fire, taxi, and U.S. Postal Service operations. These exclusions demonstrate the typical practice of not including certain types of commercial vehicles in regional surveys conducted by transportation planning agencies. Table 3.9 shows the excluded business and vehicle types.

Table 3.9: Denver Survey Excluded Businesses and Vehicle Types

Excluded Businesses	Excluded Vehicle Types
Auto and Truck Rental	Rental Cars
Police and Fire Departments	Safety Vehicles
Taxi	Private Transportation
U.S. Postal Service	Package, Product and Mail Delivery

Commercial Vehicle Category Groupings

Table 3.10 illustrates how the various commercial vehicle type are defined in the survey. The top portion of the table lists the descending order of precedence in which the vehicle types are defined. For example, it a vehicle meets the criteria to be defined as both Package Delivery and Business and Personal Services, the vehicle is classified as Package Delivery.

Table 3.10: Denver Commercial Vehicle Categories

Code	Vehicle Type	Selection
6	Package Delivery	Cargo = Mail and Purpose = Pick-up/deliver a load
2	Shuttle Services	Purpose = Drop-off/pick-up people
10	Utilities	Purpose = Fuel/service vehicle
8	Construct Transport	Land Use = Construction or trans/comm
12	Out Sales	Purpose = Service call or business meeting or Land Use = Services
7	Urban Freight	Cargo = Any and Land Use = Any and Purpose = Pick-up/deliver a load; or Return to base; or Other; or Return home/end day

Summary of Survey Results

The results from the survey are summarized in Table 3.11. These data may differ from those presented in the survey report due differences in the vehicle type groupings. The urban freight category is the largest category in this survey (45 percent of the total), and business and personal services comprise a large percentage as well (30 percent of the total). The longest average trip length is for urban freight vehicles, and the shortest average trip length is for package, product, and mail delivery.

Table 3.11: Denver Expanded Survey Data

Vehicle Type	Vehicles	Average Daily Trips	Total Daily Vehicle Miles Traveled	Trips per Vehicle	Average Daily Miles Traveled
Shuttle Service	2,204	8,098	47,819	3.67	21.70
Package/Product/Mail	5,907	12,095	57,014	2.05	9.65
Urban Freight	29,614	103,944	1,915,760	3.51	64.69
Construction/Transport	8,411	18,521	257,192	2.20	30.58
Utility Vehicles	4,935	5,038	52,881	1.02	10.72
Business and Personal Services	12,485	25,310	969,020	2.03	77.62
TOTAL	63,556	173,005	3,299,686	2.72	51.92

Source: Denver Commercial Vehicle Survey.

Detroit Commercial Vehicle Survey

The Southeast Michigan Council of Governments (SEMCOG) Commercial Vehicle Survey (CVS) collected detailed information on truck travel within the seven-county area of Southeast Michigan, for use in SEMCOG's Regional Travel Forecast Model. The information also will assist with other intermodal and freight planning activities. The universe for the commercial vehicles is from a data file from the Michigan Secretary of State containing the universe of commercial vehicles registered within the region. A supplemental business survey was conducted to determine the proportion of businesses located within the region that have commercial vehicles, registered at locations outside the region, but which operate within the region for business purposes on a regular basis. These trucks would not have been included in the main survey because the CVS sampling frame was limited to vehicles registered to locations within the region.

Businesses with vehicles operating in the Detroit region were contacted randomly for participation in the activity log portion of the survey. For participating businesses, a travel day was assigned and a trip diary was mailed.

Commercial Vehicle Category Groupings

Table 3.12 illustrates how the various commercial vehicle type are defined in the survey. The top portion of the table lists the descending order of precedence in which the vehicle types are defined. For example, it a vehicle meets the criteria to be defined as both Package Delivery and Business and Personal Services, the vehicle is classified as Package Delivery. The Detroit survey uses an Industry code on the destination end of the trip to define the vehicle type in addition to the cargo, purpose and land use.

Table 3.12: Detroit Commercial Vehicle Categories

Code	Vehicle Type	Selection
1	School Bus	Land Use = School
6	Package Delivery	Cargo = Mail/small parcels/packages
9	Safety Vehicles	Purpose = Plowing/snow removal or towing/road service
8	Construct Transport	Land Use = Construction site/job site or Industry = Construction and Cargo = Lumber, wood products, other building materials, or sand/gravel
10	Utilities	Industry = Utilities and Cargo = Tools, other materials
12	Business and Personal	Industry = Personal services; business services; or other professional services
7	Urban Freight	Cargo = Any and Land Use = Any and Purpose = Hauling heavy material; or Sales/service/maintenance work; or Landscaping; or Transports people/transportation

Summary of Survey Results

The results from the survey are summarized in Table 3.13. These data may differ from those presented in the survey report due differences in the vehicle type groupings. The urban freight category is the largest category in this survey (52 percent of the total); construction transport contains a large percent as well (22 percent of the total). The longest average trip length is for package/product and mail delivery, which is unique to Detroit since the other surveys have shorter than average trip lengths for this category. The shortest average trip length is for school buses, which is consistent with the other surveys.

Table 3.13: Detroit Expanded Survey Data

Vehicle Type	Vehicles	Average Daily Trips	Total Daily Vehicle Miles Traveled	Trips per Vehicle	Average Daily Miles Traveled
School Bus	6,467	10,345	87,189	1.60	13.48
Package/Product/Mail	5,322	38,211	456,477	7.18	85.77
Urban Freight	41,338	215,984	2,074,750	5.22	50.19
Construction Transport	5,501	22,118	279,301	4.02	50.78
Safety Vehicles	3,492	19,606	127,247	5.62	36.44
Utility Vehicles	1,380	3,301	32,094	2.39	23.26
Business and Personal Services	15,740	78,748	790,250	5.00	50.21
TOTAL	79,239	388,314	3,847,307	4.90	48.55

Source: Detroit Commercial Vehicle Survey.

Piedmont-Triad Commercial Vehicle Survey

The Piedmont-Triad Commercial Vehicle Survey was conducted to estimate truck trips and trips made by commercial cars in the Triad region (Greensboro, High Point, and Winston-Salem) of North Carolina.

A database of employers in the Triad region, including the number of employees and whether or not commercial vehicles are garaged at the employment location, was used as the universe of sampling commercial vehicles in the region. Eligible vehicles were those having a commercial license and being garaged at a non-residential location overnight. The definition of eligible vehicles eliminates company cars that are driven home by employees and effectively eliminates a large share of vehicles that may otherwise have been placed into the personal services commercial vehicle category. Also missing from the survey are non-commercially licensed vehicles that are used for commercial purposes.

Commercial Vehicle Category Groupings

Table 3.14 illustrates how the various commercial vehicle types are defined in the survey. The top portion of the table lists the descending order of precedence in which the vehicle types are defined. For example, if a vehicle meets the criteria to be defined as both Package Delivery and Utilities, the vehicle is classified as Package Delivery. The Piedmont-Triad survey uses a vehicle type field from the survey, in addition to the cargo, purpose and land use fields.

Table 3.14: Piedmont-Triad Commercial Vehicle Categories

Code	Vehicle Type	Selection
6	Package Delivery	Cargo = Mail or express traffic/small package freight/printer matter
3	Private transport	Vehicle Type = Passenger van and vehicle occupancy greater than 1
10	Utilities	Land Use = Utilities
7	Urban Freight	Cargo = Any and Land Use = Any and Purpose = Pick-up load; or Drop-off label; or Return to base; or Both pick-up load and drop-off label; or Both pick-up load and fuel/service unit; or Both pick-up load and other business; or Both drop-off label and fuel/service unit; or Both return to base and pick-up load; or Both return to base and drop-off label; or Return to base and pick-up load and drop-off label

Summary of Survey Results

The results from the survey are summarized in Table 3.15. These data may differ from those presented in the survey report due differences in the vehicle type groupings. The urban freight category is the majority category in this survey (82 percent of the total). The longest average trip length also is for urban freight, which is consistent with the other surveys. All other categories are well below the overall average trip length, which is dominated by the longer trips in the urban freight category.

Table 3.15: Triad Expanded Survey Data

Vehicle Type	Vehicles	Average Daily Trips	Total Daily Vehicle Miles Traveled	Trips per Vehicle	Average Daily Miles Traveled
Private Transportation	182	542	4,954	2.97	27.19
Package/Mail	920	3,554	25,236	3.86	27.43
Urban Freight	7,836	37,410	438,549	4.77	55.96
Construction/Transport	839	2,760	31,318	3.29	37.32
Utilities	220	394	3,181	1.79	14.47
TOTAL	9,998	44,660	503,239	4.47	50.34

Source: Piedmont-Triad Commercial Vehicle Survey.

3.2 Vehicle Registration

Eight persons were contacted who have either conducted or been responsible for research on vehicle emissions and are familiar with the experience of using state vehicle registration and/or inspection/maintenance (I/M) program data, to determine the degree to which these databases have been useful for identifying commercial vehicle activity patterns. The following people were contacted:

• Professor Randy Guensler, Georgia Institute of Technology;
• Professor Michael Rodgers, Georgia Institute of Technology;
• Professor Matthew Barth, University of California at Riverside;
• Dr. Herb Weinblatt, Cambridge Systematics (worked with West Virginia University on a NCHRP heavy-duty vehicle emissions research project);
• Ms. Coralie Cooper, Northeast States for Coordinated Air Use Management;
• Ms. Megan Beardsley, U.S. Environmental Protection Agency;
• Professor Arun Chatterjee, University of Tennessee; and
• Mr. Chris Saricks, Argonne National Laboratory.

Overall, most people conducting research into emissions modeling have been interested in activity data by vehicle weight class and fuel type, since these are the characteristics by which EPA regulates emissions and which therefore correspond most closely to emission levels. They typically have used national-level sources such as VIUS and data from R.L. Polk & Co. (described below), since they are not as concerned about area-specific fleet distributions or activity data. For example, EPA has used Polk databases to examine the number of vehicles registered and VMT per vehicle for different vehicle weight classes. VIUS also has been used as a source of VMT per vehicle for heavy-duty vehicles by various researchers.

State Motor Vehicle Departments

State environmental agencies often have experience working with state registration data for the purpose of developing vehicle age distributions for the MOBILE emissions model. Prof. Rodgers has examined vehicle databases in 12 to 14 different states and found that there are basically three different organizational approaches for collecting vehicle registration data.

First, the state Departments of Revenue may collect vehicle registration data for tax purposes, with a focus on related data (e.g., vehicle age, engine displacement, weight class). Second, the state Departments of Motor Vehicles may collect vehicle data for safety and/or registration purposes, including odometer readings, violations, and county of residence. Third, vehicle data may be collected at the county or municipality level, and consolidated at the state level by a state public service agency. County/municipality data records typically are not uniform.

At best, state registration databases contain only basic data related to the use of the vehicle (e.g., commercial versus non-commercial, or whether the vehicle is part of a public fleet). Other use information could be inferred by looking at the owner of the vehicle in conjunction with vehicle characteristics, but this level of analysis would require significant effort as well as access to confidential data. As a result, state registration databases were found to have little value for determining the numbers or usage of commercial vehicles by service use.

Many states maintain separate databases for permanent public tag vehicles. State and/or local agencies also are likely to maintain registration data for licensed services such as taxicabs, limos, and shuttle services available through state or local agencies. These databases provide total numbers of vehicles, although they may not provide miles traveled.

California Energy Commission

Vehicle registration databases that are maintained by a state, as evidenced by the experience in California, have the potential to yield useful information on the number of commercial vehicles existing within a particular geographic area. Experience has shown, though that it is time consuming, costly, and difficult to use these vehicle registration databases for reasons other than those for which they originally were developed. Consequently, the only example of a vehicle registration database that has been successfully used to produce information on commercial vehicle travel that was able to be identified was for California. Nonetheless, it is recommended that other states explore and develop the same kind of multi-year cooperative arrangement that exists in California so that, over time, vehicle registration data can be used to support transportation planning, including, but not limited to, the movement of commercial vehicles.

Processed California Department of Motor Vehicles (DMV) data was obtained from the California Energy Commission and extracted for four urban areas: San Francisco, Los Angeles, San Diego, and Sacramento. Summary data for these cities are shown in Table 3.16, Table 3.17, Table 3.18 and Table 3.19. The California DMV data has a large category of "other commercial" light duty vehicles that we have assigned to the business and personal services category. Since not all of the "other commercial vehicles" are being used for commercial purposes, we factored this category to exclude the business and personal services vehicles used for personal activities, based on the VIUS estimates of the use of these vehicles (24 percent of business and personal service vehicles are used for commercial purposes). Regarding school bus category, medium and heavy vehicles were divided into six groups based on their weights, and group "GVWR 6 Truck" was assumed to be the school bus category. This processing also included associating the average trip length for each commercial vehicle category from the VIUS data with the number of vehicles from the DMV data to calculate the VMT. These VIUS data were estimated for MSAs in California only but were not specific to an individual metropolitan area.

Table 3.16: DMV California Data Summary for the San Francisco MSA (Population: 4,022,000)

Commercial Vehicles Categories	Number of Commercial Vehicles	Average Daily Miles per Vehicle	Percentage of Total Vehicles	VMT	Percent of Total VMT
Business and Personal Services	152,263	41.3	3.01%	6,288,462	6.97%
Construction Transport	22,561	45.7	0.45%	1,031,038	1.14%
Other	55,520	59.95	1.10%	3,328,424	3.69%
Package, Product and Mail Delivery	470	76.1	0.01%	35,767	0.04%
Public Safety	5,090	52.57	0.10%	267,581	0.30%
Public Service	38,094	30	0.75%	1,142,820	1.27%
Public Utilities, Trades and Services	7,552	59.95	0.15%	452,742	0.50%
Rental Cars	89,805	43.11	1.78%	3,871,494	4.29%
School	1,510	36.2	0.03%	54,662	0.06%
Urban Freight Distribution, Warehouse Deliveries	22,484	74.5	0.44%	1,675,058	1.86%
Total Commercial Vehicles	395,349	44.8	7.82%	18,148,048	20.10%
Personal Vehicles	4,662,006	15.47	92.18%	72,121,952	79.90%
TOTAL	5,057,355	17.85	100.00%	90,270,000	100.00%

Source: California Department of Motor Vehicle registration data processed by the California Energy Commission for number of vehicles and the Vehicle Inventory and Use Survey for average daily miles traveled of trucks and the National Highway Travel Survey for average daily miles traveled of autos.

Table 3.17: DMV California Data Summary for the Los Angeles MSA (Population: 12,384,000)

Commercial Vehicles Categories	Number of Commercial Vehicles	Average of Daily Miles per Vehicle	Percentage of Total Vehicles	VMT	Percent of Total VMT
Business and Personal Services	321,445	41.3	3.01%	13,275,679	4.73%
Construction Transport	36,318	45.7	0.34%	1,659,733	0.59%
Other	142,950	59.95	1.34%	8,569,853	3.05%
Package, Product and Mail Delivery	449	76.1	0.00%	34,169	0.01%
Public Safety	11,149	52.57	0.10%	586,103	0.21%
Public Service	83,219	30	0.78%	2,496,570	0.89%
Public Utilities, Trades and Services	19,488	59.95	0.18%	1,168,306	0.42%
Rental Cars	88,217	43.11	0.83%	3,803,035	1.35%
School	5,259	36.2	0.05%	190,376	0.07%
Urban Freight Distribution, Warehouse Deliveries	69,617	74.5	0.65%	5,186,467	1.85%
Total Commercial Vehicles	778,111	44.8	7.28%	36,970,288	13.17%
Personal Vehicles	9,910,699	24.60	92.72%	243,821,712	86.83%
TOTAL	10,688,810	26.27	100.00%	280,792,000	100.00%

Source: California Department of Motor Vehicle registration data processed by the California Energy Commission for number of vehicles and the Vehicle Inventory and Use Survey for average daily miles traveled of trucks and the National Highway Travel Survey for average daily miles traveled of autos.

Table 3.18: DMV California Data Summary for the San Diego MSA (Population: 2,653,000)

Commercial Vehicles Categories	Number of Commercial Vehicles	Average of Daily Miles per Vehicle	Percentage of Total Vehicles	VMT	Percent of Total VMT
Business and Personal Services	50,488	41.3	2.55%	2,085,154	3.32%
Construction Transport	6,939	45.7	0.35%	317,112	0.50%
Other	33,059	59.95	1.67%	1,981,887	3.16%
Package, Product and Mail Delivery	41	76.1	0.00%	3,120	0.00%
Public Safety	3,364	52.57	0.17%	176,845	0.28%
Public Service	13,111	30	0.66%	393,330	0.63%
Public Utilities, Trades and Services	2,729	59.95	0.14%	163,604	0.26%
Rental Cars	12,107	43.11	0.61%	521,933	0.83%
School	1,267	36.2	0.06%	45,865	0.07%
Urban Freight Distribution, Warehouse Deliveries	8,510	74.5	0.43%	633,995	1.01%
Total Commercial Vehicles	131,615	44.8	6.65%	6,322,846	10.07%
Personal Vehicles	1,846,179	30.60	93.35%	56,486,154	89.93%
TOTAL	1,977,794	31.76	100.00%	62,809,000	100.00%

Source: California Department of Motor Vehicle registration data processed by the California Energy Commission for number of vehicles and the Vehicle Inventory and Use Survey for average daily miles traveled of trucks and the National Highway Travel Survey for average daily miles traveled of autos.

Table 3.19: DMV California Data Summary for the Sacramento MSA (Population: 1,394,000)

Commercial Vehicles Categories	Number of Commercial Vehicles	Average of Daily Miles per Vehicle	Percentage of Total Vehicles	VMT	Percent of Total VMT
Business and Personal Services	43,984	41.3	3.07%	1,816,539	6.11%
Construction Transport	8,798	45.7	0.61%	402,069	1.35%
Other	28,525	59.95	1.99%	1,710,074	5.75%
Package, Product and Mail Delivery	42	76.1	0.00%	3,196	0.01%
Public Safety	7,090	52.57	0.49%	372,721	1.25%
Public Service	36,710	30	2.56%	1,101,300	3.71%
Public Utilities, Trades and Services	5,108	59.95	0.36%	306,225	1.03%
Rental Cars	9,913	43.11	0.69%	427,349	1.44%
School	1,011	36.2	0.07%	36,598	0.12%
Urban Freight Distribution, Warehouse Deliveries	10,651	74.5	0.74%	793,500	2.67%
Total Commercial Vehicles	151,832	44.8	10.58%	6,969,571	23.45%
Personal Vehicles	1,282,838	17.74	89.42%	22,754,429	76.55%
TOTAL	1,434,670	20.718	100.00%	29,724,000	100.00%

Source: California Department of Motor Vehicle registration data processed by the California Energy Commission for number of vehicles and the Vehicle Inventory and Use Survey for average daily miles traveled of trucks and the National Highway Travel Survey for average daily miles traveled of autos.

To compare the commercial VMT with the total VMT, the total number of personal vehicles was obtained from the DMV. The average number of daily miles traveled for personal vehicles was calculated from the National Highway Travel Survey (NHTS)² for MSAs in California. These data were not available for specific cities, and so the calculation was based on MSAs between one and three million population (for Sacramento and San Diego) and MSAs over three million population (for San Francisco and Los Angeles). The total VMT calculation, therefore, was an estimate based not only on local data within each MSA.

The results of this analysis demonstrate that the commercial vehicle miles traveled are a higher percentage of the total than the number of vehicles, ranging from 10.3 to 15.3 percent of the total VMT compared to a range of 6.7 to 10.6 percent of the total vehicles, as shown in Table 3.16, Table 3.17, Table 3.18 and Table 3.19. This is an expected result based on the longer average miles traveled per day for commercial vehicles.

Inspection and Maintenance Programs

Many states collect data for their I/M programs that include the vehicle identification number (VIN) and odometer reading. A VIN decoder is a computer software program that is used to determine the make and model of the vehicle. Other emissions-related data also are collected, such as chassis, engine, emissions control system, fuel control system, etc. Odometer readings from at least two cycles of I/M inspection can be used to get vehicle activity (miles/year). I/M databases often identify whether the vehicle is commercial and include the gross vehicle weight rating (GVWR).

According to Professor Michael Rodgers and others, the following difficulties have been encountered with the use of I/M data for the purposes of classifying commercial vehicle travel:

1. The make and model of the vehicle do not necessarily indicate its type of use (service). For example, a vehicle may be identified as a "medium-duty GMC chassis" or a "Ford F-350" with a certain type of engine. What is on the back of the chassis, though, is not identified. The Ford F-350 could be used as an ambulance, delivery truck, contractor's vehicle, etc.
2. The I/M database may not be a random sample of vehicles registered in the state. For example, most states do not require public vehicles to be tested. (California is an exception). States often will encourage public fleets to test their vehicles, but the actual extent of participation may vary depending upon the jurisdiction, department, etc. Thus, an I/M database can be expected to underreport public safety, public utilities, or public transit vehicles. Also, the extent of heavy-duty vehicle testing varies. For example, New Jersey's heavy-duty testing program only tests vehicles over 18,000 pounds GVWR.
3. There can be problems with odometer matching as a result of mileage rollover. While many people have developed algorithms to deal with this, the algorithms (and data) are not perfect. This typically is a minor problem, but caution in using odometer data is required.
4. VIN decoder software may contain errors for several vehicle classifications. As documented by Prof. Guensler following the analysis of two different datasets, these limitations affect the accuracy of the predicted fleet distribution. The effect of these errors could be a bias towards newer vehicles and, therefore, an underestimation of mobile source emissions.

The Northeast States for Coordinated Air Use Management (NESCAUM) has worked with states in the northeast to implement heavy-duty vehicle emissions inspection programs. States that have done so include Massachusetts, New Jersey, and New York. However, these programs do not collect data related to service use of the vehicle, and not all vehicles are included (for example, the testing of only heavy trucks over 18,000 pounds in New Jersey). While the programs can identify the annual mileage per vehicle from this program, they do not have direct information on the use of the vehicle. A program specialist from New Jersey, however, suggested that service use could be inferred from the heavy-duty inspection program data by cross-tabulating U.S. DOT numbers with either company names or business type from the U.S. DOT census extract.

Massachusetts has a commercial vehicle inspection program that includes vehicles over 10,000 pounds or with a passenger capacity of at least 15. New York State's annual heavy-duty inspection program applies to most vehicles over 8,500 pound GVWR in the New York City metropolitan area. While information such as VIN, make, model year, and odometer reading are collected at the time of the test, the information collected is not useful in identifying the service use of the vehicle. Furthermore, it is primarily stored on paper, with limited (and non-centralized) downloading into electronic databases, and would therefore be nearly impossible to analyze.

In summary, the vehicles contained in state I/M databases reflect the characteristics of that state's underlying vehicle emissions inspection program. They rarely include information on the entire vehicle fleet, often covering only light vehicles and do not include information on how vehicles are being used. Consequently, it is recommended that the use of I/M databases should not be pursued further as a source of information on commercial vehicle travel.

R.L. Polk & Co.

R.L. Polk & Co.,³ a privately held consumer marketing information company, started motor vehicle statistics operations in 1922. Polk maintains comprehensive vehicle databases on both new and used vehicles in various formats, some of which are potentially useful for this study. Polk develops custom-built reports for customers and data are available by ZIP code, Metropolitan Statistical Area (MSA), county, state, or entire USA. However, these data are not free; they must be purchased from Polk.

Table 3.20 shows the data available from Polk. The ways in which the data could be used for our purposes are somewhat limited. Non-fleet vehicles owned by firms, which would presumably be commercial vehicles (although they could be used as well for non-commercial purposes) would be the sum of categories (a)4 through (a)6. Rental vehicles would be category (b)1, and other private commercial fleet vehicles would be the sum of categories (b)2 through (b)5. Government fleet vehicles would be category (b)6. However, it should be noted that information on vehicle type or use is not available from this source.

Table 3.20: Registration Type Data Available from Polk

(a) Retail
1. Personal
2. Participating Manufacturer Sponsored Lease - Personal
3. Participating Independent Lease - Personal
4. Number of Vehicles in the Fleet - Firm
5. Participating Manufacturers Sponsored Lease - Firm
6. Participating Independent Lease - Firm
7. Undetermined Manufacturer Sponsored Lease
8. Banks and Financial Institutions
(b) Fleet
1. Rental/Lease
2. Commercial
3. Participating Manufacturer Sponsored Lease - Fleet
4. Participating Independent Lease - Fleet
5. Independent Lease Fleet
6. Government
(c) Dealer/Manufacturer

Cambridge Systematics requested Polk to submit a cost estimate for all vehicle registration data, as shown in Table 3.20, for four states: Georgia, Colorado, Michigan, and North Carolina. It was requested that data be provided at the Census Block level. Polk submitted a cost estimate for these four states amounting to $24,500.

3.3 Vehicle Count Data

Highway Performance Monitoring System

The HPMS data as published in Highway Statistics were obtained for all metropolitan areas in the United States and summarized to identify the total VMT for all vehicles. Population and VMT data for 13 metropolitan areas are shown in Table 3.21. These data were intended to be used as an estimate of overall VMT so that commercial VMT could be assessed as a percent of the total and compared across different cities.

Table 3.21: Highway Statistics Estimates of Population and VMT for Selected Cities

City	Size	Population (in Thousands)	Daily VMT - All Vehicles (in Thousands)	VMT per Capita
Los Angeles	Large	12,384	280,792	22.7
San Francisco	Large	4,022	90,270	22.4
Detroit	Large	3,836	92,359	24.1
Atlanta	Mid	2,977	100,693	33.8
San Diego	Mid	2,653	62,809	23.7
Houston	Mid	2,487	91,883	36.9
Denver	Mid	1,993	43,996	22.1
Portland	Mid/Small	1,552	31,534	20.3
Sacramento	Mid/Small	1,394	29,724	21.3
Orlando	Mid/Small	1,160	32,288	27.8
Winston-Salem	Small	233	7,396	31.7
Greensboro	Small	223	7,654	34.3
High Point	Small	125	4,578	36.6

Source: Highway Statistics.

While it had been hoped that these data would provide a consistent dataset across all cities, it was discovered that there is some variation in these data across cities because the data are collected by individual states using different methods and assumptions. For example, the Air Resources Board (ARB) in California reports VMT in California cities that are quite a bit higher than for HPMS data in the same cities. These data are based on areas of different populations, but the ARB geographic areas are consistent with MPO and air quality planning areas whereas the HPMS areas are much smaller. The previously reported VMT data estimated from DMV records in California MSAs also are higher than the HPMS data, but they are closer to the ARB estimates. This is again most likely a difference in geographic area assumptions for each metropolitan area. Table 3.22 presents a comparison of these data and a calculation of the VMT per population from each data source for three California cities.

Table 3.22: Vehicle Miles Traveled from Different Data Sources

Data Source	Los Angeles	San Francisco	San Diego
HPMS Population (in Thousands)	12,384	4,022	2,653
HPMS VMT (in Thousands)	280,792	90,270	62,809
HPMS VMT per Population	22.7	22.4	23.7
ARB Population (in Thousands)	14,900	6,800	2,950
ARB VMT (in Thousands)	349,000	159,642	80,000
ARB VMT per Population	23.4	23.5	27.1
DMV VMT (in Thousands)	371,179	175,722	64,817
DMV VMT per Population (HPMS)	24.9	25.8	22.0
Percent Differences in VMT per Population (ARB versus HPMS)	3.3%	4.6%	14.5%
Percent Differences in VMT per Population (DMV versus HPMS)	10%	15%	-7%

Source: Highway Pavement Management System (HPMS), California Air Resources Board (ARB), and California Department of Motor Vehicles (DMV).

Freight Analysis Framework

The results of the Freight Analysis Framework (FAF) have been made available as a database file on the FHWA's FAF web site. The database file can be mapped to geographic information system (GIS) shape files of highways in the lower 48 states. The shape files allow the specification of highway links within specific urban areas. The database file includes mileage and functional classification information for each link in the FAF network. Because the links in the FAF database do not include all roadways, the FAF VMT does not represent the full universe of VMT although the FAF does include non-freight trucks.

We used this information to develop FAF freight truck and "non-freight truck" VMT and aggregated VMT by functionally classified roads within urban areas. Table 3.23 presents the FAF network data summarized for auto passenger cars, freight truck, and non-freight truck vehicles. The freight truck percentage of VMT varies from one to six percent by urban area, and the total truck percentage (including non-freight trucks) ranges from five to 18 percent by urban area.

Table 3.23: Vehicle Miles Traveled by Vehicle Type for Selected Urban Areas

Urban Area	Sum of Total VMT	Percent of Total: Auto	Percent of Total: Freight Truck	Percent of Total: Non-Freight Truck	Percent of Total: Total Truck
Atlanta, GA	47,868,419	90.7%	2.3%	7.0%	9.3%
Denver-Aurora, CO	17,443,820	94.1%	0.9%	5.1%	5.9%
Detroit, MI	48,426,905	94.3%	1.4%	4.3%	5.7%
Greensboro, NC	2,618,999	84.5%	3.4%	12.1%	15.5%
High Point, NC	864,998	82.5%	5.8%	11.6%	17.5%
Houston, TX	51,005,297	93.4%	2.2%	4.4%	6.6%
Los Angeles-Long Beach, CA	117,063,502	92.8%	1.9%	5.3%	7.2%
Orlando, FL	7,465,590	94.4%	2.0%	3.6%	5.6%
Portland, OR-WA	16,387,653	93.5%	1.7%	4.8%	6.5%
Sacramento, CA	10,254,347	92.1%	2.4%	5.5%	7.9%
San Diego, CA	30,112,972	94.8%	1.2%	4.0%	5.2%
San Francisco-Oakland, CA	29,655,627	94.5%	1.6%	3.9%	5.5%
Winston-Salem, NC	3,792,083	86.8%	3.2%	10.0%	13.2%
GRAND TOTAL	382,960,214	93.1%	1.8%	5.1%	6.9%

Source: Federal Highway Administration Freight Analysis Framework (FAF).

Table 3.24 presents the same FAF data stratified by functional classification. As expected, the freight truck and total truck percentages of VMT are higher for freeways than other facilities. The one anomaly in these data is the non-freight trucks on minor arterials, which has a very high percentage of VMT compared to expectations.

Table 3.24: Vehicle Miles Traveled by Functional Class for Selected Urban Areas

Functional Class	Sum of Total VMT	Percent of Total: Auto	Percent of Total: Freight Truck	Percent of Total: Non-Freight Truck	Percent of Total: Total Truck
Unknown Road	2,771,994	91.7%	2.4%	6.0%	8.3%
Rural Interstate	5,351,753	87.9%	5.0%	7.1%	12.1%
Rural Principal Arterial	3,773,437	91.1%	3.2%	5.7%	8.9%
Rural Minor Arterial	286,332	93.3%	1.5%	5.2%	6.7%
Rural Minor Collector	62,132	92.7%	0.3%	7.0%	7.3%
Urban Interstate	225,084,260	92.7%	2.1%	5.2%	7.3%
Urban Principal Arterial	97,720,514	93.7%	1.3%	5.0%	6.3%
Urban Principal Arterial	47,725,690	94.3%	1.1%	4.6%	5.7%
Urban Minor Arterial	184,100	87.5%	2.2%	10.3%	12.5%
GRAND TOTAL	382,960,214	93.1%	1.8%	5.1%	6.9%

Source: Federal Highway Administration Freight Analysis Framework (FAF).

Vehicle Classification Counts

Vehicle classification count data, which classifies the vehicles according to FHWA's 13 axle-based classes, are generally available from the state DOTs. Appendix E (Table E.1) contains a description of these FHWA vehicle classifications. Source information was obtained and examined for two states (Georgia and Florida) and summary information was examined on several state DOT web sites (Maine, Ohio, New Jersey, Massachusetts, Virginia, Pennsylvania, Delaware, and Indiana).

The source information includes counts by location for the 13 FHWA vehicle classes, by hour of the day and by date. This information is sufficient to develop hourly, daily, and seasonal distributions of traffic by vehicle type. In summary format this information generally presents truck volumes (defined as FHWA classes 5 through 13, six tires and above) and occasionally also includes buses (FHWA class 4). Four-tire pickup trucks, vans and SUVs (FHWA class 3), are almost always included with passenger cars.

Given that the format and derivation of these vehicle classification count data are inconsistent with our definition of commercial vehicles and their categories, we were unable to use these data for the evaluation of the magnitude and distribution of commercial vehicle travel. These data was given further consideration in the evaluation of methods to estimate commercial vehicle travel (and is documented in this task report).

3.4 Other Data Sources

National Transit Database for Paratransit Systems

Transportation systems that provide services mostly to disabled people are called paratransit systems. The Federal Transit Administration (FTA) collects and disseminates data on the state of mass transportation via the National Transit Database (NTD) program. Over 600 of the nation's transportation providers submit data to the NTD annually. Both the public and private sectors use these data to access the current state of mass transit and plan for the future. During the last two decades, large increases in the number of paratransit systems across the United States have been noticed. For example, across 198 cities with populations less than 400,000 in 1980, person trips by paratransit increased from six million in 1984 to 16.9 million in 1995.

However there are not many studies available which are based on both public and private paratransit data. The only comprehensive data source found is from the FTA NTD database and Steven Stern⁴ at the University of Virginia, who processed the Section 15 data for his research. Dr. Stern processed NTD data and reported paratransit buses, vehicles miles, and other data. Table 3.25 shows a sample of transit operating statistics for 11 urban areas. Complete data for about 300 cities are shown in Appendix A. However, it may be pointed out here that the FTA data include only those systems which reported their data to FTA. While all FTA-funded paratransit systems are required to submit their statistics, other paratransit systems, such as church service buses, which do not receive FTA funds, are not required to submit their data.

United States Postal Service Data

United States Postal Service vehicles and VMT data for seven urban areas (Atlanta, Denver, Detroit, Houston, Greensboro, Orlando, and Portland) were obtained from the United States Postal Service (USPS), as shown in Table 3.26. In these cities, postal service vehicles' total VMT as a percentage of the total VMT in the region varies from 0.05 to 0.63. The average daily VMT per vehicle is about 25 miles although it is much lower in urbanized areas (about five to six miles) and higher in suburban areas. In urbanized areas, daily postal delivery vehicles typically stop every block, after which the postal worker walks to deliver the mail.

Table 3.26: Public Package, Product and Mail Delivery Statistics

Service Area Location	Three-Digit Zip Code Range	Total VMT - All Vehicles (HPMS)	Total Number of Postal Vehicles	Daily Postal Vehicle VMT	Percent Postal Vehicle VMT
Atlanta, GA District	300-306, 311, 399	100,693,000	2,728	67,082	0.067%
Denver, CO District	800-807, 813-816, 820-831	43,999,000	3,380	57,937	0.132%
Detroit, MI District	481, 482	92,359,000	2,717	46,482	0.050%
Greensboro, NC District	270-279, 286	7,654,000	2,019	48,114	0.629%
Houston, TX District	770-778	91,883,000	4,169	78,575	0.086%
Orlando (Mid FL) District	327-329, 334, 347, 349	32,288,000	3,308	64,802	0.201%
Portland, OR District	970-979, 986	31,534,000	2,416	38,799	0.123%

Source: United States Postal Service for postal vehicles and the Highway Pavement Management System for total vehicle miles traveled.

School Bus Fleet Surveys

It has been estimated that school enrollment in the United States will increase 33 percent between 1990 and 2030. This means an additional four million children by 2005 and 15 million by 2030.⁵ The school transport industry provides 10 billion student rides annually - this is the largest form of public transportation in the United States. There are over 400,000 school buses operating each school day in the United States and school bus drivers log over four billion miles each school year. There are 50 million children in public and private schools, and yellow school buses transport half of this number every day.

Schoolbusfleet.com⁶ is an information service of the magazine School Bus Fleet, a trade publication serving school transportation professionals in the United States and Canada. School Bus Fleet provides information on the management and maintenance of school bus fleets operated by school districts, private schools, Head Start agencies and childcare centers.

In addition to management and maintenance articles, statistics on the largest 100 school district fleets also are published every year. Several districts' statistics are shown in Table 3.27. It should be noted that the school districts shown in the table do not, in general, represent all of the school districts located in the urban areas shown. The daily school bus VMT and the percentages of total VMT should therefore not be assumed to include all school bus VMT in the urban areas. The entire table of the largest fleets for the year 2000 is shown in Appendix B.

Taxi Fact Book

The National Association of Taxicab Operators was established in 1917 in Washington, D.C. In 1991 the Taxicab, Limousine & Paratransit Association (TLPA) was established with five membership divisions, including the Taxicab Division. TLPA publishes the magazine Transportation Leader quarterly and the Taxicab Division Fact Book annually. Table 3.28 presents taxi statistics by fleet size from the Taxicab Division Fact Book.

Table 3.28: Taxi Statistics

Items	Fleet Size: 1-24	Fleet Size: 25-99	Fleet Size: 100-Up	Fleet Size: Average
Average Annual Total Miles per Taxi	51,314	53,276	54,579	54,214
Average Distance per Paid Taxi Trip (miles)	5.38	5.82	6.57	5.80
Average Annual Paid Trips per Taxi	7,362	6,228	5,919	6,040
Average Annual Passengers per Taxi	10,048	8,229	7,703	7,913
Average Passengers per Paid Trip	1.36	1.33	1.3	1.31

Source: Taxicab Division Fact Book, 2002.

The complete Fact Book data are shown in Appendix C. However, the taxi statistics for selected 13 cities are presented in Table 3.29.

Table 3.29: Taxi Data by City

City	State	Population (Thousands)	Number of Licenses	Annual Taxicab VMT	Daily Taxicab VMT	Adjusted Daily Total All Vehicle VMTs (Thousands)	VMT Percent
Los Angeles	California	4,000	1,931	105,392,049	288,745	90,695	0.32
Sacramento	California	1,000	250	13,644,750	37,383	21,323	0.18
San Diego	California	1,200	910	49,666,890	136,074	28,410	0.48
San Francisco	California	775	1,381	75,373,599	206,503	17,395	1.19
Denver	Colorado	2,600	842	45,955,518	125,906	57,396	0.22
Orlando	Florida	1,200	1,000	54,579,000	149,532	33,401	0.45
Atlanta	Georgia	4,125	1,600	87,326,400	239,250	139,523	0.17
Detroit	Michigan	850	1,310	71,498,490	195,886	20,465	0.96
High Point	North Carolina	70	41	2,184,316	5,984	2,564	0.23
Winston-Salem	North Carolina	170	60	3,196,560	8,758	5,396	0.16
Portland	Oregon	1,500	400	21,831,600	59,813	38,872	0.15
Houston	Texas	1,800	2,245	122,529,855	335,698	66,502	0.50

Source: Taxicab Division Fact Book, 2002.

Airport Ground Access Planning Guide

The Airport Ground Access Planning Guide presents the results of the first phase of a project jointly sponsored by the Federal Highway Administration and the Federal Aviation Administration.⁷ It outlines the process for planning ground access to airports within the context of current laws, regulations, and procedures. This report identifies the key components of an airport access work program, discusses performance measures, and provides extensive information on alternative strategies for improving airport access conditions. The relevant portions of this report are described below.

The Airport Ground Access Planning Guide reports mode split, trip length, and trip cost data for trips to airports in 27 cities in the United States. Mode splits are presented in Table 3.30, summarized by urban area size. Data on each urban area is presented in Appendix D. Mode split and average trip length for other on-demand services, scheduled bus/van services and courtesy van services were combined to represent the fixed shuttle service commercial vehicle category for this study. These results show that as city size increases, the percent of travel using shuttle services also increases, from 11 percent in cities with less than 2,500,000 people, to 15 percent in cities with 2,500,000 to 5,000,000 people, to 21 percent in cities with more than 5,000,000 people.

Table 3.30: Summary of the Airport Access Mode Split

Statistic	Mode Split: Private Vehicle	Mode Split: Rental Car	Mode Split: Taxicab	Mode Split: Other On-Demand	Mode Split: Scheduled Bus/Van	Mode Split: Courtesy Vans	Mode Split: Other
Summary Statistics: Minimum	21.0%	2.0%	2.6%	0.0%	0.0%	0.0%	0.0%
Summary Statistics: Average	51.9%	20.7%	10.2%	8.0%	4.0%	3.2%	2.0%
Summary Statistics: Maximum	78.8%	46.2%	36.0%	24.0%	12.4%	8.0%	7.0%
Averages by City Size: <500 thousand	49.4%	29.2%	7.1%	7.2%	1.6%	2.0%	3.5%
Averages by City Size: 5-2.5 million	58.4%	22.1%	6.3%	3.7%	4.3%	3.2%	2.0%
Averages by City Size: 2.5-5 million	53.5%	17.8%	12.5%	7.7%	3.6%	3.6%	1.3%
Averages by City Size: >5 million	47.5%	18.4%	10.9%	11.6%	6.0%	3.5%	1.8%

Source: Derived from the Airport Ground Access Planning Guide for 27 cities provided in Tables 6.4-8, 10, 12, and 14.

Table 3.31 presents summary data on average trip length, fleet sizes, and vehicle miles traveled for shuttle services in the 27 cities in the Airport Access Planning Guide. The Guide presents average trip lengths for the taxi and bus modes only; the shuttle service average trip length was assumed to be five minutes or 2.5 miles longer than the average taxi trip length to account for pickup and drop-off travel time and distance. Airport shuttle services serve different kinds of trips than taxis, with some trips much shorter (for shuttles that serve airport hotels) and other trips much longer (for shuttles that serve other cities), so the average trip length for taxis is assumed to be in the range of the average for shuttle services. Average trip lengths reported as ranges were converted to the midpoints of the ranges for this analysis. Shuttle fleet sizes were estimated from the available data in the guide using the following assumptions:

• Average daily person trips by mode were estimated by applying the mode split percentages to the annual originations, factored by 365 to represent average daily originations. This factor of 365 days per year is based on the fact that most shuttle services operate seven days per week.
• Vehicle trips by mode were estimated by applying average vehicle occupancy factors to the person trips, estimated from data presented for Logan International Airport in Boston.⁸ These factors are three persons per vehicle for other on-demand shuttles and courtesy vans and 10 persons per vehicle for scheduled bus/van shuttle services.
• VMT was estimated as the product of the average fleet size (in vehicles) and the average trip length per vehicle per day.

Table 3.31: Summary of Shuttle Service Airport Access Vehicle Trips and Miles Traveled

Statistic	Average Trip Length: Minutes	Average Trip Length: Miles	Vehicle Trips	Vehicle Miles Traveled: Shuttle Service Vehicles	Vehicle Miles Traveled: All Vehicles	Percent of Total
Summary Statistics: Minimum	10.0	5.0	2	41	3,045,000	0.00%
Summary Statistics: Average	30.3	15.2	648	11,518	49,149,760	0.02%
Summary Statistics: Maximum	62.5	31.3	2,660	51,240	280,792,000	0.09%
Averages by City Size: <500 thousand	25.0	12.5	20	184	4,047,500	0.00%
Averages by City Size: 5-2.5 million	19.4	9.7	95	951	9,736,500	0.01%
Averages by City Size: 2.5-5 million	34.0	17.0	508	9,200	47,731,600	0.02%
Averages by City Size: >5 million	34.6	17.3	1,632	29,294	99,470,286	0.03%

Source: Derived from the Airport Ground Access Planning Guide from 27 cities provide in Tables 6.4-8, 10, 12, and 14.

The total VMT for shuttle services is presented in Table 3.31 for comparison across different urban areas. These data were derived from the HPMS, presented in Section 3.3. The percentage of total VMT attributed to shuttle services increases from close to zero in cities under 500,000 in population to 0.03 percent in cities with over five million people. The airport with the largest fleet size is San Francisco (2,660 vehicles), the highest VMT is Los Angeles (51.240 miles), and the highest percentage of total VMT is New Orleans (0.09 percent). The airport with the longest average trip length is Ontario (31.3 miles), within the Los Angeles metropolitan area.

The Airport Ground Access Planning Guide also presents data on taxis and rental cars. These data were analyzed and are presented for information only, since there are other data sources that provide a more comprehensive picture of taxis and rental cars. Table 3.32 and Table 3.33 present, respectively, summary data on the taxis and rental cars servicing airport trips. There are similar trends in shuttle services with respect to the percent of total VMT increasing with city size. New York's LaGuardia Airport has the largest number of vehicle trips and highest VMT for taxis, and New Orleans has the highest taxi percentage of VMT. Orlando has the largest number of vehicle trips and highest rental car percentage of VMT, and Los Angeles has the highest VMT for rental cars. Full data on taxis and rental cars for the 27 cities are presented in Appendix D.

Table 3.32: Summary of Taxi Airport Access Vehicle Trips and Miles Traveled

Statistic	Average Trip Length: Minutes	Average Trip Length: Miles	Vehicle Trips	Vehicle Miles Traveled: Taxis	Vehicle Miles Traveled: All Vehicles	Percent of Total
Summary Statistics: Minimum	5.0	2.5	20	133	3,045,000	0.00%
Summary Statistics: Average	25.3	12.7	1,415	20,850	49,149,760	0.04%
Summary Statistics: Maximum	57.5	28.8	9,480	142,200	280,792,000	0.27%
Averages by City Size: <500 thousand	20.0	10.0	53	706	4,047,500	0.02%
Averages by City Size: 5-2.5 million	14.4	7.2	235	1,499	9,736,500	0.02%
Averages by City Size: 2.5-5 million	29.0	14.5	1,484	22,538	47,731,600	0.05%
Averages by City Size: >5 million	29.6	14.8	2,954	43,644	99,470,286	0.04%

Source: Derived from the Airport Ground Access Planning Guide for 27 cities provided in Tables 6.4-8, 10, 12 and 14.

Table 3.33: Summary of Rental Car Airport Access Vehicle Trips and Miles Traveled

Statistic	Average Trip Length: Minutes	Average Trip Length: Miles	Vehicle Trips	Vehicle Miles Traveled: Rental Cars	Vehicle Miles Traveled: All Vehicles	Percent of Total
Summary Statistics: Minimum	5.0	2.5	4	32	3,045,000	0.00%
Summary Statistics: Average	25.3	12.7	1,478	20,060	49,149,760	0.05%
Summary Statistics: Maximum	57.5	28.8	6,308	106,624	280,792,000	0.24%
Averages by City Size: <500 thousand	20.0	10.0	168	1,470	4,047,500	0.04%
Averages by City Size: 5-2.5 million	14.4	7.2	552	3,479	9,736,500	0.04%
Averages by City Size: 2.5-5 million	29.0	14.5	1,329	18,151	47,731,600	0.04%
Averages by City Size: >5 million	29.6	14.8	3,178	45,815	99,470,286	0.05%

Source: Derived from the Airport Ground Access Planning Guide for 27 cities provided in Tables 6.4-8, 10, 12, and 14.

3.5 Individual Contacts

In addition to all of the data sources discussed, individual firms and agencies in both the public and private sectors and in all 12 urban areas were contacted. It was not expected to receive totals for all commercial vehicles operated by the firms contacted and commercial vehicle mileages in each city, but it was desired to capture a snapshot of the typical mileages that are driven by commercial vehicles of different industries in support of the other data sources. Although we contacted all 12 cities in some cases, only a few cities responded to our request for information. In cases where we needed to contact multiple firms in one category, we focused on collecting data in a single city. The following is a list of the individual contacts that were made:

• School departments (five cities);
• Public works departments (four cities);
• Police departments (two cities);
• Rental car companies (six companies, one city);
• Towing companies (three companies, one city); and
• United States Postal Service (one city).

Two pieces of information were asked for from each contact: the number of vehicles operated and the annual mileage that the vehicles accrued. The responses were received in many forms (e.g., average miles per vehicle for the fleet, total fleet mileage per year, mileages for the previous fiscal year) owing to the wide variety of sources contacted. Commercial vehicles were defined for the respondents as "any non-personal vehicle."

Issues

Some issues encountered during these contacts are worth briefly mentioning:

• An individual contact could result in either double-counting or under-representing the number of vehicles. Double-counting could occur, for example, in instances where the Public Works Department services part of the police department fleet and the police department also has its own vehicle fleet management. This could result in potentially double-counting some of the police vehicles. Under-representation could occur due to the lack of comprehensive inventories.
• Many agencies operate regionally (e.g., inner city plus suburbs) or are responsible for multiple cities; therefore it was difficult for them to estimate the number of miles which were traveled in each city. For example, school bus mileages were available for Winston-Salem/Forsyth County Schools combined, but not for just the Winston-Salem urban area. Therefore the mileage for the Winston-Salem urban area was simply estimated as a fraction of the mileage that the entire Winston-Salem/Forsyth County School buses operate as a total.
• Detailed data about the fleets was difficult to obtain; consistency among the data sources was therefore difficult to achieve. Very few agencies have a handy inventory of all vehicles they operate. Most agencies contacted knew the number of vehicles they owned, but mileage was more difficult to estimate, and mileage by vehicle type was even more difficult. Some agencies did not include heavier vehicles in their estimates.
• Data quality was highly variable. Some agencies could report mileage down to the 10th of a mile; other agencies could offer only an estimate of the number of vehicles in their fleet.

Results of Individual Contacts

School Departments

A total of 12 cities were contacted, but responses were received only from Detroit, Atlanta, Winston-Salem, Greensboro, and High Point. These are cities rather than the full urban area, based on the school districts, so the information is directly compared to urban area information. Information provided included the numbers of buses, food service vehicles, activity vehicles, maintenance/support vehicles, and special education buses. The information obtained is summarized in Table 3.34. School buses accrue most of their mileage during the school year.

Table 3.34: School Bus Statistics from Individual Contacts

City	Vehicles	Annual VMT/Vehicle	Source
Detroit	Buses = 430	15,000 Miles/Bus	Detroit Public Schools Garage
Detroit	Food Services = 60	13,000 Miles/Vehicle	Detroit Public Schools Garage
Atlanta	Buses = 388	12,630 Miles/Bus	Atlanta Public Schools Transportation
Winston-Salem	Buses = 123	14,090 Miles/Year	Winston-Salem/Forsyth County Schools Transportation
Winston-Salem	Maintenance and Support Vehicles = 32	8,000 Miles/Year	Winston-Salem/Forsyth County Schools Transportation
Greensboro	Buses = 203	11,516 Miles/Year	Guilford County Public Schools Department of Transportation
Greensboro	Countywide Activity Vehicles = 74	5,730 Miles/Year	Guilford County Public Schools Department of Transportation
Greensboro	Special ed Buses Countywide = 87	8,550 Miles/Year	Guilford County Public Schools Department of Transportation
High Point	Buses = 112	9,200 Miles/Year	Guildford County Public Schools Department of Transportation
High Point	Countywide Activity Vehicles = 74	5,730 Miles/Year	Guildford County Public Schools Department of Transportation
High Point	Special Ed Buses Countywide = 87	8,550 Miles/Year	Guildford County Public Schools Department of Transportation

Departments of Public Works

The Department of Public Works in each city is usually responsible for the fleets of city government vehicles. These vehicles perform functions such as solid waste collection and disposal, parks and recreation maintenance, public library support, street maintenance, traffic and parking enforcement, inspections, health department functions, and utility work. The mix of functions differs among the cities who responded to the contacts. For example, in some cities, the public works departments are responsible for maintaining the police department fleet while in other cities, the police departments maintain their own garages and fleets. Data were received from Detroit, Denver, Winston-Salem, and Greensboro and are summarized in Table 3.35. Again, these data represent cities rather than entire urban areas and are not directly comparable to data from urban areas.

Table 3.35: Public Works Department Statistics from Individual Contacts

City	Vehicles	Annual VMT/Vehicle	Source
Detroit	3,500 - 4,500 (Not Including Water and DOT)	N/A	Detroit Department of Public Works Fleet Management
Denver	3,354 (Includes Police Vehicles)	15,300 Miles	Denver Public Works Fleet Maintenance Division
Winston-Salem	1,100	8,200 Miles	City of Winston-Salem Fleet Services
Greensboro	1,500	N/A	Greensboro DOT Equipment Services

Police Departments

Most of the cities contacted did not want to share data because of security issues. Data were obtained only from Denver and Winston-Salem and represent cities rather than urban areas. The Denver Police Department maintains a total of 942 police vehicles, and the average VMT per vehicle per year is 11,300 miles. The Winston-Salem Police Transportation Department maintains 556 vehicles, and the average VMT per vehicle per year is 8,100 miles.

Rental Car Companies

Individual branches were contacted for several of the larger rental car companies. It is difficult to estimate the share of the market in each city represented by these rental car companies. The data on their fleets were available either at the main office of each city or from corporate headquarters. One company pointed out that vehicle rentals for the first part of the week can be almost double the number of vehicle rentals on the weekends.

The rental car companies have requested that their data not be released individually; therefore the information listed below represents the aggregated responses from multiple rental car companies. We were able to obtain rental car data only from a single rental car company with one of the largest fleets in Atlanta. This company has a total of 5,400 vehicles and averages 80 miles per day per customer and about 6,810 customers per week. This suggests that the average total daily VMT for rental cars in Atlanta is about 80 x 6,810/7 ~ 78,000 miles per day.

Towing Services

Towing companies are abundant in cities, but the number of vehicles owned by each company is low. Three towing companies in Denver were contacted to obtain fleet size information and the average VMT per vehicle. This information is summarized in Table 3.36. In order to expand these data to represent the entire urban area, all towing companies would need to be contacted.

Table 3.36: Sample Towing Truck Statistics from Individual Contacts

Company Name	Number of Trucks	Average VMT/Day
APT Service Inc.	20 trucks	75-100 miles/day Five days in a week
Burning Desire to Tow	Two weekday One weekend	150 miles/day/truck Less mileage on weekends
Midnight Express	One local truck Two long-distance trucks One repo truck	150 miles/day, Five days in a week

United States Postal Services

The United States Postal Service has Vehicle Maintenance Facilities (VMF) in each large city. The VMFs have the information on the number of vehicles in each city and the VMT that they travel. The Postal Service operations in all 12 urban areas were contacted, but data were received only from Houston. In Houston the USPS maintains:

• 40 tractors/trailers (intracity transportation);
• 35 cargo vans (intracity transportation);
• 2,280 light and medium vehicles (residential delivery); and
• 139 service and administrative vehicles.

The average annual mileage per vehicle is 7,160.

Comparison of Data from Individual Contacts with Those from Other Sources

In general, the data obtained from individual contacts numbers show lower total numbers of vehicles and VMT than the data from other (generally national) sources. In the case of school buses, as shown in Table 3.37, the total numbers of school buses and VMT from the individual contacts are substantially lower than the data from the School Bus Fleet survey and these represent the same geographic area (school districts). Similarly, the information on USPS vehicles and VMT obtained from the contact in Houston differs significantly from the data obtained from the national USPS source. Further evaluation of these data may indicate that the geographic coverage of these datasets are not the same, even though it is reported for the same area or it may indicate that the individual contacts are not capturing all vehicles where the national sources are better at capturing all vehicles. We were unable to determine the cause of the differences from the data available from these sources.

Table 3.37: Comparison with Other Data Sources

Cities	School Bus Fleet Size: Individual Contact	School Bus Fleet Size: School Bus Fleet Data	USPS Fleet Size: Individual Contact	USPS Fleet Size: National USPS Contact
Atlanta	388	2,885	0	0
Winston-Salem	156	345	0	0
Greensboro	364	598	0	0
Detroit	490	777	0	0
Houston	0	0	2,494	4,169
Total Number	1,671	4,605	2,494	4,169
VMT	VMT	VMT	VMT	VMT
Atlanta	27,225	97,475	0	0
Winston-Salem	11,050	27,530	0	0
Greensboro	19,476	41,973	0	0
Detroit	40,167	28,132	0	0
Houston	0	0	58,356	78,575
Total VMT	110,130	195,110	58,356	78,575

Data Sources

1. Atlanta Area Commercial Vehicle Survey. Draft Final Report. NuStats International, 1996.

2. http://nhts.ornl.gov/2001/index.shtml

3. R.L. Polk & Co., 26955 Northwestern Highway, Southfield, Michigan 48034.

4. Steven Stern, Department of Economics, Rouss Hall, University of Virginia, Charlottesville, Virginia 22903. http://www.people.virginia.edu/~sns5r/sect15stf/sect15.html

5. http://transportation.sandi.net/stats.html

6. http://www.schoolbusfleet.com

7. Airport Ground Access Planning Guide First Phase, Federal Highway Administration Intermodal Division, Washington, D.C. 20590. http://ntl.bts.gov/DOCS/AGAPP.html

8. Federal Highway Administration, Airport Ground Access Planning Guide, Intermodal Division, HEP-50, page 110.

Updated: 6/28/2017