TEMPORARY LOSSES OF HIGHWAY CAPACITY AND IMPACTS ON PERFORMANCE: PHASE 2

Title | Contents | Acknowledgements | Exec. Summary
1. Intro | 2. Approach | 3. Crashes | 4. Breakdowns | 5. Work Zones | 6. Weather | 7. Signal Timing
8. RR Crossings | 9. Toll Facilities | 10. PUD | 11. Results Summary | 12. Next Steps | 13. References

10.  Commercial Truck Pickup and Delivery (PUD) Activities in Urban Areas

10.1  Methodology

Pickup and delivery (PUD) activities of large trucks in urban areas can cause highway capacity reductions and traffic delays when these vehicles are parked illegally, partially or fully blocking a lane of traffic. Commercial drivers park their vehicles illegally for several reasons, including lack of parking near the PUD point or a desire to park their vehicle where it can be easily seen to prevent theft or vandalism.

The method used to estimate the delay from commercial truck PUD activities varies somewhat from the methods used for other capacity-reducing phenomena in this study. Rather than using a Monte Carlo simulation method to place lane-blocking PUD events on highway links at a specific time and location and calculating the resulting capacity loss and delay, this study estimates a weighted average delay value for lane-blocking PUD activities on each link, which are summed for each urban area and multiplied by the estimated number of lane-blocking PUD activities for that area. The steps used to estimate delay from commercial truck PUD activities can be summarized as follows:

Step 1. All urban arterial links were identified and assigned to an FHWA urbanized area using FHWA's Highway Performance Monitoring System (HPMS) data.

Step 2. The weighted average capacity reduction and delay resulting from a single PUD activity on each urban link was determined.

Step 3. The average delay for a single PUD activity occurring in each urban area was estimated by averaging across all links and weighting by the length of each link.

Step 4. The annual number of PUD activities occurring in each urban area was estimated based on employment statistics for each land use type (i.e., commercial, industrial, and office) and trip generation factors that apply to each land use type—PUD activities were assumed to only occur within ZIP codes where the total annual salary was in excess of $600 million per square mile.

Step 5. Delay was summed for each FHWA urbanized area.

These steps are described in more detail in the sections that follow.

10.1.1  Identifying Arterial Links and Assigning Them to the Appropriate FHWA Urbanized Area

HPMS data was used to identify all links classified as urban arterials. This included links designated as functional system code 14 (other principal arterials), 16 (minor arterials), and 17 (collectors) in HPMS. These links were also assigned to the FHWA urbanized area in which they were located.

10.1.2  Estimating Weighted Average Capacity Reduction and Delay per PUD Event on Each Link for Each Land Use Type

For each urban principal arterial, the average capacity reduction and delay that would result from a lane-blocking PUD was estimated for each of three land use types (commercial, office, and industrial). A method similar to the one used to estimate capacity reduction and delay for crashes, breakdowns, and work zones was used, with some minor changes.

First, PUD events were not broken down into time intervals with different capacity-reducing characteristics; each event was modeled as a single interval. The dwell time for each event was assumed to be 15 minutes. This was based on observed dwell times reported in studies by Aherns et al. (1977) and Habib (1980). Aherns observed an average dwell time of 17.4 minutes per delivery for all stores participating in its survey.  The Habib study also reported an average dwell time of 19.5 minutes for PUD trucks legally parked at curb-side (5,046 observations), 13.8 minutes for trucks curb-parked illegally (1,697 observations), and 11.5 minutes for those double-parking in a moving lane (1,398 observations).

It was assumed that an illegally parked truck will block one lane of traffic—the one closest to the curb. If this event occurs on a link that has only one lane in that direction, it was assumed that traffic will go around the parked vehicle by partially crossing over into the oncoming lane.

For each link, the delay was averaged across all applicable hours of the day (6:00 am to 5:00 pm) and days of the week (Monday through Friday), weighting each time slice by the probability of a PUD activity taking place. An hourly PUD arrival distribution observed by Habib was used to weight delay by time of day for each land use type. The arrival distribution used to weight delay by day of the week was taken from Aherns et al. These distributions are presented in the tables below.

The equation used to estimate weighted average delay for each link for each land use type is provided below.

Equation 27:

where

D = delay

HFactor = probability of PUD activity by hour of the day and land use type

DFactor = probability of PUD activity by day of the week

l = link

f = FHWA functional classification (14=other principal arterial, 16=minor arterial, 17=collector)

q = land use type (commercial, office, industrial)

d = weekday, Monday through Friday (i.e., 1 to 5)

t = time of the day from 6:00 am to 5:00 pm (i.e., 6 to 17)

10.1.3  Estimating Weighted Average Delay per PUD Event for Each Urban Area for Each Land Use Type

Once the average delay per link was determined for each land use type, the average delay across all principal arterial links in each urban area for each land use type was estimated. Average delay was estimated by multiplying the delay for each urban principal arterial link by the percentage of total arterial mileage categorized as other principal arterial (functional class 14). This was calculated using the following equation.

Equation 28:

where

D = delay

u = FHWA urbanized area

q = land use type (commercial, office, industrial)

l = link

f = FHWA functional classification (14=other principal arterial, 16=minor arterial, 17=collector)

L = length

10.1.4  Estimating Annual Delay for Each Urban Area Based on Number of PUD Events

The next step was to estimate the annual delay for each urban area. Annual delay was estimated by multiplying the average delay per lane-blocking PUD event for each urbanized area by the annual number of lane-blocking events for that area. The number of PUD events (including those that do not block lanes) was estimated by applying PUD trip generation rates by the number of people employed within in each land use type. To estimate the number of PUD events that blocked lanes, it was assumed that 20% of all PUD events involve illegal parking that blocks a lane. This assumption was based on the Habib study that reported that 20–25% of PUD vehicles parked illegally in a curb-side moving lane. The equation used to estimate annual delay for each urbanized area is given below.

Equation 29:

where

D = delay

E = number of employees

R = trip rate (number of PUD activities generated per number of employees)

B = percent of PUD activities that block lanes due to illegal parking

W = work days in a year (260)

u = FHWA urbanized area (400 areas designated as urban)

q = land use type (commercial, office, industrial)

The rate for PUD events was based on truck trip generation data from the National Cooperative Highway Research Program (NHCRP) publication Truck Trip Generation Data: A Synthesis of Highway Practice, NCHRP Synthesis 298 (Fischer and Han 2001). The NHCRP report contains tables that provide trip rates per employee by type of land use. The NHCRP trip rates used in the TLC2 study were taken from a 1993 study in Tampa, Florida, by Gannett Fleming, Inc. (Table 35). For each land use type, the average rate for all truck types was used.

Employment estimates for each land use type in each urban area were based on two sources: U.S. Census Bureau ZIP Code Business Patterns data and U.S. Department of Labor (U.S. DOL) Bureau of Labor Statistics (BLS) Covered Employment and Wages (CEW) Program data.

The Census ZIP Code Business Patterns dataset provides business data summarized for nine employment-size classes by hundreds of North American Industry Classification System (NAICS) codes and about 40,000 ZIP codes nationwide. The database is geo-coded based on the ZIP Code Boundary & Inventory Files developed by Geographic Data Technology and contains location data for ZIP centroids. However, data on self-employed persons, domestic service workers, railroad employees, agricultural production workers, most government employees, and employees on ocean-borne vessels or in foreign countries are beyond the scope of the dataset.

Therefore, in order to include trips generated by government sites, the BLS data was used to supplement the Census data. The BLS data includes statistics on the number of establishments, monthly employment, and quarterly wages, by NAICS industry, by county, by ownership sector, for the entire United States. County-based government employment statistics were extracted from this dataset, and it was assumed that the land use type for all such establishments was "office."

The Census and BLS employment data were mapped to FHWA urbanized areas using location data in these data sets along with the Bureau of Transportation Statistics (BTS) Federal-aid Urbanized Area Boundaries database, a geographic database of Federal-aid boundaries for urban areas with a population greater than 50,000. The database includes boundaries for urban areas in all 50 states, the District of Columbia, and Puerto Rico. The Census ZIP-code-based dataset includes location information on the centroid of each ZIP. For each urbanized area, employment for each ZIP whose centroid lay within the urbanized area boundary was added to the total for that area.

A different method was used for mapping county-based BLS data to each urbanized area. The percentage of each county's land area that lay within the urbanized area boundary was multiplied by the total employment for that county. The BLS and Census employment estimates were then summed for each urbanized area.

It was assumed that most capacity loss and delays from PUD activities would occur along roadways in central business districts where buildings are close together and parking is limited. However, within the time and funding constraints of this study, it would be quite difficult to identify these locations. Therefore, as a surrogate indicator for these conditions, capacity reduction was only assumed to occur on roadways within ZIP codes where the total annual salary of employees working within the ZIP code was $600 million per square mile—this data is included in the Census ZIP Code Business Patterns dataset.

Since the truck trip generation factors are for daily trip generation, the delay had to be multiplied by the number of workdays in a year (260) in order to generate annual totals.

The final step was to sum capacity reduction and delay for all urban areas by urban area size category.

10.2 Results

The study estimates that, in 1999, PUD activities on urban principal arterials caused a capacity reduction of about 117 million vehicles, resulting in approximately 947 thousand vehicle-hours of delay. Most of this delay (nearly 90 percent) occurred in very large urban areas. Nine (9) percent of the total delay occurred in large urban areas, with medium and small urban areas accounting for 0.1 and 1 percent, respectively. Most of the delay (67 percent) occurred during off-peak hours.

Figure 28. Nearly 90 percent of the delay from PUD activities in 1999 occurred in very large urban areas.

Figure 29. Most of the delay on urban principal arterials from PUD activities in 1999 occurred in off-peak hours.

10.3  Reliability

10.3.1  Methodology

Methodologies from the Highway Capacity Manual were used to estimate capacity loss and delay. These methodologies have evolved over the years and have been updated and enhanced continuously when new information or methodologies are made available.  They are well established and accepted within the traffic engineering community. These methodologies are qualified as having a high degree of confidence. However, the final capacity reduction and delay estimates are significantly influenced by trip generation assumptions (described below) that are qualified as having a low level of confidence. Furthermore, using annual salary per square mile as a method for identifying dense commercial development with limited parking has a low level of confidence.

10.3.2  Data & Key Assumptions

The Census ZIP Code Business Patterns dataset prepared annually by the U.S. Census Bureau is well established and is qualified as having high degree of confidence.

The Bureau of Labor Statistics (BLS) Covered Employment and Wages (CEW) Program data pertains to workers covered by State unemployment insurance (UI) laws and Federal civilian workers covered by the Unemployment Compensation for Federal Employees (UCFE) program. The data for both private sector and public sector workers are reported to the Bureau of Labor Statistics (BLS) by State employment security agencies as part of the Quarterly Census of Employment and Wages program. The CEW data are used as the benchmark source for employment by the Current Employment Statistics program and the Occupational Employment Statistics program. This information is qualified as having high degree of confidence.

Pickup and delivery trip generation rates are based on a single study cited in Truck Trip Generation Data: A Synthesis of Highway Practice, NCHRP Synthesis 298 and the ITE Trip Generation Handbook (a publication widely used by traffic engineers). However, the study is over 10 years old and only recorded observations for 30 sites (5 observations for each land use type). Thus, this information is qualified as having low degree of confidence.

Other urban truck pickup and delivery information used in TLC2, such as PUD activity percentages by day of week, percent of PUD events that block a traffic lane, and duration of lane-blocking activities, are also based on limited, outdated studies. Thus, this information is qualified as having low degree of confidence.

None of the information regarding PUD activities is differentiated by highway class. Therefore, TLC2 assumed that PUD events were equally likely to occur on all urban highway classes other than freeways. However, it is likely that a large share of lane-blocking PUD activities occur on lower-order highway classes (i.e., classes other than principal arterials). Therefore, this assumption may result in an overestimate of temporary capacity loss and delay on principal arterials.

Previous | Table of Contents | Next