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

5.  Work Zones

5.1  Methodology

The method used to estimate delays due to work zones was similar to the one used for crashes since both types of events produce localized bottlenecks.  The relevant variables defining a work zone are its location (urban or rural), the total number of lanes, the number of lanes closed, the length of the work zone, and the duration of the work.  Since 1999 data was not available, the work zone estimates in the study are for May 2001 through May 2002.

The basic process for estimating the delay from work zones includes the following:

Step 1. An inventory of work zones and their attributes was obtained from the Rand McNally website.

Step 2. Capacity loss was estimated based on the number of lanes normally open and the number of lanes closed due to the work zone, along with the length of time the lanes were closed.

Step 3. Delay was estimated based on capacity reduction, vehicle demand on the segment, and the duration of the work zone.

These steps are described in more detail in the paragraphs below.

5.1.1  Identifying Work Zones

Identifying the location and time span of work zones was problematic.  There are currently two national-level data sets that identify work zones:  the Rand McNally website and FHWA's Fiscal Management Information System (FMIS). However, each has significant limitations.

Rand McNally Construction Information Data

Rand McNally Construction Information for North America is an Internet-based searchable highway construction information system maintained to inform drivers of work zone activities.  It allows the users to find highway construction information by road type, by signed route, by beginning and ending dates, and by state.  It includes location, beginning and ending dates, and, in some cases, lane closures for existing and scheduled work zones. However, any data regarding work zones that are no longer active are purged from their database.  In addition, data on scheduled work zones are typically only accurate to about four months into the future from the time it is accessed.  Finally, this data is collected from state and local agencies by a team of Rand McNally employees.  Thus, time and resource considerations can affect the amount of data that is gathered.  ORNL downloaded work zone data for May 2001 through May 2002 and used it as a surrogate for 1999. Statistics for the number of work zone miles and mile-days by highway and area type are given in Table 18.

Fiscal Management Information System (FMIS)

The FMIS database, maintained by FHWA, tracks the allocation of funding obligated to states for federally funded highway projects.  According to FHWA, the database includes information on highway and bridge construction and maintenance projects, their locations, the type of activity funded, and other data.  The FMIS data for calendar year 1999 was acquired and evaluated as an alternate or complimentary source of information on work zone characteristics.  Initial review of the data suggested that it might be useful for developing an inventory of work zones and their characteristics.  However, further examination showed that FMIS's project time and location characteristics did not provide enough detail to estimate capacity reduction accurately. Therefore, it was decided that the Rand McNally database was more appropriate for the TLC and TLC2 studies.

5.1.2  Estimating Loss of Capacity

Work zones produce a loss of capacity that depends on the total number of lanes normally available and the number of lanes closed.  The "end-of-transition" and "activity-area" capacities of work zones' open lanes as a function of the available lanes and the environment (rural or urban) of the facility were determined using data from the Highway Capacity Manual, Special Report 209 (TRB 1985) and studies performed by Dixon (1996).  Table 19 shows the end-of-transition and activity-area capacities of work zones' open lanes as a function of the available lanes and the environment (rural or urban) of the facility.  The table was obtained by combining information from Tables 4 and 5.

Information from Table 5 was given priority in creating Table 16 above.  Where information was missing, data provided in Table 4 was used.  For some combinations (e.g., 5 available lanes and 3 lanes closed), the closest data from Table 16 that provided the same number of open lanes was used. For the previous example, 1,480 vphpl would be used, which corresponds to 4 lanes available and 2 open lanes.  When the information was derived from Table 4, the same values were adopted for the end-of-transition and activity-area capacities.

To determine the capacity of open lanes in a work zone, it was necessary to know the total number of available lanes and the number of lanes closed.  The total number of lanes on each highway segment being analyzed was obtained from HPMS.  If the work zone database contained the number of lanes closed, that information was used.  For cases where this data was not available, data from the Highway Capacity Manual, Special Report 209 (TRB 1985) was used to estimate the number of lanes closed due to the construction/maintenance work (Table 20).

Finally, to compute the travel time delays at work zones, it was necessary to know the physical length of the work zone.  This information was obtained from the Rand McNally database, where available.  If this data was missing, a work zone length of 0.5 miles was assumed for bridgework, or a length of 1.0 mile was assumed for other work.

5.1.3  Delays Due to Work Zones

Work zone delay was estimated using a method similar to the one used for crashes, since both events produce localized bottlenecks.  Delay estimates were based on (1) the capacity of the open lanes at the work zone (determined based on Table 19), (2) the traffic demand as a function of time of day and day of week, and (3) and the location of the work zone (determined using the same procedure as for crashes).  Delays were computed for each hour of each day during the entire duration of the work (using the corresponding demand) and then added to determine the total delays due to the work zone.

To estimate the queues that would form upstream of the work zones and to calculate the delays produced, the "end-of-transition" capacities were used (see Table 19), while the "activity area" capacities were used to compute the travel time delays.  One difference, however, between crashes and work zones is that travelers can often adjust their travel behavior (e.g., trip re-routing, trip re-scheduling, and trip canceling) during the duration of the construction period.  When this is a feasible alternative to drivers, the demand will be reduced.  This effect was not considered in the present study.

5.2  Results

Capacity losses due to 1,063 work zones reported by Rand McNally to be active on freeways and principal arterials during May 2001 to May 2002 amounted to an estimated 8.4 billion vehicles per year (Table 21).  Resulting delay is estimated at 889 million vehicle-hours. The study estimates that most of the delay was experienced on urban freeways (83 percent) and rural freeways (15 percent). Work zones on urban and rural principal arterials combined are estimated to only account for about 2 percent of delay. However, delay for these highways may be underestimated, since work zone data for freeways is likely to be more complete.

The reliability of capacity loss and delay estimates for work zones is unclear due to potential data gaps in the Rand McNally database (the Rand McNally database was not designed for this application). Small or short-term construction or utility work zones may not be included, and the database is less accurate at identifying work zones that will begin activity four months into the future. In addition, work zone impacts are complicated to measure since drivers often have prior knowledge of work zones and can reroute, reschedule, or cancel trips accordingly. This is especially true for long-term work zones. However, our estimates are based on the assumption that no rerouting, rescheduling or reduced mobility occurred.

Fig. 13. TLC2 estimates that 98 percent of delay from work zones was experienced on freeways.

5.3 Reliability

5.3.1 Methodology

The confidence level of the methodology used to determine capacity losses depends on the type of facility where the work zone was located. The data used to estimate capacity loss is based on only two studies of freeway work zone impacts, and these studies stress the high variability that exist from site to site in terms of lane capacities. Therefore, higher confidence is given to freeway work zones than for work zones located on arterials.

The methodology used to estimate delay for TLC2 is based on traffic flow theory and is well established in the transportation community (TRB, Highway Capacity Manual 1985). This methodology can be qualified as having a high degree of confidence for freeway work zones, and a medium level for arterials and other facilities with unrestricted access. The reliability of the delay estimates, however, is also affected by the reliability of the data and methods used to determine capacity loss and demand.

5.3.2 Data & Key Assumptions

Work Zone Inventory & Characteristics: The Rand McNally dataset used to identify and locate work zones is accorded a medium to low level of reliability for the application for which it was used. The primary short-comings include (1) the fact that May 2001 to May 2002 had to be used as a surrogate for 1999 data and (2) the likelihood that smaller, short-term work zones and work activities beginning more than four months into the future are not included in the data.

Traffic Demand & Surface Street Characteristics: The same traffic demand data and surface street assumptions used to estimate crash delays were used for breakdowns. Therefore, the same caveats apply: traffic demand data is accorded a low level of confidence and surface street characteristics are accorded a medium level of confidence (see section 3.3.2).

Previous | Table of Contents | Next