美国养护工区特征分析

美国养护工区特征分析

Wunderlich, Hardy and Hardesty

Characterizing Nationwide Work Zone Activity:

Key Findings From Website Content Analyses, 2001-2003

Dr. Karl Wunderlich

Senior Principal Analyst, Mitretek Systems

Suite 755, 600 Maryland Avenue, SW

Washington, DC 20024

Phone: (202) 488-5707

Fax: (202) 863-2988Matthew Hardy

Lead Transportation Engineer, Mitretek Systems

Suite 755, 600 Maryland Avenue, SW

Washington, DC 20024

Phone: (202) 863-2988

Fax: (202) 863-2988

matthew.hardy@

Dawn Hardesty

Senior Systems Engineer, Mitretek Systems

Suite 755, 600 Maryland Avenue, SW

Washington, DC 20024

Phone: (202) 863-3648

Fax: (202) 863-2988Words: 4,738

Figures and Tables: 5

Total Count: 5,988

Page 1

美国养护工区特征分析

ABSTRACT

In order to characterize and estimate the scope of work zone activity across the nation, the Federal Highway Administration (FHWA) initiated a periodic web site content

acquisition and analysis effort in 2001. This activity captures a snapshot of work zone activity from selected state department of transportation websites dealing with road

closures and roadwork. Five snapshots have been taken and analyzed in the effort, each comprising roughly two weeks of work zone activity, three in the peak summer roadwork season (2001, 2002, 2003) and two in the off-peak (winter) periods (2002, 2003).The report provides a synthesis of findings and preliminary trends observed in the five snapshots.

The snapshot efforts estimated that during the peak roadwork season (summer), there are 6,500-7,200 work zones covering 20-27% of the National Highway System (NHS). Off-peak (winter) work zone activity declines by more than half, with 2,000-3,000 work zones covering 7-8% of the NHS.

The vast majority (over 85%) of work zone descriptions contain information on project duration, most (over 50%) report lane closures by time of day and number of lanes closed, few (less than 15%) provide delay advisories. From analysis of this detailed

content, roughly one-third of work zones with lane closures occur primarily at night.The most likely hoursof the day for a lane closure are9-11 AM (roughly two-thirds of all work zones), the least likely period is 6-7 PM (one quarter of work zones).

KEYWORDS: Work Zone, Delay, National Highway System, Capacity Impacts, Construction, Maintenance, Bridge Work, Incidental Construction, Website, Project Duration, Lane Closure.

美国养护工区特征分析

BACKGROUND

In response to increases in work zone fatalities and evidence of growing traveler frustration with delays associated with work zones, in 2001 the Federal Highway Administration (FHWA) Work Zone Safety and Mobility Integrated Product Team

initiated an effort to measure and report national-level statistics on work zone exposure and delay. Lack of data on the number and type of work zones from year to year made the formulation of national policy on work zone safety and mobility difficult, since

increases in fatalities or delay could not be examined with respect to key factors such as a potentially corresponding increase in nationwide work zone activity or exposure.

One approach to the development of national statistics on work zone activity was the automated capture and interpretation of work zone data posted to state road closure and construction websites. In order to assess the potential value of these web-based resources as a part of a national monitoring program, researchers conducted an initial “snapshot” survey of 13 state road closure and construction websites during a two-week period in the peak summer roadwork season of 2001. The pilot snapshot [1] was followed by a two-year series of semi-annual snapshot analyses for 20 state road closure and construction websites [2,3,4,5]. Two snapshots per year are analyzed to identify trends for work zone activity in what are traditionally considered off-season (winter) and peak-season (summer) construction periods.

The objective of the series of snapshots was to explore the suitability of web-based work zone data to support the calculation of measures that could characterize the extent and nature of work zone activity across the nation. In addition to documenting findings with respect to this goal, this paper summarizes observations and trends regarding work zone activity seen in data posted to state road closure and construction websites. Data

collection techniques and analysis methods are described in the next section. The paper then provides an overview of the kinds of work zone data found on state websites as well as observed trends in the form and content of state road closure and construction

websites. Last, the paper presents highlights and trends of the nationwide statistics that could be supported by the data obtained.

APPROACH

In this section,data collection sources considered for the work zone snapshot study are described. In addition, our general approach for interpreting work zone data is presented. This is a key issue given that states do not report the same data, convey the data in a similar format, or provide detailed explanation for various ambiguous data rmation regarding the location of roadwork can be found in several web-based resources. For this effort, three potential sources were considered:

1.Rand McNally Road Construction Website

(/rmc/tools/roadConstructionSearch.jsp)

美国养护工区特征分析

2.Project-specific Websites

(e.g., )

3.National Traffic and Road Closure Information PageThe Rand McNally website covers roadwork activity from major construction projects across the nation. The data from this resource has been utilized in other studies looking at the sources of traffic congestion [6]. The data provided identifies projects by location and date but does not contain details of work zone activity such as schedules of lane closures by day of week or time of day [7].

Project-specific websites contain a wealth of information regarding primarily major urban rehabilitation and construction projects. However, there were too few of these resources to provide a representative cross-section of roadwork across the nation.

In order to collect data on some details of work zone activity and to improve the breadth of our analysis, data for this effort was collected from web-based resources linked to the National Traffic and Road Closure Information page. The level of detail was somewhere between the detail of the project-specific websites and the project-level approach of the Rand McNally resource, although this differed significantly from state to state (and even within a state). Further, collectively, the state web resources covered roughly twice the number of work zone records found concurrently in the Rand McNallyresource [7].

Of the 50 state and other resources listed, data were systematically captured on individual roadway projects and their associated work zones in 13-20 states over a two-week “snapshot” period.The pilot snapshot was conducted in thirteen states, the four semi-annual snapshots were conducted in twenty states. Figure 1 shows the statesanalyzed in all five snapshot efforts. The final sample of twenty states, kept consistent across the four semi-annual snapshot efforts was selected based on the number of work zones described and the amount of detail provided for each work zone.Since the data were organized differently state-to-state, and in many cases descriptions provided in text form, an analyst to inspect and code each record was required.

Managing and interpreting data from the disparate state resources was a significant effort. In order to avoid double-counting of projects posted twice over the same time period, automated routines were developedto speed the identification and removal of duplicate records. A systematic coding rubric was required so that interpretation of various data elements was consistent regardless of the analyst performing the interpretation. Even with partial automation, data acquisition and coding was a significant and time-

consuming element of each snapshot study.

Our experience with the acquisition and coding of data elements suggests that if states followed a consistent set of guidelines for data elements and their definition,

comprehensive national studies of work zone activity would be significantly cheaper and more valuable. A recent paper explores the possibility of establishing a set of national guidelines for posting work zone information [8].

美国养护工区特征分析

WORK ZONE WEBSITES: DATA ELEMENTS AND TRENDS

Typical data elements that were identified for each work zone include:

Location. This element typically references a roadway by name and indicates an approximate location by referencing a nearby

exit, mile post marker, or town.

Purpose.Project purpose is often described in detail, (e.g., “erosion control”).

Project Duration.Project start and end dates are typical data elements. Duration is computed as the difference in days between

these dates.

Length.The length of a work zone is most often expressed using mile post designations (e.g., “between MP 12.4 and 18.1”) Lane Closure ne closures are identified for work zones by day of week and time of day. The number of lanes closed or other

capacity impacts are also reported here.

Delay.Sometimes an advisory about potential delays are posted, and in rare cases an estimate in minutes of delay is posted. Cost.Total project costs are sometimes included.

Some data elements are reportedmore frequently than others. Every work zone record identified had some sort of location entry. Excluding the 2001 pilot study, the vast majority (86-90%) of work zone descriptions contain information on project duration, most (51-62%) report lane closures by time of day and number of lanes closed, while few (5-13%) provide delay advisories. Table 1 presents our key data elements and their frequency of reporting in each of our Snapshot analyses, including the 2001 pilot study.Some elements frequently reported while others are rarely reported. For example, expected delay, particularly quantitative estimates of delay, arerarely reported. The dearth in delay reporting reflects the fact that delays are not typically estimated in detail prior to the start of roadwork and are not typically monitored precisely once roadwork is underway. It is also clear that different state websites have different target audiences ranging from roadwork contractors and state personnel to the traveling public. In many cases data elements of interest to contractors and state personnel (e.g., project location and purpose) are known in advance of the actual roadwork, do not change, and are

frequently reported. Data elements of interest to the traveler (e.g., lane closure schedules or quantitative estimates of delay) can change from day-to-day or minute-to-minute, are harder to estimate, and are less frequently reported.

Discerning the precise meaning of some data elements is complicated by ambiguities and inconsistencies among states and within states. The snapshot study approach in such cases is to have a clear, consistent coding policy to guide the analyst. For example, when states report times of day with lane closures in a work zone, it is frequently unclear whether the intervals reported are the times of day when a traveler may encounter lane closures or times of day when a traveler is certain to encounter a lane closure. In the

美国养护工区特征分析

coding rubric, all lane closure schedules are reported as the latter –as periods when traveler is certain to encounter a lane closure.

Although the state web resources provided the best breadth of coverage of work zone activity of the three web sources considered, they cannot report every work zone

statewide--there are simply too many. Unsurprisingly, the state websites report primarily on the facilities that reflect state responsibilities: interstate, US highway and state routes. No more than 11% of the work zone records examined dealt with roadwork on local streets in any of our snapshot analyses.Clearly, additional roadwork also occurs within the boundaries of the states considered here. Therefore, the data collected in this study from state websites should not be construed as a completely representative illustration of all maintenance, construction and utility activity across the nation, but rather as a subsetfocusing on major facilities where road construction and maintenance activities are primarily state responsibilities.

From the 2001 pilot study to our more recent snapshot analyses, several clear trends have emerged in the format and content of state roadwork websites:

the pilot study, there were 15-20 states to choose from when selecting a sample for

analysis. By 2002, the number of potentially useful state websites had roughly doubled. At this point (2004), all stateshave websites reporting work zone information of some kind.and appeared to primarily target contractors and other involved directly in the roadway projects. In 2003, 25-50% of state websites have moved to a road-user perspective reporting daily lane closures.posting work zone data. For example, in the 2002 summer season, Maine reported data on two major projects. After a website redesign effort in the summer of 2002, Maine reported data on over 200 work zones.

THE NATIONAL WORK ZONE PICTURE

In this section, an overview of findings from all of the five snapshot studies is presented to provide a broad national characterization of work zone activity. First, an analysis of primary data elements is presented, such as project duration and lane closure schedules(Table 2). Second, from these basic elements, national-level measures are derived, including the total number of work zones on the National Highway System (NHS), the percentage of the NHS with work zones, and an estimate of capacity loss from work zones. The NHS is a designated collection of 163,000 miles of roadway critical to the nation’s economy, defense, and mobility. The NHS is composed of interstate highways

美国养护工区特征分析

and principal arterials, as well as connector roadways critical for defense or the

movement of goods between intermodal facilities.

()

Project Length

In summer 2003, the average work zone length was 6.1 miles and the median length was

2.9 miles. A number of long work zones associated with paving work accounts for the difference between the average and median length values and has been observed

consistently in 2001 and 2002. As indicated in Table 2, work zone length has varied from

5.3-6.8 miles in the summer, and 3.8-4.5 miles in winter. The shorter winter work zone length is related to a drop in paving work in cold weather states.

Project Duration

Duration is one of the most widely reported statistics, with an average duration of 98 daysin summer 2003 (about 3 months) and a median of 16 days (about 2.3weeks). This

reflects a consistent pattern where the average duration is higher than the median duration because of a number of multi-year projects. One trend that can be observed from Table 2 is the decline in average project duration from summer 2002 to summer 2003. One reason for this decline is the trend in state website reporting format away from a project-level perspective to a daily lane closure perspective. When a state adopts the daily lane closure perspective, often it is difficult to discern project duration, particularly for large, complex projects lasting more than one season.

Lane Closures by Time-of-Day

Roughly half of the records inspected reported the time of day when the work zone had lane closures that were in place or capacity reductions from work zones. Of these records, summertime lane closures averaged 11.0-11.9 hours per day over our five

snapshot studies in 2001-2003. On average, winter lane closures had somewhat shorter durations, averaging 9.0-10.7 hours per day.

This data was further stratified into three subgroups: Work zones with capacity impacts 18 or more hours per day, work zones active (or with lane closures) primarily at night, and work zones with capacity impacts primarily during daylight hours. Any work zone with a duration under 18 hours but more than 2.5 hours of work during presumed hours of darkness (10 PM – 5 AM) was designated as “nightwork.” All other roadwork with duration less than 18 hours per day was designated “daywork.” Roughly a quarter to a third of all roadwork meets our definition of “nightwork” in the summer. Nightwork is generally lower in the winter.

Lane closure data can also be arranged to show capacity impacts by time of day, broken down by 30-minute periods (illustrated in Figure 2 for summer 2003). Note that some work zones that are primarily nightwork can extend into the mid-morning hours.

美国养护工区特征分析

Likewise, work zones that are primarily daywork sometimes extend into the late evening.The most frequent period of work zone capacity impact has been9-11AM in all five snapshot studies, when roughly two-thirds or more of the work zones were identified as having lane closures.The least frequent hour for lane closures in our snapshot studies have been from 6-7 PM when only 26% of work zones were designated as having lane closures.

Lane closure temporal patterns, such as the one shown in Figure 2, have remained

generally consistent across the snapshot studies conducted in the same season (winter or summer). In winter, however, lane closures are weighted more towards dayworkProject Purpose

The purpose of the roadwork activity was a widely reported attribute of projects across all states. In some cases there was a state-standard set of purposes, in other cases, freeform text descriptions had to be interpreted. The records were grouped into four broad

categories (Bridge Work, Resurfacing/Paving, Roadway Construction and Rehabilitation, Other/Incidental Construction). Table 2 shows the frequency of each activity. Roadway construction and maintenance was the most likely reported purpose.

closures of long duration. Roughly one half of all bridge work activity takes

place during daylight hours. Bridge work is one of the most frequently reported

work zone purposes during the winter season.

operations had long work zone lengths and are the most likely activity for work zones with lane closures at night. Paving operations are far

more frequent in the summer season than in the winter season, dropping in

frequency by more than half.

likely because this work is largely reconstruction of existing facilities. The split of daywork and nightwork followed the average for all roadwork observed.

primarily as daywork. This category is difficult to characterize further as it

represents a disparate mix of construction and maintenance activities. National Measures of Work Zone Activity

From our analysis of work zone length, lane closure duration and lane closure type,

several high-level measures of work zone activity on the NHS are estimated. Excluding work zones on local streets, the number of work zone records was utilized to estimate a total number of work zones on the twenty states in our sample (2002-2003).Avisual inspection of NHS highway maps and records from both Arizona and Washington in

美国养护工区特征分析

2001 revealed that, in addition to all interstate and US-highway work zones, the majority of state roadways with work zones were also NHS roadways.

Based on the total number of estimated work zones in the sample states, a national

estimate is derived by dividing by the total percent of NHS roadway in the sample. For example, the twenty state sample used in the four semi-annual snapshot studies

represented 42.2% of the NHS. NHS mileage by state was obtained from .

Work Zones on the NHS

As shown in Table 3, 6,400-7,200 work zones are estimated to be on the NHS during the peak construction season (2002-2003). Note that the 2001 estimate is lower because of the pilot study’s smaller sample size. In the winter, 2,600-3,000 work zones are

estimated to be on the NHS, roughly half of the number seen during the summer peak.Multiplying these estimates of the number of work zones by the average work zone length calculated for each snapshot, one can obtain an estimate of total miles of NHS with a work zone during the two-week snapshot period in each analysis.

There are 163,732 miles of roadway on the National Highway System. Therefore, in each snapshot, the estimate of miles of NHS with a work zone is expressed as a

percentage of the over total miles of NHS. From Table 3, our estimates have ranged from 13% in the pilot 2001 study to a 21-27% range in the summer 2002-2003 snapshots. Since there are fewer work zones in the winter and they have shorter length, the percent of NHS with a work zone drops to 6.9-7.4%, roughly a third of the summer time peak value.

Capacity Impacts

States provided data on likely capacity impacts for the majority of the work zones in each snapshot. Of these, nearly all of the capacity impacts were from temporary closures of one or more lanes or complete closures of roadway facilities. A small number indicated lane narrowing or shoulder work that would have some impact on roadway ing these capacity impact indicators from the state websites, total estimated capacity loss from work zones on the NHS can be estimated. First, a baseline capacity using standard values for each facility type from the Highway Capacity Manual (2000) is established. Based on text descriptions of each work zone, each record was coded as either shoulder or indirect work, a partial closure (by number of lanes closed), or full closure. A percentage reduction in roadway capacity was estimated in each case using the HCM. From these data, an estimate of reduced capacity from all the work zones on the NHS can be generated by including factors for work zone duration.This loss of capacity is measured in vehicles, and has ranged up to 180 million vehicles per day of lost capacity (summer 2003).

A peak loss in carrying capacity during the most frequent hour of activity was also

estimatedin each snapshot, expressed in terms of a single connected work zone extending

美国养护工区特征分析

over one direction of a six lane freeway. Collectively, this capacity loss has ranged

between 5,000-5,300 miles of freeway in the summer and roughly 2,000-3,000 miles in the winter snapshots (Table 3).

Delay

No estimate of road user delay was generated from snapshot data. Delay data,

particularly quantitative estimates of delay, are extremely rare (5-6%), and cannot be used to accurately characterize delay. Although capacity loss can be estimated, without data on distributions of travel demand, alternative routes, and other data, a defensible estimate of delay could not be calculated from these measures.

KEY FINDINGS AND CONCLUSIONS

This section presents a summary of key findings from across the five snapshot studies and discusses some conclusions that can be drawn from the findings.

Key Findings

Analysis of data obtained from the state web resources contained enough detail to support During peak mid-day hours, there are 6,500-7,200 work zones covering 20-27% of the NHS has a work zone in the summer peak construction season.

Nationwide, off-peak (winter) work zone activity declines by more than half, with 2,000-3,000 work zones covering 7-8% of the NHS. About one-third of work zones with lane closures occur primarily at night.

The most likely hours of the day for a lane closure is 9-11 AM (roughly two-thirds of all work zones), the least likely is 6-7 PM (one quarter of work zones). The loss of nationwide capacity each day during the summer construction season on the NHS is about180 million vehicles – and losses in the peak hour equivalent

to the carrying capacity of over 5,000 miles of a six-lane freeway (one direction). The vast majority (80-90%) of work zone descriptions contain information on project purpose and duration, most (50-65%) report lane closures by time of day

and number of lanes closed, few (5-13%) provide delay advisories.

In comparison to a companion study of paper records of work zone activity in five DOT districts [9], it is clear that work zones posted to the web tend to be part of

the largest, most-visible projects in the state. Although the sample set of work

zones where the two studies had direct overlap was too small to analyze in any

depth, it is clear that work zones posted to websites tended to have somewhat

different attributes (most notably longer project durations).

美国养护工区特征分析

State websites are moving away from an internal project-level data reporting scheme to a more motorist-oriented daily lane closure reporting format.

In some states, the number of work zones reported on websites has risen dramatically (e.g., Maine reported 2 work zones in 2002 and 222 in 2003). This

increase reflects an increased effort to report many lane closures, not just those

associated with a few high-visibility projects.

Although a rise in the number of work zones being reported to the web can be observed, far less dramatic change has been observed in the type of data elements

provided. For example, delay advisories have remained uncommon throughout

all five snapshot studies.

ConclusionsBecause of the type of data obtained in this study, there will always be a challenge in discerning changes in “on the ground” roadwork and changes in state policies on the format and content of their road closure and construction websites. For example, average reported project duration has fell from 142 days in summer 2002 to 98 days in summer 2003. Although this may reflect a trend towards more rapid completion of roadwork projects from innovative management and contracting efforts, it is confounded to some degree by a shift in web reporting format away from a project orientation and towards a lane closure schedule reporting format.the Texas Transportation Institute (TTI) where detailed paper records were analyzed in six jurisdictions across the nation. The TTI study estimated that 19-24% of the NHS had a work zone during the peak construction season in 2001, comparable to our 2002-03 estimates of 20-27%. Similarly, characterization of lane closure management such as nightwork were also comparable, with both TTI and the snapshot studies estimating that roughly one-third of all lane closure activity occurs as nightwork.cost-effectively characterize from paper records on a routine basis. Although there are limitations in the data obtained, one can estimate key measures on a national scale that have never been systematically tracked in the past. Clearly some of the key measures needed at the national level (exposure and delay) are not yet supported from web-baseddata. This issue may be resolved over time as more uniform and more standardized work zone data is presented on the web. Efforts at the federal level to establish guidelines on work zone data reporting will significantly strengthen the snapshot approach to national

美国养护工区特征分析

work zone monitoring. Such an effort will make national work zone activity monitoring cheaper, more accurate and more valuable.

ACKNOWLEGEMENTS

The authors wish to acknowledge numerous colleagues for their contributions to this study: Tracy Scriba and Scott Battles (Office of Operations, Work Zone Integrated

Product Team) the sponsors of this work at FHWA for their comments and support. Phil Ditzler (FHWA), for his help initiating and developing the snapshot study concept.

Colleagues at Mitretek who contributed to the studies: James Larkin, Donald Roberts and Michael McGurrin.

REFERENCES

1.A Snapshot of Summer 2001 Work Zone Activity Reported on State Road Closure and

Construction Websites, Wunderlich, K. and Hardesty D., FHWA Electronic

Document Library (), #13739, February 2003.

2.A Snapshot of Work Zone Activity Reported on State Road Closure and Construction

Websites, Winter 2002, (July 2002) to appear on FHWA Electronic Document

Library.

3.A Snapshot of Work Zone Activity Reported on State Road Closure and Construction

Websites, Summer 2002, (December 2002) to appear on FHWA Electronic Document Library.

4.A Snapshot of Work Zone Activity Reported on State Road Closure and Construction

Websites, Winter 2003, (May 2003) to appear on FHWA Electronic Document

Library.

5.A Snapshot of Work Zone Activity Reported on State Road Closure and Construction

Websites, Summer 2003, (December 2003) to appear on FHWA Electronic Document Library.

6.Temporary Losses of Highway Capacity and Impacts on Performance, Chin, S.M. et

al., (), May 2002.

7.A Review and Analysis of the Rand McNally Construction Website for Utilization in

Developing National Statistics on Work Zone Activity, February 2004, to appear on FHWA Electronic Document Library.

8.Guidelines for Posting Web Based Work Zone Information: Performance

Measurement Needs, March 2004, to appear on FHWA Electronic Document Library.

9.Work Zone Exposure and Safety Assessment, Ullman, G., Holick, A., and Turner, S.

prepared for FHWA, June 2003.

美国养护工区特征分析

Figure 1. State Work Zone Website Resources Analyzed, ’01-‘03

美国养护工区特征分析

80%

Percent of Sample Work Zones Active by Time of Day

70%60%50%40%30%20%10%

0%

t

11 AMM10 PMM10 AMidnigh11 PMMMMMMMMMM7 PMMMMM1 A2 A3 A4 A5 A6 Aoon1 PM6 P8 PM7 A8 A9 A2 P3 P4 PN5 P9 P

M

Figure 2. Work Zone Activity/Lane Closures By Time-of-Day (Summer 2003)

美国养护工区特征分析

Table 1. Data Element Reported By Snapshot, 2001-2003

美国养护工区特征分析

Table 2. Work Zone Characteristics by Snapshot, 2001-2003

美国养护工区特征分析

National Estimate Summer

2001(NHS)

Number of Work Zones

Lane-Miles of Lost

Capacity

Percent of NHS with Work

Zone3,1102,64312.8%Winter20022,9823,0436.9%Summer20026,4465,09820.9%Winter20032,6751,9337.4%Summer20037,1415,29927.0%

Table 3. Estimates of Work Zone Impact on NHS, 2001-2003

本文来源：https://www.bwwdw.com/article/vkl4.html