Table of Contents

4. Analysis of Work Zone Attributes

In this section, aggregate statistics are presented on the data elements collected. First each data element is discussed, followed by a characterization of work zone attributes by four major purpose categories. We then present a methodology for the estimation of several national-level statistics from the data collected. This includes the total number of work zones on the NHS, the percent of NHS with a work zone, and a national estimate of capacity loss from work zones.

4.1 Length of Work Zones

The state web sites either directly reported length of work zones or reported data from which length can be inferred for 142 (18%) of the 789 work zones in the analysis. The data obtained by inference here was primarily from web records that report roadwork from one milepost to another on the facility. The average recorded work zone length was 6.8 miles and the median length was 4.0 miles. A histogram of the work zone length data is provided in Figure 2. Note that work zones varied greatly in length, from less than 1 mile to 20 miles or longer. The longest work zone reported in the data covered 88 miles of State Route 260 in Arizona. More than half of the work zones with posted length data came from three states: Kentucky, Arizona and Oregon. Two states (Idaho, Nevada) did not report work zone length for any projects.

 

Figure 2. Distribution of Work Zone Length (142 Work Zones)
Figure 2. Distribution of Work Zone Length (142 Work Zones) [D]

4.2 Work Zone/Project Duration

Duration is defined as the number of days between the reported start and end date, and is one of the most widely reported statistics. Start and end dates were reported for 563 (71%) of the work zones, with an average duration of 125 days (about 4 months) and a median of 65 days (about 2 months). A histogram of duration is illustrated in Figure 3. Note that the majority of the projects were 1-6 months in duration, which is not surprising given the nature of peak summer construction season activity.

Duration was widely reported in the state websites studied. It was clear that some states are reporting project durations (a year or more in some cases) while others reported the duration of currently posted lane closure or work zone activity. Sometimes it was unclear whether the indicated duration describes a single phase (e.g., a two-week demolition phase) of a larger project (e.g., a multi-year rehabilitation) or the entire project. In many cases there was no data to support categorizing the reported duration as either project, phase or lane closure duration. Because of this ambiguity, we report all durations here in a single category.

 

Figure 3. Distribution of Work Zone Duration (563 Work Zones)
Figure 3. Distribution of Work Zone Duration (563 Work Zones) [D]

4.3 Hours of Activity or Lane Closures

For a total of 175 work zones (22%), states reported the time of day when either the work zone had lane closures in place or had activity in the work zone that was likely to have an impact on roadway capacity. Like the ambiguity with respect to project and work zone duration, hours of work zone activity and lane closures are reported together here since they cannot be clearly differentiated. For simplicity, we lump these items together as capacity reductions to the roadway. These data elements were dominated by reports from Washington, Arizona and Nevada. However, each state reported lane closures by time of day for at least one of its work zones.

The average number of hours of capacity reduction reported was 11 hours per day, with a median of 10 hours. A histogram showing the various hours of lane closures/activity per day is shown in Figure 4. This data was further stratified into three work zone subgroups:

Note that daywork average duration and nightwork average duration were quite similar (roughly 9 hours) but that daywork had wider variation. That is, very short and relatively long days were common for daywork, whereas nightwork was tightly clustered around durations of 7-9 hours - indicating that if nightwork is conducted, it is conducted all night rather than just one portion of the hours of darkness.

The data can also be arranged to show capacity impacts by time of day, broken down by 30-minute periods (illustrated in Figure 5). Figure 5, again stratified by night, day and all-day subgroups, illustrates that 33% of work zones had capacity impacts primarily at night, 58% in the daytime hours, and 9% all day or nearly all day (18+ hours).

The data also show that the most frequent hour of work zone capacity impact was 9 AM to 10 AM, when 67% of the work zones were found to be either active or have lane closures, while the least frequent hour was 6 PM to 7 PM when only 19% of work zones were found to have a capacity impact. Note that some work zones that are primarily nightwork can extend into the mid-morning hours; likewise, work zones that are primarily daywork sometimes extend into the late evening.

 

Figure 4. Distribution of Daily Duration, Work Zone Activity/Lane Clsoures (175 Work Zones)
Figure 4. Distribution of Daily Duration, Work Zone Activity/Lane Closures (175 Work Zones) [D]

 

Figure 5. Work Zone Activity/Lane Closures by Time-of-Day (175 Work Zones)
Figure 5. Work Zone Activity/Lane Closures by Time-of-Day (175 Work Zones) [D]

4.4 Delay Impacts Estimation

Expected impacts on delay were reported for 26% of work zones. This included reports stating that no delay was expected. For the other 74% of work zones, the websites did not provide an indication of whether delays could be expected. Quantitative estimates of delay were quite rare, appearing in only 7% of the work zone reports. Of these 56 work zones reporting quantitative delays, the average expected delay was roughly 12 minutes. No site provided expected delays by time-of-day or separate forecasts for approaches to the work zone from various directions of travel.

4.5 Construction/Maintenance Project Purpose

The purpose of the roadwork activity was a widely reported attribute of projects across all states (78% of all work zones). In some cases, there was a state-standard set of purposes, while in other cases freeform text descriptions had to be interpreted. We grouped the records into four categories and nine sub-categories of construction and maintenance work:

Figure 6 shows the frequency of each activity. Bridge construction was the most likely purpose reported, followed by pavement maintenance and roadway reconstruction. In general, we observed that a work zone was more likely to be posted to a state website if it was a large project or one that closed a facility. This likely over-representation of big projects with road closures helps to explain the large percentage of bridge-related activity posted to the state websites. Clearly, there are many more mobile work zones or short-term maintenance activities in these jurisdictions that cannot be identified by capturing only data posted to state DOT websites.

 

Figure 6. Work Zone Purpose (619 Work Zones)
Figure 6. Work Zone Purpose (619 Work Zones) [D]

4.6 Characteristics of Work Zones by Project Purpose

Next, we examined characteristics of work zones for the four project purpose categories:

To characterize work zone attributes by general category, the data was processed to identify values for project duration, length, and hours of activity per day, as shown in Tables 4(a)-(c). Observations based on the data are presented below.

Table 4(a). Duration (number of days) By Purpose Category
Type of Work Average Median Maximum
Bridge Work 130 79 1615
Roadway Construction/Reconstruction 163 90 2193
Resurfacing/Paving 104 65 607
Incidental Construction/Other 89 43 544
Table 4(b). Work Zone Length By Purpose Category
Type of Work Average Median Maximum
Bridge Work 3.1 1.5 10.0
Roadway Construction/Reconstruction 3.5 2.5 13.0
Resurfacing/Paving 13.9 10.0 104.0
Incidental Construction/Other 4.6 3.5 20.0
Table 4(c). Activity Per Day By Purpose Category
Type of Work Average Median Maximum
Bridge Work 11.2 10.0 24.0
Roadway Construction 12.3 10.8 24.0
Resurfacing/Paving 12.3 11.0 24.0
Incidental Construction/Other 10.2 10.0 24.0

Bridge Work can be characterized as having relatively short length and long duration. One of the factors leading to a high duration for bridge work is the long cure times for concrete. Roughly two-thirds of all bridge work activity takes place during daylight hours, generally matching the overall observed pattern of two-thirds daywork vs. one-third nightwork.

Roadway Construction/Reconstruction had the longest duration, most likely because this work is largely reconstruction of existing facilities and much time is expended trying to construct new lanes in the midst of existing traffic. The physical length of these projects is relatively short, which is not surprising since these projects are generally intensive in nature. The split of daywork (two-thirds) and nightwork (one-third) followed the average for all roadwork observed.

Resurfacing/Paving demonstrated attributes different from any of the other categories. Pavement operations showed the longest length and fairly short project duration. The majority of this work was likely maintenance work where crews are able to cover a lot of territory in a short period of time. This work can often be performed at night without significant loss in quality and at reduced exposure to traffic. The ratio of nightwork to daywork was reversed for paving operations compared to the other three categories: roughly two-thirds of all pavement operations took place at night.

Incidental Construction/Other had the shortest duration and was performed primarily as daywork. This category is difficult to characterize further as it represents a disparate mix of construction and maintenance activities.

4.7 Estimating Percent of National Highway System With Work Zones

Records pertaining to roadwork on interstates, US highways and state roadways represented 97% of the work zones posted to the state websites. A visual inspection of NHS highway maps and records from both Arizona and Washington revealed that, in addition to all interstate and US highway work zones, the majority of state roadways with work zones were also NHS roadways. The exclusion of local street records yielded 762 NHS work zones. From an analysis of work zone length, an average of 6.8 miles per work zone on interstates, US highways, and state roads was obtained. Multiplying the 762 NHS work zones by the average length of 6.8 miles yielded a total of 5,115 miles of NHS roads with a work zone in the 13 states. Based on NHS statistics (http://www.fhwa.dot.gov/hep10/nhs/index.html), we determined that the 13 sample states represented 24.5% of the 163,734 miles of roadway on the NHS. During the two-week study period in the sample, the 5,115 miles of NHS under construction scaled up to an estimate of 20,876 miles (or 12.8%) of the NHS under construction nationwide. Similarly, we estimated 3,110 work zones on the NHS during the study period.

4.8 Capacity Impacts Estimation

States provided data on likely capacity impacts for 71% of the work zones in the data set (557 sites). Of these, 79% indicated lane closures ranging from a single, hour-long lane closures to the complete closures of roadway facilities. The remainder indicated intermittent lane closures, lane narrowing, or shoulder work that would have some impact on roadway capacity.

Capacity loss from work zones was first calculated by examining the facility type. Interstate and US highways were assumed to have 3 lanes in each direction and have a carrying capacity per lane of 2000 vehicles per lane per hour. State roadways were assumed to have 2 lanes in each direction with per lane carrying capacities of 1500 vehicles per lane per hour. Based on text descriptions of each work zone, each record was coded as either shoulder or indirect work, the number of lanes closed, or full closure. A percentage reduction in roadway capacity was estimated in each case using the Highway Capacity Manual (2000). For example, shoulder work on an interstate was estimated to cause a drop in capacity of 10% (600 vehicles per hour), while a one-lane closure reduced capacity by 50% (3000 vehicles per hour). Using these estimates and the data available for each work zone, these impacts are estimated to generate an average loss of capacity by work zone of 2,565 vehicles per hour. Multiplying this figure by the average work zone duration of 11 hours per day generates a total loss in carrying capacity equal to 28,215 vehicles per day per work zone. Expanding to the work zones in the sample set (excluding local roads), this scales to over 15 million vehicles per day of lost capacity. Scaling to the national level (by the proportion of NHS in each state) we obtain a lost capacity of nearly 60 million vehicles per day from work zones on the NHS.

During the most frequent hour of activity (9AM to 10AM), we estimate that 2,084 work zones (67% of 3,110 work zones) have lane closures or roadside activity reducing capacity on the NHS by an average of 2,565 vehicles per hour. Collectively, this capacity loss is equivalent to 2,672 lane-miles of interstate roadway (rated at 2000 vehicles per hour per lane), roughly the same as one direction (three lanes) of a six-lane freeway connecting Washington, DC to St. Louis, MO.

While the data obtained can be combined with some rule-of-thumb assumptions to provide a rough estimate of capacity loss, a defensible quantitative estimate of road user delay or loss of productivity caused by these work zones cannot be made from the data. It is clear that a nighttime lane closure on a stretch of rural interstate reduces capacity, but it may have little or no impact on road user delay or productivity. On the other hand, failing to restore a single lane of urban freeway in a timely manner at the start of the afternoon peak travel period may cause extensive delays that persist well into the evening. The impact of work zones on delay and productivity are directly linked to detailed attributes such as temporal travel demand and availability of alternative routes, as well as precise estimates of capacity loss by time of day. These attributes are almost never reported in the state web resources identified in this study. As part of ongoing FHWA work zone efforts, researchers are currently seeking to identify other data sources (contracting records, automated traffic data repositories, etc.) that might be utilized to estimate delay, productivity, worker/road user exposure and other key measures.

Previous: 3. Data Collection  |   Next: 5. Implications and Conclusions

Go to Office Of Operations