Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
16 The research team conducted a number of activities aimed at (1) gathering and synthesizing information on existing and promising design treatments and (2) collecting data that could be used to evaluate the effectiveness of these treatments at reducing delay due to nonrecurrent congestion. These activities included the following: ⢠Obtaining of travel time reliability models from SHRP 2 Project L03 ⢠Assembly of traffic operational, crash, and weather data- bases from various sources ⢠Review of completed and ongoing research related to design treatments to address travel time reliability, delay, and non- recurrent congestion ⢠Initial contact, through e-mail and telephone, with highway agencies to obtain relevant information about design treat- ments in use, or considered for use, to reduce nonrecurrent congestion ⢠Focus groups with select highway agencies to gather details and insights about design treatments in use ⢠Workshops with highway agencies to gather details and insights about design treatments to address nonrecurrent congestion due to weather events ⢠Meetings with highway agencies to obtain detailed infor- mation about design treatments ⢠Development of a list of design treatments for evaluation This section of the report summarizes each of these activities. reliability Models from Shrp 2 project L03 SHRP 2 Project L03, Analytic Procedures for Determining the Impacts of Reliability Mitigation Strategies, developed models to predict several points along the annual travel time distribution of a highway segment for a given time-slice (1). (A time-slice can be a 1-h or multihour period, and Project L03âs models apply to nonholiday weekdays only.) The travel time distribution is essentially the fundamental descriptor of reliability, from which other reliability indicators of interest (e.g., buffer time, planning time) can be readily derived. For the purposes of both Project L03 and Project L07, travel time is most conveniently represented by the travel time index (TTI), which is defined as the ratio of actual travel time on a segment to the free-flow travel time. The Project L03 models quantify the effect of incidents and work zones on reliability by predicting several percentiles of the TTI distribution on the basis of three key variables: ⢠Lane hours lost due to incidents and work zones, which is calculated as the average number of lanes blocked per inci- dent (or work zone) multiplied by the average duration per incident (or work zone) multiplied by the total number of incidents (or work zones) during the time-slice and study period of interest. ⢠Critical demand-to-capacity ratio, which is defined as the ratio of demand to capacity during the most critical hour of the time-slice and study period. ⢠Hours of rainfall exceeding 0.05 in. during the time-slice and study period. Project L03 developed these relationships for various time- slices of the day over an extended study period. These time- slices included peak hour, peak period, midday (the 2-h period from 11:00 a.m. to 1:00 p.m.), and weekday (all 24 h). Because nonrecurrent congestion can occur at any time of the day, Project L07 needed a relationship that covered each of the 24 h of the day. The only Project L03 model that could quantify reliability for an hourly time-slice was the peak hour model, so this was the model used in Project L07 (at least to evaluate the effectiveness of design treatments during congested con- ditions). Reliability models more applicable to uncongested conditions were developed in Project L07, as discussed in Chapter 4 of this report. C h a p t e r 3 Data Collection and Documentation of Current Design Practice
17 The Project L03 relationships have the following general functional form, as shown in Equation 3.1: TTI 3.1% LHL+ dccrit 0.05en j k l Rn n n ( )= ( )+ â²â² where TTIn% = nth percentile TTI value; LHL = lane hours lost; dccrit = critical demand-to-capacity ratio; R0.05â³ = hours of rainfall exceeding 0.05 in.; and jn, kn, ln = coefficients for nth percentile (see Table 3.1). Table 3.1 shows the coefficients used to calculate each TTI percentile, as derived by Project L03 for peak hour data. The resulting TTI percentile values can be plotted as cumulative TTI curves, as illustrated in Figure 3.1, which shows 24 cumu- lative curves for each hour of the day from actual field data for a freeway in Minnesota. As an example of interpreting such curves, the darkened curve (representing the worst, i.e., most unreliable, hour of the day) has a 50th percentile TTI of 2.3, signifying that 50% of the vehicles that travel through this roadway segment during that hour spend more than 2.3 times the amount of time that it would take to traverse this segment under free-flow conditions. assembly of Databases Traffic operational, crash, and weather data were obtained for use in several analyses in the research, including the following: ⢠Analysis of traffic operational and crash data to determine a relationship between safety and congestion for use in evaluating design treatments ⢠Analysis of traffic operational and weather data to develop reliability models that accounted for both rain and snow conditions ⢠Analysis of traffic operational data to develop reliability models that were applicable to less congested conditions (i.e., d/c < 0.8) Each of these databases is described below. Traffic Operational Data Three years (2005 to 2007) of traffic operational data were obtained from the SHRP 2 Project L03 research team from freeways in two metropolitan areas: Seattle, Washington, and MinneapolisâSt. Paul, Minnesota. The sites in Seattle included two to four directional lanes of travel and represented 200 mi of directional freeway segments. The sites in Minneapolisâ St. Paul included two to five directional lanes of travel and represented 410 mi of directional freeway segments. Each station for which traffic volume and speed data were avail- able included detectors in each lane across one direction of travel on a freeway. The original detector data collected at each station on the freeways consisted of 5-min volume data per travel lane and 5-min average speed data per travel lane. A decision was reached to exclude from the study all data in the MinneapolisâSt. Paul metropolitan area after the I-35W bridge collapse on August 1, 2007. Although this period might have been interesting (because volumes changed dramatically on many freeway segments), the changed driving conditions were new to many drivers and the Minnesota Department of Transportation (DOT) made many modifications to spe- cific roadways to increase base capacity; thus, this time period would likely include unusual flow conditions. Crash Data Crash data for each directional freeway segment were obtained through the Highway Safety Information System. The crash Table 3.1. Coefficients Used in Project L03 Reliability Models: Peak Hour (2) n (percentile)a jn kn ln 10 0.07643 0.00405 0.00000 50 0.29097 0.01380 0.00000 80 0.52013 0.01544 0.00000 95 0.63071 0.01219 0.04744 99 1.13062 0.01242 0.00000 a The coefficients used to calculate the mean TTI are 0.27886 for jn, 0.01089 for kn, and 0.02935 for ln. Figure 3.1. Cumulative TTI distribution per hour of day.
18 data included all mainline freeway crashes that occurred within the limits of each roadway section of interest dur- ing the study period. Crash severity levels considered in the research were the following: ⢠Total crashes (i.e., all crash severity levels combined) ⢠Fatal-and-injury crashes ⢠Property-damage-only crashes Weather Data The research team obtained 10 years (2001 through 2010) of hourly precipitation data across the United States from the National Climactic Data Center (NCDC). From the NCDC website, the research team downloaded quality-controlled local climatological data databases for all stations within the United States. The databases used in the analysis were the âPrecip,â âHourly,â and âStationâ files. The Precip database contained a record of total precipitation at each station for each hour of the day. The hourly database included a variable (WeatherType) that indicated specific weather conditions and that was used to classify precipitation as either rain or snow. The Station file listed information about the stations that reported during a given month, including station number, station name, lati- tude, and longitude. A subset of 387 weather stations (those with a World Meteorological Organization designation) was selected for use in the research. These 387 stations are depicted in Figure 3.2. review of Completed and Ongoing research The types of design treatments applicable to nonrecurrent congestion, and the objectives of those design treatments, were identified through a review of completed and ongo- ing research, technical articles, vendor literature, conference proceedings, highway agency and technical association web- sites, internet search engine query results, and direct contacts with highway agencies. The research team looked for the following information: ⢠Design treatments being used or considered for use in reducing congestion ⢠The applicability of design treatments to nonrecurrent congestion ⢠Design guidelines or standards for treatments ⢠Implementation policies and practices for treatments ⢠Cost estimates of treatments ⢠Traffic operational effectiveness of treatments ⢠Safety effectiveness of treatments ⢠Other key information on the use of design features for nonrecurrent congestion The research team also documented international experience with design treatments to reduce nonrecurrent congestion. The overall results of the literature review can be summa- rized as follows: ⢠There is substantial information about the effects of design treatments on recurrent congestion. ⢠There is substantial information about the effects of intel- ligent transportation system strategies on nonrecurrent congestion. ⢠There is only limited information about the effects of design treatments on nonrecurrent congestion. This lack of information about the effects of design treatments on nonrecurrent congestion showed a clear need for research on this topic. Initial Contacts with highway agencies The research team contacted highway agencies to obtain rel- evant information about design treatments in use, or consid- ered for use, to reduce nonrecurrent congestion. The research team contacted 20 state highway agencies to obtain this infor- mation. Telephone interviews were then conducted with knowledgeable engineers in the most promising agencies. The highway agencies that were contacted included those in the following states: ⢠Arizona ⢠California ⢠Florida ⢠Georgia ⢠Illinois ⢠Indiana ⢠Maryland Source: © Microsoft Streets and Trips. Figure 3.2. U.S. weather stations with data available.
19 Each focus group meeting consisted of a 2-day visit. On the first day of the visit, the research team met with several highway agency staff experienced in geometric design, traf- fic operations, traffic management, or maintenance to review implemented or planned projects to reduce nonrecurrent congestion. Issues discussed and the types of questions asked included the following: ⢠What design treatments have been used in your region to reduce nonrecurrent congestion? ⢠Is nonrecurrent congestion considered and addressed in the design phase of new projects? If so, how? ⢠Who is involved in the decision making for implement- ing a nonrecurrent congestion mitigation strategy? What agencies and departments are involved? Whose responsi- bility is it? Who takes the lead? ⢠What treatments have been used to address recurrent con- gestion that could be considered for use in nonrecurring congestion situations? ⢠What treatments are considered promising but have not yet been tried in your region? ⢠How does your agency decide whether to implement a treatment in additional locations? For each treatment identified by the highway agency as hav- ing been implemented, the research team asked the following questions: ⢠What information is available about the traffic opera- tional effectiveness, safety effectiveness, and cost of the treatment? ⢠Massachusetts ⢠Michigan ⢠Minnesota ⢠Missouri ⢠New Jersey ⢠New York ⢠North Carolina ⢠Ohio ⢠Tennessee ⢠Texas ⢠Virginia ⢠Washington ⢠Wisconsin Telephone interviews were also conducted with the follow- ing agencies: ⢠Floridaâs Turnpike Enterprise ⢠New York State Thruway Authority ⢠Port Authority of New York and New Jersey During the telephone interviews, the research team discussed with each highway agency (1) design treatments being used or considered for use in reducing congestion, (2) the applicability of those treatments to nonrecurrent congestion, (3) whether the agency would be willing to participate in a focus group as part of the research, and (4) whether the agency had any suit- able projects or sites for evaluation in the research. Focus Groups with highway agencies The research team gathered details and insights about design treatments identified through initial contacts with highway agencies by conducting focus groups in the following four metropolitan areas with active congestion reduction programs: ⢠MinneapolisâSt. Paul: Minnesota DOT ⢠Atlanta: Georgia DOT ⢠BaltimoreâWashington, D.C.: Maryland DOT ⢠New York CityâNewark: Port Authority of New York and New Jersey The four metropolitan areas were selected for conducting focus groups because they represented diverse geographic regions of the country and because, based on the telephone interviews conducted by the research team, they were actively involved in using design treatments to address delay caused by nonrecurrent (as well as recurrent) congestion. Figure 3.3 shows the states that were contacted by e-mail or telephone, as well as the metropolitan areas where focus groups were conducted. Figure 3.3. Darkly-shaded states were contacted by e-mail or telephone; Minnesota DOT (MnDOT), the Port Authority of New York and New Jersey (PANYNJ), Maryland State Highway Authority (MD SHA), and Georgia DOT (GDOT) participated in focus groups.
20 knowledge in geometric design, traffic operations, and main- tenance. Participants included Minnesota DOT staff from the MinneapolisâSt. Paul metropolitan area, as well as from vari- ous rural districts. Design treatments related to weather events that were identified during the two workshops included the following: ⢠Snow fences ⢠Road weather-information systems ⢠Anti-icing systems ⢠Flood warning systems ⢠Fog detection systems ⢠Wind warning systems ⢠Contraflow for hurricane evacuation ⢠Road closure Meetings with highway agencies to Obtain Detailed Information about Design treatments As part of the focus groups (discussed above), the research team documented highway agency experience with existing treatments and gathered basic information about their design and application. However, to conduct traffic operational assessments of design treatments, more detailed treatment information was needed. Several additional visits to high- way agencies were made to gather more detailed information about treatments that had been implemented. In particular, the research team met with three highway agencies to obtain detailed information about many geometric design treat- ments, but with a particular emphasis on crash investigation sites, for which more information was clearly needed. These agencies are the following: ⢠Minnesota DOT. The research team met with geometric designers, traffic engineers, and maintenance personnel; the meeting included representatives from Minnesota DOTâs incident response program (FIRST) and the Min- nesota Highway Patrol. ⢠Illinois DOT. The research team met with traffic engineers responsible for traffic operations and incident manage- ment in the Chicago metropolitan area. ⢠Wisconsin DOT. The research team met with traffic engi- neers responsible for traffic operations and incident man- agement in the Milwaukee metropolitan area. Crash investigation sites varied greatly among the three agencies. For example, most of the sites that Minnesota DOT has constructed in recent years would more appropriately be considered emergency pulloffs, because they have been implemented where shoulders are no longer available due to ⢠Are there any design policies and guidelines for the treatment? ⢠What application criteria are used to determine when and where the treatment should be installed? ⢠What difficulties and challenges have been encountered in implementing the treatment? ⢠What are the perceived advantages and disadvantages of the treatment? ⢠What historical data are available concerning deployment of the treatment to address nonrecurrent congestion? ⢠Are any crash data available to compare sites with and without the treatment? On the second day of the visit, the research team made field visits to several implemented treatments in the area. Workshops with highway agencies to Discuss Weather treatments Because design treatments that address weather events were added to the scope of work several months after Project L07 began, they were not initially considered in the same depth as other design treatments. To provide further consideration of weather-related treatments, the research team held two work- shops with highway agencies to ensure that the list of design treatments related to weather events was complete and that the current state of knowledge concerning the effectiveness of such treatments was fully documented. The primary focus of these workshops was on design treatments related to winter weather events; however, other weather-related events were also considered. The first workshop was held in Kansas City, Missouri, and included weather experts from the following highway agen- cies and organizations: ⢠Florida DOT ⢠Missouri DOT ⢠City of Kansas City, Missouri ⢠City of Overland Park, Kansas ⢠City of West Des Moines, Iowa ⢠McHenry County, Illinois The local agencies participated in person; the others partic- ipated by phone. The workshop included a discussion of the types of weather events experienced by each participating agency that led to nonrecurrent congestion. The primary focus of the workshop was on design treatments that partic- ipating agencies have used to reduce the impact of weather events on congestion. The second workshop was held in Minnesota with key Minnesota DOT staff with expertise in winter weather events and design treatments to address those events, as well as
21 ⢠Ramp closure ⢠Ramp metering ⢠Reversible lanes ⢠Traffic signal improvements ⢠Ramp metering ⢠Variable speed limits List of Design treatments Using the results of the initial contacts and follow-up focus groups with highway agencies, the research team identi- fied a list of design treatments to be further assessed in the research. The factors that were used as the basis for deciding which design treatments should be considered for assessment included the following: ⢠Treatment is used (or can be used) for nonrecurrent congestion. ⢠Treatment supports one or more of the objectives. (An objective is how a treatment is used to reduce nonrecurrent congestion; e.g., it reduces the duration of the incident.) ⢠Operational effectiveness of the treatment is promising. ⢠Safety effectiveness of the treatment is promising. ⢠Cost of the treatment is low to moderate. ⢠Treatment has broad application potential. ⢠Treatment is a strong candidate for inclusion in the Project L07 Design Guide. A particular design treatment did not have to meet all of these criteria to be selected for further assessment in the research. Only the first two criteriaâthat the treatment addressed nonrecurrent congestion and that it supported one or more of the objectivesâwere mandatory. All of the design treatments selected met not only those two mandatory crite- ria, but several of the other criteria, as well. Table 3.2 presents the specific design treatments that were selected for assessment in the research. The design-related treatments are those that are highway design features or func- tion through changes in highway design features. Specifically, the nonrecurrent congestion design treatments are imple- mented through physical changes in the highway design that have a direct influence on traffic flow. For example, providing a paved shoulder or widening an existing shoulder is a nonre- current congestion design treatment. Some treatments, like shoulders, affect traffic flow at all times; others, like portable incident screens, affect traffic flow only at times when the treatment is deployed or in operation. The secondary treat- ments are not intrinsically highway design features, but they have secondary implications related to highway design. For example, ramp metering is a traffic control strategy rather than a highway design feature; however, the implementation of ramp metering has implications for the design of ramps the shoulders being converted to travel lanes. The emergency pulloffs serve several purposes, including crash investiga- tion and enforcement, but they were primarily constructed to accommodate inoperable vehicles and other emergencies that would otherwise be accommodated by a shoulder. Illi- nois DOTâs crash investigation sites range in size and design and have been installed in a variety of locations, including along the right side of the freeway (beyond the shoulder), inside the median, on ramps, and underneath overpasses. The research team gathered as much detailed information as possible about the use of crash investigation sites so that the traffic operational effectiveness of such sites could be esti- mated most accurately. Key information obtained during the highway agency visits included the following: ⢠Typical number of lanes blocked during an incident ⢠Policy about moving crashes to an emergency pulloff or crash investigation site ⢠Types of crashes moved ⢠Percentage of crashes moved ⢠Average time between when a crash occurs and when it gets moved ⢠Average reduction in lane hours lost when a crash is moved ⢠Typical dimensions and design of an emergency pulloff or crash investigation site ⢠Typical signing at an emergency pulloff or crash investiga- tion site ⢠Cost issues with constructing an emergency pulloff or crash investigation site The research team used this information to develop input variables for the models developed in Project L03, and any related simulation models that were developed, to estimate the impact of crash investigation sites on nonrecurrent congestion. During the highway agency visits, detailed information was obtained for other treatments, as well, including the following: ⢠Alternating shoulders ⢠Bus pulloffs ⢠Bus shoulders ⢠Designated bus lanes ⢠Dynamic shoulders ⢠Emergency pulloffs ⢠Emergency traffic operations plan ⢠Extra-height median barriers ⢠Flood control systems ⢠Geometric improvements to alternate routes ⢠Glare screens ⢠Median crossovers ⢠Movable barriers
22 Table 3.2. Candidate Design Treatments Considered in the Research Category Nonrecurrent Congestion Design Treatment Secondary Treatment Medians Median crossovers â Movable traffic barriers Gated median barrier Extra-height median barriers Mountable or traversable medians Shoulders Accessible shoulder â Drivable shoulder Alternating shoulder Portable incident screens Vehicle turnouts Bus turnouts Crash investigation sites Crash investigation sites â Right-of-way edge Emergency access between interchanges â Arterials and ramps Ramp widening â Ramp closure Ramp terminal traffic control Ramp turn restrictions Detours Improvements to detour routes â Truck incident design considerations Runaway truck ramp â Construction Reduce construction duration Improved work site access and circulation Animalâvehicle collision design considerations Wildlife fencing, overpasses, and underpasses â Weather Snow fences Fog detection Blowing sand treatment Roadway weather-information system Anti-icing systems Flood warning system Wind warning system Lane types and uses â Contraflow lanes for emergency evacuation Contraflow lanes for work zones HOV lanes and HOT lanes Dual facilities Reversible lanes Work-zone express lanes (continued on next page)
23 storage area for vehicle breakdowns, a safe stopping place for service assistance patrols, increased flexibility for work-zone operations, as well as the potential for use when needed as an additional through lane. Second, some design treatments provided primarily to reduce recurrent congestion have the potential for use to address nonrecurrent congestion, but require an explicit decision by the highway agency or traffic management center. Such treatments were also considered within the scope of the research. Examples of such design treatments include reversible lanes and HOV lanes, which could be used to accommodate special event traffic or to route traffic around a major incident. Technology-related treatments, such as changeable mes- sage signs and demand detection systems (loop detectors or video detection), have a key role in reducing nonrecur- rent congestion, but were not considered design treatments for purposes of this research. However, technology-related treatments were considered within the scope of the research to the extent that they support a design treatment. For exam- ple, changeable message signs may be used to communicate detours, ramp closures, or ramp turning restrictions to drivers, or in conjunction with using reversible lanes for special events. Demand detection systems may be used to detect an incident so that an appropriate treatment may be implemented more quickly. The design treatments in Table 3.2 served as the basis for the traffic operational, safety, and life-cycle benefitâcost analyses presented in the following three chapters. that highway agencies must understand in order to use this treatment effectively. In developing the list of design treatments in Table 3.2, careful consideration was given to whether to include design treatments whose sole or primary function is to increase the base capacity of the roadway. Such treatments primarily address recurrent congestion, which was outside the scope of the Project L07 research. Design treatments for interchanges such as ramp braiding, adding collectorâdistributor roads, and adding auxiliary lanes were not included in Table 3.2 because they reduce congestion solely or primarily by increasing base capacity. Although any design treatment that increases the base capacity of the roadway will reduce both recurrent and nonrecurrent congestion, the primary purpose of such treatments is to reduce recurrent congestion. Furthermore, once implemented to reduce recurrent conges- tion, such treatments are available at any time to reduce non- recurrent congestion, with no intervention required by the highway agency or traffic management center when incidents, demand fluctuations, or special events occur. Two exceptions were made to the general principle that design treatments that increase base capacity should be excluded from the research scope. First, if a design treat- ment functions to relieve nonrecurrent congestion not only by increasing base capacity, but also in other ways, it was considered within the research scope. For example, add- ing shoulders or widening existing shoulders on a roadway increases the base capacity of the roadway, but also provides a Table 3.2. Candidate Design Treatments Considered in the Research Category Nonrecurrent Congestion Design Treatment Secondary Treatment Traffic signals and traffic control â Traffic signal preemption Queue jump and bypass lanes Traffic signal improvements Signal timing systems Ramp metering and flow signals Temporary traffic signals Variable speed limits and speed limit reduction Technology â Electronic toll collection Overheight vehicle detection and warning systems Emergency response notification â Reference location signs Roadside call boxes Note: HOV = high-occupancy vehicle; HOT = high-occupancy toll; â = not applicable. (continued)
