Incorporating Travel Time Reliability into the Highway Capacity Manual (2014)

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.

107 C h a p t e r 7 A corridor study, by definition, goes beyond the single-facility focus of a typical HCM2010 facility analysis. The purpose of a corridor study is to assess the ability of a subsystem of inter- related facilities to achieve a set of transportation perfor- mance objectives. The evaluation of travel time reliability at the corridor level introduces a new set of considerations that do not come into play when evaluating the reliability for indi- vidual facilities. These include • The greater geographic coverage required to evaluate reli- ability at a corridor level, which imposes greater input data and data analysis requirements than for single-facility analyses; • The interpretation of reliability results for different facility types within the same corridor; and • The effect of varying congestion on facility demands (i.e., route shifting between facilities). The latter consideration is the most likely reason for evaluat- ing a corridor as a whole, rather than evaluating the individ- ual facilities separately: • Do some, all, or none of the facilities in the corridor pro- vide at least minimum desired levels of reliability? • Is localized congestion on the arterial that creates traffic diversion to the freeway a possible source of reliability problems on the freeway (or vice versa)? • How well do the corridor’s urban street components accommodate traffic diversion generated by incidents on the freeway? This chapter provides guidance on extending the individ- ual freeway facility and urban street facility reliability meth- ods into an overall assessment of corridor reliability. Corridor Definition The HCM2010 defines a corridor as follows: Corridors are generally a set of parallel transportation facilities designed to move people between two locations. For example, a corridor may consist of a freeway facility and one or more parallel urban street facilities. There may also be rail or bus transit service on the freeway, the urban streets, or both, and transit service could be provided within a separate, parallel right-of-way. Pedestrian or bicycle facilities may also be present within the corridor, as designated portions of roadways and as exclusive, parallel facilities (TRB 2010a, 2-5). For the purposes of a reliability analysis, a corridor is defined as a freeway facility and one or more parallel urban street facili- ties. When traffic diversion occurs between facilities that make up a corridor, the freeways, highways, and urban streets that cross the corridor and provide connections between the cor- ridor’s facilities are also affected; however, those effects are beyond the scope of a reliability analysis. The focus of a corri- dor evaluation is on the parallel facilities. Methodological Considerations The freeway facility and urban street facility reliability mod- els use different methods to develop a travel time distribution for the reliability reporting period: • The freeway facility method develops scenarios based on their probability of occurrence during the reliability report- ing period. Some highly unlikely scenarios may be dropped from the analysis. • The urban streets method randomly assigns demand, weather, and incident conditions to each day, based on distri- butions of conditions likely to occur within a month. Some Corridor Applications

108 highly unlikely combinations may be included by random chance; therefore, multiple runs of the method may be needed to establish a representative travel time distribution. Importantly, the two methods have no direct link. The weather pattern generated by the urban streets method may produce more or less severe conditions over a given model run, compared with the 10-year average weather conditions used by the freeway method. An incident scenario for the free- way does not generate a corresponding high-demand scenario for urban streets. When local data are used to generate demand patterns, traffic diversion effects appear in individual days’ demands used to create month-of-day factors, but the effects of days with diversion are likely washed out by demands from all of the days without diversion. When default demand pat- tern data are used, no diversion effect occurs at all beyond that resulting from bad weather (and associated higher incident rates) which occurs more often in some months of the year than in others. Finally, the HCM2010 itself is incapable of pre- dicting where and how much traffic diversion or change in demand (i.e., trips postponed or not made) will occur in response to incidents or severe weather events. Given these differences, one could easily conclude that the freeway and urban streets reliability methods cannot be com- bined to describe corridor reliability. That would be incor- rect. However, these limitations do have to be considered as one asks questions about corridor reliability and designs an analysis to answer those questions. potential applications of Corridor reliability analysis Evaluating Overall Corridor Reliability An analysis of overall corridor reliability involves comparing selected reliability performance measures (e.g., TTI, PTI, per- cent on-time arrivals) generated for the individual facilities against either an established standard or comparative national values of reliability. Because different agencies may be respon- sible for different facilities within a corridor (or, in the case of urban streets, different portions of the facility), and because corridor analysis focuses on longer-distance travel, a regional standard may be most appropriate. In the absence of such a standard, a percentile threshold can be used so that unaccept- able performance is defined in terms of, for example, a facil- ity’s PTI being among the worst 20% of U.S. facilities. Once acceptable or unacceptable performance has been determined for each facility forming the corridor, the results can be interpreted as follows: • All facilities have acceptable performance: The corridor offers acceptable reliability and provides travelers with multiple options for reliable travel. • Some facilities have unacceptable performance: The cor- ridor offers acceptable reliability and provides at least one option for reliable travel, although it may not be the fastest option on a typical day. These corridors may be candidates for variable message signing that provides travel times via both (or all) facilities to encourage the use of alternative facilities when one facility’s travel time is considerably higher than the average. • All facilities have unacceptable performance: The corridor’s reliability is unacceptable—travelers cannot rely on any facility within the corridor to provide consistently reliable travel times. Overall corridor reliability can be evaluated using actual travel time distributions to produce the reliability perfor- mance measures (thus accounting for traffic diversion effects directly) or by using this chapter’s methods to generate travel time distributions (thus accounting for any traffic diversion effects manually, as described in the following text). Prioritizing Corridor Management Strategies This application focuses on measuring or estimating TTI, PTI, or both. These performance measures are compared with reli- ability performance measures to determine the percentile rank each facility in the corridor falls into (e.g., 25% worst freeway, 45% worst arterial). The rankings become inputs to a process for setting priorities (one that potentially incorporates other factors, such as vehicle miles traveled, or VMT). Worse- performing corridors are given higher priority to receive oper- ational or physical treatments designed to improve travel time reliability. Diagnosing Potential Recurring Diversion-Related Causes of Unreliability In this application, each facility forming the corridor is divided into sections defined by the locations of connections between facilities (i.e., locations where a freeway and an urban street facility cross, or where another facility provides a connection between a freeway and an urban street). For freeways, sections end immediately before the basic freeway segment in the middle of the interchange. For urban streets, sections end at the downstream end of the segment that includes either (1) the cross-street intersection or (2) the intersection serving the freeway off-ramp in the analysis direction, in cases where the freeway and parallel urban street cross each other, as shown in Figure 7.1. Next, the TTI is determined for each facility on a section- by-section basis. In some cases, on parallel sections of freeway and urban street, the TTI drops significantly from one section to the next on both facilities. Those sections are candidates for investigation to see whether operational problems on one

109 facility cause regular traffic diversion to the other facility, resulting in problems on the other facility. Treating the source cause of unreliability can result in reliability improvements on both facilities. Identifying the Potential for Nonrecurring Traffic Diversion Parallel arterials are potential alternate routes for freeway traffic when incidents occur on the freeway. However, motor- ists may know they are better off sitting in traffic on the free- way because taking the urban street on an average day is slower than the freeway on its 5% worst day. If motorists know this, then little diversion is likely to occur—even if they are encouraged through information strategies such as vari- able message signs or highway advisory radio. In this application, the facilities forming the corridor are divided into sections as before. This time, a planning travel time to the corridor end is determined for each freeway sec- tion based on the 95th percentile time to travel from the start of the freeway section to the end of the corridor. An average travel time to the corridor end is determined for each urban street section based on the 50th percentile travel time from section start to corridor end. Or, if data are available, the aver- age time to travel from the freeway to the urban street along the roadway connecting the two facilities can be calculated and incorporated into the urban street section’s time to cor- ridor end. Sections in which the arterial street travel time to corridor end is greater than the freeway planning time to cor- ridor end would not be candidates for operations measures that encourage traffic diversion. Modeling Freeway Traffic Diversion Effects on Urban Streets As noted previously, the freeway and urban streets methods are not linked: an incident in a freeway scenario does not gen- erate a corresponding demand increase in an urban street scenario. However, for the purposes of modeling corridor operations, freeway incidents that would cause diversion can be modeled as special events in the urban streets model. This allows the urban streets reliability results to reflect demand increases resulting from freeway incidents. Two key aspects of this method are notable. First, the analyst is responsible for estimating the amount of diversion that would occur—HCM2010 methods do not estimate demand or demand diversion. Second, the process has no feedback loop that allows the diverted demand to be subtracted from the free- way demand so that freeway reliability can be reestimated. The following steps describe a conceptual process for determining freeway traffic diversion effects on urban streets reliability. Step 1: Identify Number of Freeway Incidents During Reliability Reporting Period The scenario generation step of the freeway modeling process is performed as described previously in this report. The pro- cess determines probabilities of an incident occurrence by month and incident severity and specifies the average dura- tion. The average number of freeway incidents of severity can be determined from Equation 7.1: ( ) = × (7.1) RRP I h P I d s s s where Is = number of freeway incidents of severity s generated during the reliability reporting period; hRRP = number of hours in the reliability reporting period; P(Is) = probability of an incident of severity s; and ds = average duration of an incident of severity s. Because a large number of shoulder and one-lane-closed incidents can be generated, and because more-severe inci- dents are most likely to generate traffic diversion, analysts should focus on incidents affecting two or more lanes. Figure 7.1. Illustration of a corridor.

110 Step 2: Apply the Freeway Reliability Method Next, the freeway facility reliability method is applied. The scenario characteristics for scenarios containing incidents of interest to the analysis (e.g., two lanes closed and more severe) and the computational engine output for those scenarios should be saved for use in subsequent steps. Step 3: Generate Incident Data IncIdent day and tIme Because the freeway model develops probabilities of incident occurrence, while the urban streets model generates condi- tions day-by-day during the reliability reporting period, each freeway incident needs to be assigned to a specific day and time within the reliability reporting period. This can be done randomly. As the freeway model defines two possible incident start times—at the start of the study period or in the middle of the study period—one of these times should be assigned randomly. IncIdent LocatIon The freeway model assigns locations as the first, middle, or last segment of the freeway facility, with equal probability. The choice of freeway segment should be selected randomly. All diverted freeway demand is assumed to leave the freeway at the last cross-connection point (i.e., section boundary) before the incident and to return at the first opportunity (i.e., at the start of the following section). The corresponding arte- rial section is the one affected by the diverting traffic. correspondIng Freeway scenarIos The analyst needs to locate the input data for the scenario corresponding to the incident month, nonsevere weather, incident location, incident duration, incident start time, and average incident duration. Freeway demand and capacity data for the time the incident is in effect are required in the next step. dIversIon duratIon If the incident results in freeway traffic demand exceeding the remaining capacity, diversion occurs. Traffic starts to divert in the first analysis period after the incident occurs. amount oF dIverted traFFIc The analyst needs to specify the amount of traffic by analysis period that diverts as a result of a given incident. This amount is likely proportional to the amount to which freeway demand exceeds capacity. It also includes assumptions about the meth- ods used to communicate information about the incident to motorists. Step 4: Code Alternative Data Sets An alternative data set needs to be developed for each incident. The data set should specify changes in turning demand at the intersections that form urban street section boundaries, along with changes in through demand for segments and inter- sections within the urban street section. If special signal timing plans are used on the urban street during diversion situations, these should also be entered into the alternative data set. Step 5: Apply the Urban Streets Reliability Method Next, the urban streets reliability method is applied. Each alternative data set is supplied to the procedure as a special event, with its schedule corresponding to the selected day, start time, and duration of each diversion event. The statistics produced by the urban streets method now include the effects of diverted freeway traffic in addition to all the other sources of variability normally accounted for by the method.