Highway Capacity Manual Methodologies for Corridors Involving Freeways and Surface Streets (2020)

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.

- 27 - 6. Queue Spillback into Urban Streets Queue spillback into an urban street occurs due to insufficient merging capacity in a freeway on-ramp, and may have the following causes: • Insufficient capacity at the freeway merge segment, per guidance provided on HCM Chapter 14 – Merge and Diverge Segments; • Insufficient capacity of the on-ramp (HCM Exhibit 14-12); • Active ramp metering at the on-ramp. If insufficient capacity at the on-ramp occurs and its queue storage is also insufficient, queue spillback is expected to occur at the upstream intersection. The project has developed procedures for evaluating spillback into signalized intersections as well as unsignalized intersections (TWSC, AWSC and roundabouts). This section presents the data collection process for queue spillback into urban streets. It discusses field data and simulation, qualitative analysis, and the developed methodology to evaluate the occurrence and impacts of queue spillback from an on-ramp. Data Collection The data collection process for queue spillback into urban streets is very similar to the method used for off-ramp queue spillback. This section describes the details of data collection for spillback into urban streets. Data Requirements The data collection needs for studying spillback into urban streets are shown in Figure 25. The figure refers specifically to a signalized intersection as this type of facility has the most complex data collection requirements. Unsignalized intersections have a similar set of data collection requirements, excluding the signal controller configuration. Figure 25 – Data collection framework for queue spillback into urban streets (signalized intersection) The data collection specifics for this portion of the project are as follows:

- 28 - 1. Mainline detectors: Data were obtained from loop detectors located upstream and downstream of the merge area, providing lane-by-lane speed/volume/occupancy measurements, with resolution of 1 minute or less. 2. Queue/ramp detectors: Data from detectors along the ramp were obtained to provide the queue length and available queue storage along the ramp. If the ramp is metered, the metering rates were also provided. 3. Signal controller configuration: For signalized intersections, the following signal timing parameters were collected: pre-timed/actuated, phasing, cycle length, green/yellow/all red intervals, and actuation parameters (min green, max green, passage time, etc.) 4. Video cameras: It was essential to have video at the interchange in order to view spillback occurrence and the available queue storage along the ramp together with the approaching movements’ discharge rates. 5. Automated traffic volume counts: If available, traffic volume counts were obtained to save time and effort in data reduction. Agencies Contacted The same agencies that were contacted for off-ramp queue spillback data (Figure 5) were also contacted for potential locations to observe on-ramp queue spillback. The data collection for urban street queue spillback was significantly more challenging when compared to off-ramp queue spillback, for the following reasons: • A very limited number of DOTs have access to cameras located in urban streets. In most cases, these cameras are operated by municipalities or counties, which frequently lacked the resources to capture and record videos for our research purposes; • Identifying locations with queue spillback into unsignalized intersections was extremely difficult. The research team was able to identify a few locations where roundabout ramp terminals experience on-ramp queue spillback, but in all cases these locations did not have cameras able to capture video. For this reason, unsignalized intersections were analyzed using microsimulation. Dataset Description Table 4 summarizes the list of the four study locations where field data were collected, with a description of their key characteristics and number of video observations. Each video observation typically lasts between 1-3 hours, and it corresponds to one peak period recording, where the development and discharge of queues can be observed.

- 29 - Table 4 – Summary of study locations – field data collection – on-ramp queue spillback Location Interchange type Intersection type Number of on-ramp lanes # video observations Atlanta/GA - I-285W NB @ South Cobb Dr NB Diamond Signalized 2 9 Atlanta/GA - I-285W SB to Cobb Pkwy Dr. Diamond Signalized 1 10 Atlanta/GA - I-285W NB to Cobb Pkwy Dr. Diamond Signalized 1 10 San Diego/CA - Friars Rd. @ I- 15 NB Parclo Signalized 2 10 Schematics and sample screenshots of the study locations are presented next. Atlanta, GA - I-285W NB@ South Cobb Dr. This diamond interchange connects the I-285W NB freeway with the South Cobb Dr. arterial corridor. During the AM peak period, the freeway facility experiences heavy congestion, while ramp metering at the on-ramp limits the number of vehicles merging into the freeway. This causes queues to develop along the on-ramp, which has insufficient storage capacity and causes spillback into the upstream signalized intersection. Two movements at the intersections are affected by the queue spillback. First, the SB-L movement Figure 26a) discharges directly into the on-ramp, which reaches its capacity before the end of green. When this occurs, SB-L vehicles are held behind the stop bar as they are unable to proceed to the on-ramp. The second affected movement is the NB-Th/R (Figure 26b), which receives green immediately after the SB-L. The on-ramp is completely occupied by the queue from the previous movement, restricting the capacity of the NB-R movement. The NB-Th movement is indirectly affected, as the last vehicles discharging from the SB-L movement get trapped inside the box and partially block the movement of NB-Th vehicles at the beginning of their green.

- 30 - Source: Photo provided by the Georgia Department of Transportation and Google Maps Figure 26 – I-285 NB @ South Cobb Dr. and affected movements by on-ramp queue spillback: (a) NB-Th/R and (b) SB-L Atlanta, GA - I-285W SB @ Cobb Pkwy Dr. This diamond interchange connects the I-285W SB freeway with the Cobb Parkway Dr. arterial corridor. During the PM peak period, the freeway facility experiences heavy congestion, and the number of vehicles merging into the freeway is constrained by the capacity of the merge. The spillback pattern observed at this intersection is similar to the previously described location (I-285W NB@ South Cobb Dr). The heavy demand of the left-turn movement from the major street causes queue spillback at the on-ramp. Consequently, the opposing through movement is also impacted by the residual left-turn vehicles which block the intersection.

- 31 - Source: Photos provided by the Georgia Department of Transportation and Google Maps Figure 27 – I-285W SB @ Cobb Pwky Dr. and affected movements by on-ramp queue spillback: (a) NB-L and (b) SB-Th/R Atlanta, GA - I-285W NB @Cobb Pkwy Dr. This intersection is also part of the diamond interchange connecting the I-285W freeway with the Cobb Parkway Dr. arterial corridor, previously described. The northbound direction of the freeway also experiences heavy congestion, causing queue spillback into the signalized intersection. The intersection does not have a high demand for the SB-L movement, therefore it does not cause queue spillback. However, the NB-R movement’s demand is often higher than the on-ramp capacity and causes spillback (Figure 28). Source: Photos provided by the Georgia Department of Transportation and Google Maps Figure 28 – I-285W NB @Cobb Pkwy Dr. with NB-R movement affected by on-ramp queue spillback

- 32 - San Diego, CA - Friars Rd. @ I-15 NB This partial cloverleaf interchange connects the I-15 NB freeway facility to the Friars Rd. arterial corridor. The freeway experiences heavy congestion during the AM peak period. Active ramp metering also limits the on-ramp throughput into the freeway. As a consequence, queue spillback occurs frequently into this intersection. The only movement affected at this intersection is the WB-R (Figure 29). This is a protected-only right- turn movement with a high demand. At the onset of green, the on-ramp has considerable storage space, but it is occupied quickly as the WBR discharges into the on-ramp. After several cycles of high demand, vehicles are unable to discharge into the on-ramp during the green interval, causing queues to grow further upstream. Source: Photos provided by the City of San Diego and Google Maps Figure 29 – Friars Rd. @ I-15 NB with WB-R movement affected by on-ramp queue spillback Data Reduction Similar to the data collection for off-ramps, the process of obtaining video observations was very challenging. After the recordings were obtained from the respective agencies, the research team performed a first screening of the videos for an initial evaluation, which considered the following: • Camera shows the approaches at the intersection and at least part of the on-ramp;

- 33 - • No occurrence of incidents, disabled vehicles, lane closures or other factors that may affect the natural dynamics of vehicular traffic; During the data reduction, the following were measured: • Displayed green for each movement during each cycle; • For movements discharging into the on-ramp: the observer noted whether the subject movement experienced blockage during the given cycle. (Blockage was defined as the occurrence of queue spillback during the green, negatively impacting its discharge rate.) • For movements impacted by blockage of the intersection: the observer noted whether the through movement was impacted by trapped vehicles blocking the intersection (similar to Figure 27b), for any given cycle. • For every movement, the headways of vehicles crossing the stop bar were recorded by lane. This measurement procedure followed the guidelines provided by HCM Chapter 31 (Signalized Intersections Supplemental) for Field Measurement of Saturation Flor Rate (Section 6). Qualitative Observations The videos obtained were used to observe operations during queue spillback, in order to inform the development of analytical models. These observations are discussed in the following paragraphs. Partial or total blockage of lanes for through movement When queue spillback from an on-ramp reaches the upstream intersection, the movements that discharge into the on-ramp are directly affected. However, field observations show that other movements that do not discharge into the on-ramp can be affected by vehicles trapped inside the intersection box. The most common case occurs at intersections with leading left turn operation, as illustrated in the example of Figure 30. The SB-L movement discharges into the on-ramp that has insufficient storage to accommodate the demand. At the end of the SB-L movement, some vehicles are trapped inside the intersection box. Therefore, at the onset of green for the NB-Th movement, only the leftmost lane (L1) is clear to discharge at the saturation flow rate. The other lanes (L2, L3 and L4) cannot discharge immediately at the start of green and have to wait until the on-ramp queue clears, resulting in additional lost time for these lanes. Figure 30 – Partial blockage of northbound through movement by vehicles trapped inside the intersection box

- 34 - Simulated Sites The use of microsimulation is especially important for on-ramp queue spillback, due to the difficulties collecting data for unsignalized intersections. The same interchanges used for simulation of off-ramp queue spillback were used to simulate on-ramp queue spillback. These models were calibrated with a signalized intersection operation, and these were replaced with stop-controlled and roundabout intersections, as illustrated in Table 5. The intersection volumes were also changed to ensure on-ramp queue spillback occurs. Table 5 – Summary of simulated locations – on-ramp queue spillback Location Intersection type I-105 NB to Bellflower Blvd. Signalized Stop-controlled Roundabout I-105 SB to Bellflower Blvd. Signalized I-605 SB to Firestone Blvd Signalized Stop-controlled Roundabout I-605 NB to Firestone Blvd Signalized Queue Spillback Check Procedure The first step of the methodology evaluates whether queue spillback is expected to occur. The detailed methodology for on-ramp spillback check is described in Appendix D - On-ramp Queue Spillback Check. The proposed approach evaluates two potential capacity bottlenecks at the on-ramp (Figure 31): • Merge capacity at the freeway; • Active ramp metering, if present.