**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

**Suggested Citation:**"Appendix - Applications Guidance for Selected Quantitative Analysis Methods." National Academies of Sciences, Engineering, and Medicine. 2018.

*Guide for the Analysis of Multimodal Corridor Access Management*. Washington, DC: The National Academies Press. doi: 10.17226/25342.

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.

85 Introduction This appendix provides guidance on applying the quantitative analysis methods that appear most often in the guide. These methods can estimate performance measures that describe how access management techniques interact with the operations or safety performance of particular travel modes. The interactions of access management techniques with travel modes can be described in two ways. One way is by presenting absolute values of a performance measure (e.g., average motor vehicle speeds would increase from 26.7 to 28.2 mph following installation of a non- traversable median). Another way is by presenting the relative change in a performance measure (e.g., average motor vehicle speeds would increase by 1.5 mph following installation of a non- traversable median). Calculating relative changes is often easier than calculating absolute values. This appendix focuses on demonstrating calculations that can be performed with nothing more than a scientific calculator. However, the appendix also provides guidance on where to turn for more information for performing complex calculations that require developing spreadsheets or applying specialized software. Methods Described in this Appendix The methods described in this appendix consist of the following from the Highway Capacity Manual: A Guide for Multimodal Mobility Analysis, 6th ed. (HCM6): â¢ HCM6 intersection delay methods. These methods estimate the average motor vehicle control delay for a turning movement, approach, and/or intersection as a whole for signalized inter- sections, unsignalized intersections (i.e., minor street stop controlled), roundabouts, inter- change ramp terminals, and alternative intersection and interchange forms. They can be used to evaluate the effects of access management techniques that alter turning movement patterns, traffic volumes, or both. â¢ HCM6 arterial speed estimation methods. These methods estimate the free-flow and average midblock running speeds of motor vehicle traffic along roadway links between signalized intersections or roundabouts. They also estimate the average travel speed of motor vehicles along longer sections of roadway, including the delays that occur at intersections. These methods can be used to evaluate the effects of access management techniques that alter the roadway geometry, number of access points, and/or on-street parking provisions. â¢ HCM6 queue estimation methods. These methods provide the 95th percentile back of queue and, sometimes, other percentile queues. They can be used to determine the size of an inter- sectionâs influence area and to size turn lane lengths and driveway throat lengths. A P P E N D I X Applications Guidance for Selected Quantitative Analysis Methods

86 Guide for the Analysis of Multimodal Corridor Access Management â¢ HCM6 multimodal level of service (MMLOS) methods. These methods calculate pedestrian, bicycle, and transit level of service scores that estimate traveler satisfaction with quality of travel by these modes along a roadway section. In many cases, the effect of an access man- agement technique on other modes is indirect, resulting from changes in average midblock motor vehicle speeds caused by the technique. However, in some cases, a particular access management technique may directly affect a non-auto modeâs level of service score (e.g., the reduction or elimination of on-street parking). â¢ HCM6 pedestrian and bicycle delay methods. These methods calculate average pedestrian and bicycle delay at signalized intersections and roundabouts, along with average pedestrian delay crossing roadways at unsignalized locations. The methods can be used to evaluate the effects of access management techniques that change traffic volumes, change street widths, add pedestrian refuges, or are a combination of these. â¢ Truck level of service. This method evaluates the effect of changes in average truck speeds on overall truck level of service. â¢ Crash modification factors. These factors estimate the change in crash rate that would occur as a result of implementing a particular access management technique. CMFs are straight- forward to apply; this appendix provides guidance on selecting appropriate CMFs that may be developed following the publication of this guide. â¢ Vehicle crash models. These models estimate the crash rate or total number of crashes that would occur given a particular set of conditions. They are applicable to a small number of access management techniques that affect factors included in one of these models. Individual guide sections may present more quantitative techniques than are discussed in the appendix. Those techniques that are not discussed are considered straightforward to apply and do not require additional explanation. In addition, other analysis tools (e.g., simulation) exist that can also be used to estimate motor vehicle operations and that may be equally or more appropriate to use in a given circumstance. However, describing how to use these alternative tools is beyond the scope of this guide. HCM6 techniques are described because they are well researched, well documented, and widely used. Table A1 lists the analysis methods described in this appendix. For each group of access management techniques described in this guide, the table indicates whether an analysis method (a) can be used to calculate relative changes in operations or safety without calculating absolute values first, (b) can only calculate relative changes when an absolute value is calculated first, or (c) is not applicable to any technique in the group. A given method will typically be applicable to only one or a few of the access management techniques included in a given group. HCM6 Intersection Delay Methods Chapters 19 through 23 in the HCM6 (1) provide methods for assessing motor vehicle operations at signalized intersections (Chapter 19), two-way stop-controlled intersections (Chapter 20), all-way, stop-controlled intersections (Chapter 21), roundabouts (Chapter 22), and interchange ramp terminals and alternative intersections (Chapter 23). The effect of access management techniques that alter turning-movement patterns, traffic volumes, or both at individual access points or intersections can be evaluated using these chaptersâ methods by calculating the resulting change in average control delay and/or motor vehicle level of service. Once average delay to the major street through movement is determined at each intersection of interest, an average travel speed along the major street can also be determined, as described in the next section. With the exception of the delay method for roundabouts, which can be performed by hand or automated in a simple spreadsheet, the intersection delay methods described in the HCM6 are only

Applications Guidance for Selected Quantitative Analysis Methods 87 practical to apply by using specialized software, given the number of computations involved. A number of commercial software packages are available that perform these calculations for some or all of these intersection types. In addition, STREETVAL, a research-grade computational engine that can estimate delay at signalized intersections, is available on the HCM6 Volume 4 website (www.HCM6volume4.org, free to access but requires a one-time registration). The input data required to apply these methods are listed in Exhibits 19-11 and 19-12 in the HCM6 (signalized intersections), Exhibit 20-15 (two-way stop-controlled intersections), Exhibit 21-9 (all-way stop-controlled intersections), Exhibit 22-9 (roundabouts), and Exhibit 23-21 (signalized ramp terminals). In addition, the âRequired Data and Sourcesâ section starting on Page 23-71 describes input data requirements for certain alternative intersection forms that go beyond the data required for signalized intersections or two-way stop-controlled intersections. At a minimum, the HCM6 delay methods require the analyst to supply traffic volumes and lane configurations by movement, plus left-turn phasing for signalized intersections and special geometric conditions (e.g., presence of two-way left-turn lanes or median storage) for two-way Access Management Technique Group HCM6 Vehicle Delay HCM6 Arterial Speed HCM6 Queues HCM6 MMLOS HCM6 Pedestrian and Bike Delay Truck LOS CMFs Bowman Crash Rate Models Potts Pedestrian Crash Model Carter Intersection Safety Restrict left-turn movements ï ï ï ï ï Non-traversable medians ï ï ï ï ï ï ï ï Two-way left-turn lanes ï ï ï ï ï ï ï ï ï Frontage and service roads ï ï ï ï Unsignalized median openings ï ï ï ï ï ï ï Traffic signal spacing ï ï ï ï ï Number and spacing of access points ï ï ï ï ï ï ï ï ï Interchange areas ï ï ï ï ï ï ï ï Left-turn lanes ï ï ï ï ï ï ï Right-turn lanes ï ï ï ï ï Driveway channelization ï ï ï ï ï ï ï ï ï ï Alternative intersections and interchanges ï ï ï ï ï ï ï ï ï Parking and stopping restrictions ï ï ï ï ï ï ï ï ï Roundabouts ï ï ï ï ï ï ï Driveway sight distance ï ï ï ï ï ï ï ï ï ï One-way driveways ï ï ï ï ï ï ï ï ï Driveway width ï ï ï ï ï ï ï ï ï ï Driveway vertical geometry ï ï ï ï ï ï ï ï ï ï Driveway throat length ï ï ï ï ï ï ï ï ï ï ïï ïïï ï ï Note: ï = Relative change can be estimated without calculating absolute values. Absolute values can also be calculated. = Relative change can be estimated by calculating absolute values first. ï = Method not applicable to this group of access management techniques. ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ïïï ï ï ï ï ï ï ï Table A1. Quantitative method applicability by access management technique group.

88 Guide for the Analysis of Multimodal Corridor Access Management stop-controlled intersections. Other input data can be defaulted, but the fewer default values that are used the greater is the chance that the end result is likely to be accurate. In particular, the heavy vehicle percentage, peak hour factor, and (for traffic signals) saturation flow rate, cycle length, and minimum green time by phase are the default values with the greatest potential impact on the end result. In areas with high levels of pedestrian activity, pedestrian volumes are also important to account for. NCHRP Report 825: Planning and Preliminary Engineering Applications Guide to the Highway Capacity Manual (2) provides simplified versions of the HCM6 signalized intersection and round- about delay methods that can be applied by hand or by using one of the computational engines (spreadsheets) available in the Application Guides section of the HCM6 Volume 4 website. A simplified delay method and spreadsheet are also provided for two-way stop-controlled inter- sections; however, this method may not be appropriate for analyzing some access management techniques because it assumes the presence of left-turn lanes on the major street, no median, no U-turns, and no ability to make two-stage left turns, among other things. HCM6 Arterial Travel Speed Methods Techniques Affecting Free-Flow Speed Some access management techniques affect a roadwayâs free-flow speed, âthe average running speed of through vehicles traveling along a segment under low-volume conditions and not delayed by traffic control devices or other vehicles. It reflects the effect of the street environment on driver speed choiceâ (1). Changes in the free-flow speed affect the average mid- block motorized vehicle speed, although not on a one-to-one basis. As demonstrated in the next section, a 1 mph increase in free-flow speed typically results in a 0.85â0.95 mph increase in average midblock speed at volumes of 500 vehicles per lane or lower. But a 1 mph increase in free-flow speed but can be potentially as low as 0.65 mph in situations with a combination of high free-flow speeds, high traffic volumes, short traffic signal spacing, and many access points. Estimating Changes in Free-Flow Speed The following types of access management techniques affect a roadwayâs free-flow speed: (a) installing a non-traversable (restrictive) median, (b) reducing the number of access points, and (c) removing on-street parking. The effects of these techniques on free-flow speed are as follows: â¢ Non-traversable median. Exhibit 18-11 in the HCM6 gives the change in roadway free-flow speed resulting from a conversion from an undivided roadway or a non-restrictive median to a restrictive (i.e., non-traversable median) (1). The change in free-flow speed (in mph) equals 1.5 prm â 3.7 prm, pcurb, where prm is the proportion of the link length (decimal) with a restrictive median and pcurb is the proportion of the link length (decimal) with a curb on the right side of the roadway. When a non-restrictive median is implemented, the free-flow speed increases where there is no curb on the right side of the roadway and decreases where there is a curb. â¢ Access point density. Equation 18-3 and Exhibit 18-11 in the HCM6 give the reduction in roadway free-flow speed due to access point density (1). This reduction (in mph) equals â0.078 Da/Nth, where Da is the number of access points per mile (considering both sides of the roadway but only those accessible to or from the direction of travel), and Nth is the number of through lanes in the direction of travel. Thus, reducing the number of access points increases the free-flow speed. â¢ Parking restrictions. Exhibit 18-11 in the HCM6 indicates that roadway free-flow speed increases by 0.03 mph for each 1% reduction in the percentage of street length where on-street parking is allowed (1). Thus, eliminating on-street parking increases the free-flow speed by

Applications Guidance for Selected Quantitative Analysis Methods 89 up to 3 mph. In most cases, the speed increase will be lower, as right-turn lanes, driveways, and unsignalized intersections between traffic signals will reduce the street length where park- ing is allowed. In addition, parking would often have already been restricted in driveways. Estimating Changes in Average Midblock Running Speed from Changes in Free-Flow Speed The average midblock running speed is an input to the pedestrian and bicycle LOS methods in the HCM6, described later in this appendix. The average midblock running speed is also used in combination with delay at intersections to determine average motor vehicle travel speeds along a roadway segment or facility. When the free-flow speed changes, so does the average midblock running speed, although by a smaller amount. This section presents two methods for relating the average midblock running speed of motor vehicles to the free-flow speed. The first method is a graphical generalized approach that provides pre-calculated values for a variety of situations. The second method is a direct calculation using the HCM6âs equations. Either method can determine either (1) an absolute value for the average midblock running speed for a given set of conditions or (2) the change in average midblock running speed given a change in free-flow speed. Generalized Approach Figure A1 shows average midblock running speed as a percentage of free-flow speed for various combinations of free-flow speeds, per-lane traffic volumes, traffic signal spacing, and access spacing. To use the figure, find the graph with the conditions most closely matching the situation being analyzed. The assumptions used to develop these graphs are presented at the bottom of the figure. In many cases, the average midblock running speed is approximately 85%â95% of the free-flow speed, but in more extreme situations (i.e., high volumes, high free- flow speed, short signal spacing, and many access points) it can be as low as 65% of the free-flow speed. In the absence of other information about the free-flow speed, it can be assumed that the free-flow speed is 5 mph greater than the posted speed limit. Example 1. As an example of how to use these graphs to calculate an absolute speed value, assume a roadway has a 35-mph free-flow speed, 500 vehicles per hour per lane, Â½-mile signal spacing, and 30 access points per mile (including public street intersections). Figure A1c shows that the average midblock running speed would be approximately 90% of the free-flow speed, or about 31.5 mph. Further adjustments to free-flow speed related to median presence and on-street parking can be made by using the relationships given above. Example 2. As an example of how to use these graphs to calculate a relative change in speed, assume the same roadway. Curbside parking is allowed along this roadway, except within 30 feet of an access point. The average access width is 30 feet (i.e., the length of no-parking zone associated with each access is 90 feet). Given 15 access points on each side of the street, a total of 1,350 feet out of the 2,640 feet between traffic signals is already designated for no parking. If parking is prohibited along the remaining 1,290 feet (i.e., 49% of the section length), the free-flow speed would be expected to increase by 1.5 mph (49 Ã 0.03). From Figure A1c, the average midblock running speed would be expected to increase by 90% of this value, or about 1.35 mph. Direct Calculation Approach Equations from the HCM6 can directly estimate the average midblock running speed for a given set of conditions. To determine the relative change in running speed, calculate the speed twice, once for the before condition and once for the after condition.

90 Guide for the Analysis of Multimodal Corridor Access Management Step 1. Determine the base free-flow speed. The base free-flow speed is a function of the posted speed, median and curb presence, on-street parking provision, and access point density. The base free-flow speed is calculated by using Equation A1, which is derived from Equation 18-3 in the HCM6. = + + + (A1)0S S f f ffo cs A pk where Sfo = base free-flow speed (mph) S0 = speed constant, from Table A2 (mph) (a) (b) (c) (d) Note: Assumptions used in each graph are as follows: (a) Number of lanes = 2 per direction, Â½-mile signal spacing, 10 access points per mile. (b) Number of lanes = 2 per direction, 10 access points per mile, 500 vehicles per hour per lane. (c) Number of lanes = 2 per direction, Â½-mile signal spacing, 500 vehicles per hour per lane. (d) Number of lanes = 2 per direction, Â¼-mile signal spacing, 40 access points per mile. 35 mph FFS 45 mph FFS 55 mph FFS 1 mile signal spacing Â½ mile signal spacing Â¼ mile signal spacing 35 mph FFS 45 mph FFS 55 mph FFS 35 mph FFS 45 mph FFS 55 mph FFS Figure A1. Average midblock running speed as a percentage of free-flow speed or FFS: (a) volume and free-flow speed, (b) free-flow speed and signal spacing, (c) access density and free-flow speed, and (d) volume and free-flow speed, many access points and signals.

Applications Guidance for Selected Quantitative Analysis Methods 91 fcs = adjustment for cross-section (mph), from Table A2 fA = adjustment for access point density (mph), from Table A2 fpk = adjustment for on-street parking (mph), from Table A2 All of the values needed for Equation A1 may be obtained directly from Table A2 or from the equations provided in the notes below Table A1. Step 2. Adjust the free-flow speed for short traffic signal spacing. Shorter distances between traffic signals have been found to result in lower free-flow speeds, all other factors being equal (1, 3). Equations A2 and A3, derived from Equations 18-4 and 18-5 in the HCM6 (1), adjust the base free-flow speed to account for this effect. 1.02 4.7 19.5 max , 400 1.0 (A2)f S L L fo ( ) = â â â¤ = â¥ (A3)S S f Sf fo L pl where fL = adjustment for traffic signal spacing (unitless) Sf = free-flow speed (mph) L = segment length (ft), which is the distance between adjacent signalized intersections Spl = posted speed limit (mph) Speed Limit (mph) Speed Constant S0 (mph)a Median Type Percent with Restrictive Median (%) Adjustment for Cross Section fCS (mph) b No Curb Curb 25 37.4 Restrictive 20 0.3 â0.9 30 39.7 40 0.6 â1.4 35 42.1 60 0.9 â1.8 40 44.4 80 1.2 â2.2 45 46.8 100 1.5 â2.7 50 49.1 Nonrestrictive na 0.0 â0.5 55 51.5 No median na 0.0 â0.5 Access Density (points/mi) Adjustment for Access Points fA by Lanes Nth (mph) c Percent with On- Street Parking (%) Adjustment for Parking fpk (mph)d 1 Lane 2 Lanes 3 Lanes 0 0.0 0.0 0.0 0 0.0 2 â0.2 â0.1 â0.1 20 â0.6 4 â0.3 â0.2 â0.1 40 â1.2 50 â1.5 10 â0.8 â0.4 â0.3 20 â1.6 â0.8 â0.5 40 â3.1 â1.6 â1.0 60 â1.8 80 â2.4 100 â3.0 60 â4.7 â2.3 â1.6 Source: Adapted from Highway Capacity Manual, 6th ed., Exhibit 18-11 (1). Note: na = not applicable. a S0 = 25.6 + 0.47Spl, where Spl = posted speed limit (mph). b fCS = 1.5 prm â 0.47 pcurb â 3.7 pcurb prm,, where prm = proportion of link length with restrictive median (decimal) and pcurb = proportion of segment with curb on the right side (decimal). c fA = â0.078 Da /Nth, where Da = access point density on segment, considering both sides of the roadway but only counting those accessible to or from the direction of travel (points/mi), and Nth = number of through lanes in the direction of travel; d fpk = â3.0 Ã propor on of link length with on-street parking available on the right side (decimal). Table A2. Base free-flow speed adjustment factors.

92 Guide for the Analysis of Multimodal Corridor Access Management Step 3. Calculate the average midblock running speed. The following equations, derived from Equations 18-6, 18-7, and 18-15, respectively, in the HCM6 (1), can estimate the average midblock running speed for a given free-flow speed, traffic volume, and access point density, for streets with signalized intersections (i.e., no roundabouts or intersections where the major street is stop controlled). 2 1 1 52.8 (A4) th 0.21f v N S v m f = + âï£«ï£ï£¬ ï£¶ ï£¸ï£· 4.0 0.0025 3,600 5,280 5,280 (A5)t L L S f n L dR f v ap ap= + + 3,600 5,280 (A6)S L t R R = where fv = proximity adjustment factor tR = average segment running time (s) nap = access point density, considering both sides of the roadway, but only counting those accessible to or from the direction of travel (points/mi) dap = average through vehicle delay per access point, from Table A3 (s/access point) v = average midblock volume (veh/h) vm = average midblock 15-minute demand flow rate (veh/h) = v/PHF, where PHF = peak hour factor = v/(4 Ã peak 15-min volume) Nth = number of through lanes SR = average midblock running speed (mi/h) Example 3. Assume a roadway with the following characteristics. They are 30-mph posted speed, 5 lanes (two travel lanes in each direction and a two-way left-turn lane), Â½-mile signal spacing, 30 access points per mile, on-street parking for 50% of the segment length, curb present, Midsegment Volume (vehicles per hour per lane) Through Vehicle Delay (s/access point) by Number of Through Lanes 1 Lane 2 Lanes 3 Lanes 200 0.04 0.04 0.05 300 0.08 0.08 0.09 400 0.12 0.15 0.15 500 0.18 0.25 0.15 600 0.27 0.41 0.15 700 0.39 0.72 0.15 Source: Highway Capacity Manual, 6th ed., Exhibit 18-13 (1). Note: Based on 10% right turns and 10% left turns on average from the roadway at a typical access point. Adjust proportionally for other turn percentages. Assumes no exclusive turn lanes are provided at a typical access point. Reduce adjusted values by 50% for each movement provided with an exclusive turn lane of adequate length or prohibited (e.g., 50% for a left-turn lane provided but no right-turn lane, 100% for left- and right-turn lanes provided, and 100% for left turns prohibited and a right-turn lane provided). Table A3. Average delay due to turning vehicles at access points.

Applications Guidance for Selected Quantitative Analysis Methods 93 no right-turn lanes at access points, 1,000 vehicles per hour in the direction of travel, and a peak hour factor of 0.92. The average midblock running speed is calculated as follows: Step 1. The following values are obtained from Table A2: â¢ Speed constant S0 = 39.7 mph â¢ Cross-section adjustment fcs = â0.5 mph (nonrestrictive median + curb) â¢ Access-point density adjustment fA = â1.2 mph (2 lanes, interpolate for 30 points/mi) â¢ On-street parking adjustment fpk = â1.5 mph (interpolate for 50% parking) From Equation A1, the base free-flow speed is then 36.5 mph. Step 2. The traffic signal spacing adjustment fL is determined from Equation A2, using a segment length of 2,640 ft (Â½ mi), and found to be 0.99. The free-flow speed is then determined to be 36.1 mph, using Equation A3. Step 3. Equation A4 is used to determine the proximity adjustment, using inputs of 1,087 veh/h (= 1,000 veh/h/0.92), 2 through lanes, and a free-flow speed of 36.1; the factor fv is determined to be 1.035. Equation A5 is then used to determine the average segment running time tR. From Table A3, the average delay per vehicle per access point dap is initially determined by interpolation to be 0.32 s/point, using inputs of 2 through lanes and a 15-min demand flow rate of 544 veh/h/ln (= 1,087/2). Because there is a two-way left-turn lane but there are no right-turn lanes, this initial value is reduced by 50% to a final value of 0.16 s/access point. All other inputs to the equation have been given or previously calculated, and tR is determined to be 54.6 s. Finally, Equation A6 is used to calculate the average midblock running speed SR, which is determined to be 33.0 mph, which is 91% of the free-flow speed. Example 4. Assume the same roadway, but with on-street parking prohibited. In Step 1, the base free-flow speed increases by 1.5 mph to 38.0 mph. In Step 2, the free-flow speed becomes 37.6 mph. In Step 3, the proximity adjustment becomes 1.034, the average segment running time becomes 52.5 s, and the average midblock running speed becomes 34.3 mph. Compared with the situation in Example 3, the average running speed has increased by 1.3 mph (i.e., it increased by 87% of the increase in the base free-flow speed). Estimating Average Travel Speed Average travel speed is based on the combination of the running time for a segment with the average delay for the through movement at the downstream, signalized intersection (or another boundary intersection, such as a roundabout). Average travel speed can be calculated with Equation A7, which is based on Equation 18-15 in the HCM6 (1): 3,600 5,280 (A7)S L t d T R t( ) = + where ST is the average travel speed (mi/h) and dt is the average delay to the through movement at the downstream boundary intersection. Queue Estimation Methods in the HCM6 Chapter 19 in the HCM6 (1) provides a method for assessing average and any percentile back-of-queue length at signalized intersections. The HCM6 also provides methods for esti- mating 95th percentile queue lengths at two-way stop-controlled intersections (Chapter 20),

94 Guide for the Analysis of Multimodal Corridor Access Management all-way stop-controlled intersections (Chapter 21), and roundabouts (Chapter 22). The effect of access management techniques that alter turning movement patterns, traffic volumes, or both at individual access points or intersections can be evaluated by using the methods in these chapters, by calculating the resulting 95th percentile queue length and comparing the result with the available queue storage length. With the exception of the method for roundabouts, which can be performed by hand or by automation in a simple spreadsheet, the queue estimation methods described in the HCM6 are only practical to apply with specialized software, given the number of computations involved. However, calculating a 95th percentile queue can be done by hand using just one equation, once the capacity of a lane group or approach has been determined in the process of calculating intersection delay. A number of commercial software packages are available that can perform these calculations. The required input data are the same as described earlier for the HCM6 inter- section delay methods. NCHRP Report 825 (2) provides simplified versions of the HCM6âs methods for 95th per- centile queue estimation, which can be done by hand or in a spreadsheet. At the time of this writing, the signalized intersection computational engine available in the Application Guides section of the HCM Volume 4 website (hcmvolume4.org) could calculate queues, but the engines for other intersection forms could not calculate queues. This guideâs simplified method for two-way stop-controlled intersections, however, may not be appropriate for analyzing some access management techniques, because it assumes the presence of left-turn lanes on the major street, no median, no U-turns, and no ability to make two-stage left turns, among other things. Multimodal Level of Service Methods in the HCM6 Chapter 18 in the HCM6 (1) provides a unified set of methods for estimating pedestrian, bicycle, and on-street transit LOS. Key features of these methods are (a) the LOS determined for each travel mode, as well as each modeâs underlying level of service score, can be directly compared with each other and (b) the LOS values were developed from actual traveler percep- tions. These methods are particularly well suited for evaluating how changes in the allocation of roadway right-of-way among different modes affects each modeâs quality of service, but they can also be used to evaluate the impact of various access management techniques. In many cases, the impact of an access management technique on a modeâs LOS is indirect: the technique changes the speed, volume, or both of motorized traffic on the roadway and thereby influences the modeâs LOS. In other cases, the impact is direct: for example, changing the amount of occupied on-street parking directly influences both pedestrian and bicycle LOS, while changing the number of access points per mile directly influences bicycle LOS. The HCM6 provides multimodal LOS methods for links (between signalized intersections), signalized intersections and segments (combining links and intersections), and facilities (multiple consecutive segments). A link-based evaluation requires the least amount of data and calcu- lations and is sufficient for many applications, including evaluating the effects of many access management techniques. However, a segment-based evaluation (including evaluating modal LOS for intersections) may be appropriate for the following types of techniques: (a) techniques that result in substantial changes in turning-movement volumes that conflict with pedestrians and bicycles, (b) techniques that change the access density, and (c) techniques that change pedestrian delay crossing a street in the middle of the block. A segment-based evaluation is also necessary when desired to directly compare bicycle and pedestrian LOS scores with transit LOS scores, because transit LOS is only calculated at a segment level.

Applications Guidance for Selected Quantitative Analysis Methods 95 The following section focuses on link-based evaluations and discusses relevant aspects of segment-based evaluations. All calculations can be performed by hand or readily input into a spreadsheet. A computational engine (spreadsheet) for calculating multimodal LOS along a link is available in the Application Guides section of the HCM Volume 4 website (hcmvolume4.org). Pedestrian LOS Links Pedestrian LOS (PLOS) for a roadway link is influenced by a number of factors. They include sidewalk width and separation from traffic, traffic volumes and speeds, and presence of buffers such as street trees. The HCM6 method calculates a PLOS score, which then translates into an LOS letter as shown in Table A4. Equation A8, adapted from Equation 18-32 in the HCM6 (1), shows the component factors used to determine the PLOS score. A change in any one of these component factors changes the PLOS score on a one-to-one basis, making it easy to identify relative changes in PLOS due to particular access management techniques, without having to perform the full set of calculations. = + +PLOS 6.0468 + (A8)F F Fl w v S where PLOSl = pedestrian LOS score for link Fw = cross-section adjustment factor Fv = motorized vehicle volume adjustment factor FS = motorized vehicle speed adjustment factor The cross-section adjustment factor Fw is a function of the sidewalk width, the separation of the sidewalk from the street, the presence of physical buffers or barriers between the street and sidewalk, and the effective separation of moving traffic from the curb (including such factors as the width of any parking, shoulder, or bicycle lanes, and parking occupancy). Figure A2 shows the key dimensions used in determining PLOS. Note that the width of the gutter and curb (if present) is not included in any of the dimensions. Equation A9, adapted from Equation 18-33 in the HCM6 (1), is used to determine the cross- section adjustment factor. 1.2276ln 0.5 50 (A9)SWF W W p W f W fw v bps pk B B A( )= â + + + + LOS Segment-Based PLOS Score Link-Based PLOS Score A â¤ 2.00 â¤ 1.50 B > 2.00â2.75 > 1.50â2.50 C > 2.75â3.50 > 2.50â3.50 D > 3.50â4.25 > 3.50â4.50 E > 4.25â5.00 > 4.50â5.50 F > 5.00 > 5.50 Source: Derived from the Highway Capacity Manual, 6th ed., Exhibit 18-2 (1). Note: The HCM6 also considers pedestrian density on the sidewalk when determining pedestrian LOS. However, pedestrian density only influences the result at very high pedestrian volumes (> 1,000 pedestrians per hour) and therefore is not included. Table A4. Pedestrian LOS criteria.

96 Guide for the Analysis of Multimodal Corridor Access Management where ln = natural logarithm WT = total width of outside through lane, bicycle lane, parking lane, and shoulder (ft) WSW = available sidewalk width, not including landscape buffer or furnishing zone (ft) v = average midblock volume in the direction of travel closest to the sidewalk (veh/h) Wv = effective total width of outside through lane, bicycle lane, and shoulder as a function of traffic volume (ft) = WT if vm > 160 veh/h or WSW > 0, then Wv = WT; if not (i.e., other- wise), Wv = WT Ã (2 â 0.005 vm) Wbps = total width of bicycle lane, parking lane, and shoulder (ft) ppk = proportion of on-street parking occupied (decimal) WB = width of landscape buffer or furnishing zone (ft) fB = buffer area coefficient = 5.37 for any continuous barrier at least 3 feet high located between the sidewalk and the outside edge of roadway; otherwise, use 1.0 WA = adjusted sidewalk width = min(WSW, 10) (ft) fSW = sidewalk coefficient = 6.0 â 0.3 WA Equations A-10 and A-11, adapted from Equations 18-34 and 18-35, respectively, in the HCM6 (1), are used to determine the motorized vehicle volume and speed adjustment factors. 0.0091 4 (A10) th F v N v ma= 4 100 (A11) 2 F S s R= ï£«ï£ï£¬ ï£¶ ï£¸ï£· where vma = adjusted average midblock 15-min demand flow rate (veh/h) = max(vm, 4Nth) Nth = number of through lanes in the direction of travel closest to the sidewalk SR = average motorized vehicle midblock running speed (mph) Example 5. Assume that, as a result of an access management technique, average motorized vehicle midblock running speeds increase by 2 mph, from 42 to 44 mph. From Equation A11, Source: Adapted from NCHRP Report 825, Exhibit 99 (2). Figure A2. Key dimensions used in determining pedestrian LOS.

Applications Guidance for Selected Quantitative Analysis Methods 97 the before-and-after values of Fs are 0.71 and 0.77, respectively. Because nothing else changes that would affect PLOS, the PLOS value would go up (i.e., get worse) by 0.06 point as a result of the technique. The LOS letter would likely stay the same, as 0.06 is much less than the width of the range used to determine LOS letters (i.e., 1.00 point for links and 0.75 point for segments). Example 6. Assume the roadway used in Example 3, with the following additional assumptions. They are 80% of the available parking is occupied on average, the sidewalk width is 6 feet, there is no landscape buffer or bicycle lane, the parking lane width (not including the gutter) is 8 feet, and the outside travel lane width is 12 feet. PLOS is then calculated as follows: â¢ Cross-section adjustment factor. Because a sidewalk is provided, Wv equals WT, the total width of the outside travel lane, bicycle lane, parking lane, and shoulder (12 + 0 + 8 + 0 = 20 feet). Wbps equals the sum of the bicycle lane, parking lane, and shoulder widths, or 8 feet. The pro- portion of occupied on-street parking ppk is the proportion of the street where on-street parking is provided, multiplied by the percentage of the parking that is occupied: 0.5 Ã 0.8 = 0.4. The buffer width is 0 feet and because there are no street trees or other forms of barriers, the buffer area coefficient is 1.00. Because the sidewalk width is less than 10 feet, the adjusted sidewalk width WA is the same as the sidewalk width (6 feet); the sidewalk coefficient then computes to be 4.2. Entering these values into Equation A9 gives a value of Fw of â5.201. â¢ Motorized vehicle volume adjustment factor. From Example 3, vm = 1,087 veh/h and Nth = 2 lanes. Entering these values into Equation A10 gives a value of Fv of 1.236. â¢ Motorized vehicle speed adjustment factor. From Example 3, SR = 33.0 mph. Applying Equa- tion A11 gives a value of Fs of 0.436. â¢ Pedestrian level of service score. Entering the calculated adjustment factor values into Equa- tion A8 gives a PLOS score of 2.52. From Table A4, this score for a link corresponds to LOS C, just beyond the threshold for LOS B. Figure A3 demonstrates the sensitivity of PLOS to various factors influenced by access management techniques. Segments Midblock pedestrian crossing delay is a factor in determining PLOS at the segment level. It is determined as the lesser of (a) average pedestrian delay making a midblock pedestrian crossing (if allowed) and (b) delay diverting to the nearest signalized intersection to cross. Providing a non-traversable median along a roadway makes it possible for pedestrians to cross the street in two stages, which reduces pedestrian delay, while widening the roadway (e.g., to add a turn lane) increases pedestrian delay. Step 8 of the pedestrian LOS methodology in the HCM6 (starting on page 18-54) describes how to calculate a âroadway crossing difficulty factorâ (1), while Step 9 applies this factor to the calculation of PLOS for a segment. Chapter 19 in the HCM6 estimates pedestrian delay and PLOS at the traffic signal, while the pedestrian delay methodology in Chapter 20 is used to estimate midblock crossing delay, including accounting for driver-yielding behavior (1). See the Pedestrian and Bicycle Delay section later in the appendix for more details. Figure A4 shows the sensitivity of segment PLOS to street-crossing delay for various levels of link PLOS, under the assumption that intersection PLOS (used in calculating segment PLOS) is equal to link PLOS. The roadway crossing difficulty factor is constrained to minimum and maximum values beyond which it has no additional effect, as indicated by the horizontal lines in the figure. In addition, the street-crossing delay used in calculating the factor is capped at 60 seconds.

98 Guide for the Analysis of Multimodal Corridor Access Management (a) (b) (c) Note: Assumptions used in each graph are as follows: (a) Outside lane demand flow rate = 500 veh/h; no landscape buffer or street trees, outside travel lane width = 12 feet; parking lane width = 8 feet; and traffic speed = 35 mph. (b) Occupied on-street parking = 50%; no landscape buffer or street trees, outside travel lane width = 12 feet; parking lane width = 8 feet; and traffic speed = 35 mph. (c) Outside lane demand flow rate = 500 veh/h; occupied on-street parking = 50%; no landscape buffer or street trees, outside travel lane width = 12 feet; and parking lane width = 8 feet. Figure A3. Sensitivity of link PLOS to factors influenced by access management techniques: (a) volume and sidewalk width, (b) occupied on-street parking, and (c) average midblock traffic speed.

Applications Guidance for Selected Quantitative Analysis Methods 99 Bicycle LOS Links A number of factors influence bicycle LOS (BLOS) for a roadway link. They include bicycle lane presence, on-street parking presence, separation of bicycles from motorized traffic, traffic volumes and speeds, heavy vehicle percentage, and pavement condition. The HCM6 method calculates a BLOS score, which translates into a LOS letter as shown in Table A5. Equation A12, adapted from Equation 18-41 in the HCM6 (1), shows the component factors used to determine the BLOS score. A change in any one of these component factors changes the BLOS score on a one-to-one basis, making it easy to identify relative changes in BLOS due to particular access management techniques, without having to perform the full set of calculations. BLOS 0.760 (A12)F F F Fl w v s p= + + + + where BLOSl is the bicycle LOS score for link and Fp is the pavement condition adjustment factor. Note: Assumptions used to develop the graph include link PLOS = intersection PLOS, segment length = Â½ mile, average pedestrian delay at the traffic signal at the end of the segment = 50 seconds, and average pedestrian walking speed = 4 feet per seconds. Figure A4. Sensitivity of segment PLOS to street-crossing delay. LOS Segment-Based PLOS Score Link-Based PLOS Score A â¤ 2.00 â¤ 1.50 B > 2.00â2.75 > 1.50â2.50 C > 2.75â3.50 > 2.50â3.50 D > 3.50â4.25 > 3.50â4.50 E > 4.25â5.00 > 4.50â5.50 F > 5.00 > 5.50 Source: Derived from Highway Capacity Manual, 6th ed., Exhibit 18-3 (1). Table A5. Bicycle LOS criteria.

100 Guide for the Analysis of Multimodal Corridor Access Management The cross-section adjustment factor Fw is generally a function of the outside travel lane width, bicycle lane and shoulder width (if present), and parking presence and occupancy. Figure A5 shows the key dimensions used in determining BLOS. Note that the width of the parking lane or shoulder is included in these dimensions only when completely unoccupied and that the gutter and curb (if present) are never included. Equation A13, adapted from Equation 18-42 in the HCM6 (1) determines the cross-section adjustment factor: 0.005 (A13)2F Ww e= â with 10 0.0 4 feet 20 0.0 4 feet (A14)W W p W W W p W e v pk l v l pk l = â â¥ < + â â¥ â¥ ï£± ï£² ï£´ ï£³ï£´ 2 0.005 160 veh h or roadway is otherwise divided (A15)W W W v vv T T m m ( ) = â ï£± ï£² ï£³ï£´ > where We = effective width of the outside through lane (ft) Wv = effective total width of outside through lane, bicycle lane, and shoulder as a function of traffic volume (ft) Wl = width of bicycle lane and shoulder, from Figure A5 (ft) WT = total width of outside through lane, bicycle lane, and shoulder, from Figure A5 (ft) v = average midblock volume in the direction of travel closest to the bicycle facility (veh/h) Source: Adapted from NCHRP Report 825, Exhibit 102 (2). Figure A5. Key dimensions used in determining bicycle LOS.

Applications Guidance for Selected Quantitative Analysis Methods 101 Equations A16 through A18, adapted from Equations 18-43 through 18-45, respectively, in the HCM6 (1), determine the motorized vehicle volume and speed adjustment factors and the pavement condition adjustment factor. 0.507 ln 4 (A16) th F v N v m= ï£«ï£ï£¬ ï£¶ ï£¸ï£· 0.199 1.1199ln 20 0.8103 1 0.1038 (A17)2F S PS Ra HVa[ ]( ) ( )= â + + = 7.066 (A18) 2 F P p C where Nth = number of through lanes in the direction of travel of the bicycle facility SRa = adjusted average motorized vehicle midblock running speed (mph) = max(21, SR) PHVa = adjusted percent heavy vehicles = 50% if PHV > 50% and vm (1 â 0.01 PHV) < 200 veh/h and = PHV otherwise PHV = percent heavy vehicles Fp = pavement condition adjustment factor PC = pavement condition (present serviceability) rating, from 0 (worst) to 5 (best) Example 7. Assume that, as a result of an access management technique, average motorized vehicle midblock running speeds will increase by 2 mph, from 42 to 44 mph, and there are 5% trucks and buses in the traffic stream. From Equation A16, the before-and-after values of Fs are 1.96 and 2.01, respectively. Because nothing else changes that would affect BLOS, the BLOS value would go up (i.e., get worse) by 0.05 point as a result of the technique. The LOS letter would likely stay the same, as 0.05 is much less than the width of the range used to determine LOS letters (i.e., 1.00 point for links and 0.75 point for segments). Example 8. Assume the roadway used in Example 3, with the following additional assump- tions. They are 80% of the available parking is occupied on average, there is no bicycle lane, the parking lane width (not including the gutter) is 8 feet, the outside travel lane width is 12 feet, heavy vehicles form 5% of the traffic volume, and the pavement condition rating is 3. BLOS is then calculated as follows: â¢ Cross-section adjustment factor. Because the traffic volume is greater than 160 veh/h (as given in Example 3), Wv = WT = 12 feet (Equation A15) because there is no bicycle lane and because the parking lane width is not included in WT when the lane is partially occupied with parked cars. Because there is no bicycle lane, We = Wv = 12 feet (Equation A14). Then, from Equation A13, Fw = â0.720. â¢ Motorized vehicle volume adjustment factor. From Example 3, vm = 1,087 veh/h and Nth = 2 lanes. Entering these values into Equation A16 gives a value of Fv of 2.490. â¢ Motorized vehicle speed adjustment factor. From Example 3, SR = 33.0 mph. Applying Equation A17 gives a value for Fs of 1.691. â¢ Pavement condition adjustment factor. From Equation A18, Fp = 0.785. â¢ Bicycle level of service score. Entering the calculated adjustment factor values into Equation A12 gives a BLOS score of 5.01. From Table A5, this score for a link corresponds to LOS E. Figure A6 demonstrates the sensitivity of BLOS to various factors influenced by access management techniques.

102 Guide for the Analysis of Multimodal Corridor Access Management Segments The number of access points per mile on the right side of the road is a factor in determining BLOS at the segment level. Equations 18-46 and 18-47 in the HCM6 can determine segment BLOS. Chapter 19 can estimate bicycle delay and BLOS at the traffic signal (1). These calcula- tions may be easily done by using specialized HCM6-implementing software. Figure A7 shows the sensitivity of segment BLOS to right-side access density for various levels of link BLOS, under the assumption that intersection BLOS (used in calculating segment BLOS) is equal to link BLOS. (c) (a) (b) Note: Assumptions used in each graph are as follows: (a) Outside lane demand flow rate = 500 vehicles per hour; no landscape buffer or street trees, outside travel lane width = 12 feet; parking lane width = 8 feet; and traffic speed = 35 mph. (b) Occupied on-street parking = 50%; no landscape buffer or street trees, outside travel lane width = 12 feet; parking lane width = 8 feet; and traffic speed = 35 mph. (c) Outside lane demand flow rate = 500 vehicles per hour; occupied on-street parking = 50%; no landscape buffer or street trees, outside travel lane width = 12 feet; and parking lane width = 8 feet. Figure A6. Sensitivity of link BLOS to factors influenced by access management techniques: (a) traffic volume, (b) occupied on-street parking, and (c) average midblock traffic speed.

Applications Guidance for Selected Quantitative Analysis Methods 103 Transit LOS Transit LOS (TLOS) for a roadway link is influenced by the bus frequency on the link, passen- gersâ perceived travel time (primarily a function of bus speeds and loading), and the PLOS score for the link. The HCM6 method calculates a TLOS score, which translates into an LOS letter as shown in Table A6. Unlike the pedestrian and bicycle modes, TLOS is not computed at the link level but only at the segment level. Equation A19, adapted from Equations 18-62 and 18-63 in the HCM6 (1), shows the com- ponent factors used to determine the TLOS score. A unit change in either the headway factor (not normally influenced by access management techniques) or the travel time factor changes the TLOS score by 1.5 units. Similarly, a unit change in the link PLOS changes the score by 0.15 units. If one knows these relationships, it is possible to determine the relative change in Note: Assumptions used to develop the graph include link BLOS = intersection BLOS, segment length = Â½ mile, average bicycle delay at the traffic signal at the end of the segment = 40 seconds, and average bicycle speed = 12 mph. Figure A7. Sensitivity of segment BLOS to right-side access point density. LOS TLOS Score A â¤2.00 B >2.00â2.75 C >2.75â3.50 D >3.50â4.25 E >4.25â5.00 F >5.00 Source: Derived from Highway Capacity Manual, 6th ed., Exhibit 18-3 (1). Table A6. Transit LOS criteria.

104 Guide for the Analysis of Multimodal Corridor Access Management TLOS due to particular access management techniques, without having to perform the full set of calculations. TLOS 6.0 1.50 0.15 PLOS (A19)F Fh tt l= â + where TLOS = transit LOS score for segment Fh = headway factor Ftt = perceived travel time factor PLOSl = pedestrian LOS score for link, from Equation A7 The headway and perceived travel time factors are determined as follows, adapted from Equa- tions 18-56 through 18-60 in the HCM6 (1). 4.00 (A20)1.434 0.001F eh fb= ( )â + 1 1 1 1 (A21)F E T E T E T E T tt btt ptt ptt btt ( ) ( ) ( ) ( ) = â â + â â + with 60 2 (A22)1T a S T Tptt b ex at( )= ï£« ï£ï£¬ ï£¶ ï£¸ï£· + â [ ] ( ) ( ) ( ) ( ) = â¤ + â â¤ â¤ + â + â + â > ï£± ï£² ï£´ ï£´ï£´ ï£³ ï£´ ï£´ ï£´ 1.00 0.80 1 4 0.80 4.2 0.80 1.00 1 4 0.80 1.00 6.5 5 1.00 4.2 1.00 (A23)1a F F F F F F F F l l l l l l l l 1.3 0.2 (A24)T p p L at sh be pt = + where fb = bus frequency stopping in or adjacent to segment (bus/h) E = ridership elasticity with respect to changes in the travel time rate = â0.40 Tbtt = base travel time rate (min/mi) = 6.0 for the central business district of a metropolitan area with 5 million persons or more; otherwise, Tbtt = 4.0 Tptt = perceived travel time rate (min/mi) a1 = perceived travel time weighting factor for passenger load Sb = average bus speed (mph), including stops to serve passengers and traffic signal delay Tex = excess wait time rate (min/mi) = tex/Lpt tex = excess passenger wait time due to late bus arrivals (min) (default = 3) Lpt = average passenger trip length (mi) (default = 3.7) Tat = amenity time rate (min/mi) psh = proportion of stops in segment with shelters (decimal) pbe = proportion of stops in segment with benches (decimal)

Applications Guidance for Selected Quantitative Analysis Methods 105 Increases in average midblock motor vehicle speeds produce much smaller increases in average bus speeds, when bus stop and traffic signal delays are considered. Buses spend relatively little time per mile traveling at a streetâs running speed, and much more time stopping to serve passengers (including time spent decelerating and accelerating back to running speed). On higher-speed roadways (e.g., 45 mph) with just 4 stops per mile buses cannot acceler- ate all the way to the posted speed before they have to begin decelerating for the next bus stop. Traffic signal delays are also a factor. Table A7 shows representative changes in average bus speeds resulting from a unit change in average midblock motorized vehicle speed. For example, on a street with a midblock running speed of 25 mph and 6 bus stops per mile, a 1-mph increase in motor vehicle running speeds would result in a 0.24-mph increase in average bus speeds. Example 9. Assume that as a result of an access management technique average motorized vehicle midblock running speeds increase by 2 mph from 45 to 47 mph, average bus speeds are 17.1 mph, buses make 2 stops per mile along the roadway, a number of seats are usually available on the bus (load factor <0.80), and there are no shelters or benches at bus stops. From Table A7, average bus speeds would increase by (2 mph Ã 0.19) = 0.4 mph. There are no shelters or benches, therefore Tat = 0 min/mi (Equation A24). The perceived travel time weight- ing factor a1 = 0, because the load factor is less than 0.80 (Equation A23). From Equation A21, the perceived travel time rate is 5.13 min/mi, which would drop to 5.05 min/mi following the implementation of the access management technique. These values are used in Equation A21, along with a base travel time rate of 4 min/mi and an elasticity value of â0.40 to obtain perceived travel time factors of 0.906 and 0.911 for the before-and-after-conditions, respectively. Finally, from Equation A19, the increase of 0.005 in the perceived travel time factor will result in a reduction (i.e., improvement) in the TLOS score of (0.005 Ã 1.5), which rounds to 0.01 point. In comparison, the range covered by one LOS letter is 0.75 point. Example 10. Assume the roadway used in Example 6. Buses operate at 15-minute head- ways (i.e., 4 buses per hour) along the roadway. The scheduled bus travel speed is 12.5 mph, all seats on the bus are usually full during the peak hour (load factor = 1.00), and shelters and benches are provided at bus stops in this segment. From Equation A24, the amenity time rate Tat = 0.41 min/mi, assuming the default passenger trip length of 3.7 mi given with the equation. From Equation A23, a1 = 1.19 for a load factor of 1.00. The perceived travel time rate is then 6.92 min/mi (Equation A22) and the travel time factor is 0.81 (Equation A21). From Equation A19, the headway factor is 2.80. Finally, from Example 6, the PLOS score for the link is 2.52. Entering these values into Equation A19 gives a TLOS score of 2.98, which corresponds to LOS C. Condition Change 2 bus stops/mi, 45 mph 0.19 4 bus stops/mi, 35 mph 0.14 6 bus stops/mi, 35 mph 0.03 4 bus stops/mi, 30 mph 0.24 4 bus stops/mi, 25 mph 0.34 6 bus stops/mi, 25 mph 0.24 Source: Calculated from data in the Transit Capacity and Quality of Service Manual (4). Table A7. Representative changes in average bus speed per unit change in midblock motorized vehicle running speeds.

106 Guide for the Analysis of Multimodal Corridor Access Management Pedestrian and Bicycle Delay Methods in the HCM6 Pedestrian Delay Chapter 19 in the HCM6 provides Equation 19-70 (1) for estimating average pedestrian delay crossing one crosswalk at a traffic signal, assuming random pedestrian arrivals at the crosswalk: 2 (A25) walk 2 d C g C p ( )= â where dp = average pedestrian delay (s) C = traffic signal cycle length (s) gwalk = effective walk time for the crosswalk (typically the walk time plus 4 s). See HCM6 page 19-78 for additional guidance. Access management techniques that result in a wider intersection that in turn requires a longer cycle time to accommodate the increased pedestrian crossing distance may affect pedestrian delay at signalized pedestrian crossings. At unsignalized crosswalks, access management techniques that shorten the crossing dis- tance or split the crossing into two stages (e.g., by installing a non-traversable median) or that increase the crossing distance (e.g., by widening the roadway to add a right-turn lane) will affect pedestrian delay. The pedestrian method provided in Chapter 20 in the HCM6 allows the delay to be calculated and includes consideration of driver-yielding behavior (1). The method is too involved to describe in this report but can be readily input into a spreadsheet or applied by using specialized HCM6-implementing software. Required input data for the method include crosswalk length, median presence, roadway speed limit, vehicular flow rate, and average pedestrian speed. Bicycle Delay Chapter 19 in the HCM6 provides Equation 19-78 (1) for estimating average bicycle delay at a traffic signal: 0.5 1 1 min , 1.0 (A26) 2 d C g C v c g C b b b b b ( )= â â ï£«ï£ï£¬ ï£¶ ï£¸ï£· where db = average bicycle delay (s) gb = effective green time for bicyclists (typically the same as for motor vehicles, unless bicycle signals are used) vb = bicycle demand flow rate (bicycles/h) cb = bicycle capacity (bicycles/h) = 2,000 (gb/C) As is the case with pedestrian delay, access management techniques that result in a longer traffic signal cycle length being required can have an effect on bicycle delay. The HCM6 does not provide a bicycle delay method for unsignalized intersections, but Chapter 20 does reference a limited set of literature on the topic (1).

Applications Guidance for Selected Quantitative Analysis Methods 107 Truck Level of Service Section P in Exhibit 3 of NCHRP Report 825 (2) describes the use of a truck LOS measure developed in NCFRP Report 31 (5). Truck LOS is âa measure of the quality of service provided by a facility for truck hauling of freight as perceived by shippers and carriers. It is measured in terms of the percentage of ideal conditions achieved by the facility for truck operationsâ (2). Factors determining âideal conditionsâ consist of âa facility usable by trucks with legal size and weight loads, with no at-grade railroad crossings, that provides reliable truck travel at truck free-flow speeds, at low cost (i.e., no tolls)â (2). Equations A27 and A28, which are based on Equations 192 and 193 in NCHRP Report 825 (2), calculate a truck LOS index: %TKLOS 1 1 0.10 (A27) 200e U x( ) = + ( )â POTA 1 TTI 1 TOLL TFI 1 (A28)U x A B C D( ) ( ) ( ) ( ) ( )= Ã â + Ã â + Ã + Ã â where %TKLOS = truck LOS index as a percentage of ideal conditions (decimal) U(x) = utility of facility for truck shipments A = weighting parameter for reliability, sensitive to shipping distance = 5/ASL, where ASL = average shipment length (mi) = 200 (lower 48 states), 280 (Alaska), or 30 (Hawaii) POTA = probability of on-time arrival = 1 if TTImix is â¤ 1.33 (freeways and highways) or â¤ 3.33 (urban streets), where TTImix = mixed flow (auto and truck) travel time index, the ratio of FFS to actual speed B = weighting parameter for shipment time, sensitive to free-flow speed = â0.32/FFS, where FFS = truck free-flow speed (mph) TTI = truck travel time index for the study period, the ratio of truck free-flow speed to actual truck speed C = weighting parameter for shipment cost = â0.01 TOLL = truck toll rate ($/mile), a truck volumeâweighted average for all truck types D = weighting parameter for the facilityâs truck friendliness = 0.03 TFI = truck friendliness index, ranging from 1.00 (no constraints or obstacles to legal truck load and vehicle usage of facility) to 0.00 (no trucks may use facility) The value of TFI can be reduced below 1.00 to reflect sub-optimal conditions for trucks (e.g., weight restrictions or railroad grade crossings). NCFRP Report 31 provides guidance (5). The utility equation is weighted so a TFI of 0.60 or less will always result in truck LOS F. Equations A29 and A30, which are based on Equations 195 and 196 in NCHRP Report 825 (2), are used to determine the 95th percentile truck travel time index. The 95th percentile truck TTI is used in Table A8 to estimate the probability of on-time arrival, interpolating as necessary. TTI TTI (A29)mix fLA= Ã TTI 1 3.67 ln TTI (A30)95 ( )= + Ã

108 Guide for the Analysis of Multimodal Corridor Access Management where fLA = local adjustment factor to account for local truck driving behavior (decimal) (default = 1.00) TTI95 = 95th percentile truck LOS index The mixed-flow free-flow speed and average travel speed can be estimated by using HCM6 methods, as described earlier in the appendix. The local adjustment factor can be set to a value less than 1.00 when truck free-flow speeds are significantly below auto speeds (e.g., extended upgrades or downgrades or situations when the truck speed limit is lower than the auto speed limit and trucks comply). Once the probability of on-time arrival is determined, all of the information needed to cal- culate the facility utility and truck LOS index is available. If desired, the truck LOS index can be converted into a LOS letter, as shown in Table A9. A computational engine (spreadsheet) for calculating truck LOS is available in the Application Guides section of the HCM Volume 4 website (hcmvolume4.org). Example 11. An urban arterial has a mixed-flow free-flow speed of 36.5 mph, an average travel speed of 25.0 mph, and trucks are capable of traveling at the same speed as auto traffic (i.e., fLA = 1.00). Any truck with a legal load and dimensions can use the roadway. The road has no tolls and no railroad grade crossings. The mixed-flow TTI is (36.5/25.0) = 1.46, and the truck TTI equals the mixed-flow TTI, from Equation A29. The 95th percentile truck TTI is 2.39, from Equation A30. By interpolation in Table A8, the probability of on-time arrival is 99.58%. Next, from Equation A28, the facilityâs utility is â0.0041. Finally, from Equation A27, the truck LOS index is 81%, which corresponds to LOS B for an urban principal arterial (Class II, secondary facility) using Table A9. Truck TTI 95% Truck TTI Probability of On-Time Arrival (%) 1.20 1.67 100.00 1.40 2.23 99.89 1.60 2.72 98.93 1.80 3.16 96.51 2.00 3.54 92.67 2.20 3.89 87.70 2.40 4.21 81.91 Source: NCHRP Report 825, Exhibit 119 (2). Table A8. Look-up table for probability of on-time arrival. LOS Class I Primary Freight Facility (%) Class II Secondary Facility (%) Class III Tertiary Facility (%) A â¥ 90 â¥ 85 â¥ 80 B â¥ 80 â¥ 75 â¥ 70 C â¥ 70 â¥ 65 â¥ 60 D â¥ 60 â¥ 55 â¥ 50 E â¥ 50 â¥ 45 â¥ 40 F < 50 < 45 < 40 Source: NCHRP Report 825, Exhibit 117 (2). Note: Class I facilities include interstate highways and interregional rural principal arterials. Class II facilities include urban principal arterials and connectors to major intermodal facilities. Class III facilities include access roads to industrial areas, truck terminals, and truck stops. Table A9. Truck LOS criteria.

Applications Guidance for Selected Quantitative Analysis Methods 109 Crash Modification Factors Crash modification factors are tools for estimating the effect of selected access manage- ment techniques on a roadwayâs crash rate. They can be applied directly to a known, long-term (e.g., 5 years or more) crash history for a roadway to estimate the change in crash rate that would occur from implementing the technique. However, this approach is susceptible to regression-to- the-mean bias, in which the number of crashes during the study period happens to be higher or lower than the siteâs true long-term average. To address this issue, the Highway Safety Manual, 1st ed., recommends using, where possible, the Empirical Bayes method. This method combines a siteâs observed crash history with the predicted number of crashes for the site, based on data from other similar sites that have been incorporated into a safety performance function (SPF) for a roadway or intersection (6, 7). Two sources of CMFs are the HSM (7) and FHWAâs CMF Clearinghouse. CMFs from the HSM that are relevant to access management techniques are in the body of the guide. Newer CMFs may be available through the CMF Clearinghouse, which rates the quality of each CMF on a 5-star scale, based on the underlying studyâs design, sample size, standard error, potential study bias, and data source (8). Analysts should evaluate both the overall quality of CMFs obtained from the CMF Clearinghouse and their suitability to their study site (e.g., similar site charac- teristics or similar crash characteristics). The CMF Clearinghouse provides links to numerous resources on best practices for applying CMFs to safety analyses. Chapter 12 of the HSM (7) provides SPFs applicable to the following types of urban and suburban arterials: â¢ 2- and 4-lane undivided arterials â¢ 4-lane divided arterials â¢ 3- and 5-lane arterials with two-way left-turn lanes The SPF for predicting multiple-vehicle driveway-related crashes considers the number of driveways along a section of arterial, the volume of those driveways (major or minor), and the land use served by the driveway. Other SPFs, including those for predicting pedestrian and bicycle crashes, are not directly sensitive to access management techniques (except through the application of CMFs) but may be indirectly affected by changes in annual average daily traffic volumes that result from certain techniques. Bowman et al. Vehicle and Pedestrian Crash Models by Median Type Bowman et al. (9) studied the vehicle and pedestrian crash histories of 45 urban and suburban arterials evenly divided between the Atlanta, Phoenix, and Los Angeles regions. The authors developed six crash prediction models, one for each combination of vehicleâvehicle and vehicleâ pedestrian crashes along undivided roadways, roadways with a raised median, and roadways with a two-way left-turn lane. The models are negative binomial models with the general form of CR (A31)...1 1e a b x b xi i= ( )+ + + where CR = crash rate (crashes/100 million vehicle miles) a = model constant bi = model coefficient for independent variable i xi = value of independent variable i

110 Guide for the Analysis of Multimodal Corridor Access Management Table A10 provides the model coefficients for each combination of crash type and median type. The models are only applicable to urban and suburban arterials with characteristics within the range indicated in Table A11. The vehicleâvehicle TWLTL model has the best fit, while the vehicleâpedestrian models predict crash rate less accurately than the vehicleâvehicle models. In particular, the pedestrian models tend to underestimate the crash rate, particularly for roadways with TWLTLs. The authors also noted that the models indicate that the crash rate decreases as the posted speed increases and attributed this result to the fact that âhigher speeds . . . usually occur where development intensity, and hence vehicle interactions are less, thereby, resulting in lower accident frequency.â Example 12. A suburban arterial in an area with strip commercial land use has a 14-foot-wide TWLTL, a 35-mph posted speed, 30 driveways per mile, and 6 minor crossroads per mile. The crash-reporting threshold in this jurisdiction is $500. How might the vehicleâpedestrian and vehicleâvehicle crash rates change with the installation of a 14-foot raised median providing 6 crossovers per mile? Independent Variable Range Independent Variable Minimum Maximum Average daily traffic (veh/day) 11,500 60,000 Arterial length (mi) 0.5 5.6 Driveways per mile 4.3 90.0 Minor crossroads per mile 0.0 20.0 Crossovers per mile (raised medians only) 4.3 11.0 Median width (ft) (raised median) 3.0 40.0 Median width (ft) (TWLTL) 10.0 12.0 Traffic signals per mile 1.0 20.0 Posted speed limit (mph) 25.0 55.0 Number of lanes 2 6 Source: Bowman et al., Table 4 (9). Table A11. Ranges of independent variables used in developing Bowman et al. models. Independent Variable VehicleâVehicle VehicleâPedestrian Raised Median TWLTL Undivided Raised Median TWLTL Undivided Constant 7.20515 3.70539 1.88309 â0.88369 â0.97281 â1.10911 Accident reporting threshold ($) â0.00788 â0.00278 â0.003031 â â â Land use type = office (1 = yes, 0 = no) â0.44812 0.07227 1.06414 â1.65869 â 0.55689 Land use type = business (1 = yes, 0 = no) â â 0.65731 â â 0.73696 Area type (1 = central business district, 0 = suburban) â â 0.45652 1.03664 0.95036 1.43794 Number of lanes, excluding TWLTL â â â â â â0.25583 Median width (ft) â0.02755 0.03544 â â0.07866 â0.077121 â Number of minor crossroads per mile â â0.06057 â â â â Number of driveways per mile â 0.01294 0.01324 0.02163 â â Number of crossovers per mile 0.09615 â â â â â Posted speed limit (mph) â0.07002 â0.03389 â â0.03922 â â Note: A dash indicates a variable is not included in the model for this combination of crash type and median type. Source: Bowman et al., Table 3 (9). Table A10. Crash prediction model coefficients.

Applications Guidance for Selected Quantitative Analysis Methods 111 The predicted vehicleâpedestrian crash rate for the existing roadway is CR 0.13 crashes 100 million vehicle miles0.97281 0.95036 0 0.077121 14e= =( )( ) ( )( )( )â + â while the predicted vehicleâpedestrian crash rate with a median is CR 0.0670.88369 1.65869 0 1.03664 0 0.07866 14 0.02163 30 0.03922 35e= =( )( ) ( )( ) ( )( ) ( )( ) ( )( )( )â â + â + â Similarly, the predicted vehicleâvehicle crash rate for the existing roadway is 5.2 crashes per 100 million vehicle miles, compared with 2.7 with the median. Potts et al. Pedestrian Crash Model for Right-Turn Lanes Potts et al. (10) studied the vehicle and pedestrian crash histories of 103 four-leg signalized intersections in Toronto, Canada. In general, they found no significant difference in vehicleâ vehicle crashes when comparing channelized right-turn lanes with unchannelized right-turn lanes and no turn lanes. In a related case, a model addressing merging crashes on the cross street, the overall model showed channelized right turns had a lower crash rate than unchannelized right-turn lanes but a higher crash rate than with no turn lanes, but comparisons between the individual right-turn treatments showed no significant differences. However, a model of vehicleâ pedestrian crashes showed that the crash rate for unchannelized right-turn lanes was significantly higher than for channelized right turns or no turn lanes. Equation A32 shows the model form as follows: (A32)ped 12.13 0.02 STR 0.57 RTL 0.71 ln VOL1 0.50 ln VOL3N e= ( )( ) ( )( ) ( ) ( ) ( ) ( )( )â + + + + where Nped = predicted number of pedestrian crashes per year per approach STR = dummy variable for shared through/right lane = 1 if present; 0 otherwise RTL = dummy variable for unchannelized right-turn lane = 1 if present; 0 otherwise VOL1 = daily right-turning motor vehicle volume on the approach (veh/day) VOL3 = daily pedestrian volume on the two crosswalks conflicting with the right-turn movement (ped/day). Example 13. A right-turn lane is under consideration to be added at a signalized inter- section. The average daily right-turning volume on the approach is 1,700 vehicles per day, while the total average daily pedestrian volume on the crosswalks crossed by the right-turn movement is 400 pedestrians per day. What is the predicted number of vehicleâpedestrian crashes for each type of right-turn treatment, no right-turn lane, unchannelized right-turn lane, and channelized right-turn lane? Applying Equation A32 to the situation without a right-turn lane gives the following: 0.022 crash yearped 12.13 0.02 1 0.57 0 0.71 ln 1,700 0.50 ln 400N e= =( )( ) ( )( ) ( ) ( ) ( ) ( )( )â + + + + Similarly, the number of predicted crashes with an unchannelized right-turn lane is 0.038 crash/year and the number of predicted crashes with a channelized right-turn lane is 0.021 crash/year.

112 Guide for the Analysis of Multimodal Corridor Access Management Carter et al. Pedestrian and Bicycle Intersection Safety Indices Carter et al. (11, 12) developed safety indices that predict the safety ratings that pedestrians and bicyclists would give crossing and turning movements at intersections. Unlike other safety models presented earlier in the appendix, these indices do not predict crashes or crash rates. Rather, the indices prioritize locations for safety improvements and compare the relative safety ratings resulting from alternative improvement options. The pedestrian intersection safety index (ISI) can be applied to individual crosswalks at an intersection. The three bicycle ISIs are applied to the through, right-turn, and left-turn movements, respectively, on an intersection approach. The basis for the indices is a regression model that relates the ratings that expert panels (one for the pedestrian model and one for the bicycle models) gave to video clips to the conditions existing at the intersections shown in the video clips. However, the indices also incorporate variables found to be significant in a separate behavioral model, in which conflicts and avoid- ance maneuvers between vehicles and pedestrians and vehicles and bicyclists were observed and recorded. The video clips used for the pedestrian model included 68 intersection approaches in San Jose, California; MiamiâDade County, Florida; and Philadelphia, Pennsylvania. The video clips used for the bicycle models came from 67 intersection approaches in Gainesville, Florida; Eugene, Oregon; Portland, Oregon; and Philadelphia, Pennsylvania. Pedestrian Intersection Safety Index The pedestrian ISI is determined from Equation A33 as follows: ISI 2.372 1.867 Signal 1.807 Stop 0.335 ThruLanes 0.018 Speed 0.006 MainADT Signal 0.238 Commercial (A33) ped ( ) ( ) ( ) ( ) ( ) ( ) = â â + + + Ã + where ISIped = pedestrian intersection safety index (1 = best, 6 = worst) Signal = dummy variable for traffic signalâcontrolled crossing (1 = yes, 0 = no) Stop = dummy variable for stop signâcontrolled crossing (1 = yes, 0 = no) ThruLanes = number of through lanes on street being crossed (both directions) Speed = 85th percentile speed of street being crossed (mph) MainADT = average daily traffic volume on street being crossed, in thousands (1,000 veh/day) Commercial = dummy variable for commercial land use (e.g., retail, restaurants) being predominant in the area surrounding the crossing (1 = yes, 0 = no) The variables in the pedestrian ISI are not directly affected by access management techniques. However, both traffic speed and volume may be indirectly affected by some techniques. The model indicates that the pedestrian ISI worsens by 0.018 rating points for each 1 mph increase in traffic speeds on the street being crossed and by 0.006 points for each 1,000 increase in daily traffic volume on the street being crossed when the crossing is signalized. The âmain streetâ for the pedestrian ISI is the street being crossed, while the âmain streetâ for the bicycle ISI is the street on which the intersection approach of interest is located. Bicycle Intersection Safety Indices The bicycle ISIs are determined for each possible bicycle movement from an approach and are given by Equations A34 through A36:

Applications Guidance for Selected Quantitative Analysis Methods 113 ( ) ( ) ( ) ( ) ( ) ( ) ( ) = + + + + Ã + Ã + Ã + ISI 1.13 0.019 MainADT 0.815 MainHiSpeed 0.650 TurnVeh 0.470 RTLanes BikeLane 0.023 CrossADT NoBikeLane 0.428 Signal NoBikeLane 0.200 Parking (A34) bike,th ISI 1.02 0.027 MainADT 0.519 RTCross 0.151 CrossLanes 0.200 Parking (A35) bike,rt ( ) ( ) ( ) ( )= + + + + ISI 1.10 0.025 MainADT 0.836 BikeLane 0.485 Signal 0.736 MainHiSpeed BikeLane 0.380 LTCross NoBikeLane 0.200 Parking (A36) bike,lt ( ) ( ) ( ) ( ) ( ) ( ) = + + + + Ã + Ã + where ISIbike,th = bicycle intersection safety index for the through movement (1 = best, 6 = worst) ISIbike,rt = bicycle intersection safety index for the right-turn movement (1 = best, 6 = worst) ISIbike,lt = bicycle intersection safety index for the left-turn movement (1 = best, 6 = worst) MainADT = average daily traffic volume on the main street, in thousands (1,000 veh/day) MainHiSpeed = dummy variable for main street speed limit â¥35 mph (1 = yes, 0 = no) TurnVeh = dummy variable for presence of turning-vehicle traffic across the path of bicyclists at the intersection (1 = yes, 0 = no) (e.g., dummy variable is no with a bike-lane crossover of a right-turn lane or where right turns are prohibited) RTLanes = number of right-turn lanes on the main street approach (0, 1, or 2) BikeLane = dummy variable for the presence of a bike lane or bike-lane crossover (1 = yes, 0 = no) CrossADT = average daily traffic volume on the cross street, in thousands (1,000 veh/day) NoBikeLane = dummy variable for the absence of a bike lane or bike-lane crossover (1 = yes, 0 = no) Signal = dummy variable for traffic signal at intersection (1 = yes, 0 = no) Parking = dummy variable for on-street parking on the main street approach (1 = yes, 0 = no) RTCross = number of traffic lanes for cyclists to cross or enter to make a right turn (assumes the bicyclist is riding in a right- or left-side bicycle lane or on the right side of the street) CrossLanes = number of through lanes on the cross street LTCross = number of traffic lanes for cyclists to cross or enter to make a left turn (assumes the bicyclist is riding in a right- or left-side bicycle lane or on the right side of the street and does not make a two-stage left turn) The combination of a right-turn lane and a bicycle lane, where right-turning traffic crosses over the bicycle lane, makes the through-movement bicycle ISI worse by 0.47 rating points. Traffic signals make the through-movement bicycle ISI worse by 0.428 points when there is no bicycle lane and make the left-turn bicycle ISI worse by 0.485 points. Removing on-street park- ing on the intersection approach improves the bicycle ISI for all movements by 0.2 points. Access management techniques are unlikely to affect the other bicycle ISI components except they may indirectly affect the ones related to traffic volume.

114 Guide for the Analysis of Multimodal Corridor Access Management Carter et al. (11) noted that the following conditions appeared to result in lower survey ratings than predicted by the model; however, because these conditions appeared in only one or a few videos, the conditions were not able to be modeled. They are â¢ Slip lane or channelized right-turn lane â¢ Pavement irregularities (e.g., broken asphalt, tracks, gutters, or grates) â¢ High crossing pedestrian volumes â¢ Vehicles stopped in the bicycle travel space to load or unload â¢ Bicycle lane to the right of an exclusive right-turn lane â¢ Perpendicular on-street parking â¢ Buses entering or exiting an area where they can potentially interact with bicyclists â¢ Offset intersections â¢ Parking dimensions (e.g., width of parallel parking spaces or bike lane proximity to parking) References 1. Highway Capacity Manual: A Guide for Multimodal Mobility Analysis, 6th ed. Transportation Research Board, Washington, D.C., 2016. 2. Dowling, R., P. Ryus, B. Schroeder, M. Kyte, T. Creasey, N. Rouphail, A. Hajbabaie, and D. Rhoades. NCHRP Report 825: Planning and Preliminary Engineering Applications Guide to the Highway Capacity Manual. Transportation Research Board, Washington, D.C., 2016. 3. Bonneson, J. A., M. P. Pratt, and M. A. Vandehey. Predicting the Performance of Automobile Traffic on Urban Streets: Final Report. NCHRP Project 03-79. Transportation Research Board of the National Academies, Washington, D.C., Jan. 2008. 4. Kittelson & Associates, Inc.; Parsons Brinckerhoff; KFH Group, Inc.; Texas A&M Transportation Institute; and Arup. TCRP Report 165: Transit Capacity and Quality of Service Manual, 3rd ed. Transportation Research Board of the National Academies, Washington, D.C., 2013. 5. Dowling, R., G. List, B. Yang, E. Witzke, and A. Flannery. NCFRP Report 31: Incorporating Truck Analysis into the Highway Capacity Manual. Transportation Research Board of the National Academies, Washington, D.C., 2014. 6. Federal Highway Administration. Crash Modification Factors in Practice: Quantifying Safety in the Roadway Safety Management Process. Publication FHWA-SA-13-010. Federal Highway Administration, Washington, D.C., 2013. 7. Highway Safety Manual, 1st ed. American Association of State Highway and Transportation Officials, Washington, D.C., 2010. 8. About the Star Quality Rating. Crash Modification Factor Clearinghouse. http://www.cmfclearinghouse. org/sqr.cfm. Accessed Oct. 2, 2017. 9. Bowman, B., R. Vecellio, and J. Miao. Vehicle and Pedestrian Accident Models for Median Locations. Journal of Transportation Engineering, Vol. 121, No. 6, American Society of Civil Engineers, Washington, D.C., Nov.âDec. 1994, pp. 531â537. 10. Potts, I. B., D. W. Harwood, K. M. Bauer, D. K. Gilmore, J. M. Hutton, D. J. Torbic, J. F. Ringert, A. Daleiden, and J. M. Barlow. NCHRP Web-Only Document 208: Design Guidance for Channelized Right-Turn Lanes. Transportation Research Board of the National Academies, Washington, D.C., July 2011. 11. Carter, D. L., W. W. Hunter, C. V. Zegeer, J. R. Stewart, and H. F. Huang. Pedestrian and Bicyclist Intersection Safety Indices: Final Report. Report FHWA-HRT-06-125. Federal Highway Administration, Washington, D.C., Nov. 2006. 12. Carter, D. L., W. W. Hunter, C. V. Zegeer, J. R. Stewart, and H. F. Huang. Pedestrian and Bicyclist Intersection Safety Indices: User Guide. Report FHWA-HRT-06-130. Federal Highway Administration, Washington, D.C., April 2007.