Planning and Preliminary Engineering Applications Guide to the Highway Capacity Manual (2016)

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.

P A R T 3 High-Level Analyses The sections in Part 3 of the Guide describe high-level analysis methods that work best when evaluating highway systems at an areawide level. These methods enable the analyst to cover large geographic areas with hundreds of miles of highways very efficiently. The methods presented here can be applied to monitoring existing system performance and to forecasting future performance. Part 3 includes the following content: Q. Corridor quick estimation screenline analyses a. Screening for capacity and multimodal LOS hot spots b. Screening alternatives for capacity impacts R. Areas and systems a. Computational tools b. Data needs c. Estimation of demand model inputs d. Performance measure estimation S. Roadway system monitoring a. Travel time datasets b. Identification of problem spots through the travel time index c. Identification of multimodal problem spots d. Diagnosis of causes of mobility problems

173 Q. Corridor Quick Estimation Screenline Analysis 1. Overview Transportation planners assess future investments in a cor- ridor based on the performance of the freeways and streets that make up the corridor transportation system. The perfor- mance of the corridor system and its components are often estimated through a travel demand and analysis forecasting process combined with either a microscopic or macroscopic traffic operations model. This process requires a variety of inputs and outputs which the HCM can provide, including capacity, queues, delay, travel speeds, and level of service (LOS). The consistency of default values used across facili- ties in a corridor should be considered when conducting a corridor analysis. This section presents a high-level quick estimation method for quickly assessing available corridor capacity. More detailed corridor analyses would employ the high-level methods described next in Section R, or they would employ the medium-level methods described earlier in Part 2. 2. Screening for Capacity and Multimodal LOS Hot Spots For the purposes of quickly screening the corridor for multimodal LOS problem (hot) spots, one can divide the corridor into a set of screenlines where the demands are checked against HCM service volume tables for auto, transit, bicycle, and pedestrian LOS, as illustrated in Exhibit 120. The screenlines are located by the analyst at key points, particularly choke points in the cor- ridor. For example, in Exhibit 120, Screenlines 1 and 6 are located at key choke points in the cor- ridor with the fewest parallel facilities available to carry traffic. The other screenlines are located at spots where corridor demand may significantly change from section to section (often between freeway interchanges). The forecasted AADT for each freeway or major surface-street crossing the screenline is com- pared to the values in the appropriate service volume table for the auto mode to assess whether a facility is likely to operate at a level of service acceptable to the agency. For the non-auto modes, it is necessary to perform specific analyses of the conditions present at the screenline. Note that the use of screenlines for modal analysis will not catch intersection problem spots, so key intersections should be checked as well. The screenline analysis may indicate sections of

174 Planning and Preliminary Engineering Applications Guide to the Highway Capacity Manual the corridor where high vehicle volumes suggest that the intersections should be checked for potential LOS problems. Exhibit 121 illustrates a set of corridor screenline checks for the freeway and arterial corridor shown in Exhibit 120. 3. Screening Alternatives for Capacity Impacts A similar screening approach can be used to assess the relative effects of various capacity improvement alternatives on screenline demand-to-capacity ratios and multimodal LOS. Exhibit 122 illustrates such an analysis for the option of adding auxiliary lanes between the two freeway interchanges at screenline #3. The first table shows the before condition. The second table shows the after condition. In this case, the auxiliary lanes were estimated to increase capacity by 5% (the proportion of traffic on the freeway using the on-ramp and the off-ramp which would be likely to take advan- tage of the auxiliary lane). The increased capacity will also allow some traffic that currently used the arterial to shift to the freeway. This diversion is estimated to shift about 5% of arterial traffic to the freeway (increasing freeway traffic by 2%). The result is that the auxiliary lanes are estimated to improve the freeway v/c ratio from 105% to 102%, and the arterial v/c ratio from 57% to 54% at Screenline #3. Note that because the effects being estimated are in the 2% to 5% range, the effects are not rounded to the nearest 100 vehicles per day. Exhibit 120. Example corridor screenlines.

Q. Corridor Quick Estimation Screenline Analysis 175 AADT per Lane Mixed Flow v/c Screening Multimodal LOS Screening Freeway Arterial Screenline Freeway Arterial Freeway Arterial Corridor Auto Auto Transit Bike Ped 1 16,500 3,300 83% 24% 59% LOS D LOS C LOS F LOS D LOS E 2 16,100 4,800 81% 36% 63% LOS D LOS C LOS F LOS D LOS E 3 20,900 7,700 105% 57% 86% LOS F LOS D LOS D LOS E LOS D 4 16,600 7,600 83% 56% 72% LOS D LOS D LOS D LOS E LOS D 5 17,300 3,500 87% 26% 62% LOS D LOS C LOS F LOS D LOS E 6 13,900 N/A 70% N/A 70% LOS C N/A N/A N/A N/A Notes: N/A = not applicable (arterial not present). Freeway capacity assumed to be 19,900 AADT/lane per Exhibit 19 in Section H4. Arterial capacity assumed to be 13,500 AADT/lane per Exhibit 46 in Section K4. Transit, bicycle, and pedestrian LOS would be computed at the screenlines following the defaults given in Section O4 and the procedures given in HCM Chapter 18, Urban Street Segments. Transit LOS is F at Screenlines 1, 2, and 5 because no transit service is provided outside the city limits. Bicycle LOS is D outside the city (Screenlines 1, 2, and 5) due to the provision of paved shoulders with no parking. Inside the city, higher traffic volumes and on-street parking contribute to LOS E conditions. Pedestrian LOS is E outside the city due to the lack of a sidewalk and buffer from traffic. Sidewalks and the presence of on-street parking contribute to LOS D conditions at Screenlines 3 and 4. Exhibit 121. Example corridor screenline volume-to-capacity ratio and LOS checks. Demand (AADT/ln) Capacity (AADT/ln) Volume-to-Capacity Ratio Screenline Freeway Arterial Freeway Arterial Freeway Arterial Corridor Before 1 16,500 3,300 19,900 13,500 83% 24% 59% 2 16,100 4,800 19,900 13,500 81% 36% 63% 3 20,900 7,700 19,900 13,500 105% 57% 86% 4 16,600 7,600 19,900 13,500 83% 56% 72% 5 17,300 3,500 19,900 13,500 87% 26% 62% 6 13,900 N/A 19,900 N/A 70% N/A 70% After 1 16,500 3,300 19,900 13,500 83% 24% 59% 2 16,100 4,800 19,900 13,500 81% 36% 63% 3 21,318 7,282 20,895 13,500 102% 54% 83% 4 16,600 7,600 19,900 13,500 83% 56% 72% 5 17,300 3,500 19,900 13,500 87% 26% 62% 6 13,900 N/A 19,900 N/A 70% N/A 70% Note: N/A = not applicable (arterial not present). Exhibit 122. Screenline volume-to-capacity analysis of auxiliary lanes.

176 R. Areas and Systems 1. Overview Transportation planners assess future investments based on the performance of the freeways and streets that make up a regional transportation system. The performance of the sys­ tem and its components are often estimated through a travel demand and analysis forecasting process. This process requires a variety of inputs which the HCM can provide, including pre­ diction of travel speeds. The procedure is performed for all of the highway sub­ systems in five steps. 1. The necessary input data are assembled, 2. The free­flow speed of the links is computed, 3. The capacity of each link is computed, 4. The mean link speeds are computed, and 5. The travel time and other performance measures are com­ puted for all the links and summed for each subsystem. Look­up tables of capacity and free­flow speed defaults can be used to shortcut two of the steps (Steps 2 and 3), but poor choice of capacities, free­flow speeds, or both can significantly reduce the accuracy of the speeds estimated using this procedure. In addition, the consistency of default values applied to the facilities of the same area type (e.g., urban, rural) within the study area should be considered. 2. Computational Tools Planning analyses of multimodal transportation systems in large areas are best performed in a travel demand modeling environment, which can equilibrate the forecasted demands between facilities and modes based on the forecasted performance. The guidance provided here is on the use of HCM procedures to generate the key performance analysis inputs required by typical demand models. These procedures are generally performed manually with spreadsheet assis­ tance to facilitate and document the calculations. 3. Data Needs Exhibit 123 lists the required input for the analysis of areawide systems of facilities. Individual performance measures may require only a subset of these inputs.

R. Areas and Systems 177 4. Estimation of Demand Model Inputs The HCM can support the estimation of two key demand model inputs related to the highway network: the free-flow speed of a link and its capacity. Free-Flow Speed Estimation The free-flow speed of a facility is defined as the space mean speed of traffic when volumes are so light that they have negligible effect on speed. Free-flow speed excludes intersection control delay. Options for estimating free-flow speed include the following: • The best technique for estimating free-flow speed is to measure it in the field under light traffic conditions. Such observations can be obtained for highways on the National Highway System (http://www.fhwa.dot.gov/planning/national_highway_system/) from travel time reliability databases such as the National Performance Management Research Data Set (http://www.ops. fhwa.dot.gov/freight/freight_analysis/perform_meas/vpds/npmrdsfaqs.htm). Commercial datasets of travel times and speeds may also be available. One caution is that free-flow speeds must be measured during low-flow conditions when sample sizes may not be large. • When and where direct observations of free-flow speeds are not available, or are difficult to obtain, the next-best technique is to use the procedures defined in the HCM to estimate the free-flow speeds. Locally developed look-up tables of HCM-estimated free-flow speeds can be generated using default inputs by facility type and area type to automate the process. • If posted speed limits are available, the posted speed limit may be adjusted by the analyst to estimate the free-flow speed. One approach is to assume the free-flow speed is 5 miles per hour greater than the posted speed limit. Freeway Subsystem The free-flow speeds for all freeway subsystem links (weaving, merge, diverge, and basic seg- ments for general purpose lanes, and their equivalents for managed lanes) can be measured in the field or estimated using the procedures described in HCM Chapter 12, Basic Freeway and Required to Estimate Input Data (units) FFS Cap Spd Que Rel Default Value Facility type • • • • • Defaults by area and facility type Segment design geometry • • • • • Defaults by area and facility type Terrain type • • • • Must be provided Percentage heavy vehicles (%) • • • • 10% (rural), 5% (urban) Peak hour factor (decimal) • • • • 0.88 (rural), 0.95 (urban) CAF for driver pop. (decimal) • • • • 1.00 Number of directional lanes • • • • Must be provided Segment length (mi) • • • Must be provided Directional demand (veh/h) • • • Output of travel model Notes: FFS = free-flow speed (mph), Cap = Capacity (veh/h/ln), Spd = Speed (mph), Que = Queue (veh), Rel = travel time reliability. Facility type = freeway, arterial by control type (e.g., signal, roundabout), multilane highway, or two-lane rural highway. Segment design geometry varies by facility type but often includes average lane widths, shoulder widths, and access point density. Terrain type = level, rolling, mountainous. CAF for driver pop. = capacity adjustment factor for driver population, used to reduce capacities due to unfamiliar drivers. Exhibit 123. Required roadway segment data for area and roadway systems analysis.

178 Planning and Preliminary Engineering Applications Guide to the Highway Capacity Manual Multilane Highway Segments (see HCM Equation 12­1). The procedures require information on lane widths, lateral clearances, number of lanes, and interchange spacing. Default values for lane widths and lateral clearance are provided in HCM Exhibit 12­18. Rural Highway Subsystem The free­flow speeds for two­lane and multilane highway links can be measured in the field or estimated using the procedures described in HCM Chapter 15, Two­Lane Highways (HCM Equa­ tion 15­2), and Chapter 12, Basic Freeway and Multilane Highway Segments (HCM Equation 12­1), respectively. The procedures require information on lane widths, lateral clearances, number of lanes, median type, and access point density. Default values for missing data are provided in HCM Exhibit 15­5 for two­lane highways and HCM Exhibit 12­18 for multilane highways. Arterial/Collector Urban Street Subsystem The free­flow speed for arterial and collector streets can be measured in the field, or estimated using the procedures described in HCM Chapter 18, Urban Street Segments. Default values are provided in HCM Exhibit 18­5. Note that the HCM also defines a “base free­flow speed” (HCM Equation 18­3) which must be converted to “free­flow speed”(HCM Equation 18­5) before it can be used in planning analyses. Creating a Look-up Table of Free-Flow Speed Defaults The analyst may wish to develop a look­up table of free­flow speeds based upon local surveys and the functional class and the area type in which a link is located in order to simplify the estimation of free­flow speeds. Depending upon local conditions, the analyst may wish to clas­ sify links by area type (e.g., downtown, urban, suburban, rural); terrain type (i.e., level, rolling, mountainous); and frontage development types (e.g., commercial, residential, undeveloped). An illustrative example is provided as Exhibit 124. The accuracy of the speed estimation procedure is highly dependent on the accuracy of the free­flow speed and capacity used in the computations. Great care should be taken in the Facility Type Area Type Default Free-Flow Speed (mph) Freeway Downtown 55 Urban 60 Suburban 65 Rural 70 Arterial Downtown 25 Urban 35 Suburban 45 Rural 55 Collector Downtown 25 Urban 30 Suburban 35 Rural 40 Note: Facility types, area types, and default speed values are illustrative. Where the analyst has ready access to link-specific posted speed limits, the method of adding a fixed adjustment (such as 5 mph) to the posted speed limit may be appropriate. Other categories and values may be more appropriate for a particular study area. Exhibit 124. Illustrative look-up table of free-flow speed defaults.

R. Areas and Systems 179 creation of local look­up tables so that they accurately reflect the free­flow speeds present in the locality. Capacity Estimation Unlike travel demand models, where a roadway link represents all intersections and segments within the specified length of roadway, the HCM deals with segments (between intersections) and intersections separately, before combining them into a facility analysis. The discussion herein, therefore, combines the separate HCM segment and intersection procedures for estimating capac­ ity into a single “mini­facility” approach able to accommodate the combined effects of segment and intersection capacity on the total link capacity. In general, a link’s capacity will be determined for demand modeling purposes by the inter­ section or segment with the lowest through capacity within that link. Options for estimating link capacity include: • The best technique for estimating capacity is to measure it in the field at the bottleneck. • Field measurements of capacity are often not feasible, so the next­best technique is to employ the HCM’s procedures. The HCM’s capacities are computed in terms of passenger cars per hour and must be converted to mixed vehicle capacities. This conversion is needed to allow the use of actual vehicular demand values in the queuing and delay calculation steps, rather than passenger car equivalents. The conversion is performed by applying the HCM’s recommended demand adjustment factors to the passenger car capacity. The following equations for freeways, multilane highways, two­lane rural roads, and arterials illustrate the application of the demand adjustment factors to the passenger car capacity. Freeway Subsystem The following equation, adapted from HCM Equation 12­9 to yield capacity adjustment rather than volume adjustment, is used to compute the mixed vehicle capacity of a freeway link at its critical point. The critical point is the point on the link with the lowest throughput capacity. c PCCap N f PHF CAFhv= × × × × Equation 197 where c = capacity (veh/h), PCCap = HCM passenger car capacity from Exhibit 125 (pc/h/ln), N = number of through lanes, ignoring auxiliary and “exit only” lanes, fhv = heavy vehicle adjustment factor from Equation 14 (freeways) or Equation 38 (multilane highways), Free-Flow Speed Freeway Section Type (mph) Basic Ramp Weaving 75 2,400 2,400 2,160 70 2,400 2,400 2,160 65 2,350 2,350 2,115 60 2,300 2,300 2,070 55 2,250 2,250 2,025 Source: HCM (2016), Exhibit 12-4. Note: Table entries are passenger car capacities (pc/h/ln). Exhibit 125. HCM passenger car capacities (pc/h/ln) for freeway general purpose lanes.

180 Planning and Preliminary Engineering Applications Guide to the Highway Capacity Manual PHF = peak hour factor, and CAF = capacity adjustment factor (locally developed and applied to match field measure­ ments of capacity, when available), default is 1.00. Exhibit 125 provides HCM passenger car capacities for the general purpose lanes within different types of freeway sections. The passenger car capacities are reduced 10% for weaving sections. Exhibit 126 provides HCM passenger car capacities for managed lanes (e.g., carpool or high­occupancy toll lanes) on the basis of the form of separation between the managed and general purpose lanes. HCM Chapter 12 provides procedures for determining the adjustment factors used in Equation 197. Exhibit 21 in this Guide (Section H5) provides suggested default values for the adjustment factors. Rural Multilane Highways Equation 197 for freeways is also used to compute the mixed vehicle capacity of a multilane high­ way (i.e., a highway where the traffic signal spacing exceeds 2 miles). Different HCM passenger car capacities are used for multilane highways and the adjustment factors may take on different values. Exhibit 32 in this Guide (Section I5) provides default adjustment factor values that can be used with Equation 197. Exhibit 127 provides the HCM’s passenger car capacities for multilane highways. Free-Flow Speed Managed Lane Separation Type (mph) Continuous Access Buffer 1 Buffer 2 Barrier 1 Barrier 2 75 1,800 1,700 1,850 1,750 2,100 70 1,750 1,650 1,800 1,700 2,050 65 1,700 1,600 1,750 1,650 2,000 60 1,650 1,550 1,700 1,600 1,950 55 1,600 1,500 1,650 1,550 1,900 Source: HCM (2016), Exhibit 12-11. Notes: Table entries are passenger car capacities (pc/h/ln). Continuous access separation allows vehicles to enter or leave the managed lane at any point. Buffer types separate the managed lane(s) from the general purpose lanes by paint stripes; vehicles can only enter or leave the managed lane(s) at designated points. Buffer 1 provides one managed lane and Buffer 2 provides two managed lanes. Barrier types separate the managed lane(s) from the general purpose lanes by physical barriers; vehicles can only enter or leave the managed lane(s) at designated points. Barrier 1 provides one managed lane and Barrier 2 provides two managed lanes. Exhibit 126. HCM passenger car capacities (pc/h/ln) for freeway managed lanes. Free-Flow Speed (mph) HCM Passenger Car Capacity (pc/h/ln) 70 2,300 65 2,300 60 2,200 55 2,100 50 2,000 45 1,900 Source: HCM (2016), Exhibit 12-4. Exhibit 127. HCM passenger car capacities for rural multilane highways.

R. Areas and Systems 181 Rural Two-Lane Highways and Roads Equation 198 (adapted from HCM Equation 15­3) is used to compute the mixed vehicle capacity in one direction of a two­lane road (one lane each direction) that has traffic signals (or other intersection control such as all­way stops or roundabouts that slow down through move­ ments) spaced more than 2 miles apart. c PCCap f f PHFhv g= × × × Equation 198 where c = capacity (veh/h), PCCap = HCM passenger car capacity = 1,600 for a single direction (pc/h/ln), fg = grade adjustment factor for average travel speed (unitless), fhv = heavy vehicle adjustment factor for average travel speed (unitless) from Equation 48, and PHF = peak hour factor. Exhibit 38 in this Guide (Section J5) provides suggested default values for percentage of heavy vehicles and peak hour factor. HCM Exhibit 15­9 (grade adjustment factor) and HCM Equation 15­4 (heavy vehicle adjustment factor) are used to compute these factors. To reduce computational effort, the analyst may apply the HCM’s adjustment factors for average travel speed, rather than perform a second computation of the adjustments using the HCM’s adjust­ ment factors for percent time­spent­following. Urban Arterial and Collector Streets The capacity of an urban arterial or collector street link with multiple choke points (signals, all­way stops, lane drops, roundabouts, etc.) is determined by examining the through move­ ment capacity at each choke point on the arterial link. The choke point with the lowest through capacity determines the overall capacity of the arterial link for demand modeling and high­level planning analysis purposes. (The term “link,” as commonly used in demand modeling, refers to a collection of road segments and intersections that are together represented in the model by a single free­flow speed and capacity, and for which the demand model produces a single estimate of demand and average speed.) Equation 199, adapted for peak hour factor and signal timing adjustments from HCM Equa­ tion 19­8, is used to compute the through capacity of one direction of travel at a signal. c S N f f f f f f f f f f f f f PHF g Co w hvg p bb a LU LT RT Lpb Rpb wz ms sp ( )= × × × × × × × × × × × × × × × × Equation 199 where c = capacity (veh/h), So = base saturation flow rate (pc/h/ln) = 1,900 for metropolitan areas with populations of 250,000 or greater and 1,750 otherwise, N = Number of Lanes in the lane group, fw = adjustment factor for lane width (decimal), fhvg = adjustment factor for combined effect of grade and heavy vehicles in the traffic stream (decimal), fp = adjustment factor for existence of a parking lane and parking activity adjacent to lane group (decimal),

182 Planning and Preliminary Engineering Applications Guide to the Highway Capacity Manual fbb = adjustment factor for blocking effect of local buses that stop within intersection area (decimal), fa = adjustment factor for area type (decimal), fLU = adjustment factor for lane utilization (decimal), fLT = adjustment factor for left­turn vehicle presence in a lane group (decimal), fRT = adjustment factor for right­turn vehicle presence in a lane group (decimal), fLpb = pedestrian–bicycle adjustment factor for left­turn groups (decimal), fRpb = pedestrian–bicycle adjustment factor for right­turn groups (decimal), fwz = adjustment factor for work zone at intersection (decimal), fms = adjustment factor for downstream lane blockage (decimal), fsp = adjustment factor for sustained spillback (decimal), PHF = peak hour factor (decimal), and g/C = ratio of effective green time per cycle. Exhibit 19­11 in HCM Chapter 19, Signalized Intersections, provides suggested default values for the inputs needed to compute the saturation flow adjustment factors for signalized inter­ sections. For arterials where all­way stops or roundabouts control the link capacity, the proce­ dures in HCM Chapters 21 or 22, respectively, should be used to estimate the through movement capacity at each intersection. Capacity Look-up Table The accuracy of the speed estimates produced by a demand model are highly dependent on the accuracy of the estimated capacity for the facility. Consequently, it is recommended that the analyst use capacities that are specific to each link whenever possible. However, it is recognized that this link­specific approach is not feasible when evaluating thousands of links in a metropolitan area. The analyst may select sets of default values for the various capacity adjustment factors that vary by functional class (e.g., freeway, highway, arterial, collector, local), area type (e.g., downtown, urban, suburban, rural), terrain type (i.e., level, rolling, mountainous), and other conditions. These default values may be substituted into the above capacity equations to develop a set of look­up tables of link capacities that vary by functional class, area type, general terrain, and number of lanes. The effects of the heavy vehicle, constrained geometry, peaking, and other factors generally reduce the base capacity (expressed in terms of passenger cars per hour per lane) by 10% to 20%. The 10% reduction is typical of facilities on level terrain that meet agency design standards, carry modest volumes of heavy vehicles (5% or less), and have typical peak hour factors in the range of 92% to 97%. The 20% reduction is typical of facilities with geometric constraints, relatively high heavy vehicle use, or higher demand peaking. The saturation flow rates for signalized arterials must be first discounted by the g/C ratio (per­ cent green time) for the through lanes on the arterial. Research in Florida (Florida DOT 2013) suggests that g/C ratios of 41% are a practical maximum for suburban arterials with left­turn phases and typical left­turn volumes at major intersections. Higher values may be achieved for the mainline through lanes at intersections without left­turn phases, on one­way streets, and at intersections of major streets with a minor cross street. Other values can be (and should be) selected based on local experience. Exhibit 128 illustrates the construction of a per­lane capacity look­up table from which the analyst can select capacity values from the 90% and 80% columns depending on the analyst’s general assessment of facility conditions. Unique situations may warrant greater capacity reduc­ tions than shown in this illustrative table.

R. Areas and Systems 183 5. Performance Measure Estimation Performance measure estimation is accomplished mostly within the travel demand model environment. This section focuses on the use of HCM procedures to estimate the roadway­ related performance, plus two performance measures not typically produced by travel demand models: queuing and reliability. The discussion is split between the estimation of auto­related performance measures and multimodal performance measures (truck, transit, bicycle, and pedestrian). Auto-Related Performance Measures Demand-to-Capacity Ratio The demand­to­capacity ratio for each link is typically output by the travel demand model, based on the analyst’s input capacity. Average Travel Speed and Average Travel Time The mean vehicle speed for through trips on a link is computed by the travel demand model during a traffic assignment process using either a speed–flow equation or a more­sophisticated approach that combines link delay with an estimate of mode delay. Class Area or Facility Type Free-Flow Speed (mph) Assumed g/C HCM Capacity (pc/h/ln) 90% HCM Capacity (veh/h/ln) 80% HCM Capacity (veh/h/ln) Freeway Downtown 55 N/A 2,250 2,000 1,800 Urban 60 N/A 2,300 2,100 1,800 Suburban 65 N/A 2,350 2,100 1,900 Rural 70 N/A 2,400 2,200 1,900 Arterial Downtown 25 0.45 860 800 700 Urban 35 0.45 860 800 700 Suburban 45 0.41 780 700 600 55 N/A 2,100 1,900 1,700 Rural Two-lane 55 N/A 1,600 1,400 1,300 Collector Downtown 25 0.41 780 700 600 Urban 30 0.41 780 700 600 Suburban 35 0.37 700 600 600 45 N/A 1,900 1,700 1,500 Rural Two-lane 45 N/A 1,600 1,400 1,300 Notes: N/A = not applicable. capacity value, as well as a conversion from passenger car capacity (pc/h/ln) to mixed vehicle The 90% and 80% HCM capacity values incorporate 10% or 20% reductions, respectively, from the HCM capacity (veh/h/ln). The 90% HCM capacity column is used where the effects of substandard geometry, heavy vehicles, and demand peaking are expected to be negligible to minor. The 80% column is used where these factors are expected to have greater effects on capacity. suburban arterial and collector values would be reduced by 8% for a smaller metropolitan or urban Table prepared for a metropolitan area with a population greater than 250,000. Downtown, urban, and area. Downtown, urban, and suburban arterial and collector values can assumptions by the proportion of the analyst’s g/C value to the value shown in the table. Categories and values are illustrative. Other categories and values may be more appropriate. be adjusted for different g/C Exhibit 128. Illustrative per-lane capacity look-up table.

184 Planning and Preliminary Engineering Applications Guide to the Highway Capacity Manual Demand Models That Compute Mode Delay. If mode delay is used by the demand model, the following equation is used to compute average link speed, including mode delay. 3,600 Equation 200= + S L R D where S = mean link speed (mph), L = link length (mi), R = link travel time (h), and D = mode delay for link (s). The mode delay D is computed only for the traffic signal or stop­ or yield­controlled inter­ section at the end of the link (all other intersection related delays that occur in the middle of the link are incorporated in the link travel time calculation). The mode delay procedures described in Section L5 (signalized intersections), M5 (stop­controlled intersections), or N4 (roundabouts) can be used. The calculation requires information on all of the intersection approaches at the mode so that the mean approach delay for each link feeding the intersection can be computed. Demand Models that Do Not Compute Mode Delay. If the available travel demand model software package is unable to compute mode delay, the delay can be approximated by using the mode approach capacity rather than the link capacity in the computation of travel time T. In this situation, the mode delay is set to zero in Equation 200. When a model does not model mode delay separately, it is necessary to include the intersection control delay with zero volume in the link’s estimated free­flow speed. This diverges from the HCM practice of excluding intersection control delay from the free­flow speed for urban streets. The following equation, commonly called the BPR or Bureau of Public Roads equation (Cambridge Systematics 2010), can be used to quickly estimate approximate link travel times. T T AxB( )= +1 Equation 2010 where T = link travel time (h), T0 = link travel time at low near­zero volumes (h), A = ratio of speed at capacity to free­flow speed, minus one (standard value = 0.15), B = parameter that affects the rate at which speed drops (standard value = 4.0), and x = the link demand­to­capacity ratio (unitless). The calibration parameter A is selected so that the travel time equation will predict the mean speed of traffic when demand is equal to capacity. Substituting x = 1.00 in the travel time equa­ tion and solving for A yields: A S S f c = −1 Equation 202 where A = BPR speed at capacity calibration parameter, Sc = mean speed at capacity (mph), and Sf = mean free­flow speed (mph). The calibration parameter B is selected to predict the approximate delay when demand exceeds capacity for a target range of demand­to­capacity ratios (generally in the range of 1.7 to 1.9).

R. Areas and Systems 185 The BPR curve (Equation 201) estimates the proportional increase in travel time at a given demand­to­capacity ratio. As speed and travel time are inversely related, a given proportional increase in travel time produces an identical proportional decrease in speed (i.e., doubling the travel time from free­flow conditions implies halving the speed). Therefore, Equation 201 is modified as shown in Equation 203 to work with free­flow speed as an input. Equation 203 is entered with the free­flow speed S0 (in mph) and the A and B parameters defined previously, and the link speed S (in mph) is computed. S S Ax B( )= +1 Equation 2030 Exhibit 129 shows recommended capacities, speeds at capacity, and values of the A and B parameters that were selected to reproduce the travel times at capacity predicted by the HCM’s analysis procedures. It is important to note that there are many other speed–flow functions besides the BPR curve that can be and are used in demand modeling to predict the impacts of changes in demand on traffic speeds. These other functions may have properties superior to that of the BPR for the planner’s needs. Presentation, discussion, and demonstration of these other potential speed– flow functions are beyond the scope of this Guide. Vehicle-Hours and Person-Hours of Delay Vehicle­hours and person­hours of delay are typically output by the travel demand model using thresholds specified by the analyst. Vehicle­hours and person­hours of travel time may be compared to an agency specified minimum speed goal for each link. The speed goal may be the link free­flow speed, or some other value. Level of Service The HCM provides LOS measures for road segments, freeway segments, and intersections. LOS measures are also provided for freeway facilities and urban street facilities, but they are Facility Type Area Type Free-Flow Speed (mph) Capacity (veh/ln) HCM Base Speed at Capacity (mph) BPR A Parameter BPR B Parameter Freeway Downtown* 55 1,800 50.0 0.10 7 Urban 60 1,800 51.1 0.17 7 Suburban 65 1,900 52.2 0.24 7 Rural 70 1,900 53.3 0.31 7 Principal Highway 55 1,700 46.7 0.18 8 Rural Two-lane 55 1,300 42.5 0.29 8 Minor Highway 45 1,500 42.2 0.07 9 Rural Two-lane 45 1,300 32.5 0.38 9 Arterial Downtown 25 700 6.7 2.71 3 Urban 35 700 11.0 2.19 2 Suburban 45 600 11.4 2.95 2 Collector Downtown 25 600 6.7 2.71 3 Urban 30 600 10.4 1.89 3 Suburban 35 600 11.0 2.19 3 Note: *The speeds and capacities shown here for downtown freeways may not be appropriate for more modern central business district and downtown areas. Rural Multilane Rural Multilane Exhibit 129. Recommended speed–flow equation parameters.

186 Planning and Preliminary Engineering Applications Guide to the Highway Capacity Manual applicable to only those individual facility types and begin to lose their meaning (masking out problem spots within the facility) when applied to facilities more than 10 miles in length. Aggregate performance measures, such as vehicle-miles or person-miles traveled (VMT or PMT), vehicle-hours or person-hours traveled (VHT or PHT), and vehicle-hours or person-hours of delay (VHD or PHD) are generally the best basis for comparing future system performance to existing conditions or to future investment alternatives. However, it is often difficult to convey the significance or meaning of numerical results to the general public or decisionmakers. Analysts may use A–F levels of service to try to convey the degree of acceptability of the performance results; however, one should use care when simplifying numerical results to a few letter grades. Rather than reporting a single letter grade for the entire system or mode of travel, area-wide or system-wide LOS reports should report the distribution of the segment and intersection LOS results, weighted by the VHT or PHT experiencing each specific segment and intersection LOS. The link LOS is computed from the travel demand model’s link output by facility type and intersection control type. Service volume tables may be used to estimate link LOS. The analyst may then choose to report the percentage of daily VHT experiencing each LOS by mode within the system. Further value can be gained by breaking down the road system results by facility type and area type. Exhibit 130 shows an illustrative system LOS report. Exhibit 131 shows how the system report for non-freeway facilities might be displayed in a dashboard format. Note that at this point, the focus has been on results for a single average weekday with fair weather and no incidents, such as is typically produced by a travel demand model analysis. The reliability of that result under varying incident and weather conditions over the course of the year is addressed by post-processing the model’s single-day results, as described below. Area Type Facility Type Mode LOS A-C LOS D LOS E LOS F Total Urban Freeways Auto 7% 24% 38% 31% 100% Truck 4% 20% 38% 38% 100% Nonfreeway Auto 16% 34% 34% 16% 100% Truck 5% 22% 38% 34% 99% Transit 10% 29% 38% 24% 101% Bicycle 12% 31% 37% 21% 101% Pedestrian 31% 38% 24% 7% 100% Note: Values and categories are illustrative. Other area types, facility types, and modes may be appropriate. Exhibit 130. Illustrative system LOS report—typical weekday peak period. Note: Exhibit 131. Illustrative system LOS dashboard—typical weekday peak period.

R. Areas and Systems 187 Density Density can be computed for each roadway link by dividing the demand model’s predicted directional volume in vehicles per hour per through lane by the demand model’s predicted aver­ age link speed in miles per hour. The result is average number of vehicles per mile per through lane per hour. Queuing Total system vehicle­hours in queue can be estimated. First, the predicted demand for a link is multiplied by the link’s average travel time to obtain VHT. If the predicted link directional v/c ratio is greater than 1.00, or the average link speed is estimated to be below the speed at capacity, then the link is assumed to be in queue and the link’s VHT is added to the tally of system vehicle­ hours in queue (VHQ). Note that the queue may exceed the link’s length. Demand models do not typically propagate queues upstream of the bottleneck link. If intersection delay is not included in the estimate of average link travel times and speeds, then the intersection delay for the approach specific to the link is multiplied by the approach volume on the link and added to the estimated VHQ for the link. Travel Time Reliability The travel time reliability for the freeways in the study area can be estimated using the pro­ cedures described in Section H7. These procedures produce the 95th percentile highest travel time index (TTI) and the percent of trips under 45 miles per hour for each freeway link. The procedure is illustrated in Case Study 3, Long­Range Transportation Plan Analysis (Section V). Truck LOS Truck level of service can be estimated using the procedures described in Section P. Transit, Bicycle, and Pedestrian LOS The level of service for bus transit, bicycles, and pedestrians can be estimated using the pro­ cedures described in Section O. 6. References Cambridge Systematics, Inc. Travel Model Validation and Reasonableness Checking Manual, 2nd ed. Report FHWA­HEP­10­042. Federal Highway Administration, Washington, D.C., Sept. 24, 2010. Florida Department of Transportation. 2013 Florida Quality/Level of Service Handbook. Systems Planning Office, Tallahassee, 2013. Highway Capacity Manual: A Guide to Multimodal Mobility Analysis. 6th ed. Transportation Research Board, Washington, D.C., 2016.

188 S. Roadway System Monitoring 1. Overview Transportation planners monitor the performance of the freeways and streets that make up a regional transportation system in order to identify problem spots and to assess the impacts of previous invest- ments in transportation operations and capacity improvements. The performance of the system and its components are measured using recently available archived data on roadway travel times. The value of this monitoring process can be significantly enhanced through the use of various HCM relationships to identify and diagnose the causes of travel time reliability and capacity problems. 2. Travel Time Datasets The methods described in this section assume that an agency has access to archived average travel times by road segment by time of day, similar to the National Performance Management Research Data Set (NPMRDS) (FHWA 2015). In the NPMRDS, each travel time observation is the average travel time for all vehicles on a traf- fic message channel (TMC) segment over a minute period. TMC segments are generally defined between driver navigation decision points on the network (e.g., between ramp gore points on a freeway or between intersections on an urban street) (ITS America 2010). 3. Identification of Problem Spots Through the Travel Time Index The ratio of the actual travel time to the free-flow travel time, the travel time index (TTI), is a useful indicator of congestion problem spots (and times of day when congestion is present) on the roadway network. Estimation of Free-Flow Travel Time The free-flow travel time is obtained from the archives by finding the 95th percentile travel time in the archives for the selected TMC segment. Alternatively, the free-flow travel time may be estimated from the posted speed limit (with an analyst-selected adjustment, such as 5 mph, if appropriate) for the TMC segment. The analyst may choose to apply an adjustment to the posted speed limit to reflect local compliance with Source: Florida DOT (2015).

S. Roadway System Monitoring 189 the speed limit. The TMC segment length divided by the adjusted posted speed limit gives the free-flow travel time. TT TT L PSL UserAdj FF = × + or 60 Equation 20495 where TTFF = free-flow travel time (min), TT95 = 95th percentile highest observed travel time (min), L = TMC segment length (mi), PSL = posted speed limit (mph), and UserAdj = user adjustment (mph) to account for local differences between the posted speed limit and the free-flow speed, and effects of intersection controls (if any) on maxi- mum travel speeds under free-flow conditions. Computation of TTI The TTI for a TMC segment is the ratio of the observed travel time to the free-flow travel time. Equation 205TTI TT TTFF = where TTI = travel time index (unitless), TT = observed travel time (min), and TTFF = free-flow travel time (min). The TTI for a specific percentile condition (such as the 95th highest hour of the year) is computed using the travel time for the specific percentile condition. Thus, if the 95th percentile highest TTI is desired, the 95th percentile travel time is used in the computation. Identification of Congestion Problems The identification of congestion problems consists of determining whether the TTI falls above a limit indicative of demands greater than capacity. It is possible to apply the planning methods described in the earlier chapters to identify the values of TTI when demand is likely to exceed capacity. This limit varies by facility type. A table of free-flow speeds and speeds at capacity, such as illustrated in Exhibit 129 in Section R5, can be constructed based on assumed default free-flow speeds by functional class, facility type, and area type. The TTI threshold above which congestion (defined as demand greater than capacity) is present is then computed by dividing the speed at capacity into the free-flow speed. Exhibit 132 shows an example of such a table that can be used for congestion monitoring purposes. From this table one can construct some general rules for interpreting TTIs. For uninterrupted- flow segments (freeways, rural multilane highways, and rural two-lane highways): • If the observed TTI is 1.40 or greater, there is a high probability that the demand for the seg- ment exceeds its capacity and the segment is congested. • If the TTI is 1.05 or less, there is a high probability that the demand is less than capacity and the segment is uncongested.

190 Planning and Preliminary Engineering Applications Guide to the Highway Capacity Manual • If the TTI falls between 1.05 and 1.40, then it is uncertain whether the link is over capacity or not. Further information on the facility’s speed at capacity is needed to make a determination. For interrupted-flow streets (i.e., streets with traffic signals, all-way stop-controlled intersec- tions, and roundabouts), a TTI of 1.30 may be indicative of intersection delays rather than any controlled intersection exceeding its capacity. Therefore, in the case of interrupted-flow segments: • If the TTI is 4.00 or greater, there is a high probability that the demand for the segment exceeds its capacity and the segment is congested. • If the TTI is 2.50 or less, there is a high probability that the demand is less than capacity and the segment is uncongested. • If the TTI falls between 2.50 and 4.00, then it is uncertain whether the link is over capacity or not. Further information on the facility speed at capacity is needed to make a determination. 4. Identification of Multimodal Problem Spots Motor vehicle travel times are not a suitable indicator of bicycle and pedestrian problem spots. Motor vehicle volumes and speeds must be monitored in the context of the physical design of the bicycle and pedestrian travel ways to identify non-motorized-vehicle problem spots. For transit, the motor vehicle TTI can be used as an indicator of likely transit problem spots, in the absence of exclusive lanes for transit. For trucks, both the motor vehicle (auto plus truck) TTI and the truck specific TTI are inputs to the calculation of truck LOS (see Section P). 5. Diagnosis of Causes of Mobility Problems Once auto, truck, transit, bicycle, or pedestrian mobility problems are identified, the causes of the problems can be diagnosed through the planning application of the HCM as described in Part 2 of this Guide. Diagnosis will generally require some additional information beyond the travel times, such as volumes, geometry, and signal controls. Facility Type Area Type Free-Flow Speed (mph) Capacity (veh/h/ln) HCM Base Speed at Capacity (mph) TTI at Capacity Freeway Downtown 55 1,800 50.0 1.10 Urban 60 1,800 51.1 1.17 Suburban 65 1,900 52.2 1.24 Rural 70 1,900 53.3 1.31 Principal Highway Rural Multilane 55 1,700 46.7 1.18 Rural Two-lane 55 1,300 42.5 1.29 Minor Highway Rural Multilane 45 1,500 42.2 1.07 Rural Two-lane 45 1,300 32.5 1.38 Arterial Downtown 25 700 6.7 3.71 Urban 35 700 11.0 3.19 Suburban 45 600 11.4 3.95 Collector Downtown 25 600 6.7 3.71 Urban 30 600 10.4 2.89 Suburban 35 600 11.0 3.19 Note: Values in this table are illustrative and may not be applicable to a specific jurisdiction. Exhibit 132. Illustrative TTI thresholds for monitoring congestion.

S. Roadway System Monitoring 191 6. References Federal Highway Administration. National Performance Management Research Data Set: Technical Frequently Asked Questions. http://www.ops.fhwa.dot.gov/freight/freight_analysis/perform_meas/vpds/npmrdsfaqs. htm. Accessed April 11, 2015. Florida Department of Transportation. Florida Multimodal Mobility Performance Measures Source Book. Trans- portation Statistics Office, Tallahassee, 2015. ITS America, North American Traffic Working Group. Traffic Information Benchmarking Guidelines. Version 1.0. Washington, D.C., April 22, 2010.