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 1 Overview The first three sections in Part 1 describe the different levels of planning and preliminary engi- neering analyses and the potential role of the HCM in supporting these analyses. These sections serve as gateways to the remainder of the Guide: A. Introduction a. Scope of the Guide b. Target audience c. How to use the Guide d. The hierarchy of analysis methods B. Medium-level (facility-specific) analyses a. Project traffic and environmental impact studies b. Applications of default values C. High-level analyses a. Screening and scoping studies b. Long- and short-range areawide transportation planning c. System performance monitoring The remaining four sections in Part 1 provide reference information applicable to many of the planning and preliminary engineering applications described in the Guide: D. Working with traffic demand data a. Selecting an analysis hour b. Converting daily volumes to shorter timeframes c. Seasonal adjustments to traffic volumes d. Rounding traffic volumes e. Differences between observed volumes and actual demand f. Constraining demand for upstream bottleneck metering g. Generating turning-movement volume estimates from link volumes E. Predicting intersection traffic control a. Manual on Uniform Traffic Control Devices b. Graphical method F. Default values to reduce data needs a. When to consider default values b. Sources of default values c. Developing local default values G. Service volume tables to reduce analysis effort a. Description b. When to consider service volume tables c. Sources of generalized service volume tables

3 A. Introduction 1. Overview The Highway Capacity Manual (HCM) is commonly used by transportation agencies to evaluate the current or forecast operations of roadway facilities. Less well known is that the HCM can also be used to cost-effectively and reliably sup- port agencies’ planning, programming, and management decisions. This Planning and Preliminary Engineering Applications Guide to the HCM (“Guide”) is intended as a reference and educational resource on best practices for applying HCM methods to a variety of planning and preliminary engineering applications. It is designed to improve planning practice by identifying appropri- ate techniques for utilizing the HCM in planning and preliminary engineering analyses and to illustrate these techniques through the use of case studies. It is intended to be used by planners, engineers, and system analysts at various stages of the system management, operation, planning, and project development process. 2. Scope of the Guide Definitions The HCM defines planning analyses as those “generally directed toward broad issues such as initial problem identification (e.g., screening a large number of locations for potential operations deficiencies), long-range [needs] analyses, and regional and statewide [system] performance monitoring.” It defines preliminary engineering analyses as those supporting (1) planning deci- sions on roadway design concept and scope, (2) alternatives analyses, and (3) proposed system- wide policies. Applications The Guide can support statewide and local application of HCM methods to planning and preliminary engineering evaluations of current and future traffic operations and multimodal level of service. Topics covered in the Guide include: • The potential application of HCM and HCM-consistent methods to a broad spectrum of plan- ning and preliminary engineering applications (including different stages of project planning and development, various study area sizes, under and over capacity conditions, and system performance monitoring); • The appropriate use of default values when applying HCM methods, along with techniques for developing and using local default values;

4 Planning and Preliminary Engineering Applications Guide to the Highway Capacity Manual • The coordinated use of the HCM with simulation models, travel demand forecasting models, mobile source emissions models, multimodal transportation analysis tools, and other plan- ning tools; • The ability to incorporate and test more factors in an analysis than traditional planning tools allow, by integrating HCM methods with existing tools; and • The simplification of calculations to produce a more transparent, quicker evaluation and review process, while not sacrificing the accuracy of the conclusions that are drawn. Similar to the HCM, the Guide describes methods for estimating a variety of multimodal trans- portation performance measures, including traffic speed, travel time, delay, density, and queues, as well as auto, truck, bus, bicycle, and pedestrian level of service (LOS). Unlike the HCM, the Guide focuses on methods appropriate for the amount and quality of data typically available to planning analyses, as well as the available computational resources. The Guide is not intended to replace the HCM, nor to specify what constitutes good planning and preliminary engineering analysis. In many cases, current local practice may be superior to the guidance included in the Guide because local practices have been validated for local condi- tions (all of which cannot be reasonably anticipated in any single national guide). Levels of Analysis Planning and preliminary engineering covers a wide spectrum of possible levels of analysis. At the highest level (visualize a plane flying at high altitude), the area covered by the analysis is large, but the degree of detail (precision) for any particular segment of road is low. This is a typical characteristic of regional areawide studies, regional plans, statewide plans, and sketch planning and screen- ing studies. HCM analyses using a mix of default and measured inputs are examples of mid- or medium-level analysis. The area covered is significantly reduced, to that of a single roadway facility, segment, or intersection, but the degree of precision in the estimated performance is much improved. Even so, the performance estimates are still at a macroscopic level (i.e., the estimates describe traffic operations averaged over a period of time and do not consider individual vehicles in the traffic stream). A microsimulation analysis provides an extremely low-level (highly focused but highly detailed) performance analysis. In general, the data and time requirements to conduct a low-level analysis, as well as the high precision of the results, are incompatible with the needs of a typical planning study. Consequently, this Guide focuses on high- and medium-level applications of the HCM to planning and preliminary engineering. Relationship of the Guide to the Project Life Cycle Exhibit 1 illustrates that a roadway project goes through many stages from concept to con- struction to operation. Initially, the potential need for a project is identified through a long- or short-range areawide or corridor-based plan. These studies cover relatively large areas, and the level of precision for any given roadway element is relatively low. The Guide describes how to apply HCM methods in support of these types of plans. Later, if selected for further development and if funding is available, a project will move into the project initiation and project clearance stages, and facility-specific project and environmental plans will be developed. These studies cover more focused areas and have a higher level of preci- sion. Again, the Guide describes how to apply HCM methods in the development of these plans.

A. Introduction 5 Once the project moves into final design, it moves out of the realm of planning and preliminary engineering. However, once the project is constructed and in operation, it becomes part of the overall transportation system and a subject for system performance monitoring. As performance monitoring covers large areas at low levels of precision, planning and preliminary engineering techniques for estimating roadway operations performance measures again become applicable. 3. Target Audience The range of potential users for the Guide includes every technical professional involved in estimating the need for, and feasibility of, highway capacity, monitoring, management, and operations investments. This audience includes all current HCM users, plus planners and travel demand modelers who may not consider themselves HCM users but who have used pieces of the HCM in the past. University students in transportation planning and transportation engineering programs are also part of the target audience. 4. How to Use the Guide The Guide is intended primarily as a resource for practitioners. As such, it is not intended to be read cover to cover. Instead, its organization is designed to help practitioners quickly find information on how to apply the HCM to a particular planning need. The Guide is divided into four parts: • Part 1, Overview, describes typical planning and preliminary engineering analysis needs, iden- tifies points where an HCM analysis can provide useful inputs to the analysis, and points the reader to the appropriate part of the Guide for guidance on how to apply, and adapt as necessary, HCM methods for use in the analysis. Part 1 also contains several sections that are cross-referenced throughout the Guide. These sections address: working with traffic demand data, predicting future intersection traffic control, using default values to reduce data needs, and using service volume tables to reduce computational effort. Exhibit 1. Scope of the Planning and Preliminary Engineering Applications Guide to the HCM.

6 Planning and Preliminary Engineering Applications Guide to the Highway Capacity Manual • Part 2, Medium-Level Analysis Methods, presents guidance on applying and adapting HCM methods for medium-level planning analyses, those planning analyses that focus on a single facility and its component interchanges, intersections, and segments. The sections in Part 2 are organized according to the system elements (e.g., freeway facility, signalized intersection) used by the HCM and are cross-referenced from the HCM. These sections describe typical plan- ning needs for these system elements and present simplified methods for calculating a variety of performance measures commonly used in planning and preliminary engineering studies. • Part 3, High-Level Analyses, presents guidance on extending HCM methods for high-level planning analyses involving roadway corridors, large areas, and entire transportation systems. Part 3 also covers the use of service volume tables and volume-to-capacity ratios to quickly identify the needed geographic scope of an analysis. • Part 4, Case Studies, illustrates the application of the HCM techniques described in the Part 2 and 3 sections to three types of studies: (1) a freeway master plan, (2) the development of bus rapid transit service on an urban street, and (3) a long-range countywide transportation plan. To help distinguish cross-references within this Guide with cross-references to the HCM, the Guide is organized into lettered sections (e.g., Section A, Introduction) to distinguish them from the HCM’s numbered chapters (e.g., Chapter 1, HCM User’s Guide). Although having access to the latest edition of the HCM is certainly helpful for learning about the supporting research, theory, and computational details of HCM methods, in many cases the information needed to apply the HCM in a particular planning context is provided within the Guide itself. 5. The Hierarchy of Analysis Methods In some cases the Guide provides one or more alternative methods that supplement the stan- dard HCM method for estimating a particular performance measure. These alternative methods are designed to better balance the required analysis resources against the accuracy requirements of different levels of planning analysis. For example, at a high, sketch-planning level, or for regional demand modeling purposes, it may be satisfactory to estimate free-flow speeds (i.e., average vehicle speeds under low-volume conditions) for all facilities on the basis of the posted speed limits. For environmental clearance analyses of specific improvements to specific facilities (an example of a preliminary engineering analysis), it may be more appropriate to use the HCM methods for estimating free-flow speeds. Thus, the Guide may provide several methods for esti- mating performance measures and will provide advice on which level of planning or preliminary engineering analysis a given method is most suitable for, given the particular analysis objectives. Generally, when one can measure a performance measure directly in the field, it is usually (but not always) better than estimating that measure indirectly using the methods in the HCM or this Guide. When conditions make it difficult to accurately measure performance in the field, then the Guide takes the perspective that an HCM analysis using field-measured inputs is most accurate, followed by an HCM analysis using a mix of default values and field-measured inputs, followed by the alternative analysis methods described in the Guide. The general hierarchy of methods is shown in Exhibit 2. In general: • Field measurement is most reliable if it can be done cost-effectively and accurately. Note that the resources required to directly measure performance in the field can vary widely, depend- ing on the performance measure and the geographic and temporal scope of the measurement. • Microsimulation modeling of performance is the next most accurate approach if adequate resources are invested in calibrating and validating the model. • HCM estimates of performance using field-measured inputs are generally the next most accurate.

A. Introduction 7 • HCM estimates of performance using a mix of default values and field-measured inputs are usually the next most accurate. • Alternative planning methods described in the Guide for estimating performance will usually be the least accurate, but will be among the most cost-effective methods for obtaining estimates of existing and future performance. It may be infeasible, or require a disproportionate amount of resources, to employ more detailed analysis approaches such as microsimulation. For example, the analyst may need to screen many possible scenarios or solutions (20, 30, or more at times) prior to conducting a more detailed simu- lation analysis. Even if the analyst has all of the data available to conduct a simulation analysis, it may not be practical or useful to use the microsimulation process for a screening-level analysis. In such circumstances, the more detailed analysis methods may not be practical for the initial analysis process. The analyst may use the higher level methods to screen and document the analysis of the initial alternatives and then apply the more detailed methods to the two or three scenarios that pass the initial screening process. Exhibit 2. Relative effort and precision of traffic performance estimation methods.

8 B. Medium-Level (Facility-Specific) Analyses 1. Overview This section describes planning and preliminary engineering analyses that are performed at the medium level of analysis. This level of analysis typically focuses on a specific facility, or specific segments, interchanges, and intersections on that facility. Examples of these types of studies include preliminary or conceptual design studies to determine the number of required lanes and traffic, transit, and environmental impact studies required to obtain project approval and environ- mental clearance. While the data requirements of this level of analysis can be relatively extensive, the HCM, the Guide, and other publications such as the Transit Capacity and Quality of Service Manual (Kittelson & Associates et al. 2013) provide default values for some of the required inputs to assist in a planning or preliminary engineering analysis. 2. Project Traffic and Environmental Impact Studies A project traffic and environmental impact study focuses on predicting the impacts of one or more specific transportation improvement or land development projects. Examples of typical analysis guidelines for these types of studies include Oregon Department of Transportation (2005); CH2M Hill (2006); and Association of Environmental Professionals (2014). Typical Project Impact and Alternatives Analysis Process Typical project traffic and environmental impact analyses employ comparatively simple analysis techniques to add project-generated traffic onto existing or forecasted future traffic, and then evaluate the impacts on highway facility performance. The impact analysis may extend to other travel modes, such as trucks, buses, bicycles, and pedestrians, and may extend to include vehicle emissions analysis for air quality analyses, and a noise analysis. The objectives of these impact studies are to identify the project’s performance impacts by travel mode, to determine whether those impacts are significant, to generate mitigation measures for those impacts, and to assess whether those mitigations can reduce the project impacts to a less-than-significant level. Typical Tools Used In Project Impact and Alternatives Analysis Traffic, transit, and environmental impact analyses typically employ relatively simple manual traffic forecasting techniques and invest most of their effort in employing HCM-type analysis tools for predicting the resulting highway performance for each travel mode (auto, truck, bus, bicycle, and pedestrian).

B. Medium-Level (Facility-Specific) Analyses 9 Microsimulation modeling may be employed for the operations analysis of more complex projects in which the interactions between queuing and operation performance are expected to be significant. If a regional demand model is used to assist in the demand forecasting, then some of the methods described in Part 3 of the Guide on high-level methods (Section R, Areas and Systems) may be useful for improving the demand model forecasts used in the project impact analysis. Air quality and noise analysis models may use the forecasted traffic volumes as inputs for estimating the project’s air and noise impacts. Basic Data Needs for Project Impact and Alternatives Analysis The basic data needs for an impact analysis include: • Project description • Expected influence area for project impacts • Existing and forecasted demands at key intersections, freeway mainline sections, and ramps • Highway network data – Segments (length, facility type, lanes, geometric cross section) – Intersections (turn lanes, geometric cross-section, signal control settings) • Transit data – Routes, frequencies, bus stop characteristics • Bicycle and pedestrian data – Street and intersection cross-sections, bicycle, and pedestrian facility characteristics How the HCM Can Support Project Impact and Alternatives Analyses The HCM can be used to support the project impact analysis tasks shown in Exhibit 3. This exhibit lists the sections of Parts 2 and 3 of the Guide where the specific methods are described. Project Impact and Alternatives Analysis Task Parts 2 and 3 Reference Part 4 Case Studies Input to travel demand models (if used) Estimate highway capacities and free-flow speeds Section R Case Study 3.1 Traffic assignment module within travel demand model (if used) Apply volume–delay functions for estimating congested speeds Section R Case Study 3.2 Input to microsimulation model (if used) Estimate free-flow speeds Sections H-N None Microsimulation model validation and error checking (if used) Estimate capacity for error checking simulated bottlenecks Sections H-N None Project impact and alternatives analyses Estimate segment speeds for air quality and noise analyses Sections H-N Case Studies 1.3, 2.4 Estimate auto intersection utilization (v/c ratios) Sections H-N Case Studies 2.2, 2.3 Estimate delay Sections H-N Case Study 2.4 Estimate queuing Sections H-N Case Studies 1.5, 2.5 Interpret results Sections H-N Case Studies 1, 2 Analyze travel time reliability Sections H, K Case Study 1.6 Estimate multimodal quality of service for transit, bicycles, and pedestrians Section O Case Study 2.6 Estimate truck LOS Section P None Corridor analyses Section Q None Exhibit 3. Project impact analysis task cross-reference table.

10 Planning and Preliminary Engineering Applications Guide to the Highway Capacity Manual The exhibit then lists the Part 4 case studies where the applications of the methods to typical planning analyses are illustrated. 3. Applications of Default Values Default values can be used for many traffic characteristics parameters (e.g., percentage heavy vehicles, peak hour factor) required in a typical HCM analysis. Other defaults may be used to characterize the facility’s geometric design (e.g., lane widths, lateral clearances) when the analyst is confident that the facility generally meets (or will meet) agency standards. Default values for less-critical inputs to HCM analyses are provided in each procedural chapter of the HCM. Additional values are provided in NCHRP Report 599 (Zegeer et al. 2008). Both docu- ments provide sensitivity analyses of the effects of different input values on the analysis results. The Guide provides guidance on the selection and use of default values in Section F, Default Values to Reduce Data Needs. In addition, tables in the individual Part 2 and Part 3 sections of the Guide provide suggested default values for different system elements (e.g., areas, freeway facilities, signalized intersections). 4. References Association of Environmental Professionals. California Environmental Quality Act, 2014 CEQA Statute and Guidelines. Palm Desert, Calif., Jan. 2014. CH2M Hill. Best Practices for Traffic Impact Studies. Oregon Department of Transportation, Salem, June 2006. 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. Oregon Department of Transportation. 2005 Development Review Guidelines. Salem, 2005. Zegeer, J. D., M. A. Vandehey, M. Blogg, K. Nguyen, and M. Ereti. NCHRP Report 599: Default Values for High- way Capacity and Level of Service Analyses. Transportation Research Board of the National Academies, Washington, D.C., 2008.

11 C. High-Level Analyses 1. Overview This section describes planning and preliminary engineering analyses that are performed at a high level of analysis. These analyses typically cover large areas and systems of facilities. These high-level analyses may also be performed as “screening analyses,” when one is attempting to determine what the geographic and temporal limits should be for a more detailed level of analysis. Examples of these types of studies include long- and short-range regional trans- portation plan analyses and transportation system performance monitoring studies. These types of studies cover a large number of roadway miles for a given investment in data collection and analysis resources. 2. Screening and Scoping Studies Scoping studies seek to quickly determine the geographic and temporal limits required for more detailed analyses. Alternative screening studies seek to quickly identify which improve- ment alternatives may be worthy of further consideration and analysis. Role of the HCM in Screening and Scoping The service volume tables in the HCM can be used to identify facilities, segments, and inter- sections not meeting (or not likely to meet) the agency’s LOS standards for autos, trucks, transit buses, bicyclists, and pedestrians. Tables of capacities by facility type can be constructed for local facility conditions using local defaults and the HCM procedures. These tables then can be used to quickly identify volume-to-capacity (v/c) problems for individual facilities, segments, and inter- sections, as well as to evaluate the reserve capacity available in a corridor. Improvement alterna- tives can be quickly compared based on their effect on facility or corridor v/c ratios to identify those alternatives delivering a target v/c ratio. Exhibit 4 lists the specific tasks that can be supported by the HCM. How to Use the Guide for Screening and Scoping Exhibit 4 lists the sections of Parts 2 and 3 where the specific methods are described. This table then lists the Part 4 case studies where the applications of the methods to typical planning analyses are illustrated.

12 Planning and Preliminary Engineering Applications Guide to the Highway Capacity Manual 3. Long- and Short-Range Areawide Transportation Planning Long-range areawide transportation planning, specifically the production of state and regional long-range transportation plans (LRTPs), defines the vision for the region’s or state’s transportation systems and services for a 20-year or longer period. Short-range areawide plan- ning focuses on just as large an area, but on a shorter time frame. “In metropolitan areas, the LRTP is the official multimodal transportation plan addressing no less than a 20-year planning horizon that is developed, adopted, and updated by the metropolitan planning organization (MPO) through the metropolitan transportation planning process (MTPP)” (Federal Highway Administration 2007). Typical Areawide Planning Analysis Process In the long-range transportation planning process, planners assess future investments on the basis of the performance of the freeways and streets that make up a regional transportation system. The performance of the system and its components is often estimated through a travel demand and analysis forecasting process. This process requires a variety of inputs and analytical methodologies, which the HCM can provide. Typical Tools Used in the Areawide Planning Analysis A combination of specialized travel demand models, geographic information systems (GIS), and spreadsheets are typically used when conducting analyses for LRTPs. The region to be modeled is divided into zones and highway and transit networks are coded. The GIS, in combination with a land-use model, is used to develop forecasts of socioeconomic activity (population, employment, etc.) for the region. Basic Data Needs for Areawide Planning Analysis Data needs are kept relatively simple (in terms of different types of data), but end up being massive in size because of the large areas often covered in regional transportation plans. The basic data needs for LRTPs include: Exhibit 4. Screening and scoping task cross-reference table. Screening and Scoping Task Parts 2 and 3 References Part 4 Case Studies Identify potential level of service (LOS) hot spots Screen for auto LOS problems Sections H-N Case Studies 1.4, 2.4 Screen for truck LOS problems Section P None Screen for transit, bicycle, and pedestrian LOS problems Section O Case Study 2.6 Identify potential capacity problems: auto Sections H-N Case Studies 1.1, 1.2, 2.1, 2.2, 2.3 Preliminarily evaluate improvement alternatives Auto improvements Sections H-N Case Study 1.7 Truck improvements Sections H-N None Transit, bicycle, and pedestrian improvements Section O Case Study 2.6

C. High-Level Analyses 13 • Socioeconomic data by traffic analysis zone (e.g., population, employment) • Highway network data – Segments (e.g., length, facility type, lanes, capacity, free-flow speed) – Connectivity • Transit network data – Segments, routes, frequencies, transfer points How the HCM Can Support Areawide Planning Analyses The HCM can be used to support the LRTP planning analysis tasks shown in Exhibit 5. This exhibit lists the sections of Part 3 of the Guide where the specific methods are described. This exhibit then lists the Part 4 case studies where the applications of the methods to typical planning analyses are illustrated. 4. System Performance Monitoring Highway system performance monitoring is the measurement of highway use and operating characteristics under existing conditions (Federal Highway Administration 2014a). Performance Monitoring Context The Moving Ahead for Progress in the 21st Century Act (MAP-21) established a “performance and outcome based program for states to invest resources in projects that collectively will make progress toward the achievement of national goals” (Federal Highway Administration 2014b). MAP-21 requires the Federal Highway Administration to work with stakeholders to identify per- formance measures tied to seven goal areas for the federal-aid highway program: • Safety, • Infrastructure Maintenance, Exhibit 5. Areawide planning analysis task cross-reference table. Areawide Planning Analysis Task Part 3 Reference Part 4 Case Study Input to travel demand models Estimate highway segment capacities and free-flow speeds Section R Case Study 3.1 Traffic assignment module within the travel demand model Apply volume–delay functions to estimate congested speeds Section R Case Study 3.2 Post-processing travel demand model outputs Obtain more accurate speed estimates for air quality analyses Section R Case Study 3.3 Spot auto volume-to-capacity and LOS hot spots (quick screening) Section R Case Study 3.3 Estimate delay based on agency policy Section R Case Study 3.3 Estimate queuing Section R Case Study 3.3 Interpret results Section R Case Study 3.3 Analyze travel time reliability Section R Case Study 3.4 Estimate multimodal quality of service for autos, trucks, transit, bicycles, and pedestrians Section R None Corridor analyses Section Q None

14 Planning and Preliminary Engineering Applications Guide to the Highway Capacity Manual • Congestion Reduction, • System Reliability Improvement, • Freight Movement and Economic Vitality, • Environmental Sustainability, and • Reduced Project Delivery Delays. Of these seven goal areas, the HCM can assist agencies in monitoring highway performance relevant to the three goal areas of Congestion Reduction, System Reliability Improvement, and Freight Movement. Role of the HCM in Performance Monitoring The HCM can be used to compute the performance measures not directly monitored at a monitoring site. It can be used to spot data errors and inconsistencies. It can be used to impute missing performance data. Exhibit 6 lists the specific performance monitoring tasks that can be supported by the HCM. How to Use the Guide for Performance Monitoring Exhibit 6 lists which methods are described in Sections R and S of the Guide. This table then lists the example problems in the Part 4 case studies where the applications of the methods to typical planning analyses are illustrated. 5. References Federal Highway Administration. The Transportation Planning Process Briefing Book: Key Issues for Transportation Decisionmakers, Officials, and Staff. FHWA-HEP-07-039. Washington, D.C., 2007. Federal Highway Administration. Highway Performance Monitoring System website. https://www.fhwa.dot. gov/policyinformation/hpms.cfm. Accessed September 2, 2014(a). Federal Highway Administration. MAP-21 Performance Management Fact Sheet website. https://www.fhwa. dot.gov/map21/factsheets/pm.cfm. Accessed September 2, 2014(b). Exhibit 6. Performance monitoring task cross-reference table. Performance Monitoring Task Part 3 Reference Part 4 Case Study Estimate monitoring site capacities and free-flow speeds Section R4 Case Study 3.1 For volume-only monitoring sites Estimate speeds Section R5 Case Study 3.2 For travel time–only monitoring segments Estimate congestion Section S3 None Performance analyses Auto and truck VMT by LOS Section R5 None Estimate delay Section R5 Case Study 3.3 Estimate queuing Section R5 Case Study 3.3 Analyze travel time reliability Section R5 Case Study 3.4 Estimate multimodal LOS for transit, bicycles, and pedestrians Section R5 None Estimate truck LOS Section R5 None

15 D. Working with Traffic Demand Data 1. Overview The traffic demand data available for a planning or preliminary engineering analysis may require adjusting before it can be used with an HCM planning method. For example, annual average daily traffic (AADT) volumes may need to be converted to hourly volumes repre- sentative of the conditions of interest to the analysis (e.g., peak hour, peak season volumes). This section provides guidance on these types of demand volume adjustments. The analyst should be aware that state and local traffic forecasting and analysis guidelines and policies often specify the methods that should be used to adjust demand volumes, as well as the analysis hour(s) that should be analyzed. It is important for planning and preliminary engineering analyses to follow these local guidelines, in part because any subsequent operational analyses will apply the same guidance. The goal is for the more detailed operations study to focus on the specific issues identified by the earlier, more-general planning study, and not to have to redo prior work because the wrong procedures were used. Therefore, it is recommended that the analyst check whether state and local guidelines already exist prior to applying the guidance found in this section. NCHRP Report 255 (Pedersen and Samdahl 1982) and NCHRP Report 765 (CDM Smith et al. 2014) are good references on processing demand model forecasts for use in traffic analyses. 2. Selecting an Analysis Hour One important decision when performing a traffic analysis is the selection of an analysis hour. This choice balances a transportation agency’s desire to provide adequate operations during the large majority of hours of the year and its need to use its limited resources as efficiently as pos- sible. AASHTO (2009) recommends the use of the 30th-highest hour of the year as a design hour, resulting in a few hours per year with (sometimes substantially) higher volumes, and many hours per year with lower volumes. Some agencies choose other analysis hours for cost-efficiency reasons; for example, Florida uses a combination of the 100th-highest hour (for areas under 50,000 population) and a typical weekday peak hour (for larger areas) (Florida DOT 2014). In some cases, the needs of the analysis may require using a non-weekday peak hour (e.g., special event planning, transportation planning for recreation areas). The choice of an analysis hour will affect the way traffic volumes may need to be adjusted for use with HCM methods. 3. Converting Daily Volumes to Shorter Timeframes HCM methods work with hourly directional demand volumes as a starting point and typi- cally analyze traffic flows during the peak 15 minutes of an analysis hour. Sometimes, however, the traffic demand volumes available for a planning analysis consist of AADTs. These must be

16 Planning and Preliminary Engineering Applications Guide to the Highway Capacity Manual converted into peak hour directional flows. Three factors are used in this process: the K-factor (the proportion of AADT occurring during the analysis hour); the D-factor (the proportion of traffic in the peak direction during the analysis hour); and the peak hour factor (PHF, which con- verts design-hour volumes to the equivalent hourly flow that occurs during the peak 15 minutes). In the case of “base-year” (existing) demands, the hourly and peak direction flows can often be directly measured. In the case of “future-year” demands (and base-year demands when demands could not be directly counted), it may be necessary to adopt peaking and directional factors to convert daily traffic forecasts to the required hourly demands by direction. K-Factor The K-factor converts AADT to analysis hour volumes. It is the percentage of AADT occurring during the analysis hour. The selection of an appropriate K-factor is very important, as selecting a value that is too high can result in too many locations being identified as not meeting roadway operations standards (as the resulting estimated hourly volumes are too high), while selecting a value that is too low can result in some problem locations not being identified (because the estimated hourly volumes are too low). The former may result in unnecessary follow-up work and potentially too bleak a picture of future conditions, while the latter may result in potentially important problems going undetected. For many rural and urban highways, the K-factor falls between 0.09 and 0.10, but it also can fall outside this range. For highways with strongly peaked demand, the K-factor may exceed 0.10. Conversely, for highways with consistent and heavy flows for many hours of the day, the K-factor is likely to be lower than 0.09. In general, • The K-factor decreases as the AADT on a highway increases; • The K-factor decreases as development density along a highway increases; and • The highest K-factors occur on recreational facilities, followed by rural, suburban, and urban facilities, in descending order (HCM 2016). In addition, the K-factor will be higher when a 30th-highest hour is chosen as the analysis hour (K30) than when the 50th- (K50) or 100th-highest hour (K100) is used. The K-factor should be determined, if possible, from local data for similar types of facilities with similar demand characteristics. Data from the automatic traffic recorders maintained by state DOTs and other transportation agencies are good sources for determining K-factors. Exhibit 7 presents illustrative K30 values, on the basis of average data from Washington State that demonstrate how K-factors decrease as AADT increases (HCM 2016). Exhibit 7 also shows standard K-factors specified by the Florida DOT (2013) for analyses of state highways. Note that K-factors can and do change as traffic congestion changes. Thus, base-year and future- year K-factors may differ. K-factors may also vary between urban and rural areas. The analyst may Washington State DOT Florida DOT AADT Average K30 Area Type Standard K-Factor 0–2,500 0.151 Urbanized/Transitioning 0.090 2,500–5,000 0.136 Large Urbanized 0.080-0.090 5,000–10,000 0.118 Urban Freeway 0.105 10,000–20,000 0.116 Urban Highway 0.090 20,000–50,000 0.107 Urban Arterial 0.090 50,000–100,000 0.091 Rural Freeway 0.105 100,000–200,000 0.082 Rural Highway 0.095 >200,000 0.067 Rural Arterial 0.095 Sources: Washington State DOT (2008) in HCM (2016), Exhibit 3-11; Florida DOT (2013), p. 80. Exhibit 7. Illustrative K values.

D. Working with Traffic Demand Data 17 consider sensitivity analyses to address uncertainty in future-year K-factors. Toll facilities may have different K-factors than similar untolled facilities. Additional references on K-factors can be found in the literature (e.g., Dykstra et al. 2011). D-Factor The D-factor represents the proportion of traffic in the peak direction on a roadway during the peak hour. Radial roadways into a city center and recreational and rural routes are often subject to strong directional imbalances during peak hours. In contrast, circumferential roadways and routes connecting major cities within a metropolitan area may have very balanced flows dur- ing peak periods. Exhibit 8 presents illustrative directional distributions derived from selected California freeways (HCM 2016). Note that D-factors can and do change as traffic congestion changes. Thus, base and future- year D-factors may differ. The analyst may consider sensitivity analyses to address uncertainty in future-year D-factors. Directional Design-Hour Volume The directional design-hour volume (DDHV) is the starting point for many HCM-based analyses. It can be calculated by multiplying the AADT by the K- and D-factors, as shown in Equation 1 (HCM 2016). = × × Equation 1DDHV AADT K D where DDHV = directional design-hour volume (veh/h), AADT = annual average daily traffic (veh/day), K = proportion of AADT occurring in the peak hour (decimal), and D = proportion of peak hour traffic in the peak direction (decimal). Peak Hour Factor Most HCM methods analyze conditions during the peak 15 minutes of the peak hour. Although this may seem to be a fairly short timeframe on which to base roadway design and control deci- sions, it should be kept in mind that the effects of roadway operations breaking down at a single location can last for much longer periods of time (potentially hours in larger metropolitan areas) and that the ripple effects of a breakdown can extend to other roadway segments and intersections. Therefore, the HCM analyzes the peak 15 minutes, to evaluate the worst 15-minute period within the analysis hour that can lead to facility breakdowns. In the absence of direct measurements of peak 15-minute volumes (a common situation for planning analyses), a PHF is used to convert hourly demand volumes into an hourly flow Freeway Type D-Factor Rural–intercity 0.59 Rural–recreational and intercity 0.64 Suburban circumferential 0.52 Suburban radial 0.60 Urban radial 0.70 Intraurban 0.51 Source: 2007 Caltrans data in HCM (2016), Exhibit 3-12. Exhibit 8. Illustrative D-factor values.

18 Planning and Preliminary Engineering Applications Guide to the Highway Capacity Manual rate equivalent to the peak-15-minute volume being sustained for an entire hour. The PHF is calculated as shown in Equation 2, with the peak-15-minute flow rate calculated as shown in Equation 3 (HCM 2016). = ×4 Equation 2 15 PHF V V = Equation 3v V PHF where PHF = peak hour factor (decimal), V = hourly volume (veh/h), V15 = volume during the peak 15 min of the analysis hour (veh/15 min), and v = flow rate for a peak 15-min period (veh/h). As with the K-factor, the selection of an appropriate PHF strongly influences the accuracy of the analysis results. For high-level planning analyses it is often appropriate, given the amount of uncertainty in some of the inputs (e.g., demand), to evaluate average hourly conditions. In these cases, the PHF is set to 1.00. For medium-level preliminary engineering studies, it may be more appropriate to use either field-measured PHFs or the default PHFs suggested in the HCM or NCHRP Report 599 (Zegeer et al. 2008). 4. Seasonal Adjustments to Traffic Volumes Seasonal adjustments to traffic volumes may be appropriate for roadways showing high “sea- sonality” in their demand. Sometimes when peak hour or peak-15-minute traffic counts are available for a planning or preliminary engineering analysis, the time of year when the counts were made may not correspond to the desired analysis hour. While it is preferable to avoid using counts where large seasonal adjustments are required, in cases when the available count falls outside the peak season, the count may need to be adjusted to represent analysis hour volumes. The basic adjustment process is to factor the count by the ratio of (1) the average monthly volume for a month reflective of the analysis hour to (2) the average monthly volume during the month when the count was made. Data from the automatic traffic recorders maintained by state DOTs and other transportation agencies are good sources for average monthly traffic volumes. Alternatively, tables of monthly factors (the ratio of monthly average volume to AADT) for each month of the year for specific count stations or for particular types of facilities may be available from transportation agencies (again, based on automatic traffic recorder data). In these cases, a count can be factored by the ratio of the monthly factor for a month reflective of the analysis hour and the monthly factor for the month when the count was made. 5. Rounding Traffic Volumes The traffic volumes used for planning and preliminary analyses are often estimates. Therefore, to avoid giving the impression of a greater degree of accuracy than is warranted, AASHTO (2009) recommends rounding traffic volumes as follows: • Volumes under 1,000 should be rounded to the nearest 10. • Volumes between 1,000 and 9,999 should be rounded to the nearest 100. • Volumes of 10,000 or more should be rounded to the nearest 1,000.

D. Working with Traffic Demand Data 19 6. Differences Between Observed Volumes and Actual Demand HCM methods typically require demand volumes: the traffic volume that would use a road- way during an analysis hour in the absence of any capacity constraints (i.e., bottlenecks). Field measurements of traffic volumes produce observed volumes: the traffic volume that is capable of using a roadway during an analysis hour. When demand is less than capacity (undersaturated flow) and no bottlenecks exist upstream, then the demand volume can be assumed to be equal to the observed volume. When demand exceeds capacity (oversaturated flow), then determining demand requires a count of the traffic joining the queue upstream of the bottleneck, as opposed to a count of traffic departing the bottleneck (Pedutó et al. 1977). However, it may not be easy to determine how much of the traffic joining the queue is bound for the bottleneck location once the queue extends past the previous intersection or interchange (as some traffic may intend to exit the roadway at that point) (HCM 2016). 7. Constraining Demand for Upstream Bottleneck Metering Transportation planning models produce demand volume estimates. However, when a model does not account for the metering effect of bottlenecks (i.e., is not capacity-constrained), it will produce estimates of demand downstream of a bottleneck that are higher than would actually be observed. This can result in HCM-based methods predicting LOS F for situations in which the traffic physically cannot arrive at the study area. The following procedure, adapted from Appendix F of FHWA’s Guidelines for Applying Traf- fic Microsimulation Modeling Software (Dowling et al. 2004) can be used in a post-processing analysis of demand model outputs to constrain demand forecasts for segments downstream of a bottleneck. Step 1: Identify Gateway Capacities The analyst should first identify the capacities of the facility or facilities at the gateways delivering traffic to the study HCM facility, segment, intersection, or area. A gateway is defined as a point where traffic enters or leaves the study area. These gateways cannot physically feed traffic to the HCM facility at a higher rate than their capacity. Any forecasted demands greater than the inbound capacity of a gateway should be reduced to the inbound capacity of the gateway. Step 2: Estimate Excess Demand at Inbound Bottlenecks If the forecasted hourly demand in the inbound direction at a gateway exceeds its capac- ity, the proportion of the demand that is in excess of the available hourly capacity should be computed: = − Equation 4P D C C where P = proportion of excess demand (decimal), D = forecasted demand (veh/h), and C = estimated capacity (veh/h).

20 Planning and Preliminary Engineering Applications Guide to the Highway Capacity Manual Step 3: Reduce Forecasted Demand within HCM Study Area The forecasted hourly demands for the facilities and segments within the HCM study area that are downstream from the bottleneck should also be reduced. However, the reduction must take into account the traffic entering and exiting the facility within the study area. It is suggested that the forecasted downstream demands be reduced in proportion to the reduction in demand that can get through the gateway, assuming that the amount of reduction in the downstream flows is proportional to the reduction in demand at the bottleneck. If the ana- lyst has superior information (such as an origin–destination [O-D] table), then the assumption of proportionality should be overridden by the superior information. The gateway-constrained downstream demands are then obtained by summing the constrained gateway, off-ramp, and on-ramp volumes between the gateway and the downstream segment. ( )= × −1 Equation 5D D Pc u where Dc = constrained demand for a downstream off-ramp or exit point (veh/h), Du = unconstrained demand forecast (veh/h), and P = proportion of excess demand (decimal). Exhibit 9 illustrates how the proportional reduction procedure would be applied for a single inbound gateway constraint that reduces the peak hour demand from 5,000 veh/h to 4,000 veh/h. Starting upstream of the gateway, there is an unconstrained demand for 5,000 veh/h. Because the gateway has a capacity of 4,000 veh/h, the downstream capacity-constrained demand is reduced from the unconstrained level of 5,000 veh/h to 4,000 veh/h. Thus, 1,000 vehicles are stored at the gateway during the peak hour. Because it is assumed that the stored vehicles are intended for down- stream destinations in proportion to the exiting volumes at each off ramp and freeway mainline, the downstream volumes are reduced by the same percentage as the percentage reduction at the bottleneck (20 percent). A 20-percent reduction of the off-ramp volume results in a constrained demand of 800 veh/h. The on-ramp volume is unaffected by the upstream gateway bottleneck, so its unconstrained demand is unchanged at 500 veh/h. The demand that enters the segment downstream of the interchange is equal to the constrained demand of 4,000 veh/h leaving the gateway bottleneck, minus the 800 veh/h leaving the freeway on the off ramp, plus 500 veh/h entering the freeway at the on ramp, which results in a constrained demand of 3,700 veh/h for the downstream segment. Notes: Du = unconstrained demand, Dc = constrained demand. Exhibit 9. Capacity-constraining demands entering and within HCM study facility.

D. Working with Traffic Demand Data 21 Step 4: Caution When Working With Constrained Demands The analyst should recompute the capacity-constrained demands for future scenarios that change one or more capacity constraints (such as adding a lane to a bottleneck). A scenario that eliminates a bottleneck may release demands that create new bottlenecks downstream. Once the constraint is changed in some way, the analyst should check downstream to see if the increase in volume is creating new bottlenecks. A related hazard with using constrained demands in planning analyses is that unanticipated improvements in the coming years might release one or more constraints that were presumed to be in place for the planning analysis. 8. Generating Turning-Movement Volume Estimates from Link Volumes The HCM’s intersection analysis methods require turning-movement volumes. However, this information may not be available for planning and preliminary engineering analyses (e.g., when only link volume data are available, or when the turning movements produced by a transporta- tion planning model are not considered to be reliable). In these cases, the following methods for estimating turning movements can be applied, originally documented in NCHRP Report 255 (Pedersen and Samdahl 1982) and updated in NCHRP Report 765 (CDM Smith et al. 2014). The analyst will need to manually check the results of these methods for reasonableness. Procedures for estimating turning movements from travel model link volumes or intersection approach volumes are: • Factoring procedures – Ratio method – Difference method • Iterative procedures – Directional method – Non-directional method • “T” intersection procedures – Directional method – Non-directional method Factoring Procedures Factoring procedures are used to estimate future turning movements based on the relation- ship between existing or base-year intersection turning-movement counts and base-year turning- movement assignments from a travel model. They require an accurate turning-movement count as a starting point. The assumption is that future-year turning movements will be similar to existing turning movements. Future-year turning movements can be predicted by comparing either relative ratios or differences between base-year and future-year travel model assignments. Ratio Method The ratio method produces a future-year turning-movement estimate by applying the ratio of the future-year model assignment and the base-year model assignment to the existing or base-year count. The method form is given by Equation 6: = ×   Equation 6FF BC FA BA ri i i i

22 Planning and Preliminary Engineering Applications Guide to the Highway Capacity Manual where FFri = future-year forecast volume for turning-movement i (veh/h), BCi = base-year count for turning-movement i (veh/h), FAi = future-year model assignment for turning-movement i (veh/h), and BAi = base-year model assignment for turning-movement i (veh/h). Turning-movement estimates or forecasts are computed individually and then summed to obtain approach volumes. Data needed to perform the procedure are the following: • Base-year intersection turning-movement counts, • Base-year traffic model turning-movement assignments, and • Future-year traffic model turning-movement assignments. In the following example, the count for a particular turning movement is 200 veh/h, the base-year travel model assignment for this movement is 260 veh/h, and the future-year travel model assignment is 500 veh/h. Applying the procedure yields the following future-year turning- movement forecast: = ×   = × =200 500 260 385 veh hFF BC FA BA ri i i i Difference Method The difference method produces a future-year turning-movement estimate by applying the relative difference between the base-year and future-year travel model assignment to the existing or base-year count. The method form is given by Equation 7: ( )= + − Equation 7FF FA BC BAdi i i i where FFdi = future-year forecast volume for turning-movement i (veh/h), BCi = base-year count for turning-movement i (veh/h), FAi = future-year model assignment for turning-movement i (veh/h), and BAi = base-year model assignment for turning-movement i (veh/h). Turning-movement estimates or forecasts are computed individually and then summed to obtain approach volumes. Data needed to perform the procedure are the following: • Base-year intersection turning-movement counts, • Base-year traffic model turning-movement assignments, and • Future-year traffic model turning-movement assignments. Returning to the data from the previous example, where the count is 200 veh/h, the base-year travel model assignment for this movement is 260 veh/h, and the future-year travel model assign- ment is 500 veh/h, the future-year turning-movement forecast is computed as: ( ) ( )= + − = + − =500 200 260 440 veh hFF FA BC BAdi i i i Comparing the Results The ratio method produces a future turning-movement estimate of 385 veh/h, while the difference method, using the same data, produces an estimate of 440 veh/h. NCHRP Report 255 (Pedersen and Samdahl 1982) recommends averaging the two to reduce the extremes, but aver- aging is not advised in NCHRP Report 765 (CDM Smith et al. 2014). The belief is that averaging

D. Working with Traffic Demand Data 23 will reduce the accuracy of one method or the other. Instead, advice is given that the analyst should evaluate the results from each method within the context of existing traffic volumes and turning-movement forecasts and then select a preferred method. A fundamental assumption of both methods is that future turning movements will be of simi- lar nature to existing turning movements. This assumption can be applied to land use, general development patterns, and resulting traffic patterns within the study area. Iterative Procedures Iterative procedures are applied to produce either directional or non-directional (two-way) turning volumes; the typical application of this Guide will utilize directional turning movements. Iterative procedures are useful when it is important to preserve link entry and departure volumes. They require an initial estimate of turning percentages. Existing turning-movement counts are often used as the initial input, but an estimate of turning proportions can be used as well. Iterative procedures can use approach and departure link volumes directly, or they can be estimated by applying K- and D-factors to AADT or design-hour volumes. Directional Method The directional method uses the initial estimate of turning movements, then alternatively bal- ances approach (inflow) and departure (outflow) volumes in a turning-movement matrix until an acceptable level of convergence is reached. The future-year link volumes are fixed and the turning movements in the trip matrix are adjusted until they match the link volumes. The number of iterations required depends on the desired level of convergence and the difference between exist- ing and future link volumes. Where large differences in link volumes occur (in an area where high growth is predicted, for example), several iterations may be required. Volumes normally converge within 6 to 10 iterations using this method. The directional method consists of five steps, as described in NCHRP Report 765 [CDM Smith et al. (2014) on pages 116–122]. The following notation is used: n = number of intersection legs, b = base-year, f = future-year, O = inflows (“from origin”), D = outflows (“to destination”), i = inflow (origin) link number, j = outflow (destination) link number, T = traffic volume, P = estimated percentage of traffic flow (expressed in decimal form), and * = adjusted value in each iteration. The notation is combined to define the elements used in the method: Oib = base-year inflow to the intersection on link i, Oif = future-year inflow to the intersection on link i, Djb = base-year outflow from the intersection on link j, Djf = future-year outflow from the intersection on link j, Tijb = base-year traffic flow entering through link i and departing through link j, Tijf = future-year traffic flow entering through link i and departing through link j, and Pijf = future-year estimated turning-movement percentage (expressed in decimal form) of traffic flow entering through link i and departing through link j. These elements are illustrated in Exhibit 10. Exhibit 10. Iterative method elements. Source: NCHRP Report 765 (CDM Smith et al. 2014), Figure 6-3.

24 Planning and Preliminary Engineering Applications Guide to the Highway Capacity Manual Step 1: Turning-Movement Matrix Construction. The first step in the process is to con- struct a turning-movement matrix. This is a square matrix, with one row and one column for each intersection leg, as shown in Exhibit 11. The inflows (approach volumes) are arranged in matrix rows and the outflows (departure volumes) are arranged in matrix columns. Each matrix cell represents the corresponding turning-movement “from link i to link j.” Unless U-turns are allowed, diagonal cells (i = j) always will be zero. An illustrative example of the intersection turning movements and corresponding matrix is shown in Exhibit 12. This example applies when base-year turning move- ments are known. When unknown, percentages are substituted for actual turning volumes and the corresponding matrix values are shown as a proportion. When percentages are used, all row totals of Pijf must equal 1.0. Column totals will not equal 1.00 except by coincidence, but the sum of all columns (and the sum of all rows) should equal the total number of inter- section legs. For example, for a four-legged intersection, SProws = SPcolumns = 4 × 1.00 = 4.00. If initial turning movements are unknown, the turning volume movement matrix cells are populated by multiplying future link inflows (Oif) by the corresponding turning-movement per- centage (Pijf), as shown in Equation 8: = ×* Equation 8T O Pij f i f ij f where all variables are as defined previously. Source: NCHRP Report 765 (CDM Smith et al. 2014), Figure 6-4. Exhibit 11. Turning-movement matrix structure. Source: NCHRP Report 765 (CDM Smith et al. 2014), Figure 6-5. Exhibit 12. Turning-movement matrix structure.

D. Working with Traffic Demand Data 25 Exhibit 13 shows an example turning-movement matrix after the first step has been completed. Step 2: Perform First Row Iteration. In the second step, base-year inflows Oib are replaced with future-year inflows Oif . Each matrix cell is adjusted according to Equation 9: =* Equation 9T O O Tij f i f ib ij b where T*ijf is the adjusted future turning volume for this iteration. A new matrix is constructed containing the future-year origin inflows (rows) Oif. New destina- tion outflows (columns) D*jf are created by summing the adjusted turning movements T*ijf in each column j, as indicated by Equation 10: ∑= = * * Equation 10 1 D Tj f ij f i n Column totals D*jf from the adjusted turning-movement matrix are compared with the origi- nal column totals Djf from the first step. If the difference between them is acceptable, then the method is complete and no further iterations are necessary. For most applications, a difference of ±10% is acceptable. If the difference is greater than the desired limit, then further iterations are required. Step 3: Perform First Column Iteration. In the third step, turning movements from the pre- vious stage are adjusted further. The previous matrix is used, but the adjusted outflows D*jf are replaced with the original outflows Djf (i.e., the outflow forecasts). Individual turning movements then are adjusted by the ratio of the original outflow forecasts to adjusted outflows as given by Equation 11: * Equation 11,new ,oldT D D Tij f j f j f ij f= where Tijf,old = T*ijf matrix value from the previous step, and Tijf,new = Adjusted turning-movement matrix value Tijf after this column iteration. Source: NCHRP Report 765 (CDM Smith et al. 2014), Figure 6-6. Exhibit 13. Example turning-movement matrix and future link volumes.

26 Planning and Preliminary Engineering Applications Guide to the Highway Capacity Manual Should subsequent iterations be necessary, values of Tijf,new created in this step become Tijf,old in the next step. A new matrix of adjusted turning movements Tijf,new and future destination outflows (columns) Djf is created. Adjusted row totals O*if are computed by summing T*jjf,new in each row, as shown in Equation 12: * * Equation 12,new 1 O Tif ij f j n ∑= = Similar to the previous step, adjusted row totals O*if are compared with original inflows Oif . If the difference between these two totals is acceptable using the same convergence cri- terion, then the method is done. If the discrepancy is greater, then further iterations will be necessary. Step 4: Repeat Row Iteration and Step 5: Repeat Column Iteration. The fourth and fifth steps involve repeating the procedure. For row iterations (Step 2), new values for T*jjf,new are cal- culated, then D*jf is compared with Djf . For column iterations, new values for T*jjf,new and O*if are computed, then O*if is compared with Oif . Row and column iterations should be continued until acceptable differences between D*jf and Djf and O*if and Oif are obtained. When those differences are deemed acceptable, values in the final matrix will be the estimated turning volumes. Detailed documentation of the iterative directional method, including step-by-step iterations of an example, is provided in NCHRP Report 765 (CDM Smith et al. 2014) and also in NCHRP Report 255 (Pedersen and Samdahl 1982). Non-Directional Method The non-directional method is intended for general planning purposes where non-directional (i.e., two-way) turning movements are desired. As HCM methods rely on directional volumes as inputs, this method is not applicable within the context of this Guide. The iterative non- directional method is fully documented in NCHRP Report 765 (CDM Smith et al. 2014) and also in NCHRP Report 255 (Pedersen and Samdahl 1982). “T” Intersection Procedures Turning-movement estimates at three-legged or “T” intersections can be developed using sim- pler procedures than for four-legged-intersections. Directional Method The directional method uses basic mathematical relationships among link volumes for estimat- ing turning movements. To apply the method, one of the turning movements must be known or estimated, along with the approach and departure volumes for all three legs. If only two-way AADT volumes are known, hourly approach volumes can be estimated using appropriate K- and D-factors. The directional method configuration and notation are shown in Exhibit 14. There are six potential scenarios where one of the turning movements is known or estimated. The computa- tions for each scenario are shown in Exhibit 15.

D. Working with Traffic Demand Data 27 Non-Directional Method The non-directional method is intended for general planning purposes where non-directional (i.e., two-way) turning movements are desired. As HCM methods rely on directional volumes as inputs, this method is not applicable within the context of the Guide. The iterative non-directional method is fully documented in NCHRP Report 765 (CDM Smith et al. 2014) and NCHRP Report 255 (Pedersen and Samdahl 1982). Spreadsheet demonstrations of the methods for estimating turning movements from link volumes are included on the CD bound into NCHRP Report 765; an image file of the CD, which can be used to burn a CD containing the spreadsheets, is available at http://www.trb. org/Publications/Blurbs/170900.aspx. Florida DOT Method The Florida DOT uses a method originally described by Hauer et al. (1981). An initial estimate of the proportion of an approach’s traffic turning right, turning left, or continuing straight can be provided by the user (for example, from existing turning movements), or the method can Source: NCHRP Report 765 (CDM Smith et al. 2014), Figure 6-21. Exhibit 14. Directional method configuration and notation. Known or Estimated Turning Volumes A B C D E F Co m pu te d Tu rn in g V ol um es A –B –E –B –B –E B –A –A –C –C –A C –B –D –D –D –B D –C –F –C –F –C E –F –A –F –A –F F –D –E –D –E –E Source: NCHRP Report 765 (CDM Smith et al. 2014), Table 6-6. Exhibit 15. “T” intersection directional turning-movement computations.

28 Planning and Preliminary Engineering Applications Guide to the Highway Capacity Manual create its own first-guess proportions from the approach volumes. Once the turning propor- tions have been specified, the method goes through a series of iterations, similar to the previ- ously described iterative directional method, to develop the turning-movement estimates. As with the iterative method, the Florida DOT method is useful when it is desirable to preserve the link entry and exit volumes in the analysis. FDOT has developed the “TURNS5” spreadsheet (http://teachamerica.com/tih/PDF/turns5-V02_XML.xls) to assist analysts with implementing this method (Florida DOT 2014). 9. References American Association of State Highway and Transportation Officials. AASHTO Guidelines for Traffic Data Programs. Washington, D.C., 2009. CDM Smith, A. Horowitz, T. Creasey, R. Pendyala, and M. Chen. NCHRP Report 765: Analytical Travel Fore- casting Approaches for Project-Level Planning and Design. Transportation Research Board of the National Academies, Washington, D.C., 2014. Dowling, R., A. Skabardonis, and V. Alexiadis. Traffic Analysis Toolbox Volume III: Guidelines for Applying Traffic Microsimulation Modeling Software. Report FHWA-HRT-04-040. Federal Highway Administration, Washington, D.C., June 2004. Dykstra, L., D. McLeod, and A. Piszczatoski. Standard K-Factors for Transportation Planning and Design. ITE Journal, Vol. 81, No. 11, Nov. 2011, pp. 20–25. Florida Department of Transportation. Project Traffic Forecasting Handbook. Tallahassee, 2014. Florida Department of Transportation. 2013 Quality/Level of Service Handbook. Tallahassee, 2013. Hauer, E., E. Pagitsas, and B. T. Shin. Estimation of Turning Flows from Automatic Counts. In Transportation Research Record 795, Transportation Research Board, National Research Council, Washington, D.C., 1981, pp. 1–7. Highway Capacity Manual: A Guide to Multimodal Mobility Analysis. 6th ed. Transportation Research Board, Washington, D.C., 2016. Pedersen, N. J., and D. R. Samdahl. NCHRP Report 255: Highway Traffic Data for Urbanized Area Project Planning and Design. Transportation Research Board, National Research Council, Washington, D.C., Dec. 1982. Pedutó, F., G. Cioffi, and R. Albertin. Documentation of Selected Programs of the NYSDOT’s Simulation System. New York Department of Transportation, Albany, June 1977. Washington State Department of Transportation. Peak Hour Report: Year 2007. Transportation Data Office, Olympia, 2008. Zegeer, J. D., M. A. Vandehey, M. Blogg, K. Nguyen, and M. Ereti. NCHRP Report 599: Default Values for Highway Capacity and Level of Service Analyses. Transportation Research Board of the National Academies, Washington, D.C., 2008.

29 E. Predicting Intersection Traffic Control 1. Overview Analyzing the operation of an urban street using the HCM requires some knowledge of the type of traffic control used at the intersections along the street. When analyzing future conditions as part of a planning or preliminary engineering analysis, decisions may not have been made about the type of traffic control used at an intersection, or the purpose of the analysis may be to determine the type of traffic control that would likely be needed in the future under a particular analysis scenario. This section provides guidance on forecasting which type of traffic control may be needed at an intersection in the future, for use in preparing inputs to an HCM planning analysis. The analyst should be aware that state and local policies may often specify the conditions under which particular types of intersection traffic control should or should not be considered. These policies supersede the guidance presented in this section. 2. Manual on Uniform Traffic Control Devices FHWA’s Manual on Uniform Traffic Control Devices (MUTCD 2009) provides warrants and criteria to help determine whether a traffic signal or all-way stop control may be justified at an intersection. Meeting one or more warrants does not automatically mean a particular type of traffic control is justified, but not meeting the warrants generally means that type of traffic control would not be justified. State supplements to the MUTCD, or state or local policies, may specify that certain warrants found in the MUTCD should not be used, and planning studies should respect those policies. Determining 8th- and 4th-Highest Hour Volumes The most commonly applied MUTCD warrants require the analyst to determine the 8th- or 4th-highest hour traffic volumes. The decision to install a traffic signal would normally be based on actual traffic counts, but when a planning or preliminary engineering analysis is being performed, future volumes are being estimated and typically exist only in the form of AADTs or peak hour volumes. Therefore, some other means is required to estimate what the 8th- or 4th-highest volume would be. Possible methods for doing so include the following, in order of preference: • Calculate the ratio of 8th- (or 4th-) highest hour traffic volumes to peak hour traffic volumes using recent traffic counts from the intersection or a similar intersection. • Calculate the ratio of 8th- (or 4th-) highest hour traffic volumes to peak hour traffic volumes using data from a permanent traffic recorder in the area.

30 Planning and Preliminary Engineering Applications Guide to the Highway Capacity Manual • Apply a factor to the peak hour traffic volume. The specific factor will depend on how peaked the peak hour is. For example, when peak hour traffic represents 7.8% of AADT, the 4th-highest hour volume is approximately 90% of the peak hour volume, while the 8th-highest hour is approximately 80% of the peak hour volume (May 1990). On the other hand, when peak hour traffic represents 10.6% of AADT, the 4th-highest hour volume is approximately 67% of the peak hour volume, and the 8th-highest hour volume is approximately 55% of the peak hour volume (ITE 1982). In both cases, the 4th-highest volume represents approximately 7% of AADT, while the 8th-highest volume represents approximately 6% of AADT. Applying MUTCD Warrants The basic information needed to apply the MUTCD warrants is listed in Exhibit 16. Once the required data are available, the appropriate sections of the MUTCD are consulted to determine whether the traffic volumes would satisfy one or more warrants, given the other conditions (e.g., number of lanes, major street speed) existing at the intersection. These are: • Section 4C.02 for the 8-hour traffic signal warrant, • Section 4C.03 for the 4-hour traffic signal warrant, and • Section 2B.07 for the all-way stop control criteria. 3. Graphical Method As an alternative to evaluating the MUTCD traffic signal warrants, graphical methods can be used to predict the future intersection traffic control for use in a planning analysis. (MUTCD warrants, supplemented with state or local practice and engineering judgment, should always be used in making final decisions about intersection traffic control.) Graphical methods have the advantage of requiring fewer data than a signal warrant evaluation does, but have the dis- advantage of employing built-in assumptions that may not be appropriate for a given location. Exhibit 17 can be used to determine the likely future intersection traffic control, using only peak hour two-directional volumes for the major and minor streets and the directional distri- bution of volumes (50/50 or 67/33) as inputs. The signal warrant incorporated in the exhibit is the basic MUTCD eight-hour minimum hourly volume warrant for locations with populations of 10,000 or greater, major street speeds of 40 mph or less, and single-lane approaches. If other Input Data (units) For 8HR For 4HR For AWS Default Value 8th-highest vehicular volume by approach (veh/h) • • 6% of AADT Number of lanes on major street approach • • Must be provided Number of lanes on minor street approach • • Must be provided Major street speed (mph) • • • Posted speed City population < 10,000 (yes/no) • • • Must be provided 4th-highest vehicular volume by approach (veh/h) • 7% of AADT Peak hour minor street delay (s/veh) • Must be provided; see Section N5 for guidance Notes: See MUTCD Section 4C for definitions of the required input data and additional guidance. 8HR = 8-hour signal warrant, 4HR = 4-hour signal warrant, AWS = all-way STOP warrant. Exhibit 16. Required data for MUTCD warrant analysis.

E. Predicting Intersection Traffic Control 31 (a) 50/50 Volume Distribution on Each Street (b) 67/33 Volume Distribution on Each Street Source: Calculated from MUTCD 8-hour signal warrant, MUTCD all-way STOP warrant, and HCM methods for roundabout capacity and STOP-controlled intersection delay. Notes: Assumes eighth-highest-hour volumes = 55% of peak hour volumes, peak hour factor = 0.92, 10% left turns and 10% right turns on each approach, and a single lane on each approach as the base case. See text for an explanation of how boundaries between regions in the graphs were determined. Exhibit 17. Intersection control type by peak hour volume.

32 Planning and Preliminary Engineering Applications Guide to the Highway Capacity Manual warrants or conditions are desired to be evaluated, then the MUTCD method described in Section E2 should be used instead. As indicated in Exhibit 17, a roundabout is a potential option in many cases, in lieu of stop or traffic signal control. In deciding which option to use in the analysis, the analyst should consider local policies favoring or disfavoring roundabouts, as well as potential right-of-way or other constraints at the location. Chapter 3 of NCHRP Report 672 (Rodegerdts et al. 2010) provides planning-level considerations for making choices between roundabouts and other forms of intersection control. The upper boundary for two-way stop control in these graphs occurs when all-way stop warrants are met or when demand on the higher-volume minor street approach exceeds its capacity, whichever comes first. The lower boundary for a single-lane roundabout is set at the LOS C/D threshold of 25 seconds of average delay per vehicle for the higher-volume minor street approach; the upper boundary is set at 85% of the capacity of an entry to a single-lane round- about. The boundary between single-lane and multilane all-way stop control is set at the HCM’s LOS E/F threshold for a single-lane all-way stop intersection (HCM 2016). The lower-right portion of the graphs includes the label “restrict left turns.” In this region, major street volumes may be too high to provide sufficient capacity for side-street left turns, but side-street volumes are too low to meet all-way stop or traffic signal warrants. In this case, the side street might need to be restricted to right turns out only. (This region of the graphs may also indicate the need for access management measures for minor streets and driveways along an extended length of the major street.) 4. References Highway Capacity Manual: A Guide to Multimodal Mobility Analysis. 6th ed. Transportation Research Board, Washington, D.C., 2016. Institute of Transportation Engineers. Traffic and Transportation Engineering Handbook. Washington, D.C., 1982. Manual on Uniform Traffic Control Devices. Federal Highway Administration, Washington, D.C., 2009. May, A. Traffic Flow Fundamentals. Prentice Hall Publishing, Englewood Cliffs, N.J., 1990. Rodegerdts, L., J. Bansen, C. Tiesler, J. Knudsen, E. Myers, M. Johnson, M. Moule, B. Persaud, C. Lyon, S. Hallmark, H. Isenbrands, R. B. Crown, B. Guichet, and A. O’Brien. NCHRP Report 672: Roundabouts: An Informational Guide. 2nd ed. Transportation Research Board of the National Academies, Washington, D.C., 2010.

33 F. Default Values to Reduce Data Needs 1. Overview Many HCM computational methods require a number of input parameters. For a detailed operations analysis, this can be an advantage, as the performance measure output by the method reflects many different factors that can influence the result. However, for planning and preliminary engineering analyses, the num- ber of inputs can pose a challenge. The desired information may not yet be known, the level of effort required to gather the data may be out of proportion to the aims of the analysis, or a combination of these and other considerations can make it difficult to supply a particular input value. One solution to applying HCM methods to planning and preliminary engineering analyses is to substitute default values for those inputs that cannot be measured directly. Using default values instead of field-measured values may introduce some error into the analysis results, but other data used for planning analyses (particularly forecast demand volumes) may have much greater uncertainties associated with their values and, consequently, much greater impact on the results. Furthermore, the goal of these types of analyses is not to make final decisions about roadway design and control elements, but rather to identify potential problems or to screen large numbers of alternatives; in these cases, precise results are neither required nor expected. It is important to recognize that HCM input data have a hierarchy that varies according to the context of the planning and preliminary engineering application: There are applications where certain input data can be and must be measured. (These data are identified as “required inputs” in subsequent sections.) There are planning and preliminary engineering applications where certain input data can and should be estimated sensibly based on local and planned conditions; Section F4 addresses this situation. Finally, as discussed in Section F2, there are applications where certain data need not be measured and a general default value can be used instead. Parts 2 and 3 of the Guide provide simple default values for analysis situations where the analyst has deemed a locally measured value is not necessary. This section provides guidance on applying default values to HCM methods and on developing local default values to use in place of the HCM’s national defaults. 2. When to Consider Default Values The decision to use a default value in place of a field-measured value should consider a num- ber of factors, including: • The intended use of the analysis results. In general, the less precisely that analysis results will be presented (e.g., under, near, or over capacity versus a particular LOS versus a specific travel speed estimate), the more amenable the analysis is to using default values, or tools based

34 Planning and Preliminary Engineering Applications Guide to the Highway Capacity Manual on default values, such as service volume tables. Similarly, the farther away a final decision is (e.g., identifying potential problem areas for further analysis versus evaluating a set of alter- natives versus making specific design decisions), the less potential exists for incorrect decisions to be drawn from the analysis results due to the use of a default value. • The scale of the analysis. The larger the geographic scale of the analysis (i.e., the greater the number of locations that need to be analyzed), the greater the need to use default values due to the impracticality of collecting detailed data for so many locations. • The analysis year. The farther out into the future that conditions are being forecast, the more likely that information will not be known with certainty (or at all), and the greater the need to apply default values. • The sensitivity of the analysis results to a particular input value. Sections H through O of this Guide provide information about the sensitivity of analysis results to the inputs used by a given HCM operations method. Input parameters are characterized as having a low, moderate, or high degree of sensitivity, depending on whether a method’s output changes by less than 10%, 10% to 20%, or more than 20%, respectively, when an input is varied over its reasonable range. The lower the result’s sensitivity to a particular input, the more amenable that input is to being defaulted. • Ease of obtaining field or design data. According to the HCM (2016), input parameters that are readily available to the analyst (e.g., facility type, area type, terrain type, facility length) should use actual values and not be defaulted. • Inputs essential to an analysis. A few inputs to HCM methods, such as demand volumes and number of lanes, are characterized as “required inputs” and should not be defaulted. When the purpose of the analysis is to determine a specific value for a required input (e.g., the maximum volume for a given LOS), the HCM method is run iteratively, testing different values of the input until the desired condition is met. • Local policy. State and local transportation agencies’ traffic analysis guidelines may specify that particular inputs to HCM methods can or should not be defaulted. 3. Sources of Default Values Once a decision has been made to use a default value for a particular methodological input, there are several potential sources for obtaining a default value. These are, in descending order of desirability according to the HCM (2000): • Measure a similar facility in the area. This option is most applicable when facilities that have not yet been built are being analyzed and the scope of the analysis does not require measuring a large number of facilities. • Local policies and standards. State and local transportation agencies’ traffic forecasting guide- lines may specify, or set limits on, default values to assume. Similarly, these agencies’ roadway design standards will specify design values (e.g., lane widths) for new or upgraded roadways. • Local default values. When available, local default values will tend to be closer to actual values than the HCM’s national defaults. Heavy vehicle percentage, for example, has been shown to vary widely by state and facility type (Zegeer et al. 2008). The next subsection provides guidance on developing local default values. • HCM default values. If none of the above options are feasible, then the HCM’s national default values can be applied. 4. Developing Local Default Values This section is adapted from HCM (2016), Chapter 6, Appendix A. Local defaults provide input values for HCM methods that are typical of local conditions. They are developed by conducting field measurements in the geographic area where the values

F. Default Values to Reduce Data Needs 35 will be applied, during the same time periods that will be used for analysis, typically weekday peak periods. For inputs related to traffic flow and demand, the peak 15-minute period is recom- mended as the basis for computing default values because this time period is most commonly used by the HCM’s methodologies. When an input parameter can significantly influence the analysis results, it is recommended that multiple default values be developed for different facility types, area types, or other factors as appropriate, as doing so can help reduce the range of observed values associated with a given default and thus the error inherent in applying the default. The K- and D-factors used to con- vert AADT volumes to directional analysis hour volumes are two such parameters. For urban streets, other sensitive parameters include peak hour factor, traffic signal density, and percent heavy vehicles. For freeways and highways, sensitive parameters include free-flow speed and peak hour factor. 5. References Highway Capacity Manual: A Guide to Multimodal Mobility Analysis. 6th ed. Transportation Research Board, Washington, D.C., 2016. Highway Capacity Manual 2000. Transportation Research Board, National Research Council, Washington, D.C., 2000. Zegeer, J. D., M. A. Vandehey, M. Blogg, K. Nguyen, and M. Ereti. NCHRP Report 599: Default Values for Highway Capacity and Level of Service Analyses. Transportation Research Board of the National Academies, Washington, D.C., 2008.

36 G. Service Volume Tables to Reduce Analysis Effort 1. Overview One typical planning application of the HCM is to estimate the existing or future LOS of a large number of roadway links. For example, this activ- ity might be performed as part of a screening evaluation (to identify links requiring more detailed analysis) or as part of an agency’s roadway system monitoring program. Generalized service volume tables, which estimate the maximum daily or hourly volume that a roadway can serve under an assumed set of conditions, can be useful tools for performing these types of evaluations. This section describes how service volume tables can be incorporated into a planning analysis to reduce the overall analysis effort. 2. Description Service volume tables are look-up tables that estimate the maximum daily or hourly volume for a given LOS under a specific combination of conditions. For ease of use, generalized service volume tables require a minimum of user inputs—typically, key design parameters that have the greatest influence on a facility’s capacity and LOS, such as the number of lanes. Given these inputs, a user can then read the maximum volume (service volume) for a given LOS directly from the table and compare it with the actual or forecast volume for the facility. A volume greater than the service volume for the desired LOS indicates the need for further analysis (HCM 2016). The area type (e.g., urban, rural) often serves as a proxy for many default values (for example, driver population, percentage heavy vehicles, peak hour factor). As such, the area type often has a significant effect on the service volumes. It is unlikely that any given roadway’s characteristics will exactly match the default values used in creating the table. Therefore, conclusions drawn from the use of service volume tables should be con- sidered to be, and presented as, rough approximations. In particular, generalized service volume tables should not be used to make final decisions about important roadway design features—this activity requires a full operational analysis. However, as long as the analyst recognizes and respects the limita- tions of this tool, generalized service volume tables can be a useful sketch-planning tool for developing quick estimates of LOS and capacity, especially for large numbers of facilities (HCM 2016). 3. When to Consider Service Volume Tables The decision to use a service volume table should consider a number of factors, including: • The scale of the analysis. The larger the geographic scale of the analysis (i.e., the greater the number of locations that need to be analyzed), the more applicable service volume tables become

G. Service Volume Tables to Reduce Analysis Effort 37 due to the impracticality of collecting detailed data for so many locations. When a small number of locations is being analyzed, other analysis tools will likely provide more accurate results, as it becomes more feasible to collect data, apply less-generalized default values, or both. Neverthe- less, service volume tables can be applied to smaller sets of locations when the outcome of the analysis does not require a higher level of accuracy. • The intended use of the analysis results. Service volume tables are well-suited to analyses where the identification of a potential operational problem will lead to a follow-up, more detailed analysis using more accurate tools and input data. They are also well-suited to perfor- mance management applications involving LOS or capacity calculations (e.g., calculating the number of miles of state highway operating at LOS E or worse during peak periods). They are not suitable for making final decisions about roadway design or control elements, nor for making final assessments about the adequacy of a roadway to accommodate additional demand (such as might be done as part of a traffic impact study). • Availability of a suitable table. The accuracy of the results from a generalized service volume table depends greatly on how well the default values used to generate the table match condi- tions on the roadway being analyzed. The next subsection discusses potential sources of service volume tables and their respective advantages and disadvantages. Assumptions common to any service volume table include: (1) uniform roadway cross-section, (2) uniform roadway demand, (3) no queue spillback (e.g., from a left-turn lane, from an off-ramp, from one freeway segment to another), and (4) traffic signal timing that adequately accommodates all turning movements (HCM 2016; Florida DOT 2013). The more that actual conditions vary from these assumptions, the less suitable a service volume table will be. • Local policy. State and local transportation agencies’ traffic analysis guidelines may specify that a particular service volume table should be used for particular types of analyses, or that service volume tables should not be used in particular circumstances. The analyst should also be cautious when the estimated LOS is near or at LOS F. The actual operations of the intersection, segment, or facility fluctuate a great deal at the LOS E/F boundary. Consequently, service volume tables cannot be relied upon when approaching this boundary. More detailed analyses are required to better pinpoint the actual operations. 4. Sources of Generalized Service Volume Tables There are three main sources for generalized service volume tables, which are discussed in more detail in the remainder of this section: 1. The HCM’s generalized service volume tables, 2. Florida DOT’s generalized service volume tables, and 3. Local service volume tables. HCM Generalized Service Volume Tables The HCM (2016) provides generalized service volume tables for the following system elements: • Basic freeway segments (Chapter 12, Exhibits 12-39 and 12-40), • Multilane highways (Chapter 12, Exhibits 12-41 and 12-42), • Two-lane highways (Chapter 15, Exhibit 15-46), • Urban street facilities (Chapter 16, Exhibit 16-16), and • Signalized intersections, as an illustration (Chapter 19, Exhibit 19-36). The assumptions (e.g., default values) used to develop the tables are provided with each table and explained in the accompanying text in the HCM. The default values used to develop the tables are based on the HCM’s national average values, which may be different from local conditions in the area being analyzed. In particular, the default values for percentage heavy vehicles, peak hour

38 Planning and Preliminary Engineering Applications Guide to the Highway Capacity Manual factor, and free-flow speed are recommended to be compared to local conditions, if possible, when evaluating the suitability of the HCM tables for a particular analysis. For urban streets, through traffic g/C ratio (the percentage of time through traffic receives a green signal at a traffic light) and traffic signal spacing are additional parameters that are recommended to be compared to local conditions when possible. Except for the signalized intersection table, all of the HCM’s tables are daily tables (i.e., they present maximum AADTs for a given LOS) and the user must select appropriate K- (analysis hour) and D- (directional) factors that convert AADT to an analysis hour directional volume when applying the table. The signalized intersection table presents maximum hourly volumes for a given LOS; users can convert these to AADTs by applying K- and D-factors. Other inputs required by the HCM tables are: • Number of travel lanes, • Terrain type (freeways and highways), • Area type (freeways and multilane highways), • Highway class (two-lane highways), • Posted speed (urban streets), and • g/C ratio (signalized intersections). FDOT Generalized Service Volume Tables The Florida DOT (FDOT) is one of the leading users of generalized service volume tables and has sponsored a considerable body of research related to them. FDOT’s Quality/Level of Service Handbook (2013) describes the assumptions and methodological extensions used in developing the FDOT tables; the tables themselves also list the input values used to develop them. The default values used by FDOT’s tables are based on typical Florida values. In particular, the daily Florida tables apply default K- and D-factors for specific combinations of facility geometries (e.g., four-lane undivided arterials), ensuring consistent application of the tables across the state. As with the HCM tables, key default values that can significantly affect results should be compared to local conditions when possible. These include percentage heavy vehicles, peak hour factor, and free-flow speed; for urban streets, these also include through traffic g/C ratio, saturation flow rate, and traffic signal spacing. The FDOT tables assume level terrain. Both daily and peak hour service volume tables are provided for the following facility types and travel modes: • Signalized arterial streets, • Freeways, • Uninterrupted-flow highways (multilane and two-lane highways use the same table), • Bicycles on urban streets, • Pedestrians on urban streets, and • Public transit buses on urban streets. Input data required by these tables consist of: • Signalized arterials: state or non-state roadway, number of lanes, posted speed, median type, presence of exclusive left- and right-turn lanes, and one- or two-way facility; • Freeways: number of lanes, auxiliary lane presence, and ramp meter usage; • Uninterrupted-flow highways: number of lanes, median type, presence of exclusive left-turn lanes, and (for two-lane highways only) passing lane percentage; • Bicycles: percent of facility with a paved shoulder or bicycle lane (three categories correspond- ing to nearly all, more than half, less than half);

G. Service Volume Tables to Reduce Analysis Effort 39 • Pedestrians: percent of facility with a sidewalk (same categories as for bicycles); and • Buses: percent of facility with a sidewalk (two categories of “nearly all” and “all others”) and bus frequency. Although FDOT uses the HCM as the starting point for the computations used to develop its tables, there are some important differences in the methodologies that mean that the FDOT tables will produce different results than a “pure” application of the HCM method. Key differences include: • Signalized arterials: methodological extension for auxiliary lanes through intersections (i.e., extra through lanes on the approach and exit to an intersection), and use of arterial classes for determining LOS thresholds; • Freeways: treatment of capacity reductions in interchange areas, maximum capacity values for different area types, and a methodological extension for ramp metering effects; • Uninterrupted-flow highways: methodological extension for left-turn lane provision, and a different two-lane highway method than used by the HCM; • Bicycles and pedestrians: slightly different computations than the HCM methods; and • Buses: a method adapted from the Transit Capacity and Quality of Service Manual (TCQSM) 2nd Edition (Kittelson & Associates et al. 2003), while the HCM (2016) and the TCQSM 3rd Edition (Kittelson & Associates et al. 2013) use a different method that consider some of the same factors. Local Service Volume Tables Developing local service volume tables is a way to address a key issue with applying service volume tables—namely, that the assumptions used to develop the tables may not necessarily match local conditions. In addition, local service volume tables can be developed that allow the user to vary other parameters than those used by the HCM or FDOT tables. The effort taken to develop local tables can pay off with the creation of an easy-to-apply set of service volume tables that produces reasonable results. The HCM (2016) describes a method for developing local service volume tables in Appendix B of Chapter 6. The analyst needs to develop a default value for each input parameter used by the applicable HCM method. When the HCM method is particularly sensitive to a particular parameter, or when the range of local observed values varies greatly, a set of default values should be considered for that parameter. Section F of the Guide provides guidance on selecting appropriate default values. Once the default values are selected, the analyst uses a computational engine or HCM-implementing software to back-solve for the maximum volume associated with a particular LOS, using the analyst’s selected set of default values. As an alternative, FDOT’s LOSPLAN planning software package provides table generators that build service volume tables from a set of user-specified input values (Florida DOT 2013). The user should be aware of the differences between the FDOT and HCM methods, highlighted above, before applying these service volume table generators. 5. References Florida Department of Transportation. 2013 Quality/Level of Service Handbook. Tallahassee, 2013. Highway Capacity Manual: A Guide to Multimodal Mobility Analysis. 6th ed. Transportation Research Board, Washington, D.C., 2016. Kittelson & Associates, Inc.; KFH Group, Inc.; Parsons Brinckerhoff Quade and Douglass, Inc.; and K. Hunter- Zaworski. TCRP Report 100: Transit Capacity and Quality of Service Manual, 2nd ed. Transportation Research Board of the National Academies, Washington, D.C., 2003. 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.