Traffic Analysis Guidelines

Scoping is the critical first step to any traffic analysis, and the following section presents key items that directly affect selecting the best modeling tool (Step 2) and collecting the most relevant data (Step 3) for the analysis. While this section is not an exhaustive list for scoping a project, it does represent important considerations related to the analysis goals, project context, model limits, and study area characteristics. The consultant should meet with the UDOT project team to review and discuss the items that are best suited for the type and level of analysis to be include in the project.

Figure 3 - A continuous flow intersection requires the use of more complex traffic modeling software that needs to be considered when scoping a project

1.1 Goals & Desired Outcomes
The goal of any traffic analysis is to provide actionable information that decision-makers can use to improve the transportation system. It is essential that this goal is understood so scoping can be adequately defined.

RANGE OF ALTERNATIVES TO BE CONSIDERED

Consultants, the Region Traffic & Safety groups, and the Traffic Management Division are all great resources to identify alternatives worth considering. The consultant can identify alternatives based on various needs, including budget constraints, right-of-way limitations, and environmental requirements. Understanding where and how a project may affect traffic patterns are also important factors when scoping a project, as these impacts often vary. The consultant should keep in mind that for an interchange project FHWA requires specific alternatives be considered.

HORIZON YEARS

Understanding the horizon years for the traffic analysis is also an important part of the scoping process. Virtually every analysis will consider existing conditions. For some projects there may be only year analyzed; however, most analyses will also include one or more future years. The most common future year is one of the horizon year(s) from the most recent long-range transportation plan. (This is a requirement for projects that require FHWA review) Some projects will also evaluate an interim year(s) to best represent when improvements will be needed or to determine how a project might be phased over time. Every year analyzed involves work in developing traffic volumes and performing the actual analysis, which makes establishing the years an integral part of scoping.

1.2 Project Context

PROJECT TYPE

Project type may affect the traffic scope. For example, an interchange project may require an Interstate Access Change Request (IACR), or controversial projects (e.g., projects with substantial impacts) may necessitate stakeholder outreach. Environmental studies may require additional documentation or data to inform air quality or noise analyses. It is important to note that when a project includes an IACR, both FWHA and UDOT Region Traffic Engineers are involved in the review and approval of the models and analysis.

PREVIOUS & CONCURRENT PROJECTS

Knowing about nearby previous or concurrent projects providesan opportunity to share data or analysis results to avoid duplication of effort. The consultant may have to coordinate assumptions, such as future land use, to ensure consistency between analyses. Adjacent construction may also affect scope, such as schedule, traffic patterns, data collection, and traffic conditions.

PROJECT SENSITIVITY

Depending on the limitations of alternative solutions considered and the modeling tool selected, the consultant may need to address traffic volume fluctuations or other unknown study area elements. The consultant may also have to perform sensitivity tests around several assumptions, including alternate land use scenarios, different horizon years, and the presence or phasing of nearby planned projects. Of note, analyzing the project’s sensitivity for these types of factors could lead to a more robust solution to address unknown elements.

1.3 Model Limits 

The consultant should carefully consider the purpose of the analysis and the range of alternatives when determining the traffic analysis study area limits. Project scoping elements should address both the immediate project area and the larger traffic influence area for a project.

PROJECT AREA

The project area generally includes the extents of the project’s design elements (i.e., where changes might be made). The size of the project area may vary dramatically depending on the type of project being analyzed. An intersection improvement project would generally have modifications at just one intersection, while a corridor widening could extend for several miles and include numerous intersections.

Figure 4 - The highlighted area shows an example of the project area for the study

INFLUENCE AREA

The influence area includes a larger area where traffic patterns would be affected by a project. It generally extends beyond the project area to include adjacent intersections, interchanges, or corridors. Typically, the size of an influence area is based on the regional significance of a project, with bigger projects often requiring larger influence areas. UDOT generally follows FHWA requirements for the area of influence, even if the project is not submitted to FHWA for review. At the least, the consultant should consider the first adjacent interchange on either side of the proposed change in access. The area of influence along a local roadway network could then extend at least to the first adjacent signal in either direction or to the first major intersection. The consultant could also expand the area of influence beyond these limits based on the impact of the proposed change in access, particularly if a coordinated signal system is involved. The consultant should consider the influence area (and the study area characteristics described in Section 1.4 of this guideline) when establishing the traffic analysis study area. The consultant should confirm with the UDOT Project Manager that this study area will meet the needs of the project.

Figure 5 - The highlighted area shows an example of the potential influence area for a project area, which extends beyond the project area.

LAND USE FORECAST

Depending on the analyzed horizon year(s), the projected land use, and the resulting socioeconomic forecasts may play an important role in the traffic analysis. Land use forecasts are key inputs to the travel demand models (TDMs) used to estimate future traffic volumes. The consultant can generally obtain land use forecasts from the organization that owns and maintains a particular travel model (i.e., the local Metropolitan Planning Organization [MPO] or UDOT, depending on whether the project is in an urbanized or rural area). These forecasts are prepared at the traffic analysis zone (TAZ) level and comprise of residential and employment data. For some projects, a consultant may need to revise the land use data based on feedback from a city or other stakeholders. Theconsultant may also have to refine the TAZ structure in the project and/or influence area to better match local travel patterns and the existing or future transportation network.

Figure 6 - Understanding land use characteristics is important for future growth projections.

1.4 Study Area Characteristics

The consultant should consider the following general items during the scoping and analysis process:

• Existing Geometry (lane configuration, shoulders, etc.)

• Travel Speeds (speed limits & free-flow speeds)

• Traffic Volumes

• Functional Class

• Intersection Control Type

However, UDOT has also identified some specific characteristics that may demand special attention throughout the process.

FUTURE VOLUMES

How and where existing volumes are converted to future volumes are important elements to review with UDOT at the time of scoping. The consultant can use several methods to grow existing volumes based on the horizon years. While UDOT does not have a preferred method, some typical methods include historical growth, the TDM based on socioeconomic growth, and ITE trip generation, and others. The choice on method will depend on the size and scope of the project area and influence area. Other future growth items include converting daily or period volume to daily turning movements, truck percentages, and peak hour factors. The consultant should review and discuss methods with UDOT for best results

GRADES

The grade or slope of a roadway can significantly impact the way vehicles travel through an area. This is particularly the case in areas with a high percentage of heavy trucks or roadways with long stretches of incline.

TRUCK PERCENTAGES/ROUTES

The percentage of truck traffic often changes the dynamic of traffic patterns, which, in turn, impacts data collection and modeling tool selection. Trucking routes or locations with a large percentage of trucks may increase congestion by limiting the speeds on a corridor.

DRIVER BEHAVIOR

The consultant should consider driver characteristics in the study area as part of the analysis process. For example, there may be more commuter traffic, recreational traffic, or inexperienced drivers based on the location of the project, and this may affect driver behavior differently.

PEAK PERIODS

Another important scoping consideration is the time periods analyzed. This is most often the weekday AM and PM peak hours, but the consultant can also evaluate weekend peak hours or a daily volume level. For example, most schools will have mid-afternoon peaks; industrial areas may have shift work with peaks that vary from the traditional 8 AM to 5 PM; and large retail areas may have Saturday afternoon peaks. As traffic volume and congestion increase, analyzing a single peak hour may be insufficient, and the consultant may need to include a peak period of two or three hours. UDOT has several different data sources (see section 5 of this guideline) that may help in identifying the peak periods before more data is collected.

PEDESTRIANS & BIKES

Pedestrian and bicycling amenities are an increasingly important part of the transportation system and, therefore, are important to consider with the project design and traffic analysis. The consultant should account for how pedestrians, bicyclists, and motorized vehicles interact with one another.

Figure 7 - Turning Movement Counts will show that some areas may have unique characteristics requiring different study tactics similar to this location showing a peak hour starting at 3:45 PM. This would require traffic data collected earlier than typical locations in Utah.

ACCESS LOCATIONS & DENSITY

In some locations, it may be beneficial to include driveways or unsignalized intersections as part of the traffic analysis. This is most common for access points to large traffic generators or on corridors with closely spaced signalized intersections, where access may be constrained or may influence traffic operations. When accesses have little impact upon the evaluation, it may be appropriate to combine or eliminate accesses from an evaluation.

ORIGIN & DESTINATION (O-D)

For some traffic analyses, it is important that the consultant has a detailed understanding of the traffic patterns through some or all of the study area. For example, an interchange with nearby intersections may have a substantial number of weaving vehicles as traffic turns onto a main street and then needs to change lanes to turn onto a freeway. Understanding the origin and destination (O-D) data for the various movements may help in solving the problem.

Selecting the right analysis tool is essential for the type of evaluation, and this section summarizes UDOT’s preferences to ensure consistency around the tools and methods used for a traffic analysis. It is important to discuss the analysis tool, project/influence area, and measures of effectiveness (MOEs) with the UDOT Project Manager/reviewer as a part of scoping (see Chapter 1 of this guideline).

2.1 Intersection

A single intersection analysis involves the most basic form of traffic an alysis. This can include proposed changes to a lane configuration, control type, and other elements within the intersection boundaries. The consultant should also include any adjacent driveway or other features that may be in the general influence area of an intersection.

MODEL EXTENTS

The study area would generally entail the intersection itself and enough roadway on each leg to account for queuing. The consultant should also consider adjacent intersections to represent more realistic traffic conditions by metering vehicles arriving at the intersection based on the coordination of upstream signals, rather than the default random arrivals produced by the analysis tools.

Figure 8 - Queuing like shown on this figure needs to be accounted for in an intersection analysis and may require expanding the model.

TYPICAL TOOLS

To analyze a traditional 4-leg intersection, Synchro/SimTraffic or Vistro is generally acceptable. The ability to analyze various signal types, account for pedestrian interaction, and record delay and queuing data makes these tools effective. However, when analyzing innovative intersections, such as a continuous flow intersection (CFI) or ThrU Turns, Vissim is required. Synchro/SimTraffic have built-in analysis for roundabouts that can be used based on the complexity of the project; however, Vissim may be required based on the projected needs. Synchro/SimTraffic and Vistro do not have the ability to create or analyze non-traditional intersections, specifically those with unique lane configurations or complicated signal timing. In all, the consultant should use engineering judgment when determining which tool is the right choice for intersection analysis.

TYPICAL MOEs

Typical MOEs for this type of analysis are average control delay per vehicle, level of service (LOS), percent demand served, and 95th percentile queue lengths. The consultant can extract these MOEs at the movement, approach, and intersection levels, using total intersection delay to calculate intersection LOS at signalized intersections.

2.2 Interchange

An interchange analysis evaluates control delay at the ramp terminals and often includes proposed changes to the interchange configuration within boundaries that exclude the freeway mainline. (See Section 2.4 of this guideline for an interchange analysis that deals with changes outside of these localized boundaries and extends into the freeway mainline.)

Figure 9 - Interchange analyses may require more robust analysis tools to model the often complex traffic patterns found at interchanges.

MODEL EXTENTS

Analyzing an interchange is similar to an intersection, although the area of influence is often larger. The modeling extents should include adjacent intersections and interchanges.

TYPICAL TOOLS

Although a consultant can use Synchro/SimTraffic or Vistro to model a traditional diamond interchange, UDOT’s preferred tool is Vissim because interchange alternatives often include innovative designs (e.g., a diverging diamond interchange [DDI]). The consultant should consider the range of alternatives and use engineering judgment when determining which tool is the right choice for interchange analysis.

Figure 10 - Synchro is a analysis tool that often works well when analyzing an arterial corridor.

TYPICAL MOEs

Typical MOEs for interchange analyses include average control delay in seconds per vehicle, LOS, and 95th percentile queue lengths at the interchange. When comparing interchanges, it is important to note that the HCM has specific thresholds and calculation methods from an intersection and comparisons should use consistent methods and thresholds. Travel times through the interchange may also be relevant, and percent of demand served is a helpful MOE when dealing with congested conditions.

2.3 Arterial Operations

An arterial is a collection of intersections and/or interchanges with its analysis being very similar to its component parts. Arterial projects can include coordination and progression, additional lanes, removal of lanes, or facility type changes. Design features will vary by projects, thus the arterial analysis should include all changes within a study area.

MODEL EXTENTS

The consultant should use a similar approach with that of both intersections and interchanges for an arterial project. Adjacent intersections are often included in arterial analyses when metered traffic may affect arterial performance, or if specific signals have heavy side-street traffic volume from coordinated signals.

TYPICAL TOOLS

Depending on the style of intersections or interchanges in the arterial, Synchro/SimTraffic or Vistro is often the best tool. UDOT requires Vissim when innovative intersections, such as CFIs or DDIs, are being considered. The consultant can use Vissim to provide more detailed results, such as travel time for specific routes through the network and detailed queuing. (Refer to Section 2.1 and 2.2 of this guideline for discussion on the best tool for arterial analysis depending on the composition of the arterial’s intersections and/or interchanges.) The consultant should consider the alternatives and use engineering judgment to determine which tool is the right choice for arterial analysis..

TYPICAL MOEs

Typical MOEs for arterials often include arterial LOS, average travel times, average speeds, daytime base speeds, average control, delay in seconds per vehicles, percent served, and 95th percentile queue lengths.

Figure 11 - VISSIM is a tool that can be used to evaluate complicated freeway systems and to understand complex queuing issues.

2.4 Freeway Operations

A freeway operations analysis is an evaluation of an interchange and the connected freeway segments.

MODEL EXTENTS

The consultant should include at least the adjacent interchange both upstream and downstream of the interchange being evaluated, in addition to the adjacent signals in either direction or up to the first major intersection in the study area. The consultant can use the adjacent intersections to provide effective metering to the interchange. The model may need to extend even further, if a coordinated signal system is involved.

TYPICAL TOOLS

UDOT requires Vissim for freeway operations analyses.

TYPICAL MOES

Typical MOEs for interchanges are average control delay in seconds per vehicle, LOS, percent served, and 95th percentile queue lengths for the peak 15 minutes at the interchange. Important MOEs for freeway analyses are volume throughput, speeds, and density, which attributes are used to calculate freeway LOS. Travel times can also be an effective MOE for presenting freeway performance and 95th percentile queue lengths that back onto the freeway mainline or characteristics of freeway bottlenecks. The LOS and travel times through the interchange, along freeway, merge/diverge, and weave segments, are also relevant.

2.5 Transit

Transit analyses are generally in one of two categories: ridership forecasting or detailed operations.

Figure 12 - Various travel modes may need to be evaluated, which have different MOEs specific to that mode, like ridership for transit.

RIDERSHIP FORECASTING

The typical tools used by UDOT and UTA for transit ridership forecasting are the Transit Boardings Estimation and Simulation Tool (TBEST), Simplified Trips-on-Project Software (STOPS), or the MPO’s TDMs. 

TBEST is a GIS-based model that uses socioeconomic, land use, and transit network data to produce ridership estimates based on a number of scenarios as defined by the analyst. TBEST produces stop-level boardings by time of day and day of week (weekday, Saturday, Sunday) for near and mid-term forecasts. The tool also has an empirically derived adjustment factor for bus rapid transit (BRT) forecasts that adjusts ridership based on vehicle, station, travel way, and branding/marketing characteristics. The Florida DOT developed TBEST as currently used by transit agencies across the country. UTA currently maintains a TBEST model for the Wasatch Front that a consultant can use as a starting point for transit analysis.

STOPS is a simplified transit ridership forecasting tool developed by the Federal Transit Administration (FTA) to meet New Starts and Small Starts project requirements. The tool is only applicable for fixed guideway transit (e.g., light rail, commuter rail, or BRT) to provide stop and route level ridership. STOPS is built off of General Transit Feed Specifications (GTFS) files, which is a common format used for sharing transit schedules with mapping service providers. GTFS files are available through UTA. 

The MPO DMs contain a series of “line” files that represent transit lines for each year of a regional transportation plan (RTP) for all modes of transit (local bus, BRT, LRT, and CRT). These line files define each transit line, stop location, and frequency. Commonly used MOEs include regional mode share information (including total ridership by transit mode) and ridership information at the stop and route level. Operational information also includes estimated transit vehicle speed (peak and off-peak), number of vehicles needed for a route, revenue miles, and revenue hours. One benefit of using the TDM for a transit analysis is the ability to measure impacts of roadway network changes on transit ridership and operations (and vice-versa).

TRANSIT OPERATIONS

Some projects may require detailed analyses of transit operations,particularly as each relates to traffic signal operations. Vissim is UDOT’s preferred tool for these types of analyses, since it allows for analysis of transit vehicles, including simulation of stop activity (boardings and alightings) and signal priority at traffic signals.

2.6 Regional Planning

Regional planning is a higher-level analysis that requires a unique set of skills and understanding not readily transferable from most of the traffic analyses discussed above. The MPO and UDOT’s TDMs are the typical tools for regional planning analyses to measure the larger impact of region, county, district, or corridor-wide improvements.

Each MPO (Cache, WFRC, MAG, & Dixie) owns and maintains its own TDM. Additional models have been developed for some areas outside of the MPOs, such as Tooele Valley, Heber Valley, and Summit County. UDOT also maintains and operates a statewide model (known as the Utah Statewide Travel Model [USTM]) for non-urbanized areas throughout the state. Each TDM is run on Citilabs’ Cube software, and the model structure is coordinated between agencies to maintain model consistency.

Travel models measure impacts due to changes in items such as roadway capacity, roadway classifications, network connectivity, socioeconomic changes, managed lanes, and toll costs. The TDMs are typically not effective in measuring impact due to intersection-level changes (CFI vs. traditional intersection), interchange-level changes (DDI vs. diamond), or operational changes (signal timing or corridor optimization). When using a TDM the consultant should coordinate with the model owner.

Typical MOEs for regional planning tools include roadway traffic volumes and associated volume-to-capacity (V/C) ratios, vehicle-hours of delay, vehicle-miles and vehicle-hours traveled, average speed, and person trips by mode. While some MOEs can be aggregated at a variety of geographic levels to better understand changes in traffic conditions, the models themselves contain a series of post processing scripts that automatically generate some MOEs for user-defined geographic regions. Knowledgeable travel modelers can develop scripts for additional project-specific MOEs.The consultants should review the scripts with the model owner prior to use.

Data collection is the foundation of every traffic analysis. Simply put, it is impossible to perform a good analysis using bad data. The following discusses two areas of data collection: the different types of data (including where and what data is available) and the data requirements for different types of analyses. It is expected that regardless of the data or collection methods, the consultant will verify and validate all data.

3.1 Types of Data

FIELD OBSERVATIONS

Field observations are a critical part of a traffic analysis, and the consultant should consider relevant study area conditions to accurately reflect traffic conditions. The types of data that a consultant could collect during a field observation are volumes, speeds, queuing, travel patterns, and lane changing behavior. It is important to match the type and quality of field observations to the type and scope of the traffic analysis.

Field observations should generally illustrate driving behavior of a study area. While the consultant can use video recordings to supplement a field observation, it is important that the consultant physically visits the site before finalizing the model calibration. The level of quality for a field observation must be sufficient for accurate model calibration. Viewing the site only through aerial imagery is generally not sufficient for most traffic analyses. UDOT may also request field notes to verify that calibration has been completed correctly.

GEOMETRY

Roadway geometry may include lane configuration, turn-lane storage lengths, roadway alignment, and stop bar locations. In many circumstances, aerial imagery is an acceptable data source, with 6-inch accurate aerials available on UPLAN. A consultant may also derive geometric layouts from the project’s concept designs, construction plans, or field observations. One limitation to using aerial imagery is the temporal accuracy of the image. 

The consultant should document the date the photo was taken and the timing of projects in the area to ensure that the assumed conditions reflects the existing geometry of the site. The consultant should also consider the grade of any existing and future geometric layout, calculating this using GIS-based datasets available on the UDOT Data Portal (data.udot.utah.gov).

SPEEDS

Vehicle speeds are important for creating an accurate traffic analysis model. The consultant can use speeds to calibrate models and classify roadway performance. In many cases, using the posted speed limit for the roadway is appropriate for a traffic analysis. However, there is one limitation in that posted speed limits may not accurately reflect the actual speeds of a roadway.

Other sources of speed data could include tube counts, Freeway PeMS and iPeMS, manual collection with radar or lidar guns, and Bluetooth detectors. The consultant can create speed profiles using the PeMS data and manually collected speeds. Speeds used in the traffic models should accurately simulate actual speeds within a study area.

O-D

O-D data determines travel patterns between intersections on a roadway or roadway network. The sources for the data could vary considerably based on a project’s scope. The most common collection method is deploying drones, Bluetooth or Wi-Fi detectors to record unique IDs from enabled devices, such as cars and phones. By matching the IDs from the different detector sites, the consultant can determine O-D patterns. For smaller projects, the consultant can verify O-D patterns via manual counts. For example, a consultant can view how many vehicles turn left at one intersection, and then turn right at the next intersection.

TDMs can be a useful source of O-D information, but the consultant should use this data carefully because a TDM represents modeled data rather than observed data. A common use of this data is to estimate interchange-to-interchange travel patterns for freeways, particularly future freeways.

The consultant can also use the UDOT-purchased vehicle probe data for high-level O-D data. UDOT’s probe O-D data is typically appropriate for regional-level studies. It is not aggregated for small scales. The consultant can aggregate the information to specific geographies for which O-D can be arrived at based on actual GPS data. However, the limitations of this dataset are the size of the geographies for which the data is available and the fact that the data does not necessarily apply to any one transportation facility. For example, the probe data can define a percentage of vehicle trips from Provo that travel to Park City, but it does not distinguish whether a driver used I-80 and Parley’s Canyon or US- 189 and Provo Canyon to get there.

Based on the importance of O-D data to the analysis, it may be necessary to collect O-D data for critical movements to ensure accuracy in a model.

Figure 14 - Collecting traffic volumes can be tricky at locations with unmet traffic demand as shown along Redwood Road approaching 6200 S.

AUTOMATED TRAFFIC SIGNAL PERFORMANCE MEASURES (ATSPMs)

UDOT’s ATSPM website provides signal operations data, including signal phase, pedestrian activity, split failure, and vehicle progression data. The consultant can use ATSPMs to evaluate existing conditions and, in some instances, determine methods of improvement for current conditions without the need for a traffic simulation models. Additionally, ATSPMs can measure the realtime effects of signal/intersection-related improvements in lieu of modeling.

The following are expanded descriptions for the uses of ATSPMs.

Turning Movement Counts (TMC)

What it does: Reports the number of vehicles making a specific turning movement.

Where to use it: The consultant can usethe TMC data to gain an order-of-magnitude understanding of intersection traffic volumes, determining the most appropriate peak hour(s) for collecting turning movement counts, or developing seasonal adjustment factors. The consultant may use TMC data for traffic analysis purposes when no other option is available; however, the consultant should understand that the volume accuracy is suspect and decreases as volume increases. Unless approved by UDOT, the consultant should not use TMC data for important operational analyses. 

Approach Volumes

What it does: Reports number of vehicles approaching the intersection upstream of the signal as determined by the advanced detection. 

Where to use it: The consultant can use approach volumes for a better approximation of the upstream volume approaching an intersection. The approach volumes metrics can also assess the impacts of the signal on upstream roadway performance when used in combination with approach speeds. The consultant should understand that the volume accuracy is suspect and decreases as volume increases. Unless approved by UDOT, the consultant should not use the approach volumes for important detailed analyses.

Split Monitor

What it does: Reports the actual length of time for each signal phase.

Where to use it: The Consultant can use Split Monitor to establish the average split time of each signal phase. When compared to the programmed phase split, this information helps determine if a phase is receiving extra time from other phases or if it is not using all of its programmed time, thus allowing the consultant to reallocate time at an intersection.

The Split Monitor also reports how frequently each phase terminates by force off, max out, gap out, or if the phase is skipped. Phases that consistently force off or max out may need a longer split time to adequately serve the traffic demand. Alternately, phases that consistently gap out may allow for time in other phases that have more demand.

This metric can also be used to determine if there are possible detection issues by identifying phases that consistently use their full split, even in the middle of the night, or if the phase is skipped/ends early when vehicles are in the detection zone.

UDOT maintains the ATSPM site http://udottraffic.utah.gov/atspm. This site includes many additional links on SPM methods & data. A copy of the Purdue SPM methods can be found here, http://dx.doi.org/10.5703/1288284315333 For more assistance with ATSPMs, contact the UDOT Signal Desk at 801-887-3702.

Figure 15 - An example of the Split Monitor chart from the ATSPM website.

Pedestrian Delays

What it does: Reports pedestrian activations for each phase and the delay from the time of the activation until the call is served.

Where to use it: In addition to gathering pedestrian delay data, this ATSPM identifies locations with frequent pedestrian crossings or periods of the day with the highest pedestrian activity.

Figure 16 - An example of the Pedestrian Delay chart from the ATSPM website.

Purdue Split Failure

What it does: Reports the number of split failures or the number of times that a phase ends with vehicles still waiting to enter the intersection.

Where to use it: The Purdue Split Failure identifies phases that do not have adequate green time to serve the vehicle demand. Frequent split failures indicate that the movement is consistently over capacity. This information can validate traffic simulation models by comparing the actual split failures to the modeled results to ensure the model is replicating real world conditions. For empirical models, it can verify average vehicle delay times.

Figure 17 - An example of the Purdue Split Failure chart from the ATSPM website.

Purdue Coordination Diagram

What it does: Reports the arrival of vehicles relative to the start of the green, yellow, and red phases for through movements on each approach.

Where to use it: The Purdue Coordination Diagram determines how efficient the green time is by verifying how many vehicles are arriving on green (AoG). A high percentage of vehicles arriving on green indicates that the signals are coordinated well.

This ATSPM also isolates instances of offset changes to help improve coordination with adjacent signals because it displays when the vehicles arrive during each cycle. However, since the detector is generally only about 350 feet upstream of the traffic signal, if the queue extends past the detector, the detector does not record the information, which is only displayed once the queue has started to move.

Figure 18 - An example of the Purdue Coordination Diagram from the ATSPM website.

Approach Delay

What it does: Reports the calculated time between the vehicle’s arrivals (as determined by detector activations) and the start of the green phase.

Where to use it: The Approach Delay identifies the average wait time for vehicles arriving at the intersection on red for each approach. A longer wait time might exist because the traffic signals are not coordinated well. Longer wait times could also indicate vehicles are entering the corridor from driveways or unsignalized intersections.

Figure 19 - An example of the Approach Delay chart from the ATSPM website.

Arrivals on Red

What it does: Reports the arrival of vehicles relative to the start of the red phase (the inverse of the Purdue Coordination Diagram).

Where to use it: The Arrivals on Red metric establishes areas where a higher percentage of vehicles are arriving on red, which could indicate poor coordination issues or over-saturated conditions.

Figure 20 - An example of the Arrivals on Red chart from the ATSPM website.

Approach Speeds

What it does: Reports speeds of vehicle approaching theintersection upstream of the signal as determined by the advanced detection.

Where to use it: This provides a better approximation of link speeds than just using the posted speed limit. The Approach Speed metrics can also assess the impacts of the signal on upstream roadway performance. Since the speed measurements are taken during the green phase, a lower speed might insinuatethat vehicles are not getting up to speed as they pass through the intersection because of downstream congestion or queuing. 

Of note, one of the limitations to using ATSPMs is that not all data types are available for every signal. The consultant should only use volume counts gathered from the ATSPM website for trends such as peak periods and weekly distribution. It is UDOT’s goal to continue to enhance and improve the ATSPM system. As the review of available ATSPM is conducted, any opportunities to improve and add hardware are requested of the department.

QUEUES

One important calibration metric for modeling is queue lengths, which is the distance of queued vehicles measured from the stop bar of an intersection or from the beginning of a roadway bottleneck to the end of the last vehicle in the line. The consultant should collect queue lengths during a field observation or via video recording. Although queue lengths do not need to be physically measured, the consultant should compare queue lengths to roadside landmarks to estimate length.

TRAVEL TIMES

Travel times are the time that it takes for a vehicle to travel a corridor under prevailing conditions. Data sources for travel times include manual data collection through vehicles equipped with stopwatches or GPS units that record vehicle location at set intervals. A relatively new source of travel time data is the probe data available via the iPeMS website, which allows a consultant to create specific travel time routes and record travel times during a specified time period. When dealing with probe data, a consultant should remember this is a sample of vehicle data, so care should be taken to ensure that the recorded travel time period is sufficiently long enough to capture a large enough sample to represent existing conditions.

SIGNAL TIMING

The consultant should incorporate actual signal timing into their existing conditions traffic models, and signal timings should use the appropriate time period being modeled. Signal timings include the correct sequence, phases, cycle length, offset, overlaps, pedestrian crossing times, minimum green times, yellow and all-red times, and split times.

For UDOT projects, the consultant can acquire signal timing data from the UDOT Signal Desk for signals along state routes. The consultant should verify with the desk that they are using the latest signal timing summary templates. For traffic impact studies, the consultant can obtain signal timing data from the Region Traffic Engineers. Signal timing data for most non-UDOT signals will require direct coordination with city or county staff. The consultant can acquire other useful documents for signal timing, such as from the Controller Ring Detector and Guidelines for Traffic Signal Timing by contacting the UDOT Traffic Management Group.

TRANSIT

If transit routes will impact a roadway(s) within a study area, the consultant should include transit vehicles within the model. Transit data may include the route, transit stop locations, ridership data, dwell time, headway, and vehicle type. For UTA projects, the consultant can obtain basic transit data from their website, while detailed data may require direct coordination with UTA staff.

CRASH DATA

The consultant can use crash data to supplement the operations analysis with safety information. A general rule of thumb being three to five years of data will represent safety consideration on a roadway. The primary data source for crash data is the Numetric website. One limitation to this dataset is that it only has a database of reported crashes. Although crash data is not used directly in operations analysis, it can be used in the alternative selections for further analysis, based on crash reducing countermeasures.

SOCIOECONOMIC DATA

Socioeconomic (SE) data includes population, households, employment, household income, and school enrollment data. Socioeconomic data is available from the MPOs and UDOT for their respective TDMs. New socioeconomic projections are typically developed every four years following a process where the governor’s office, in coordination with the MPOs, produces county control totals for key SE categories, which the MPOs disaggregate to the TAZ level. These forecasts are usually published in 5 to 10 year increments that at a minimum correspond with RTP horizon years.

In summary, Table 3.1-2 lists some of the most commonly used data sources for an analysis. As technology changes, additional types and sources may become available, and the consultant should continue to verify data type and sources as part of the scoping process.

Table 1 Data References: Data Sources & Types of Data

3.2 Data Requirements

Based on the type of analysis performed, different types and data sources will be required. This section presents typical requirements for the various types of traffic analysis, including intersection, arterial operations, freeway operations, interchange, and regional planning. Table 2 lists common requirements for the types of traffic analysis.

Note that certain data types are required for multiple types of traffic analysis, and the method by which they are collected is not equivalent. For example, volumes are required for all types of traffic analyses. However, an intersection analysis typically requires turning movements, while an arterial operations analysis may only require approach volumes.

Regional planning data requirements generally consist of the data needed to run the UDOT/MPO TDM. This includes socioeconomic data by TAZ, highway network lanes and facility types, and transit network routes, stops, and headways. Depending on the scope of the analysis, the consultant may also need route or system level transit ridership information. Additionally, the consultant should perform a calibration of a TDM at the sub-regional or corridor level to better represent the study area. Calibration data needs to include daily or peak period traffic volumes and potentially average travel speeds and/or O-D information.

The following presents tool-specific guidance based on what UDOT expects from a consultant’s model development, calibration, and MOEs. This guideline’s ultimate goal is to facilitate a clean and quick review that minimizes consultant requests for additional information. While not a “how-to” guide for modeling, this section assumes the consultant is knowledgeable on using the analysis tools and will be responsible for the quality of the models produced. This section is not a substitute for solid engineering judgment and internal quality control.

The following also details 1) modeling techniques that account for UDOT’s preferences on specific inputs and parameters (e.g., seeding techniques, peak hour factors, trucks, and speeds), 2) recommendations on which parameters UDOT prefers for calibration, and 3) an overview of the MOEs from each analysis tool that will require specific techniques or calculation methodologies to ensure Department-wide consistency.

4.1 Synchro/SimTraffic

4.1.1 Modeling Techniques

MODELING INPUTS & PARAMETERS

A Synchro/SimTraffic analysis relies on three basic inputs: intersection geometry, signal timing(s), and traffic volumes.

The consultant should replicate current intersection geometry in the existing conditions model. The consultant can then modify the geometry to evaluate alternatives. The consultant should also input current geometry for any future no-build scenarios, unless there are currently planned projects that will affect the geometry and are not associated with the project being evaluated. The consultant may need to include profile grades and lane widths if observed to impact traffic performance. For each turning lane, the consultant should measure available vehicle storage, typically rounded to the nearest 10 feet. The links’ speed limits should reflect the posted speed limit, unless better data is available (e.g., a speed study that calculated the 85th percentile speed).

The consultant should update signal timing parameters using the most recent signal timing(s) obtained from UDOT’s Traffic Operations Center (TOC). These timing sheets include the settings for cycle length, yellow and red time, overlaps and sequences, and pedestrian values. These should apply to all future models, unless significant changes are being made to the lane count or speeds. For applications were significant changes are made or where a new intersection is added, the consultant should refer to the Guideline for Signal Timing Manual for preferred settings. The “D. P + P” turn type will mimic flashing yellow arrow leftturn phasing, so that the permitted left-turn phase is associated with the opposing through movement rather than the adjacent through movement. UDOT’s ATSPM website includes information that can help the consultant determine whether the signals are performing as described in the signal timing database.

Figure 22 - Synchro’s graphical interface provides a visual representation of the geometric input parameters, which can aid in model review.

Volumes should match the balanced traffic volume counts and should include the peak hour factor. Estimated future volumes should typically be rounded to the nearest 10 vehicles.

The consultant should adjust the percentage of heavy vehicles by referencing the best information available or as deemed appropriate for the roadway type. The consultant should also collect pedestrian and cyclist volumes for locations where they would have an impact on intersection performance (e.g., near a school, in the central business district, or along a major bike route).

Figure 23 - SimTraffic is a good tool to help in the calibration, validation, and identification of errors in a Synchro model.

CALIBRATION & VALIDATION

The consultant should calibrate a Synchro model so that the model reported queue lengths match with corresponding observed or field-measured queue lengths. Engineering judgment will also help determine if the delay reported from the model is reasonable based on observed traffic conditions. The consultant should also consider the following when calibrating and validating the model:

• Modify intersection geometry based on how vehicles are driving the intersection (e.g., add a right-turn lane for locations with wide shoulders).

• Confirm signal timing inputs match how the signals are operating based on a review of UDOT’s ATSPM website and consultant field observations of timings.

• Mimic lane utilization issues for movements with multiple lanes based on observed or measured traffic patterns

• Review the saturation flow rate generated by Synchro based on the methodology described in the Highway Capacity Manual (HCM).

The consultant should calibrate a SimTraffic model so that observed traffic conditions in the model closely match real-life conditions. This includes queue lengths, traffic patterns, and general delay times. The consultant should also consider the following when calibrating and validating the model:

• Adjust the input volumes to account for the peak hour factor (PHF). The consultant should either directly replicate observed traffic volumes in 15-minute increments or increase the volume in one 15-minute period and correspondingly reduce the volume in the other three periods.

• Modify intersection geometry based on how vehicles are driving the intersection (e.g., adding a right-turn lane for locations with wide shoulders).

• Confirm signal timing inputs match how the signals are operating based on a review of UDOT’s ATSPM website and consultant field observations of timings.

• Mimic lane utilization issues for movements with multiple lanes based on observed or measured traffic patterns.

• Set mandatory and positioning lane change distances. 

• Determine lane alignment through the intersections and when lanes are added or dropped.

• Set vehicles turning speeds when appropriate (e.g., for free right turns, left turns at a single-point urban interchanges [SPUIs] or sharp corners).

• Include a headway factor to simulate roadway capacity reductions because of features not captured in the model (e.g., blind corners, heavy weaving, or uneven pavement).

• Verify that in most cases the default vehicle and driver types are used in the analysis.

4.1.2 MOEs

INTERSECTION/APPROACH DELAY

Intersection delay is one of the main outputs from a Synchro or SimTraffic analysis. Synchro reports a calculated delay value, while SimTraffic measures the delay in simulated traffic.

Typically, the main MOE for a Synchro analysis is average vehicle delay and LOS. The average reported delay should be from the Synchro HCM report. When an HCM analysis is not available due to an intersection’s geometry or because of signal phasing, the consultant can use the Synchro percentile delay. The consultant should base the reported intersection LOS on the threshold associated with that intersection type as described in the HCM. 

SimTraffic measures delay based on simulation runs performed. The consultant should base delay calculations on a minimum of 10 runs. The consultant must use a seeding period, which should be longer than it would take any single vehicle to travel through the entire network. For a single intersection, five minutes may be appropriate, whereas larger models may require more than fifteen minutes. The consultant must turn off ‘Record Statistics’ for the seeding period. In all cases, the simulation time should include the peak hour being evaluated, but the consultant may extend the simulation time based on project needs.

For interchange analyses in both Synchro and SimTraffic, the consultant should calculate total interchange delay by aggregating the delay for each movement at each intersection that makes up the interchange. Each movement’s delay should be a sum of the delay added by each node that the vehicles performing that movement are required to pass through. The consultant would then calculate overall interchange delay by taking the average of the movement delays weighted by the number of vehicles making each movement.

Figure 24 - MOEs like this representation of intersection and turn movement delays help in understanding issues and in making decisions about improvements.

QUEUE LENGTHS

Synchro calculates a queue length for each movement at each intersection in the model. The consultant should confirm that the queue length reported for a Synchro analysis is the HCM 95th percentile queue where available. Not all intersection configurations are supported by the HCM. The consultant should document in the model submittals the situations where Synchro reports volume exceeding capacity, the 95th percentile volume exceeding capacity, or upstream metering is in effect as denoted by the footnote symbols (~, #, m). In these instances, the reported 95th percentile queue may not be adequate, and additional analysis may be required (e.g., modifying the Synchro files or performing a SimTraffic analysis or other queue calculation method). 

The consultant should base the 95th percentile queue lengths reported from SimTraffic on a minimum of 10 simulation runs. Unlike Synchro, which reports the 95th percentile queue by lane group, SimTraffic reports a separate queue length for each lane. For movements with multiple lanes, the consultant should report the longest of the measured queues.

Queue lengths are an important factor in roadway design, particularly under the latest UDOT standards that essentially measure ramp lengths from the back of the queue rather than the stop bar. As such, a recommended check to determine the reasonableness of the queue lengths calculated in either Synchro or SimTraffic is to compare them with the calculated queue lengths using the following formula, which is based on a Poisson distribution. Note: if factors that influence the queues are not found in the formula, then the results may not fully represent field conditions.

Where:

L = Vehicle length in feet, assume 25 ft

D = Peak hour traffic volume

C = Traffic signal cycle length in seconds

V = Lane Factor

for 1 lane, use 1.00

for 2 lanes, use 1.03

for 3 lanes, use 1.06

Z = 95th Percentile Confidence Level Z-score (1.645)

Ln = Number of lanes

Rt = Right turn factor (for right turn queues),

When right turns are allowed on red use 0.85

Otherwise use 1.00

4.1.3 Submittal Requirements

The consultant should include the following items, along with the model, when submitting for UDOT review. These items only pertain to model reviews and not to the overall traffic or study reports.

REQUIRED DOCUMENTS

The consultant should summarize the observed traffic issues included with the model. At a minimum, this should define general observations about the queue lengths, lane utilization, traffic congestion, and unique intersection features (e.g., nonstandard signal operations or the presence of land uses with major traffic impacts such as a school).

The consultant should also describe the calibration efforts undertaken, detailing any deviation from the modeling recommendations outlined in this guideline or from the Synchro default values.

All submittals should be easy to read and understand, noting both existing and future results.

SUMMARY OF RESULTS

The consultant should submit Synchro/SimTraffic reports that show the intersection/approach delay and 95th percentile queue lengths with each model submittal. Where available, the consultant should provide the existing queue storage capacity, as well as the needed storage lengths.

4.2 Vistro

4.2.1 Modeling Techniques

MODELING INPUTS & PARAMETERS

Vistro uses the HCM calculations to determine the traffic performance of intersections and roadways. The consultant should discuss with UDOT the best ways and times to use Vistro for your project. The primary inputs required for this analysis are intersection geometry, signals timings, and traffic volumes. 

The consultant should replicate current intersection geometry in the existing conditions model. The consultant can then modify the geometry to evaluate alternatives. The consultant should also input current geometry for any future no-build scenarios, unless there are currently planned projects that will affect the geometry and are not associated with the project being evaluated. The consultant may need to include profile grades and lane widths if observed to impact traffic performance. For each turning lane, the consultant should measure available vehicle storage, typically rounded to the nearest 10 feet. The links’ speed limits should reflect the posted speed limit, unless better data is available (e.g., a speed study that calculated the 85th percentile speed).

The consultant should update signal timing parameters using the most recent signal timing(s) obtained from UDOT’s TOC. These timing sheets include the settings for cycle length, yellow and red time, overlaps and sequences, and pedestrian values. These should apply to all future models, unless significant changes are being made to the lane count or speeds. For applications were significant changes are made or where a new intersection is added, the consultant should refer to the Guideline for Signal Timing Manual for preferred settings. UDOT’s ATSPM website includes information that can help the consultant determine whether the signals are performing as described in the signal timing database.

Volumes should match the balanced traffic volume counts and should include the peak hour factor. The consultant can round the estimated future volumes to the nearest 10 vehicles. 

The consultant should adjust the percentage of heavy vehicles by referencing the best information available or as deemed appropriate for the roadway type. The consultant should also collect pedestrian and cyclist volumes for locations where they would have an impact on intersection performance (e.g., near a school, in the central business district, or along a major bike route).

Figure 26 - Vistro results are based on HCM calculations as shown in this report.

CALIBRATION & VALIDATION

The consultant should calibrate a Vistro model so that the reported queue lengths correspond with observed or fieldmeasured queue lengths. Engineering judgment will also help determine if the delay reported is similar to the observed conditions. The consultant should consider the following when calibrating and validating the model:

• Modify intersection geometry based on how vehicles are driving the intersection (e.g., adding a right-turn lane for locations with wide shoulders).

• Confirm signal timing inputs match how the signals are operating based on a review of UDOT’s ATSPM website and consultant field observations of timings.

• Mimic lane utilization issues for movements with multiple lanes based on observed or measured traffic patterns.

• Review saturation flow rate based on the methodology described in the HCM.

4.2.2 MOE

INTERSECTION/APPROACH DELAY

Intersection delay is one of the main outputs from Vistro. The consultant should ensure that the reported delay is based on the most recent HCM and should only consider earlier versions of the HCM analysis when the current version is not available due to the geometry of the intersection or because of the signal phasing. The consultant should base LOS reported for each intersection on the threshold associated with that intersection type as described in the HCM.

For interchange analyses in both Synchro and SimTraffic, the consultant should calculate total interchange delay by aggregating the delay for each movement at each intersection that makes up the interchange. Each movement’s delay should be a sum of the delay added by each node that the vehicles performing that movement are required to pass through. The consultant would then calculate overall interchange delay by taking the average of the movement delays weighted by the number of vehicles making each movement.

4.2.3 Submittal Requirements

The consultant should include the following items, along with the model, when submitting for UDOT review. These items only pertain to model reviews and not to the overall traffic or study reports.

REQUIRED DOCUMENTS

The consultant should summarize the observed traffic issues included with the model. At a minimum, this should define general observations about the queue lengths, lane utilization, traffic congestion, and unique intersection features (e.g., nonstandard signal operations or the presence of land uses with major traffic impacts such as a school). 

The consultant should also describe the calibration efforts undertaken, detailing any deviation from the modeling recommendations outlined in this guideline or from the Vistro default values. 

All submittals should be easy to read and understand, noting both existing and future results.

SUMMARY OF RESULTS

The consultant should include Vistro reports that show the intersection/approach delay and 95th percentile queue lengths with each model submittal.

4.3 Vissim

4.3.1 Modeling Techniques

MODELING INPUTS & PARAMETERS

The consultant should base all Vissim models on the UDOT template file obtained from the UDOT website at www.udot.utah.gov/go/VISSIM. UDOT maintains Vissim templates for each version release. It is recommended that the consultant periodically check with UDOT to verify the latest template, as these models contain default values created for Utah’s driving conditions and vehicles. While the following subsections are not intended to replace engineering practice and judgment, adjustments during the modeling process may be necessary, and the consultant should discuss any changes with UDOT, fully documenting any modifications and reasons why the changes are needed.

Figure 27 - Vissim is a powerful tool that is extremely customizable to any situation as shown in this example how links can be coded with site specific parameters.

Links & Connectors

In most circumstances, the consultant should use separate links for all turning lanes at an intersection. Some exceptions include shared thru and right-turn lanes or closely spaced intersections where vehicles are still maneuvering into the correct lane near an intersection. The consultant should account for profile grades on all links and connectors where observations show an impact traffic performance.

Connectors and links should not overlap for extended lengths. This will lead the model to run vehicles on top of each other, which is not a realistic condition. It is also good practice to make transitions smooth and uniform. 

UDOT uses the connector number as the primary key in node evaluations to collect volumes and percent served. Although it is not required, UDOT prefers that the links be broken at the intersection, so that UDOT can pull vehicle counts and percent served using existing tools.

Input Volumes & Routing

For vehicle inputs, the consultant should include hourly flow rates in 15-minute intervals, except for the seeding period where a single time interval can be adjusted as appropriate. The consultant may need to vary the 15-minute input values according to the peak hour factor or based on the variation in the collected volume counts.

Static routes should typically extend through the length of the model. In general, it is not realistic to assume that vehicles make new decisions in short increments. If necessary, the consultant can use routing breaks only for locations when the signal spacing in conducive (e.g., signal spacing is close to a mile). The consultant can, at times, break routing at the freeway ramps, but both the freeway and arterial routing should be continuous. The consultant should consider separate routing decisions for heavy vehicles, if 1) their travel patterns do not match those of passenger vehicles or 2) if there are high occupancy vehicle (HOV) lanes in the model. The consultant should coordinate with UDOT Traffic Management Division if deviating from this routing methodology. 

The consultant should always input vehicle routing as hourly flows rather than as percentages, and the consultant should verify that pedestrians are included in the traffic counts and in the Vissim models.

Vehicle Speeds

The consultant should adjust vehicle speed based on the posted speed limit unless better data is available. When no speed data is available, the consultant should use the values in the template. 

Desired speed decision points should line up; otherwise, vehicles can change lanes between them and continue through the model without updating to the appropriate speed.

Reduced speed areas should generally be less than 30 feet in length, unless specifically observed otherwise. Of note, the Vissim algorithm to adjust speed for a vehicle passing through a reduced speed area already anticipates the slow down ahead of time, allowing for the vehicle to transition appropriately.

Figure 28 - Reduced Speed Areas are a simple but important tool for calibrating Vissim models.

Traffic Signals

For each signalized intersection included in the model, the consultant should obtain current signal timings from the TOC for the existing conditions models. The consultant should consult UDOT’s ATSPM website to ensure that the signals are performing as described in the signal timing database. Typically, UDOT prefers signals and detection to be simply coded with the phase they activate. However, the consultant is free to code innovative intersections according to the channels in the field. In these cases, the consultant must follow the red book associated with that signal. Red books can be obtained through the UDOT TOC.

The consultant should code loop detection according to UDOT’s old loop detection standard (included at the end of this guideline). VISSIM’s logic and the setup in UDOT’s base files mimic that of loop technology rather than radar. Radar detection should follow the field setup with additional passage time to account for latency. Since UDOT uses radar for advanced detection, the movements with this technology in the field should have 2.0 seconds added to the extension time in the Ring Barrier Cycle (RBC) to account for latency.

The consultant should include pedestrian phases in RBC files, even if the pedestrian volumes are not included in the project’s scope. Pedestrian crossing times often restrict a signals flexibility. This restraint cannot be artificially removed even if pedestrians are relatively few.

Public Transportation

All public transportation lines included in the existing conditions model should match the current UTA schedule. The consultant should also match how the line operates at signals and stops along the corridor.

Figure 29 - Priority Rules are another tool in Vissim to help calibrate the traffic to behave as life-like conditions.

Priority Rules & Conflict Areas

The consultant should apply priority rules and conflict areas only where appropriate. Mixing priority rules and conflict areas can also cause unexpected behaviors so it is typically best to use just one. Additionally:

• Protected left turns typically do not need priority rules or conflict areas unless used to prevent blocking of an intersection.

• Excessive conflict areas can create unintended behaviors in the model.

• Protected-permissive left turns should be coded with a priority rule that  allows the flexibility to adjust observance based on which signal phases are active and which are not.

• Conflict areas are generally recommended for all other movements unless a priority rule is needed. Conflict areas and priority rules are to mimic field conditions, and the consultant should verify that they perform accordingly.

• UDOT requires roundabouts be coded according to the examples used in the Vissim user manual for priority rules and conflict areas to ensure a realistic representation of roundabout operations.

• If vehicles are queuing in a “keep clear” zones (e.g., in a signalized intersection), the consultant must apply a priority rule to enforce the keep clear zone.

Other Inputs

While Vissim has several other input variables that can be modified based on project needs, the consultant should generally use the default values from the UDOT template. The consultant should discuss changes with UDOT and fully document the modifications and reasons why the changes are needed.

CALIBRATION & VALIDATION

Number of Runs

A minimum of 10 runs is recommended to determine the MOEs for all Vissim analysis. If outliers are excluded from these runs, then the consultant should use additional runs. More than 10 runs may be required to reach the calibration target, and the consultant could determine this using the following GEH equation:

Where:

M = Total Traffic Volume Served

  C = Total Traffic Volume Demand

                         If the value is >5, more model runs are required

       GEH Statistic Equation source: Geoffrey E. Havers

Queue Lengths

The consultant should evaluate queue lengths to determine if traffic congestion in the model accurately represents real-life conditions. Models should reflect observed and/or measured queues for the existing conditions models.

Queue lengths are an important factor in roadway design, particularly under the latest UDOT standards that essentially measure ramp lengths from the back of the queue rather than the stop bar. As such, a recommended check to determine the reasonableness of queue lengths calculated in Vissim is to compare them with the calculated queue lengths from the following formula, which is based on a Poisson distribution.

Where:

L = Vehicle length in feet, assume 25 ft

D = Peak hour traffic volume

C = Traffic signal cycle length in seconds

V = Lane Factor

for 1 lane, use 1.00

for 2 lanes, use 1.03

for 3 lanes, use 1.06

Z = 95th Percentile Confidence Level Z-score (1.645)

Ln = Number of lanes

Rt = Right turn factor (for right turn queues),

When right turns are allowed on red use 0.85

Otherwise use 1.00

Figure 30 - Vissim has a myriad of parameters that can be set to collect specific data during the model runs

Travel Times

The consultant may need to use travel times for calibration or MOE purposes. The travel time segments in Vissim should match the location of the segments used to determine existing travel times. In general, when travel times are used to calibrate a Vissim model, the consultant should confirm that the Vissim travel times along the length of the model are within 10% of the measured travel times.

Volume Served

The consultant should calibrate the existing models to allow each modeled intersection or freeway segment to serve the same number of vehicles counted as served in the field. Since these vehicles were recorded passing through the intersection and/or freeway segment, the model should accommodate the same volume over the analysis period.

While some variation is expected because of the nature of the Vissim software, the consultant should verify that each intersection is serving at least 95% of the traffic demand. This does not include unmet demand (such as queues that arrived during the peak hour, but were not served in the analysis period).

Driving Behaviors

The consultant should apply the default driving behaviors from the UDOT template as a starting point for the existing Vissim model. However, the consultant may need to modify this along sections of the model to mimic field conditions. 

The waiting time before diffusion should typically be set to 60 seconds. More than this and the model can overreact to anomalies, and less time can mask issues in the model.

Car Following Model Factors

The consultant should use the default car following model factors from the UDOT template as a starting point for the existing Vissim model. However, the consultant may need to modify these sections of the model to change the capacity. 

For the Wiedemann 74 car following model, which is used for surface streets, the consultant should only adjust the additive part of safety distance and the multiplicative part of the safety distance factors. Table 3 shows the suggested range of values for these two factors.

Table 3 Recommended Wiedemann 74 Values

For the Wiedemann 99 car following model, which is used for freeways, the consultant should adjust the CC0, CC1, and CC2 values, as needed, to best represent real-world conditions. Table 4 show recommended values based on the freeways section type.

Table 4 Recommended Wiedemann 99 Values

Situations may arise where additional factors may need to be adjusted. The consultant should discuss the changes with UDOT and fully document the modifications and reasons why the changes are needed.

4.3.2 MOEs

INTERSECTION DELAYS (LOS)

The consultant should collect intersection delay at each intersection in the Vissim model with a node. In general, the delay collected at each node should extend to the upstream node. The default setting in Vissim is too short to capture the queues that the state experiences. Typically, the consultant would set the start of the delay segment to at least 800 feet before the node (which can be adjusted as necessary) to ensure the queue is fully captured. The reported average delay per vehicle should include each movement at an intersection. For innovative intersections, if the vehicle is required to stop at multiple signals (e.g., for a CFI) or travel out of direction for that movement (e.g., for a ThrU turn), the consultant must aggregate the delay for each movement to include the delay at each signal and out-of-direction travel time.

The consultant should base the reported LOS for each intersection on the threshold associated with that intersection type as determined in the HCM.

The consultant can calculate interchange delay by aggregating the delay for each movement at the intersections that make up the interchange. This should be done at the movement level. Each interchange movement’s delay should be a sum of the specific movements through each intersection that a vehicle performing that interchange movement is required to pass through. The consultant would then calculate overall interchange delay by taking the average of the movement delays weighted by the number of vehicles making each movement.

Figure 31 - For innovative intersections, aggregate the delay for each movement to include the delay at each signal.

4.3.2 MOEs

INTERSECTION DELAYS (LOS)

The consultant should collect intersection delay at each intersection in the Vissim model with a node. In general, the delay collected at each node should extend to the upstream node. The default setting in Vissim is too short to capture the queues that the state experiences. Typically, the consultant would set the start of the delay segment to at least 800 feet before the node (which can be adjusted as necessary) to ensure the queue is fully captured. The reported average delay per vehicle should include each movement at an intersection. For innovative intersections, if the vehicle is required to stop at multiple signals (e.g., for a CFI) or travel out of direction for that movement (e.g., for a ThrU turn), the consultant must aggregate the delay for each movement to include the delay at each signal and out-of-direction travel time.

The consultant should base the reported LOS for each intersection on the threshold associated with that intersection type as determined in the HCM.

The consultant can calculate interchange delay by aggregating the delay for each movement at the intersections that make up the interchange. This should be done at the movement level. Each interchange movement’s delay should be a sum of the specific movements through each intersection that a vehicle performing that interchange movement is required to pass through. The consultant would then calculate overall interchange delay by taking the average of the movement delays weighted by the number of vehicles making each movement.

QUEUE LENGTH CALCULATIONS

There are several methodologies in practice to collect the 95th % queue and each produce different results. In light of this ambiguity, the Department has adopted the following methodology. The consultant would calculate the 95th queue length for each movement at all intersections in the study area using the “Percentile.Inc” function found in Excel. The formula should include all “QLenMax” values for all simulation runs collected in 90-second intervals. 

A recommended check of the 95th percentile queue length can be calculated for each intersection movement in the study area, using the following equation:

Where:

QAve = Average of the maximum queue lengths for all simulations runs collected in 90 second intervals

             S = Standard deviation of the maximum queue lengths for all simulations runs collected in 90 second intervals

The built-in percentile functions are not recommended for use in the report 95th percentile queues, which is based on limited data Vissim uses to create the aggregate percentile.

TRAVEL TIMES

Travel times can measure the impacts of an alternative on a particular route, which can extend through multiple intersections. While the distance traveled may vary according to the alternative, the consultant should verify that the start and end point for all travel times used to compare alternatives is consistent between alternatives to allow for reliable comparisons.

FREEWAY DENSITY

For a freeway analyses, the consultant should collect vehicle densities for each basic, weave, and merge/diverge freeway segment. The determination of the segment classification and which lanes should be included in the density calculation is provided by the HCM. The consultant should also report LOS based on the thresholds provided in the HCM.

4.3.3 Submittal Requirements

The following items need to be included, along with the model, when submitted for review by UDOT personnel. The items only pertain to model reviews and not to overall traffic or study reports.

REQUIRED DOCUMENTS

A summary of the calibration process is required with the submission of an existing conditions model. This summary should include the calibrations methods, calibration results, and a description of any changes made to the default UDOT template values.

Vissim results that are submitted directly to Region staff are required to include a recorded video clip of the existing conditions model that highlights the critical traffic operations issues. This allows Region staff to review the effectiveness of the calibration without having the Vissim software or expertise. The scope of this video clip does not need to be complicated. The level of effort for this is expected to be 1 day or less.

Figure 31 - The Travel Demand Model contains a TAZ structure that can be modified to meet specific needs of project, but should be discussed with the owner of the model prior to completion.

SUMMARY OF RESULTS

A summary of the model results is required. This summary must include 1) the demand and the percent of vehicles served at each intersection or along each freeway link, 2) the intersection delay and 95th queue length for each movement at every intersection in the model, and 3) the density for each freeway segment being analyzed. Travel time summaries may also be needed if they are being used as MOEs. Where available, the consultant should provide the existing queue storage capacity as well as the needed storage lengths.

4.4 Travel Demand Model

4.4.1 General Inputs

Regardless of the TDM being used there are three main inputs (listed below) that need to be reviewed. Typically, any analysis will be focused on a sub-area of the larger model. The consultant should complete a review of the network, TAZ, and socioeconomic data within the influence area of the analysis. Below is a list of review points that might be considered. The list s not exhaustive and should not be a substitute for meeting and discussing the analysis with the owner of the model being used.

• TAZ Structure – The consultant should resize the zones to allow appropriate distribution onto a network within the study area. Boundaries should consider physical or practical barriers that might exist, the transportation infrastructure, and the uses of the land in the TAZ. If the zones are split, the consultant should update data fields, such as acres and developable acres, as well as other TAZ-based input files.

• Network – Links and nodes should allow for demand to represent existing and future travel patterns. Links should have correct coding of lanes, functional type, distance, speeds, and travel times. In the case of transit modeling, nodes should have correct coding of park-and-ride lots.

• Socioeconomic Data – The TAZ ID should have correct coding of the number of households, population employment by type of employment, and median TAZ income. The consultant should review school enrollment fields for reasonableness. If the zones are split, correct proportioning of socioeconomic data into will need to be done.

WASATCH FRONT (WFRC/MAG) REGIONAL MODEL

The regional model is a model of the Wasatch Front metropolitan area and includes most of Davis, Salt Lake, and Utah counties. More recent versions of the model extends north to Brigham City. The TAZ, network, and socioeconomic data s designed for a regional analysis and may need to be refined for more focused study areas.

UTAH STATEWIDE TRAVEL MODEL (USTM)

The USTM includes the entire state of Utah. It focuses on the nonurban area of Utah and is used to help provide demand volumes at the external points of the regional models.

LOCALIZED MODELS (CACHE, DIXIE, & OTHERS)

These models are typically smaller and more focused to a specific area. Review of the study area is still recommended to verify that it can be properly represented with the model.

4.4.3 Calibration & Validation

The TDMs are regional and state-level models that are calibrated at a model level. The models may need localized calibration to account for certain project extents. Some typical calibrations tools are turn penalties, speed, and capacity factors. For best results, the consultant should work with the owner of the model for validation methods and practices.

4.4.4 MOEs

The TDM models available are setup to produce several default MOEs that can evaluate alternatives. Below is a partial list of MOEs, which the consultant can aggregate at different geographic levels (screenlines, segments, study area, etc.) to better understand changes in traffic conditions.

• Roadway traffic volumes – specific comparisons between alternatives as well as with the base model.

• V/C ratios – used to compare alternatives, with a focus on specific roadways or an overall study area.

• Vehicle-hours of delay – good data to compare the alternatives at an entire study area level, regardless of the level of changes made.

• Vehicle Miles Traveled and vehicle-hours traveled – used similarly to vehicle-hours of delay.

• Average speed – useful in identifying specific locations where issues might be along a corridor or within a study area by comparing it to the base speed.

• Person trips by mode – an important MOE when modeling includes a transit component.

4.4.5 Submittal Requirements

The consultant should include the following items, along with the model, when submitting for UDOT review. These items only pertain to model reviews and not to the overall traffic or study reports.

REQUIRED DOCUMENTS

When using a TDM for analysis, the consultant should submit the following items in figures and/or documents:

• TAZ splits

• Socioeconomic data changes

• Network modifications, including number of lanes, functional type, speed and capacity factors, and link connectivity

• Transit changes

• MOE results

Consistency in using the calibrated models and comparing MOEs is vital during the alternatives analysis process. The following discusses opportunities for promoting consistency and when inputs and parameters should change based on an alternative’s influence on driver behavior and travel patterns.

5.1 Consistency

To the extent possible, the consultant should verify that the only differences between models is the alternative being analyzed. Changing model settings between alternatives makes it difficult to understand the effect of unique alternative features on the study area operations. Generally, calibration factors, such as roadway capacity or a roadway’s speed profile, are significant factors that a consultant should not modify during the alternative development. Typical alternatives consist of changes to the geometry and will often include items like a new interchange, additional lanes, revised approach geometry and storage lengths, or modifications to traffic control devices, such as the addition of a signal or roundabout.

Changes to Inputs Based on Alternatives

MODEL CALIBRATION

Some alternative concepts warrant changes to the calibration of a model. This is most likely when an alternative is changing the facility type of a roadway, such as transitioning an arterial corridor to a freeway or modifying a SPUI interchange to a system-to-system interchange. In these cases, the parameters used to calibrate existing conditions are no longer valid, and the consultant will need to use new parameters for the new facility type.

VOLUMES

When the alternative evaluated results in a change in travel patterns, the consultant should create new volume estimates for that alternative. An alternative concept that changes the fundamental functionality of the roadway model inputs may require developing future traffic volumes specific to that alternative. Make sure to coordinate such assumptions with the Traffic Operations & Analysis Reporting Group.

SIGNAL TIMINGS

The consultant may need to update signal timings for alternatives to reflect changing intersection geometry or volumes.

UDOT is committed to support its consultants by providing clear and useful feedback during the modeling process. While a review is not always required, this section describes 1) how and when to submit a model, 2) the review time frames based on the model type and complexity, and 3) what feedback consultants can expect from UDOT when a review is needed. Of note, UDOT’s review is not a substitute for the consultant’s quality control process. During any review, communication is crucial for thorough and timely completion.

6.1 Submittal Content

Based on the type of analysis and tools being used, the following lists required as parenthetically noted and recommended items for submitting a model for UDOT review. The list is not exhaustive, and UDOT may request additional items based on the scope and type of model.

• Fully functioning model equipped with all MOE collection methods, including:

* Calibrated existing conditions models

* No-build models

* Future alternatives

• Models for which the consultant would like UDOT’s input

• A copy of field observation notes (Required)

• A summary of all calibration methods and settings used, with an explanation of calibration (Required)

• A summary of all deviations from recommended methods listed in this guideline, with reasoning for the deviations (Required)

• A digital copy of MOE outputs of the models

6.2 Review Periods

UDOT’s goal is to complete its review within two weeks of receiving a model. The consultant should communicate with the Traffic Operations & Analysis Reporting Group early in the process to prevent delays. The following lists circumstances that may lead to a longer UDOT review.

• Size and scope of the model – Larger models that take longer to run may have features or intricacies that require additional time

• Project importance and significance – Some projects involving FHWA review may require additional time

• Uniqueness of the alternatives – Some alternatives are more straightforward than other. Complex or unique alternatives may require additional time

• UDOT workload – There will be times when current workload makes it difficult to complete reviews in two weeks

6.3 Feedback

The most important part of any review will be UDOT’s feedback for the consultant. This feedback is not directly intended to cover quality and accuracy of the modeling, but instead identifies any concerns uncovered during the review process. UDOT feedback will primarily focus on: 

• Agreement and concurrence to assumptions

• Methods and values used in calibration

• Modeling techniques per this guideline, including reasoning for variance

• The selection and calculation of MOEs

• Operation of designs within the models

• Results summary (as needed) 

UDOT will typically provide feedback in Microsoft Word format, grouped by model review. The consultant should not expect that UDOT will list common issues inherent to multiple models in its feedback for each model submitted. It is good practice to check all models for any issue listed in UDOT’s feedback.