Traffic Analysis Tools Volume IX: Work Zone Modeling and Simulation
A Guide for Analysts

Glacier National ParkGoing to the Sun Road Rehabilitation Project

Glacier National Park
Going to the Sun Road Rehabilitation Project^[6]

Work Zone Characteristics
Transportation Analysis:
Approach	Sketch-Planning
Modeling Tools	QuickZone
Work Zones:
Type	Type III
Network Configuration	Pipe
Geographic Scale:
Work Zone Size	Large
Analysis Area	Local

Overview

The Going-to-the-Sun Road (GTSR) in Glacier National Park (GNP), Montana, is a prime attraction for visitors to the park and the only east-west link within the park.  The scenic 50-mile roadway, completed in the 1930s, traverses the park and provides access to the Logan Pass Visitors Center from either the St. Mary’s Entrance (East) or the West Entrance.  While portions of the GTSR are open throughout the year, in the higher alpine sections, the roadway is closed and often snow-covered throughout the winter months.  The roadway offers the visitor a number of incredible vistas as well as access to trailheads and other facilities along its length.  Given the rugged terrain, the GTSR is the only roadway within the park that connects the eastern and western sections of the park.  Alternative east-west connections outside the park do exist, however, the GTSR itself or one of the facilities along its length are the destinations of the vast majority of vehicles visiting the park.

The GTSR is slated for an extensive multi-year rehabilitation project over a seven to eight year period.  Since the GTSR is a key reason for visiting GNP and has no alternative route, the roadway must remain open throughout the project to both allow access to facilities and to remain open as a scenic roadway with minimal disruption and delay for park visitors.  The steep terrain, the complexity and duration of the work to be performed, and the limited construction season for roadwork in the summer season (coinciding with peak visitor travel demand) are factors that need to be considered in the planning stage to determine if the GTSR Rehabilitation Project can be completed in a timely manner while still maintaining an acceptable level of delay to visitors using the GTSR.

Because of the key role the GTSR plays within the park as a destination itself and in providing east-west access in the park, there was significant concern that extended and onerous delays from work zones would impact park visitation as well as the segments of the local economy that depended on tourism. An agreement was reached between the National Park Service, Western Federal Lands Highway Division (WFLHD) and representatives of the local community outlining the extent to which closures and delays on the GTSR would be tolerated.  One key tenet of this agreement was to limit end-to-end delays on the GTSR in one direction to no more than 30 minutes (total).

The role of FLH-QuickZone in the GTSR Rehabilitation Project was to assess likely travel delays (and in particular, the end-to-end delays from multiple work zones) expected over the course of the multi-year project. Since the project was, at the time, in the relatively early planning phase, the details of roadwork phasing and staging could not be identified in detail.  However, given the outline of a likely phasing plan and expected traffic control developed by WFLHD, FLH-QuickZone was used to identify projected delays and queue length over the course of the project.  FLH-QuickZone was also employed to assess the likely effects of actuated signal control for 2-way, 1-lane operations, as well as the impact of reduced travel demand. Reduced travel demand is projected during roadwork, as well as a result of the provision of a transit alternative.

Work Zone Alternatives

Eight alternatives were coded and analyzed using FLH-QuickZone for the GTSR Rehabilitation Project (see Figure 17).  The eight alternatives (Table 7) are combinations of four expectations for travel demand and the utilization of actuated signal control for nighttime and weekend 2-way, 1-lane operations.  The reduction in demand is from two potential sources.  The first reflects an estimate in the EIS that the presence of major roadwork will cause overall GTSR travel demand to decline by approximately 6 percent.^[7]  The second is a planning level target by planners to reduce travel demand by 2 to 7 percent by shifting visitors into transit buses.  The 6 percent demand reduction case assumes just the overall GTSR decline from the presence of roadwork.  The -8 percent and -13 percent demand reduction cases reflect a combination of general decline with the high and low estimates of transit demand shift.

D

Figure 17 GTSR Rehabilitation Project

The use of a single fixed timing plan for the control of short work zones has been a typical practice on GTSR, as is the use of varying fixed timing plans over the course of the day for longer work zones. In the Glacier case study, WFLHD and NPS staff wanted to investigate the complete range of delay from the simplest (single fixed plan) to the most complex (all actuated) and to estimate how likely the 30 minute maximum user delay threshold might be exceeded over the life of the rehabilitation project. Base signal timing in this case study represents a single fixed plan used in the night and weekend periods when flaggers are not present.

Table 7  GTSR Alternatives Evaluated with FLH-QuickZone

			Work Zone Traffic Control
	Alternative Name	Travel Demand	Weekday Day 7am-7pm	Weekend (all hours) and Weekday Night 7pm-7am
1	Base	2004 Level	Flaggers	Fixed Time Signals
2	Actuated Signal	2004 Level	Flaggers	Actuated Signals
3	-6% Demand	-6%	Flaggers	Fixed Time Signals
4	-6% Dem + Actuated	-6%	Flaggers	Actuated Signals
5	-8% Demand	-8%	Flaggers	Fixed Time Signals
6	-8% Dem + Actuated	-8%	Flaggers	Actuated Signals
7	-13% Demand	-13%	Flaggers	Fixed Time Signals
8	-13% Dem + Actuated	-13%	Flaggers	Actuated Signals

Application

For the FLH-QuickZone analysis of the GTSR project, there were two critical measures of effectiveness to consider, both related to the designation of 30 minutes or more of delay in either direction as ‘unacceptable” throughout the project.  In order to describe the worst delay seen in a phase, maximum user delay was used..  This measure reflects the longest possible delay on the GTSR from all work zones encountered in one direction.  During 2-way, 1-lane operations, the assumption implies that some unlucky traveler will arrive at a work zone to experience the longest wait possible in that time period.  The second measure of effectiveness is the number of hours per week that the maximum delay exceeds the 30 minute threshold in one or more directions.  This measure provides insight into how long unacceptable delays are in effect throughout the week.

Base Case--In general, the FLH-QuickZone results indicate that delays in the base roadwork case are frequently in the unacceptable range, and remain in that range for a significant portion of the week for several phases.  In the base case, the delays are particularly noteworthy during night operations and in specific high-volume (July) time frames.  During high-volume time frames, flaggers are not expected to be able to allow all cars in the queue to pass through before a maximum time of 30 minutes is reached.  The result is oversaturation and queue development.  This development is relatively short-lived, however—no delay exceeds 44.9 minutes.

During night operations, delay is high because a single fixed plan must be used to cover high demand weekend patterns as well as low-demand night patterns.  The result is that relatively long cycle lengths are imposed to prevent saturation on weekends, but cause long waits at red lights at night (up to roughly 12 minutes).  It is also clear that 12 minute wait times under low flow conditions may be unsafe since the road user may presume that the signals are broken after a few minutes, proceed and eventually encounter oncoming traffic in the single open lane of the work zone. Before such unrealistic situations were to actually occur, it is likely that timing plans that change periodically throughout the evenings and weekends would be put into place.  This option was not analyzed in QuickZone-FLH as a part of this study at the direction of WFLHD and NPS staff because of a lack of time to explore every possible variant within the limitations of the case study.  Given the preliminary nature of the case study, WFLHD and NPS staff wanted to concentrate effort on identifying the likely worst case and best case delay outcomes. 

Signal Actuation—The impact of signal actuation is significant in the reduction of delays during night operations.  By allowing the signal timing to vary with actual travel demand, more appropriate (short) cycle lengths can be provided at night, and longer ones during weekend and weekday evenings (after flaggers are finished for the day).  Results indicated that the use of actuated signals eliminates unacceptable delays at night, although delays up to 25.3 minutes are predicted.  Signal actuation alone is not enough to eliminate unacceptable delays during the July and Late Summer peak, however, particularly in 2011 and later.  Here, travel demand exceeds the capacity of the expected work zones, regardless of signal timing.

Reduced Travel Demand—The reduction in demand is effective in reducing daytime delay during flagger operations but has no effect on night operations when fixed timing plans are in place.  Results indicated that this effect is highest with the -13 percent demand case and somewhat lower in the other two cases.  Overall, the demand reduction is only a critical factor when the highest seasonal travel demand is expected:  July.  In other months, the delay from the flagger operations is often at acceptable levels. Actuated signals, in combination with reduced demand (-13 percent), eliminates unacceptable delays in all phases except for July 2011, when five concurrent work zones are in place.

Table 8 provides a summary of both the fixed signal timing plan impacts and the actuated signal impacts respectively. Again, using fixed signal timing plans results in unacceptable delay in more than half (21) of the nighttime FLH-QuickZone phases. Switching to actuated signals eliminates all delays of greater than 30 minutes. Switching to actuated signals during the daytime FLH-QuickZone phases helps to reduce the occurrences of unacceptable delays but does not eliminate them as was done with the nighttime operations. Nor does the use of actuated signals reduce the severity of the delay. The minimum and maximum duration of the delay during daytime operations remains around 35 minutes with actuated signals. Other mitigation measures beyond those studied here, or a reduction in work zone activity in the peak of the summer visitation season, will likely need to be considered if all potential instances of user delay exceeding 30 minutes are to be eliminated.

Table 8  Fixed and Actuated Signal Timing Impacts

Signal Operation		Number of QZ Phases		Hours/Weeks >30 min Delay (hours)				Delay (minutes)
				Low	High	Low	High	Min	Max	Min	Max
		Fixed	Act.	Fixed		Act.		Fixed		Act.
Base Alternative	Day	8	5	3	40	7	35	30.2	44.9	31.1	36.3
	Night	21	0	48	48	0	0	30.3	37	0	0
-13% Demand Alternative	Day	1	1	-	30	-	30	-	32.4	-	32.4
	Night	21	0	48	48	0	0	30.3	37	0	0
-6% Demand Alternative	Day	4	3	8	30	20	30	32.3	36.1	30.5	34.3
	Night	21	0	48	48	0	0	30.3	37	0	0
-8% Demand Alternative	Day	4	3	5	30	10	30	30.6	33.7	30.2	33.7
	Night	21	0	48	48	0	0	30.3	37	0	0

Based on this planning-phase analysis of the GTSR, it is clear that Alternative #1 will result in unacceptable delays on a regular basis during both nighttime and day work zone operations.  The predicted value of actuated signals is large given the high variability of travel demand expected during the course of the project.  If actuated signals cannot be reliably implemented, it is possible that a fixed plan that varies by time of day could also be implemented to mitigate delays, although such a plan has not been evaluated in FLH-QuickZone.  The case for demand reduction is less clear – the problem lies primarily with reducing (or potentially redistributing) travel demand in the peak month of July, rather than in all time frames.

---

[6] This case study was adapted from the report FLH-QuickZone Case Studies: The Application of FLH-QuickZone in Six Federal Lands Projects available from Federal Lands Highway Division.

[7] The EIS also reported that without the presence of major roadwork overall park visitation was expected to increase by 0.6 percent. The estimated decrease in visitation of 6.4 percent due to major roadwork is detailed in the EIS for Alternative 3 (Preferred Alternative).

Table of Contents | Previous | Next