Chapter 1. Introduction

Drivers are more cautious during heavy rain and snow.  These adverse road conditions cause vehicles to travel and accelerate more slowly.  Normal signal coordination plans become unsuitable during adverse weather.  This is because the traffic flow parameters used to develop the dry weather plans change.  Ideally, a traffic signal coordination system would adapt to changing traffic flows and travel conditions as they occur. A compromise between the ideal control system and one that does nothing to accommodate signal coordination in adverse weather is to develop an inclement weather plan.

In the past, the implementation of such a specialized plan was impractical because it required an operator to visit each traffic signal and manually change the signal timing plan at the signal controller.  The new Utah Department of Transportation (UDOT) Advanced Traffic Management System (ATMS) soon will have real-time communication with many of the traffic signal controllers throughout the Salt Lake Valley.  This communication link makes the possibility of using a specialized signal coordination plan simple. 

An inclement weather signal coordination plan must be created using the traffic flow parameters specific to inclement weather.  This research seeks to quantify these traffic flow parameters. The modified parameters can then be compared to dry condition data to determine the percent change.  Generalized results can then be used to modify dry condition data collected for other arterials.  Thus, inclement weather plans can easily be created for other arterial streets in the area.  The modified signal timing plans are expected to generally have different offset values than the dry weather plans.

To identify the change in traffic flow parameters, data was collected over a range of weather conditions. Specifically, saturation flow, free flow speeds, and start-up lost times were collected at two intersections during dry weather and various intensity levels of rain and snow.  This data is then compared to identify the percent change from the dry weather condition.  Comparisons of the data collected with other similar research done in Alaska and Minnesota provides validation for the Utah findings.

Chapter 2. Literature Review

Limited research exits that focuses directly on inclement weather signal timing.  Such papers address how inclement weather affects: saturation flows, capacities, pedestrian walking speeds, design of freeway interchanges, and others.  Presented here is the most relevant research.

Bernardin, Lochmueller and Associates, 1995, assess the changes in speeds and saturation flows during extreme winter weather on a 24-signal network.  This study measured several traffic flow parameters in summer, winter, and severe winter conditions.  During each of these conditions, they measured saturation flow, vehicle speeds, lost time, and capacity.  The study asserts that summer signal timing parameters are inappropriate in winter and extreme conditions because of slower vehicle speeds and unreachable detectors which often are covered by packed snow.  The study defines summer as the time when temperatures are above 14º F, on dry roads, or above 32º F on wet roads (without ice).  Winter is defined as the time of year when temperatures range from -22º F to 14º F with dry pavement or with well-sanded hard-packed snow on the road. "Extreme" is when the air temperature is below -22º F or during snowfall, blizzard, and/or freezing rain, resulting in slippery roads and reduced visibility.  The existing timing plans were assumed to be appropriate for dry/summer conditions. Therefore, winter and extreme conditions were the main areas of focus for modifying traffic flow parameters. The traffic signal optimization packages SIGNAL 85, and TRANSYT-7F were used to develop an optimized signal timing plan.  Signal 85 was used to run chosen cycle lengths to generate final phase sequences and splits.  TRANSYT-7F was used to generate offsets that yield better arterial progression in the network.  The relevant results of the study are provided in Table 2.1

An important conclusion reached in this study was that all-red and amber times should not be changed during winter conditions.  It states that although much research has been done in this area, there still is much disagreement on how the parameters should be timed.  Anchorage traffic signals generally have 4-5 second yellow intervals and 1-3 second all-red intervals.  It said that if the parameters were to change, it would be based on reduced speeds, increasing the all-red time and reducing the amber time.  The study's final recommendation was to calculate these parameters based on current standards to protect the municipality from liability.

Table 2.1 MOE Improvements from the Anchorage Winter Signal Timing Modification

(Maki 1999) describes a study for the Minnesota Department of Transportation (MnDOT) to evaluate the feasibility of implementing a traffic signal timing plan for inclement weather.  The following data was collected from 3-8 p.m. on several weekdays: current signal timing, intersection geometry, turning movement counts, travel time, volume and occupancy (system detectors) start-up delay, and saturation flow rates.  They also collected weather-related data from road weather information system (RWIS) devices, including air temperature, pavement temperature, relative humidity, and roadway condition (i.e. icy, plowed), and dew point.  All of this was done on Trunk Hwy 36 in Minneapolis, Minn. The study defines inclement weather as a storm with accumulation of three inches of snow or more.  Data also was collected on fair-weather days for comparison purposes.  A street network, (Hwy 36 in Minneapolis), was simulated in existing normal conditions to establish a basis of comparison.  The SYNCHRO III traffic signal optimization software was used to create optimized signal timings for inclement conditions.  In the simulation, adverse conditions were created by modifying the saturation flow rates, average speeds, and lost times of the traffic. Comparison data was then gathered from the software output under the signal timings in use and the signal timings optimized for inclement weather.  The simulation yielded a six percent overall improvement of MOE's with an optimized signal-timing plan on a 125-second cycle.

Table 2.2 Measures of Effectiveness for Minnesota Study

An interesting ancillary conclusion of the study was that volumes during inclement weather was 15-20 percent lower than volumes collected during the same time period (3-8 p.m.) on a normal day and 15-30 percent during peak hour (5-6 p.m.).  The speeds were about 40 percent lower during inclement weather, falling from 44 mph to 26 mph.  The start-up delay increased from two to three seconds.  The saturation flow rate also dropped by 11 percent from 1,800 vplphg to 1,600 vplphg.

The study also explored the possibility of having the inclement weather plan automatically activated by several RWIS sensors located near the intersections.  The study concluded that there is not enough correlation between RWIS data and the actual road conditions to do this reliably.

Parsonson (1992) also discusses the principles of signal timing for adverse weather.  This study relates traffic in adverse weather conditions to traffic in a congested state.  The study recommends that a snowy corridor be "flushed" by setting all corridors signals to green at the same time.  This is a similar management scheme for some heavily congested corridors.

The design and/or operation of a transportation system, according to Jones and Goolsby (1970), may be based on assumptions of "normal environmental conditions." However, to have a comprehensive (system) design or control plan, the operation also must be predictable under degraded environmental conditions. The report is intended to assess the quantitative effects of rain on the design capacity of freeways.  They also defend the need for designing transportation systems, which minimize the effects of environmental disturbances such as rain or snow. 

FHWA (1977) assesses the economic impacts of adverse weather on all types of highways.  Some of these related impacts are: extra fuel consumption, operating costs (mechanical maintenance etc.), fixed costs (insurance, depreciation, taxes, etc.), and work delay.  As supportive evidence to their findings, they also measured interstate speeds of vehicles in varying degrees of inclement weather.  Their findings are provided in Table 2.3.

Table 2.3 Inclement Weather Speed Reductions (FHWA 1977)

Botha and Kruse (1992) show how inclement weather reduces saturation flow rates after collecting extensive headway data.  The study assesses the effects of residual ice and snow on saturation flow rates and start-up lost times at signalized intersections in Fairbanks, Alaska.  The winter data collection and subsequent analysis are reported and compared with the saturation flow rates suggested in the Highway Capacity Manual (HCM).  The winter saturation flows measured were much less than those suggested in the HCM.  It was found that when snow and ice were prevalent at signalized intersections, saturation flow rates are reduced by about 20 percent.  Table 2.4 shows the results of the study.

Table 2.4 Comparison of Saturation Flow Rates in Botha-Kruse Study

Knoblauch, Pietrucha, and Nitzburg (1996) study pedestrian crossing characteristics during inclement weather.  The study finds that as the severity of the weather increases, the walking speed of the pedestrians also increases.

Gilliam (1992) demonstrates how to measure the increase in congestion of a signal network due to inclement weather.  Levels of congestion can easily be found when road conditions are wet and travel times are greater than that under normal weather conditions.  Also, when a reduction in saturation flow is found on corridors, it becomes possible to relieve these areas of congestion due to poor weather conditions using a SCOOT system.  To accomplish this, specially developed wet weather parameters (such as decreased saturation flow and longer travel times) can be input into SCOOT.  It is expected that the SCOOT system will perform more appropriately and optimize traffic flow when wet parameters are applied. The precise monitoring of traffic conditions has allowed more attention to be focused towards the operation of a SCOOT system, and its perceived ability to handle varying traffic conditions over long periods of time and under different weather conditions.

Chapter 3. Current Practice by Other State DOT Agencies

As part of this project, several other state DOT agencies were contacted to find out current practices to accommodate signal timings during inclement weather.  In February 2000, a letter was e-mailed to 11 state DOT agencies to see if inclement weather signal timing had been addressed.  The states contacted are: (Alaska, Colorado, Idaho, Kansas, Maryland, Michigan, Minnesota, Ohio, Pennsylvania, Vermont, and Virginia).  The text of the letter is shown in Figure 3.1.

Figure 3.1 Letter to other DOT agencies

Four of the 11 state agencies responded that they have looked at inclement weather signal timing.  Below is a summary of the key information that Colorado, Maryland, and Ohio provided in response to the survey.  The Minnesota DOT referred us to current research by Maki (1999).  Although we have a research report from Alaska, their DOT agency did not respond to this survey.  The survey responses are included in Appendix A.

The Colorado Department of Transportation (CDOT) reported that they have not studied signal timings in inclement weather.  They have, however, modified signal timings at a few intersections on steep grades.

The Maryland State Highway Agency (Maryland SHA) currently is requesting funding from the FHWA to develop inclement weather signal timing plans similar to ours.  The FHWA first stage comments to the proposal indicate that implementation of this concept has not been widely sought by local highway agencies and that they welcomed development of appropriate hardware.  Maryland SHA  currently is working with their existing signal and weather detection equipment vendors for incorporation of this concept as part of an "adaptive" timed demonstration signal system project.

The Ohio Department of Transportation (ODOT) agency commented that if a signal system is capable of operating in "traffic responsive" mode, then it is possible to setup multiple timing plans, which are activated when the operating speed at selected locations falls below pre-determined thresholds.