Data Mining and Gap Analysis for Weather Responsive Traffic Management Studies

6.0 Findings and Recommendations

This goal of this project was to identify gaps in weather and traffic data that can be used to help implement weather-responsive traffic management strategies. Several activities were undertaken to accomplish this, including:

1. A comprehensive search and documentation of weather-related traffic studies was conducted. This effort built upon previous literature searches but focused heavily on international studies. A number of test datasets were identified and requested.
2. Representatives of the client and consulting team attended the 2010 TRB Annual Meeting and through session attendance and committee meeting attendance were able to identify additional studies. These efforts and those of the literature search were documented in a technical memorandum.
3. The project team conducted a review of available weather and traffic data from major metropolitan areas. A set of criteria were established, including proximity of weather and traffic stations and a variety of weather conditions. Weather datasets were drawn from a variety of sources, including National Weather Service, Clarus, and state DOTs. Traffic data were obtained from datasets collected for FHWA’s Mobility Monitoring program and directly from state DOTs.
4. Datasets were screened based on the criteria established and three cities were selected for further testing; Atlanta, Salt Lake City and Minneapolis‑St. Paul.
5. An analysis of the impact of weather on travel speeds and lane usage in the three selected cities using the weather and traffic datasets obtained. Locations were selected where ESS and traffic detectors were found in close proximity. Dry and rainy conditions were tested in all three cities while data on snow conditions also were tested in Salt Lake City.
6. A survey of Traffic Management Centers (TMC) regarding the collection and use of weather and traffic data was conducted. TMC managers were asked what types of weather and traffic data were collected, whether data were used or being considered for use in Weather-Responsive Traffic Management and whether any research activities were planned or underway. Seventeen TMCs responded to the questionnaire, representing a good cross-section of urban areas and climatic conditions.
7. A similar survey of research institutions regarding ongoing studies of the impact of weather on traffic flow. Thirty-nine institutions responded providing information on a variety of studies. Eighteen agreed to a follow-up survey.
8. Two members of the project team conducted a site visit to Utah DOT which is a national leader the integration of weather and traffic management. Meetings were held with TMC management and members of the meteorological staff, which provides forecasting services to UDOT and also maintains their RWIS system. Lessons learned were documented in detail in this report.
9. Based on the activities above, data gaps were identified and incorporated into the findings and recommendations documented below. The activities in this section are referenced by number below in the findings and recommendations.

6.1 Findings

1. There are many efforts underway attempting to establish relationship between weather and traffic flow variables, particularly in research institutions. Studies involve rates of precipitation and surface condition, both of which are important to WRTM strategies (activities 1, 2, 7).
2. Research studies tend to be limited in scope, however. Many of the studies are subject to confidentiality agreements, which limit their ability to be used for operational purposes. These limitations seem to be most common in international studies (activities 1, 2, 7).
3. The analysis conducted for this project indicated that relationships between weather conditions and travel speeds could be established with readily available data and relatively basic analysis tools. However, greater sophistication will be required to establish operating procedures for WRTM strategies such as variable speed limits and weather-related traffic signal plans (activities 3, 4, 5).
4. Compared to three to four years ago, more agencies are collecting both traffic and weather data, due to gradual expansion of ITS and RWIS. However, funding is a major hurdle in expansion as well maintenance of existing systems and equipment. Legacy ESS often does not collect precipitation intensity data, which are important to WRTM. Many agencies are now purchasing travel time data from private vendors, a change that enables them to greatly increase the road-miles covered. However some of these agreements include confidentiality restrictions on this information, making it more difficult to use for research purposes (activity 6).
5. In conducting research on weather impacts on traffic, particularly fast changing events such as snowstorms and thunderstorms, it is important to have weather data at frequent temporal intervals, preferably five minutes. Conditions can change dramatically during these events over very short periods of time. The sources of surface weather data that are generally most reliable, ASOS stations are located primarily at airports which may not be optimal for roadway condition analysis (activities 3, 4, 5, 6).
6. The activities of Utah DOT provide a good prototype demonstrating how investment in road weather information and staff can be leveraged effectively. UDOT staff includes a full-time staff meteorologist. The meteorologist is responsible for procurement and supervision of a private contractor that provides meteorological staff and services, including both forecasting services and maintenance of ESS statewide. UDOT been able to leverage this capability and reduce costs through actions such as an innovative ESS design that utilizes components from various vendors. In addition adjacent States have contracted with UDOT’s vendor, reducing costs to UDOT through resource sharing and providing more timely information on weather conditions in adjacent states. It also should be noted that UDOT is one of the few agencies to have implemented a WRTM program; modification of traffic signal timing along several major corridors during winter storms (activity 8).

6.2 Recommendations

1. FHWA has been effectively disseminating information on its research activities through brochures, technical papers, conferences, and webcasts. It is important to stay current with other research efforts, both domestic and international, and try to leverage these studies wherever possible. A recent example was provided in the FHWA study “Microscopic Analysis of Traffic Flow in Inclement Weather,” where a dataset on icy road conditions developed by Hokkaido University was obtained by FHWA and evaluated by the Virginia Tech Transportation Institute (activities 1, 2, 7).
2. It is recommended that a separate research effort be conducted into confidentiality agreements, focusing on Europe, where this seems to be common practice. This effort should identify the source and reasons for confidentiality agreements, and identify any initiatives that can be taken to make datasets more available (activities 1, 2, 7).
3. FHWA should continue its efforts to develop weather-related parameters for microsimulation models and should conduct tests of WRTM strategies with those models. After an initial set of tests the models could be installed in several TMCs and simulations conducted in parallel with actual TMC operations. A subsequent step would be to use the simulation results and test actual strategies implemented by transportation operating agencies. Agencies involved in the tests would require a “before” period to collect detailed data on traffic flow during weather events so that the impact of the strategies could be identified (activities 3, 4, 5).
4. FHWA has several active projects in this area, including guidelines development of WRTM and microscopic analysis of traffic flow during adverse weather. The latter research project has been conducted at arterial intersections using video and Environmental Sensor Stations. A data collection program is recommended for several freeway corridors. Changes in traffic flow parameters need to be monitored continuously over a period of time so the impacts of congestion and changing demand can be identified. Several corridors could be selected that already have good coverage of traffic detectors and ESS and then supplemented where necessary. It is possible that some of the Integrated Corridor Management projects currently being funded by FHWA could be used to leverage funding (activity 6).
5. FHWA should look at ways to leverage data from NWS, other government agencies, the IntelliDrive program and private weather sources to obtain higher-quality, more detail observations on road weather and surface conditions. While state DOT RWIS are expanding, this expansion is relatively slow due to funding constraints. In addition, many agencies are finding it difficult to maintain legacy ESS due to funding constraints. Development of a plan would be an important step in accomplishing this. Much of the input data required already has been collected by FHWA and its contractors (activities 3, 4, 5, 6).
6. These best practices of UDOT and other agencies should be disseminated through brochures and other forums such as conferences and webcasts (activity 8).

Previous | Next