Executive Summary
Inclement weather can significantly degrade roadway traffic operations, reducing service levels and creating unsafe conditions. Advances in sensor technologies and continuing deployment of intelligent transportation system (ITS) architectures provide an important opportunity for traffic management agencies to anticipate, mitigate, and intervene through various advisory and control measures to better manage conditions in the presence of inclement weather. Achieving this potential requires tying weather forecasting and traffic management capabilities together in an integrated framework that captures the effect of weather and weather-related measures on traffic system performance.
Traffic analysis tools used in practice typically ignore the effect of weather, and hence lack essential features to support weather-related traffic management. This study overcomes this deficiency by developing weather-sensitive traffic prediction and estimation models and incorporating them in Traffic Estimation and Prediction (TrEPS) tools intended for online operation in traffic management centers (TMC) as well as for offline evaluation of contemplated measures.
The development of weather-sensitive TrEPS is built on (1) a synthesis of existing cross-disciplinary knowledge on traffic responses to weather conditions and the application of weather-responsive advisory and control strategies, and (2) a thorough review of existing corridor and network traffic estimation and prediction models and systems that incorporate weather impacts or that can be adjusted to account for weather conditions. The synthesis addresses both the impact of weather events on traffic system performance (supply side), as well as traveler behavioral responses to weather events and related traffic advisory information (demand side). Both aspects are incorporated in the models developed in the study.
The principal supply-side and demand-side elements affected by adverse weather are systematically identified and modeled in the framework of traffic estimation and prediction systems, in order to account for changing weather conditions, as well as the availability of traveler information systems and weather-responsive traffic control devices. Where possible, the models and relations developed have been calibrated using available observations of traffic and user behavior in conjunction with prevailing weather events. The proposed weather-related features have been implemented in the DYNASMART TrEPS, and demonstrated through successful application to a real world network, focusing on two aspects: (1) assessing the impacts of adverse weather on transportation networks; and (2) evaluating effectiveness of weather-related advisory/control strategies in alleviating traffic congestion due to adverse weather conditions.
The procedures implemented provide immediately applicable tools that capture knowledge accumulated to date in the growing body of literature regarding weather effects on traffic. The application to a real world network shows that the proposed model can be used to evaluate weather impacts on transportation networks and the effectiveness of weather-related variable message signs.
The high level framework for incorporating weather impacts in TrEPS, presented in this study, provides a direction for future development towards a modern approach to traffic management under adverse traffic that recognizes modern technological developments (e.g. weather sensing/forecasting, weather responsive traffic management). The work accomplished in this study advances the state of the art in incorporating weather effects in network analysis tools. Additional effort in two main areas is necessary to translate these advances into practice. The first entails actual implementation in the context of a regional planning and/or traffic operations agency to establish the model and calibrate it for application under a variety of local conditions and traffic patterns. The second area of development would focus on weather-related traffic management and control measures, and interfacing their actual deployment with the decision-support tools developed in this project.