UC Berkeley

Congestion on freeways occurs when demand exceeds the available capacity. Common causes of recurring congestion, also known as freeway bottlenecks, include lane drops, on-ramp merges, and weaving sections. Adverse weather reduces traffic speeds and the maximum queue discharge flow at freeway bottlenecks. However, the impact of weather characteristics on bottleneck discharge flows has not been systematically investigated. This research investigated the relationship between bottleneck queue discharge flow and weather characteristics including rainfall intensity, wind speed, and visibility.

Queue discharge rates at four isolated merge bottlenecks within Orange County, California were measured utilizing an established methodology of cumulative count and occupancy curves. An analysis of how queue discharge varied by rainfall intensity revealed reduced discharge ranging from 5% in drizzle (rainfall <0.02 inches/hour) up to 27% in heavy rainfall (rainfall >0.1 inches/hour). However, variation in this single weather characteristic only accounted for a small percentage of the variability in discharge flow, particularly in light rain. Several hypotheses were proposed and tested utilizing the two additional variables of wind speed and visibility and dividing the periods of discharge flow into three groupings. Analyses based on these hypotheses better described the variation in queue discharge flow than the analysis with rainfall intensity alone. A model was developed to predict bottleneck discharge flow by combining data points from all sites. This model predicted that an increase in rainfall intensity of 0.1 inches per hour reduced queue discharge by approximately 1.8% at all sites after the onset of congestion.

This research shows that weather characteristics are an important predictor of bottleneck queue discharge rates. Forecasted weather patterns could be used to predict reductions in bottleneck capacity. Complementary research building on this work by examining changes in trip start time during adverse weather would allow an improved prediction of vehicle delay and travel time reliability. This information would allow traveler information services to incorporate weather characteristics in order to provide more accurate predicted route times for commuters.