Influence of rainfall on red-light running at signalised intersections and the service quality implications.

Abstract

When drivers approach a signalised intersection stop-line, they must decide whether to stop or proceed and clear the intersection before the end of a green phase. The driver behaviour at the intersection affects signalised intersections' performance especially in terms of safety (red-light violation) and efficiency (throughput and delay). Drivers are affected by the state of the traffic lights, prevailing traffic conditions, road conditions and prevailing yellow light laws. When rain falls driver behaviour is affected, and this could, in turn, affect the performance of signalised intersections. This study aims to determine the impact rainfall has on red-light violations and what implications that could have on intersection service quality. In Durban South Africa, four (4) signalised intersections were selected for a “dry” versus “rainy” study carried out using traffic, and rainfall data collected over eight weeks covering the rainfall season. The probability of red-light running (RLR) was found to decrease with an increase in rainfall intensity. The probability reduced from about 42% on average under dry conditions to 17% under light, 7% under moderate and 2.5% for heavy rainfall intensity. It implies that it becomes nearly impossible to violate a red light under rainfall conditions due to speed reduction and hence increase in travel time. The average time needed to safely cross the stop line at the onset of yellow time interval also increased from 3s during dry weather to 3.6s for light rain, 3.9s for moderate rain, and 4.5s for heavy rain. Thus, approaching vehicles cannot safely enter the signalised intersection and must wait at the stop line for a green signal. Therefore, it can be summarised that rainfall has a mitigating effect on red light violations especially under heavy rain where it is near impossible to run a red light. South Africa does not have a highway capacity manual (HCM) and relies on USA-HCM for signalised intersection assessment. HCM uses delay as the sole determiner of signalised intersection quality of service; this was found inadequate and not a complete reflection of driver perception of the level of service; this study proposed a criterion that incorporated degree of saturation in addition to the delay. With the developed criteria, analysis was done to determine the rainfall influence on signalised intersection performance. Through and right-turning traffic were considered separately in this study. For through traffic, saturation flow rate reduced by 3.9% under light rainfall, 8.68% under moderate and 10.88% under heavy. It led to a capacity loss of 4.25% under light, 9.18% under moderate and 11.5% under heavy. For right-turning traffic, saturation flow rates decreased by 7.07% under light rainfall intensity, 13.44% under moderate and 17.88% under heavy. The capacity loss was also recorded where light rainfall caused a 7.38% loss, moderate 14.5% and heavy 19.15%. For the degree of saturation, delay, and queue length, all three increased. The degree of saturation increased by 1.55% under light, 7.23% under moderate and 9.4% under heavy. The overall impact on service quality was mixed; for through traffic lanes, few instances where heavy rainfall caused a deterioration in SQ by one level were recorded. For right-turning traffic lanes, the results were more consistent with expectations. There was an increase in both degrees of saturation and delay. Overall, the SQ level deteriorated especially under heavy rainfall conditions. Right-tuning lanes showed higher SQ deterioration attributed to their higher saturation level.