Inclement weather influences roadway safety, mobility, and productivity. It affects roadway safety through higher crash risk. Federal Highway Administration (FHWA) defines inclement crashes as “those crashes that occur in adverse weather (i.e., rain, sleet, snow, and/or fog) or on slick pavement (i.e., wet pavement, snowy/slushy pavement, or icy pavement)”. Drivers tend to adapt their driving behaviour to adjust the conditions presented by inclement weather. Depending on the surroundings and visibility issues, drivers drive more vigilantly by keeping longer headways, reducing operating speeds, or being more cautious. Approximately 24 percent of the U.S. traffic crashes happen due to inclement weather, resulting in 7,130 fatalities and 629,000 injuries. According to FHWA, inclement weather contributed to over 13 percent of injury crashes and over 10 percent of fatal crashes in 2013. Moreover, there is a perception that transportation planners and managers can do little about weather induced crash prevention. Visibility conditions at the time of a crash are rarely documented at a high level of detail in the conventional police reports. While vision is a key component of how drivers acquire information, a direct relationship between quantified levels of visibility related issues and safety (in terms of crashes) needs to be identified. This research seeks to understand the ramifications of inclement weather on safety from a perspective of visibility and other key issues. The key objective of this study is to examine safety impacts of reduced visibility and associated issues during inclement weather driving. To perform a robust analysis on inclement weather crashes, the research team performed a systematic literature review to determine the key factors associated with inclement weather related crashes. The research team used the second Strategic Highway Research Program’s (SHRP2) Roadway Information Database (RID) database to determine the suitable states for data analysis. After performing quality check on the availability of the variables and percentage of missing information in each of the variables, the research team selected two states for analysis: Florida and Washington. To examine the safety impacts of inclement weather crashes, this study used three methods: 1) parametric model (ordinal logistic regression) development to quantify visibility issues by using Florida SHRP2 RID, 2) non-parametric analysis (multiple correspondence analysis [MCA]) to identify key associated factors for inclement weather crashes by using Washington SHRP2 RID, and 3) topic model development by analysing inclement weather related crash narratives from SHRP2 Insight website. This report is organized as follows. Chapter 1 discusses research problem, key objective, and research approach. Chapter 2 provides the literature review on the existing body of research. Chapter 3 describes parametric statistic models, while Chapter 4 documents non-parametric analysis. Chapter 5 provides description on the text mining performed on the crash narratives. Chapter 6 describes conclusions. (Author/publisher)
Samenvatting