Exploring factors contributing to injury severity at work zones considering adverse weather conditions

Identifying risk factors for road traffic injuries can be considered one of the main priorities of transportation agencies. More than 12,000 fatal work zone crashes were reported between 2000 and 2013. Despite recent efforts to improve work zone safety, the frequency and severity of work zone crashes are still a big concern for transportation agencies. Although many studies have been conducted on different work zone safety-related issues, there is a lack of studies that investigate the effect of adverse weather conditions on work zone crash severity. This paper utilizes probit–classification tree, a relatively recent and promising combination of machine learning technique and conventional parametric model, to identify factors affecting work zone crash severity in adverse weather conditions using 8 years of work zone weather-related crashes (2006–2013) in Washington State. The key strength of this technique lies in its capability to alleviate the shortcomings of both parametric and nonparametric models. The results showed that both presence of traffic control device and lighting conditions are significant interacting variables in the developed complementary crash severity model for work zone weather-related crashes. Therefore, transportation agencies and contractors need to invest more in lighting equipment and better traffic control strategies at work zones, specifically during adverse weather conditions.

Objective: The objective of this study was to identify the factors that influence the frequency and severity of rear-end crashes in work zones because rear-end crashes represent a significant proportion of crashes that occur in work zones. Methods: Truncated count data models were developed to identify influencing factors on the frequency of read-end crashes in work zones and ordered probit models were developed to evaluate influencing factors on the severity of rear-end crashes in work zones. Results: Most of the variables identified in this study for these 2 models were significant at the 95 percent level. The statistics for models indicate that the 2 developed models are appropriate compared to alternative models. Conclusions: Major findings related to the frequency of rear-end crashes include the following: (1) work zones for capacity and pavement improvements have the highest frequency compared to other types of work zones; (2) work zones controlled by flaggers are associated with more rear-end crashes compared to those controlled by arrow boards; and (3) work zones with alternating one-way traffic tended to have more rear-end crashes compared to those with lane shifts. Major findings related to the severity of the rear-end crashes include the following: (1) rear-end crashes associated with alcohol, night, pedestrians, and roadway defects are more severe, and those associated with careless backing, stalled vehicles, slippery roadways, and misunderstanding flagging signals are less severe; (2) truck involvement and a large number of vehicles in a crash are both associated with increased severity, and (3) rear-end crashes that happened in work zones for bridge, capacity, and pavement are likely to be more severe than others.

A principal safety issue at highway-rail grade crossings (HRGCs) is the severity of crashes. Although many studies have analyzed crash severity at HRGCs, they often rely on national datasets or a narrow set of variables, frequently overlooking region-specific factors such as roadway design, driver behavior, and local environmental conditions. However, this study contributes to the existing body of literature by providing additional insights into the factors associated with injury severity in HRGC crashes. This study aimed to model HRGC crash severity using statistical and machine learning methods, specifically Ordinal Logistic Regression (OLR) and Random Forest (RF) algorithms, to determine significant factors associated with severe injury HRGC crashes. The statistical modeling and analyses were based on five years of HRGC crash data (2017–2021) at state-maintained HRGCs in Florida. Based on the OLR statistical model, ten variables were significant at a 95% confidence interval: crashes that occurred in the morning peak hours, no lighting condition, adverse weather conditions, railway vehicle (i.e. train or train engine), driver action (i.e. disregarded signs, signals, markings as well as other contributing actions), a speed limit of greater than 45 mph, four-lane highways, driver younger than 25, female drivers, crashes that occurred at the railroad crossings, and estimated vehicle damage of more than $1,000. Results from the OLR model indicate that all significant variables increase the likelihood of an HRGC crash being more severe, except for the time of crash occurrence (morning peak), adverse weather conditions, and drivers under 25 years of age. According to the RF model, the most important (top five) factors affecting the injury severity of HRGC crashes include estimated vehicle damage, posted speed limit, type of shoulder, driver action, and crash type. Except for the type of shoulder and crash type, the RF model results are consistent with those of the OLR model. Finally, based on the model results, potential countermeasures to mitigate fatalities and injuries at HRGCs were presented.

Adverse weather conditions are one of the primary causes of motor vehicle crashes. To identify the factors contributing to crashes during adverse weather conditions and recommend cost-effective countermeasures, it is necessary to develop reliable crash prediction models to estimate weather-related crash frequencies. To account for the variations in crash count among different adverse weather conditions, crash types, and crash severities for both rain- and snow-related crashes, crash data on freeways was collected from the State of Connecticut, and crash prediction models were developed to estimate crash counts by crash type and severity for each weather condition. To account for the potential correlations among crash type and severity counts due to the common unobserved factors, integrated nested Laplace approximation (INLA) multivariate Poisson lognormal (MVPLN) models were developed to estimate weather-related crashes counts by crash type and severity simultaneously (four MVPLN models were estimated in total). To verify the model prediction ability, univariate Poisson lognormal (UPLN) models were estimated and compared with the MVPLN models. The results show that the effects of factors contributing to crashes, including median width, horizontal curve, lane width, and shoulder width, vary not only among different adverse weather conditions, but also among different crash types and severities. The crash types and severities are shown to be highly correlated and the model comparison verifies that the MVPLN models significantly improve the model prediction accuracy compared with the UPLN models. Therefore, the MVPLN model is recommended to provide more unbiased parameter estimates when estimating weather-related crashes by crash type and severity.

This study investigated the impact of traffic violations on crash injury severity on Wyoming's interstate highways. A random parameters multinomial logit (MNL) model with heterogeneity in means was estimated as an alternative to the mixed logit model. This was done to better account for unobserved heterogeneity in the crash data. As per the results, the random parameters model with heterogeneity in means not only exhibited a better fit but also uncovered more insights regarding the factors influencing crash injury severity. The advanced model showed that traffic violations, crash characteristics and environmental characteristics among other factors impact crash injury severity on Wyoming's interstate highways. With regards to traffic violations, driving too fast for prevailing conditions and driving under the influence of alcohol and drugs were identified as the main violations that significantly influenced crash severity. Among other useful insights, the heterogeneity in mean specification indicated that the likelihood of severe injury crashes is increased by the interactive effect between non-trucks (vehicles not classified as trucks) and driving too fast for conditions. This is a significant implication that high speed behavior by non-truck drivers in adverse weather conditions is ranked as one of the hazardous traffic violations on Wyoming's interstates. This study provided for the first time important information on the impact of traffic violations on crash severity of crashes that occurred on challenging roadways that characterized by mountainous terrain and severe weather conditions. Results from the study will help enforcement agencies in the state to better identify appropriate countermeasures to mitigate the impact of violations on crash severity.

Multivariate poisson lognormal modeling of crashes by type and severity on rural two lane highways

Analysing the Severity and Frequency of Traffic Crashes in Riyadh City Using Statistical Models

This paper quantified the effects of freeway ramp spacing and auxiliary lane presence on crash frequency and crash severity. Crash frequencies were predicted with a safety performance function, and crash severities were estimated with what was termed a “severity distribution function.” The paper then demonstrated how to combine quantitative knowledge related to the effects of ramp spacing and auxiliary lane presence on both crash frequency and severity into a framework for assessing the overall crash cost for different ramp configurations. Geometric features, traffic characteristics, and crash data were collected for 404 freeway segments in California and Washington State. Negative binomial regression models and multinomial logit regression models were used to estimate the effects of ramp spacing and auxiliary lane presence on expected crash frequencies and crash severities, respectively. Results showed that expected multiple-vehicle crash frequency increased as ramp spacing decreased. Meanwhile, there was a decrease in the proportion of severe crashes (fatal, incapacitating injury) with a decrease in ramp spacing, even though the overall frequency of these severe crashes remained relatively unchanged. Providing an auxiliary lane was expected to decrease crash frequency, although this reduction appeared to be primarily in crashes that were less severe (possible injury and property damage only). The findings appeared to effectively capture the complex relationships between geometric designs and operations and the high sensitivity between speed and crash severity. The paper provided quantitative tools for making informed freeway and interchange design decisions where ramp spacing and auxiliary lanes were considerations.

Objectives Understanding the factors influencing crash severity of autonomous vehicles is important for increasing road safety. This study focuses on a multi-source accident dataset of vehicles equipped with autonomous driving systems to explore the endogenous relationship between manual takeover of autonomous vehicles and the severity of crash, as well as the influencing factors. Methods By screening and summarizing data on autonomous vehicle accidents. We choose self-driving car takeover and crash severity as potential variables to build a structural equation model to explore the influences of crash severity through continuous variable updating and path improvement. We select autonomous vehicle takeover and crash severity as potential variables and designed a structural equation model to explore the factors affecting crash severity through continuous variable updating and path improvement. Meanwhile, we establish a generalized linear logit model to analyze the factors affecting manual takeover. Finally, the intrinsic link between crash severity and manual takeover is discussed through path analysis and comparison of model results. Results Cloudy and rainy weather, left rear of vehicle contact area, and daylight lighting significantly impact manual takeover and crash severity. Specifically, wet road surface, rainy weather, and daylight have relatively more significant effects on takeover in the structural equation model. And takeover, roadway type including non-freeway and intersection can significantly impact crash severity. Additionally, the study demonstrates the endogeneity between crash severity and takeover at the time of autonomous vehicle crash. Conclusions This study analyzes the potential relationships and influencing factors between takeover events of autonomous vehicles and crash severity. It is found that the frequency of takeover events significantly increases when driving in rainy weather and at night. It is suggested that a real-time monitoring module for adverse weather or lighting conditions should be added to the autonomous driving system to provide early warnings and reduce the occurrence of takeover events, thereby enhancing the safety and reliability of autonomous vehicles.

Study of work zone traffic safety under adverse driving conditions with a microscopic traffic simulation approach

The attributes of work zones have significant impacts on the risk of crash occurrence. Therefore, identifying the factors associated with crash severity and frequency in work zone locations is of important value to roadway safety. In addition, the significant loss of workers’ lives and injuries resulting from work zone crashes indicates the emergent need for a comprehensive and in-depth investigation of work zone crash mechanisms. The cost of work zone crashes is another issue that should be taken into account as work zone crashes impose millions of dollars on society each year. Applying innovative construction methods like Accelerated Bridge Construction (ABC) dramatically decreases on-site construction duration and thus improves roadway safety. This safe and cost-effective procedure for building new bridges or replacing/rehabilitating existing bridges in just a few weeks instead of months or years may prevent crashes and avoid injuries as a result of work zone presence. The application of machine learning techniques in traffic safety studies has seen explosive growth in recent years. Compared to statistical methods, MLs are more accurate prediction models due to their ability to deal with more complex functions. To this end, this study focuses on three major areas: crash severity at construction work zones with worker presence, crash frequency at bridge locations, and assessment of the associated costs to calculate the contribution of safety to the benefit-cost ratio of ABC as compared to conventional methods. Some key findings of this study can be highlighted as in-depth investigation of contributing factors in conjunction with the results from statistical and machine learning models, which can provide a more comprehensive interpretation of crash severity/frequency outcomes. The demonstration of work zone crashes needs to be modeled separately by time of day for severity analysis with a high level of confidence. Investigation of the contributing factors revealed the nonlinear relationship between crash severity/frequency and contributing factors. Finally, the results showed that the safety benefits from a case study in Florida consisted of 43% of the total ABC implementation cost. This indicates that the safety benefits of ABC implementation consist of a considerable portion of its benefit-cost ratio.

Closure of Safety Rest Areas (SRAs) increases the risk of fatigue-related crashes by reducing opportunities for long distance traveling drivers to rest, recover alertness, and access essential facilities that support safe driving. Utilizing Cluster Correspondence Analysis (CCA), this study aims to identify and analyze patterns and associations in crash data by examining changes in crash characteristics and severity, including factors and attributes related to vehicular, driver, roadway, and contributing factors, before, during, and after the SRA closure of the Snake Riverview Welcome Center on I-84 from 2018 to 2021. The findings indicated significant shifts in crash types and contributing factors, particularly under adverse weather conditions and on wet or snowy surfaces. Rear-end collisions and improper lane changes were persistent issues. There was a rise in crashes involving commercial vehicles, such as tractors with trailers, and incidents occurring during nighttime hours. Additionally, the incidence of vehicle overturns and sideswipe collisions increased during adverse weather conditions. The results underscore the role of SRAs in mitigating crashes and highlight the necessity for targeted interventions, such as maintaining SRAs and implementing safety measures during adverse weather and high-risk hours.

Objective: Work zone safety is one of the top priorities for transportation agencies. In recent years, a considerable volume of research has sought to determine work zone crash characteristics and causal factors. Unlike other non–work zone–related safety studies (on both crash frequency and severity), there has not yet been a comprehensive review and assessment of methodological approaches for work zone safety. To address this deficit, this article aims to provide a comprehensive review of the existing extensive research efforts focused on work zone crash-related analysis and modeling, in the hopes of providing researchers and practitioners with a complete overview.Methods: Relevant literature published in the last 5 decades was retrieved from the National Work Zone Crash Information Clearinghouse and the Transport Research International Documentation database and other public digital libraries and search engines. Both peer-reviewed publications and research reports were obtained. Each study was carefully reviewed, and those that focused on either work zone crash data analysis or work zone safety modeling were identified. The most relevant studies are specifically examined and discussed in the article.Results: The identified studies were carefully synthesized to understand the state of knowledge on work zone safety. Agreement and inconsistency regarding the characteristics of the work zone crashes discussed in the descriptive studies were summarized. Progress and issues about the current practices on work zone crash frequency and severity modeling are also explored and discussed. The challenges facing work zone safety research are then presented.Conclusions: The synthesis of the literature suggests that the presence of a work zone is likely to increase the crash rate. Crashes are not uniformly distributed within work zones and rear-end crashes are the most prevalent type of crashes in work zones. There was no across-the-board agreement among numerous papers reviewed on the relationship between work zone crashes and other factors such as time, weather, victim severity, traffic control devices, and facility types. Moreover, both work zone crash frequency and severity models still rely on relatively simple modeling techniques and approaches. In addition, work zone data limitations have caused a number of challenges in analyzing and modeling work zone safety. Additional efforts on data collection, developing a systematic data analysis framework, and using more advanced modeling approaches are suggested as future research tasks.

Work zones are widely recognized as major contributors to road fatalities and traffic congestion. Although extensive research has explored the relationship between work zone crashes and contributing factors, a comprehensive systematic review and meta-analysis remain absent. This study addresses this gap by exploring four key research questions: (i) Which elements of work zones are most crash-prone? (ii) What factors affect work zone severity and crash frequency? (iii) Which methods are used to predict crash occurrences and crash severity? (iv) How does the traffic volume affect crash occurrence with different severity levels in the work zone? The review identifies factors influencing crashes, including work zone characteristics, environmental conditions, roadway features, temporal aspects, driver characteristics, and crash attributes, and evaluates various modeling approaches. Moreover, a meta-analysis quantifies the association between traffic volume and crash severity, highlighting key findings for safety measures and developing targeted strategies for improving work zone safety.

Both crash count and severity are thought to quantify crash risk at defined transport network locations (e.g. intersections, a particulate section of highway, etc.). Crash count is a measure of the likelihood of occurring a potential harmful event, whereas crash severity is a measure of the societal impact and harm to the society. As the majority of safety improvement programs are focused on preventing fatal and serious injury crashes, identification of high-risk sites—or blackspots—should ideally account for both severity and frequency of crashes. Past research efforts to incorporate crash severity into the identification of high-risk sites include multivariate crash count models, equivalent property damage only models and two-stage mixed models. These models, however, often require suitable distributional assumptions for computational efficiency, neglect the ordinal nature of crash severity, and are inadequate for capturing unobserved heterogeneity arising from possible correlations between crash counts of different severity levels. These limitations can ultimately lead to inefficient allocation of resources and misidentification of sites with high risk of fatal and serious injury crashes. Moreover, the implication of these models in blackspot identification is an important, unanswered question.While a joint econometric model of crash count and crash severity has the flexibility to account for the limitations mentioned previously, its ability to identify high-risk sites also needs to be examined. This study aims to fill this research gap by employing the joint model for blackspot identification. Using data from state-controlled roads in Queensland, Australia, a new risk score is developed based on predicted crash counts by severity, weighted by the cost ratio of severity levels. This weighted risk score is then used for identifying road segments with high risk of fatal and injury crashes. Results show that the joint model of crash count and crash severity has substantially improved prediction accuracy compared to the traditional count models. The correlation between crash counts of different severity levels captures the unobserved heterogeneity caused by the extra-variation in total crash counts and moderates the parameters in the joint model. In comparison with the traditional approaches, the proposed weighted risk score approach with the joint model of crash count and crash severity leads to the identification of a higher number of fatal and serious injury crashes in the top ranked sites flagged for safety improvements.