Почетна Преглед Помоћ
Пријавите се
ardo
/
Obsidian
1
0
Креирај огранак 0
Датотеке Дискусије 0 Захтеви за спајање 0 Вики
Дрво: 94c9e4e42b
Гране Ознаке
Obsidian
/
Zotero
/
001_artiklid
/
abuzwidahassessingimpactexpresslanes2024.md

abuzwidahassessingimpactexpresslanes2024.md 64 KB
Историја Датотека

category: literaturenote citekey: abuzwidahassessingimpactexpresslanes2024 title: Assessing the impact of express lanes on traffic safety of freeways authors: "Abuzwidah, Muamer; Abdel-Aty, Mohamed" year: 2024 date: 2024-00-00 2024 doi: 10.1016/j.aap.2024.107718 publication: "Accident Analysis & Prevention" url: "https://doi.org/10.1016/j.aap.2024.107718" zotero_key: CBC5H45Q zotero_storage: KAS9EB5I collections: transpordi_planeerimine folder: 001_artiklid firstAuthor: "Abuzwidah, Muamer"

status: converted

Contents lists available at ScienceDirect

Accident Analysis and Prevention

journal homepage: www.elsevier.com/locate/aap

Assessing the impact of express lanes on traffic safety of freeways

Muamer Abuzwidah a,* , Mohamed Abdel-Aty b

  • a Department of Civil and Environmental Engineering, University of Sharjah, Sharjah, United Arab Emirates
  • b Department of Civil, Environmental and Construction Engineering, University of Central Florida, Orlando, FL 32816-2450, United States

ARTICLE INFO

Keywords: Express Lanes Crash Modification Factors High Occupancy Toll Lanes Highway Safety Manual Managed Lanes Traffic Safety Toll Roads and Expressway

ABSTRACT

The rise of Express Lanes also known as High Occupancy Toll (HOT) Lanes and Managed Lanes, signifies a major leap in traffic management and transportation funding. Despite their increased deployment to ensure reliable travel times through dynamic tolling during peak traffic periods, a comprehensive evaluation of their safety impact is notably lacking. Presently, the Crash Modification Factors Clearinghouse, a vital resource, only lists two case studies related to Express Lanes, one of which is our own research. This lack of data highlights the critical need for more extensive studies to thoroughly assess the safety benefits of Express Lanes and to improve their application. This study aims to rigorously evaluate the safety impact of express lanes on freeways, presenting a first-of-its-kind, in-depth analysis of their specific effects on both Express Lanes and General-Purpose Lanes (GP-Lanes) individually. The analysis utilized data from 55 miles of Express Lanes across various locations in Florida, comparing them to High Occupancy Vehicle (HOV) lanes. The results demonstrate that converting HOV lanes to Express Lanes or introducing new ones does not compromise overall freeway safety. In fact, safety within Express Lanes improves, as evidenced by a decrease in crash occurrence and Crash Modification Factors for Express lanes, which are below "1" across all crash categories. This underscores the effectiveness of Express lanes in enhancing roadway safety. In contrast, incidents in GP-Lanes have increased, indicating a shift of crashes to these lanes, and thus making Express lanes relatively safer. This underlines the importance of continued research into the safety impact of express lanes and calls for further studies to refine traffic management strategies, aiming at enhancing travel efficiency while ensuring traffic safety, especially for the GP-Lanes amid the expansion of express lanes.

1. Introduction

The evolution of highway systems through the implementation of Express Lanes represents a significant leap forward in traffic management and urban mobility enhancement. This innovative approach, transitioning from traditional or High Occupancy Vehicle (HOV) lanes into Express lanes, aligns with a wider trend aimed at elevating the commuting experience. By providing a paid option to bypass congestion, Express lanes leverage dynamic tolling that adjusts fees based on traffic density, thus ensuring more consistent travel speeds (FHWA, 2024; Cho et al., 2011; Mahmoud et al., 2022; FHWA, 2003). This model of variable pricing, increasing costs to mitigate congestion during peak times, parallels strategies used across various industries, including utilities and hospitality, making Express lanes a premium travel choice (FT, 2024).

The Florida Department of Transportation (FDOT) has been instrumental in integrating Express lanes into the state's highway infrastructure. The objective is to alleviate commuter stress by offering quicker, more reliable travel options (FDOT, 2024; Burris and Sullivan, 2006). The Managed Lanes Comprehensive Traffic and Revenue Study by FDOT, concluded in 2007, highlighted the potential benefits of implementing Express lanes, forecasting substantial reductions in travel times.

Implementing electronic toll collection dramatically enhances safety, decreasing crashes by 65 %, affirming its precision and improving security on roadways (Abuzwidah et al., 2014), a technology also utilized in Express Lanes. Investigations into Toll roads facility have underscored the significance of optimizing safety through forwardthinking design and technological integration such as electronic toll collection systems (Abuzwidah and Abdel-Aty, 2018). In the realm of traffic management, the implementation of High Occupancy Toll (HOT) Lanes, also identified as Express or Managed Lanes, stands as a critical innovation aimed at enhancing traffic flow and safety. Notably, microsimulation scenarios have pinpointed optimal configurations for lanes designated for connected vehicles (CVs), indicating that tailored lane arrangements can markedly elevate traffic dynamics and safety (Abdel-Aty et al., 2020; Elawady et al., 2022). Park et al. (2023) proposes an autonomous vehicles managed-lane strategy to enhance safety and

E-mail addresses: mabuzwidah@sharjah.ac.ae (M. Abuzwidah), m.aty@ucf.edu (M. Abdel-Aty).

* Corresponding author.

efficiency in mixed-vehicle traffic, validated through simulation and value function analysis. Additional research has illuminated the pivotal role of access design in managed toll lanes on freeways, emphasizing the necessity of establishing the safest accessibility levels and weaving lengths to curtail traffic conflicts and bolster roadway performance (Saad et al., 2018). The influence of managed lane geometry on operating speeds has been documented, showcasing that design directly impacts speed consistency, thus affecting overall traffic mobility (Fitzpatrick et al., 2017). The introduction of bus priority lanes has also been linked to a reduction in traffic conflicts at intersections and bus stops, affirming the safety advantages of bus lane prioritization (Goh et al., 2014).

Moreover, the safety implications of weaving distances on freeways equipped with managed lanes were thoroughly examined, advocating for optimal lane change lengths to minimize safety hazards (Cai et al., 2018; Saad et al., 2019). Data from California and Texas have indicated that enlarging managed-lane envelopes can substantially decrease freeway crashes (Avelar and Fitzpatrick, 2018). The safety performance of HOT facilities has been scrutinized using Bayesian networks, identifying crash-contributing factors and suggesting enhancements to lane safety (Kitali et al., 2021). The impact of demographic factors and post-HOT installation perceptions on commuters' choices in UAE and Atlanta offers insights into the determinants of HOT lane and carpool utilization (Shabib et al., 2020; Guensler et al., 2020). HOT lanes enhance efficiency and sustainability, with studies indicating age and income impact usage willingness, especially in emergencies, highlighting broad public acceptance (Shabib et al., 2022). Assessing the safety impact of dedicated lanes for CVs has yielded strategic recommendations based on vehicle penetration rates and traffic conditions (Miqdady et al., 2023). Furthermore, the review of methodologies for identifying high-risk accident locations advocates for proactive safety management strategies, leveraging surrogate safety measures to manage road safety effectively (Washington et al., 2018). Traffic congestion significantly contributes to driver fatigue. Therefore, the utilization of express lanes can provide a viable solution for fatigued drivers by allowing them to bypass congestion in the general-purpose lanes (Elbaz et al., 2020). Express lanes have improved traffic management by providing options between general-purpose and express lanes, reflecting Florida's commitment to traffic efficiency. Table 1 in the 2022 FDOT report details 57.2 miles of express lanes projects across Florida.

A wide range of express lane projects, from construction to early planning, aims to strengthen Florida's transport network. Focused on projects with post-implementation data, this analysis excludes those in

initial stages. These initiatives, integral to mitigating urban congestion and improving highway efficiency, reflect FDOT's commitment to optimizing transport through variable tolling express lanes, enhancing mobility and reducing travel times. Express lane deployment significantly improved traffic dynamics and commuter satisfaction. Analysis of the 95-Express lanes revealed increases in weekday traffic by 9.5 % and peak evening traffic by 15.7 %, with express bus route usage rising by 33.5 %. This shift also enhanced speeds in both Express and General-Purpose Lanes, boosting overall network efficiency. Fig. 1(a and b) illustrates Express lanes in the USA.

Fig. 2 presented showcases a range of lane management techniques, collectively referred to as managed lanes. Initiating on the left side of the figure, it features the implementation of singular operational strategies.

Moving from simple to complex, managed lane frameworks increase in sophistication, integrating varied strategies for more advanced configurations. This shift emphasizes the adaptability of managed lane solutions to different contexts, including transitions to Express lanes. Fig. 3 in the NCHRP Research Report 835 displays Managed Lanes facility types and management strategies. It details four main project delivery options—from design-bid-build to design-build-finance-operate-maintain—offering crucial insights for enhancing road safety and efficiency.

Express lanes, through dynamic tolling, speed up commutes, cut congestion, and fund transport projects, boosting travel efficiency and supporting sustainable transit by reinvesting toll revenues. Fig. 4 outlines these benefits.

Across the U.S., express lanes adapt to regional transport needs, like Minnesota's 2005 Interstate 394 HOV-to-E-Lane conversion (Cao et al., 2010). This change, evaluated using Empirical Bayes methodology, resulted in a 5.3 % drop in crashes, highlighting safety benefits and toll revenue potential. Yet, it remains one of the few studies examining such conversions' safety impact across entire road segments, without comparing Express lanes and General-Purpose Lanes directly.

Express lanes in states like Minnesota, California, Colorado, Texas, Utah, and Washington, along with Stockholm's peak-hour tolling, have reduced traffic congestion by over 20 %, enhancing road safety by lowering crash rates and boosting public transportation use. Express lanes offer smoother commutes and environmental benefits, yet their safety impacts remain poorly quantified. This study fills this gap by analyzing their effects on traffic safety in Florida, using various analytical methods. The results, depicted in Fig. 5(a,b,c,d), guide the expansion of these lanes, enhancing both commuter experience and network safety.

In summary, the proliferation of Express Lanes across the United

Table 1 Locations and length of Florida's Express lanes ().

2022 FDOT Express Lanes
District 2
Roadway ID BMP EMP Length State road number Local name Toll name Owning authority
72,010,000 0.733 5.524 4.791 SR 9A I-295 389 − I-295 Express FDOT
District 4
Roadway ID BMP EMP Length State road number Local name Toll name Owning authority
86,070,000 0.000 4.700 4.700 SR 9 I-95 302 − I-95 Express FDOT
86,095,000 0.000 8.800 8.800 SR 862 I-595 340 − I-595 Express FDOT
86,075,000 0.000 11.546 11.546 SR 93 I-75 387 − I-75 Express FDOT
District 6
Roadway ID BMP EMP Length State road number Local name Toll name Owning authority
87,075,000 0.000 5.455 5.455 SR 93 I-75 387 − I-75 Express FDOT
87,200,000 4.331 13.690 9.359 SR 9A I-95 302 − I-95 Hot Lanes FDOT
87,270,000 13.690 17.199 3.509 SR 9 I-95 302 − I-95 Hot Lanes FDOT
87,260,000 7.105 16.142 9.037 SR 826 Palmetto EXPY 393 − Palmetto Express FDOT

Source: (2022 Report - (Florida-DOT, 2024))

Fig. 1. (a and b): examples of express lanes designs in USA.

Fig. 2. Typical section of express lanes (). Source: Parsons Brinckerhoff

States represents a crucial strategy for mitigating urban congestion and improving mobility. By dynamically adjusting toll rates, these lanes not only enhance traffic flow but also contribute essential funding for transportation infrastructure projects. Despite their increasing adoption, the detailed examination of their safety impact is lacking, with the Crash Modification Factors (CMF) Clearinghouse listing only two case studies on this topic, including our previous research. This highlights the urgent need for further, comprehensive studies to grasp the full scope of Express lanes' safety benefits and refine their implementation for optimal use. Our study addresses this research gap by evaluating the safety effects of implementing new Express lanes and transitioning HOV lanes to Express lanes. The analysis excludes data from the transition period, which spans six months before and after the changes, to prevent skewed results. This investigation marks the first comprehensive assessment of their impact on both Express lanes and General-Purpose Lanes (GPL), employing various analytical methods.

2. Methodology

This study employed Cross Sectional and Before-After methods, enhanced by a Comparison Group and the Empirical Bayesian approach (Hauer, 1997), to assess the safety effects of implementing new Express lanes and transitioning HOV lanes to Express lanes in Florida. By developing Safety Performance Functions (SPFs) for precise crash

frequency predictions, the research thoroughly examined the safety impacts on both Express lanes and General-Purpose Lanes. Utilizing both simplified and detailed SPF models, the study provided a detailed analysis of traffic safety dynamics post-conversion. It included a meticulous comparison of Crash Modification Factors (CMFs) derived from various methodologies, highlighting the effectiveness of Express lanes in improving roadway safety. The Comparison Group method compared treated sites with control sites that remained unchanged, ensuring similar characteristics. It calculated expected post-intervention crash frequencies by analyzing crash ratios before and after interventions at control sites versus actual post-intervention data. This rigorous approach is crucial for attributing changes in crash rates directly to the interventions, thus confirming the effectiveness of Express Lanes in reducing crashes and enhancing road safety. The concept map in Fig. 6 visualizes the interconnected factors contributing to the Safety of Express Lanes and Managed Lanes, as well as High-Occupancy Toll (HOT) lanes. The concept map highlights the interconnectedness of safety, operational performance, and traffic control in Express Lanes. It showcases how operational effects, risk management, and crash risk assessment are pivotal for safety enhancements.

Moreover, it underlines the significance of using technology and traffic simulations for optimizing flow and devising strategies for accident prevention and effective traffic management, aiming to elevate safety within express lanes. This map suggests a holistic approach to

Fig. 3. Managed lanes facility types and strategies (). Source: (NCHRP Research Report 835, 2016

Fig. 4. Benefits of using the express lanes ().

Source: Florida-DOT, 2023

safety on express lanes, considering various aspects of risk assessment, operational effects, and the integration of technology and human factors in managing and analyzing traffic to improve overall safety.

  • Analysis Methods
  • Empirical Bayes Method (Before-After study)

The Empirical Bayes (EB) method is renowned in safety studies for its predictive power, estimating crash frequencies at modified sites as if no interventions were made, using historical and comparative site data to enhance accuracy. The methodology relies on three fundamental assumptions: crash incidents at sites adhere to a Poisson distribution, average crash rates among sites conform to a Gamma distribution, and year-to-year fluctuations remain uniform across all reference sites. The EB method meticulously corrects for regression-to-the-mean biases, ensuring changes in crash rates post-intervention reflect true effects rather than statistical artifacts, offering an enhanced approach over traditional comparisons that may not fully account for evolving traffic and safety conditions. By utilizing a weighted amalgamation of data from both treated sites and their counterparts, as outlined in references (Hauer et al., 2002; Persaud et al., 1997), the EB approach delivers a

robust and accurate estimation of the expected impact of safety interventions on crash frequencies.

$$\widehat{E}_i = (\gamma_i \times y_i \times n) + (1 - \gamma_i)\eta_i \tag{1}$$

In this context, γi represents a weighting coefficient derived by considering both the negative binomial regression's over-dispersion parameter and the anticipated crash frequency during the pretreatment phase at the specific site, as outlined in Equation (2):

$$\gamma_i = \frac{1}{1 + k \times y_i \times n} \tag{2}$$

yi = The expected annual crash count for a specific type, derived from the Safety Performance Functions (SPF), reflects the data gathered from reference sites. Here, ηi denotes the actual crash count observed at the site undergoing treatment throughout the pre-intervention phase, with (n) representing the duration in years of this phase, and)k) symbolizing the over-dispersion parameter, which indicates the extent of variance in crash frequencies from the average. The reference site data, serving as the basis for the SPF output, is crucial in forecasting crash occurrences on a particular road segment.

a. Express Lanes in the Northea st of Florida b. Express Lanes in the Central Florida

c. Express Lanes in the West Central Florida d. Express Lanes in the Southeast Florida

To estimate crash frequencies on road sections, negative binomial regression models, as referenced in (Persaud et al., 1997; Abdel-Aty and Radwan, 2000), are typically employed, making them suitable for configuring the SPF based on historical crash data and both the geometric and traffic characteristics of the reference sites. An SPF generally adopts the following structure:

$$y_i = e^{(\beta_0 + \beta_1 x_1 + \dots + \beta_n x_n)} \tag{3}$$

In this scenario, the βi's = values represent regression coefficients, with x1 corresponding to the logarithmic transformations of the Annual Average Daily Traffic (AADT), and the subsequent x1 terms for (i > 1) encompassing various other traffic-related and geometric factors of significance.

The variability of the prediction in Equation 1 is quantified by its standard deviation (̂σi), as delineated in Equation (4). This standard deviation is critical for understanding the precision of the crash number estimates during the pre-intervention phase.

$$\widehat{\sigma}_i = \sqrt{(1 - \gamma_i) \times \widehat{E}_i} \tag{4}$$

To derive the projected crash counts at the site following the implementation of safety measures, it's essential to adjust the initial estimates from Equation 1. These adjustments account for variations in traffic volume and the distinct characteristics of the pre- and post-intervention periods, as specified in Hauer et al. (Hauer et al., 2002). The coefficients for these adjustments are detailed as follows:

$$Adjustment for AADT(\rho{AADT}): \rho{AADT} = \frac{AADT_{after}^{a1}}{AADT{before}^{a_1}}$$ (5)

In this context, the term represents the (AADT) observed at the site undergoing treatment after the implementation of safety measures, while indicates the AADT recorded before these measures were introduced. Additionally, (α1) is identified as the regression coefficient for

Fig. 6. Concept map of safety of express lanes.

AADT within the Safety Performance Functions (SPF).

An adjustment factor, denoted as (ρtime), is applied to account for the variability between the pre- and post-intervention periods:

$$\rho_{time} = \frac{y_a}{y_b} \tag{6}$$

In this case, (ya) signifies the duration, measured in years, following the implementation of safety measures, while (yb) represents the number of years preceding these interventions.

The adjusted estimate of crash occurrences at the site of intervention for the period after the introduction of safety measures (̂πi) is calculated by taking into account variations in traffic volumes and the disparity in time frames between before and after the intervention:

$$\widehat{\pi}_i = \widehat{E}i \times \rho{AADT} \times \rho_{time} \tag{7}$$

The effectiveness index (̂θi)for the implemented safety measure is determined through the following formula:

$$\widehat{\theta}_i = \frac{\widehat{\lambda}_i/\widehat{\pi}_i}{1 + (\widehat{\sigma}_i^2/\widehat{\pi}_i^2)} \tag{8}$$

Where, ̂λ i= The actual crash count recorded at the site postintervention. The crash type-specific reduction rate (̂τi) at each location (i) is calculated as follows:

$$\widehat{\tau}_i = (1 - \widehat{\theta}_i) \times 100\% \tag{9}$$

The average safety effectiveness (̂θ), or Crash Reduction Factor (CRF), across all evaluated sites is derived as follows (Hauer et al., 2002):

$$\widehat{\theta} = \frac{\sum{i=1}^{m} \widehat{\lambda}{i} / \sum{i=1}^{m} \widehat{\pi}{i}}{1 + \left( var(\sum{i=1}^{m} \widehat{\pi}{i}) / (\sum{i=1}^{m} \widehat{\pi}{i})^{2} \right)}$$ (10)

Where, m = total number of treated sites,

$$var(\sum_{i=1}^{k} \widehat{\pi}i) = \sum{i=1}^{k} \rho{AADT}^2 \times \rho{time}^2 \times var(\widehat{E}_i)$$ (11)

The standard deviation (σ) for the overall treatment effectiveness is determined by analyzing the variance between estimated and observed crashes, as described in Equation (12).

$$\widehat{\sigma} = \sqrt{\frac{\theta^{2} \left[ \left( var(\sum{i=1}^{k} \widehat{\pi}{i}) / (\sum{i=1}^{k} \widehat{\pi}{i})^{2} \right) + \left( var(\sum{i=1}^{k} \widehat{\lambda}{i}) / (\sum{i=1}^{k} \widehat{\lambda}{i})^{2} \right) \right]^{2}}$$

$$\left[ 1 + \left( var(\sum{i=1}^{k} \widehat{\pi}{i}) / (\sum{i=1}^{k} \widehat{\pi}{i})^{2} \right) \right]^{2}}$$ (12)

Where,

$$var(\sum_{i=1}^{k} \hat{\lambda}i) = \sum{i=1}^{k} \lambda_i$$ (13)

Equation 1 served as the basis for predicting post-intervention crash numbers at the treatment sites. These projections were then juxtaposed with the actual crash counts recorded post-intervention to ascertain the treatment's impact in terms of crash number reduction percentage.

• Safety Performance Functions

Safety Performance Functions (SPFs), leveraging Negative Binomial regression, predict crash frequencies by examining traffic and road characteristics. This study formulated both basic SPFs, focusing solely on traffic volume, and detailed versions, integrating additional variables for a comprehensive safety analysis. Specific SPFs were designed for the 1) complete roadway system, 2) Express Lanes only, and 3) solely General-Purpose Lanes, facilitating crash predictions as if Express lanes had not been introduced, allowing for a nuanced understanding of postimplementation safety impact. The NB regression models, tailored to various crash types and severity levels, were formulated to facilitate this evaluation, as delineated by the subsequent equation (Abdel-Aty and Radwan, 2000):

$$N_{\text{predicted}} = \exp(\beta_0 + \beta_1 * \ln(AADT) + \beta_2 S + \beta_3 L)$$ (14)

Where,

N predicted = Expected crash frequency without treatment,

βi= Coefficients of the significant variables,

AADT=Annual Average Daily Traffic,

S=Posted Speed.

L=length of Segment.

• Cross Sectional Method

The cross-sectional analysis was utilized to assess the impact of Express lanes on traffic safety along Florida's Interstate 95, including sections converted before 2012 and other locations completed around 2018. This method ensured consistency in traffic characteristics postCOVID-19, enhancing the accuracy of infrastructure studies. Some I-95 E-Lane segments predated the 2012 overhaul of FDOT's reporting system, necessitating the Cross-Sectional method for estimating Crash Modification Factors (CMFs) based on post-implementation data.

To calculate the CMFs pertinent to a particular roadway environment and its treatment, a series of SPFs were established using the Negative Binomial (NB) distribution. These SPFs quantify crash occurrences by examining a variety of explanatory variables, such as the implementation of Express lanes, Annual Average Daily Traffic (AADT), and the lengths of the roadway segments, thereby providing a comprehensive framework for safety evaluation:

$$F_i = \exp(\alpha + \beta_1 *E - Lanes_i + \beta_2 *Length_i) *AADT_i^{\beta_3}$$ (15)

where,

Fi = crash number on a freeway segment i;

Express lanes i = indicates the presence of Express lanes on road segment i (=1 if the Express lanes of a segment i is implemented, = 0 if the Express lanes of a segment i is not implemented);

Lengthi = length of a road segment i (mi);

AADTi = represents the average annual daily traffic on road segment i (veh/day);

α = constant.

β = variables' coefficients.

CMFs form cross sectional method are then calculated using the following equation:

$$CMF = \exp(\beta_1 * (1 - 0)) = \exp(\beta_1)$$ (16)

This model is versatile, allowing for predictions concerning overall crash frequencies or the frequencies of specific types of crashes and their severities. The calculation of the standard error (SE) for the CMF is conducted using the formula presented in reference (Bahar, 2010):

$$SE = \frac{\exp(\betak^*(x{kt} - x{kb}) + SE{\beta_k}) - \exp(\betak^*(x{kt} - x{kb}) - SE{\beta_k})}{2}$$ (17)

where,

SE=CMF' standard error.

SEβk = coefficient of the variable' standard error (Express lanes).

3. Data description

This investigation into Florida's express and high occupancy toll lanes thoroughly examined both the development of new Express Lanes and the transformation of High-Occupancy Vehicle (HOV) lanes into Express Lanes across 55 miles in various state locations. It analyzed road segments based on Annual Average Daily Traffic (AADT), speed, and geometric data for an in-depth safety evaluation, reviewing and identifying about 132 reference segments for comparative analysis. Crash data from July 2012 to December 2022 was used, excluding the sixmonth construction phase surrounding their implementation, to understand the safety impact of Express Lanes using data from the Florida Department of Transportation. It is worth mentioning that the Florida Department of Transportation (FDOT) instituted new criteria for documenting PDO crashes from July 2012, which significantly affected the compilation of both total and PDO crash records.

The study extended its scope to encompass all of Florida's Express lanes and HOT lanes implemented until 2018, (to ensure the availability of after Express lanes implementation data) utilizing diverse data sources including the Florida Department of Transportation database, such as the Roadway Characteristics Inventory (RCI) system and the FDOT Project Information database, with additional verification through Google Earth and Florida Turnpike reports. By examining an extensive range of data over an extended period, the research provided fresh insights into the safety effects and efficacy of Express lanes across the state, indicating significant trends and outcomes crucial for future transportation planning. It highlighted the pivotal role of Express Lanes

in improving road safety and managing traffic flow, with Fig. 7 from the FDOT 2018 report illustrating the progress and reach of these projects across the state. Furthermore, given the COVID-19 impact on traffic patterns, the analysis employed the Cross-Sectional method for locations completed around 2018 to maintain consistency in traffic characteristics in the after period. This meticulous approach emphasizes the necessity of adapting research methodologies to external factors that may distort data, thereby ensuring the accuracy of conclusions regarding infrastructure studies.

Table 2, according to the 2023 FDOT Roadway Characteristics Inventory (RCI) report, outlines the latest Express Lanes' locations across Florida, highlighting the state's strategic enhancement and spatial allocation of these lanes. The 2022 FDOT report further underscores the operational span of these lanes, revealing an extension of 55 miles dedicated to facilitating smoother traffic flow and improved transportation efficiency across the state.

4. Analysis and results

The analysis was divided into two main segments. Firstly, it evaluated the safety status of sections featuring Express Lanes by employing both the Control Group (CG) and Empirical Bayes (EB) methods, using data encompassing various crash types for all locations with available before-after E-Lane implementation data. Subsequently, the study examined the safety influence on these E-Lane sections by employing the Cross-Sectional (C-S) approach for analyzing I-95 crash data, including sections converted into Express lanes before 2012 and those completed around 2018. The investigation was organized into three distinct components: 1) evaluating the entire lanes, 2) analyzing exclusively the Express lanes, and 3) examining the General-Purpose Lanes (GPL). The analysis employed the Before-After method for evaluating crash data, excluding the 95-Express segments that transitioned to Express lanes before the 2012 update in reporting practices, as well as those completed around 2018. This exclusion was due to the influence of the Covid-19 pandemic on the after data, rendering the before-after methods unusable for these locations. Conversely, the Cross-sectional method served to analyze safety outcomes for these particular segments, integrating both Express lanes and General-Purpose lanes. This methodological distinction ensured a comprehensive safety evaluation across different roadway sections, accounting for changes in data documentation standards.

• Comparison Group (Before and After study)

This approach utilized data from all treated sites converted from HOV-Lanes to Express lanes in Florida, as well as newly implemented Express lanes before 2018. These treated sites were then compared with

Fig. 7. Status of express lanes implementation in Florida (Source FDOT, 2018 express lanes report).

Table 2 FDOT express lanes locations ().

Code Toll Facility Name Ownership
388 I-275 (MANAGED LANES) FDOT
389 I-295 EXPRESS FDOT
386 I-295 EXPRESS FDOT
339 I-4/SELMON CONNECTOR FDOT (Turnpike)
340 I-595 EXPRESS FDOT
387 I-75 EXPRESS FDOT
302 I-95 HOT LANES (NORTH/SOUTH EXPY) FDOT
393 PALMETTO EXPRESS (SR-826) FDOT

Source: (Florida-DOT-RCI - 2023 report)

an equal number of untreated sites with similar characteristics for analysis. Based on FDOT 2022 report, Florida witnessed a series of transformations, converting HOV lanes into E- lanes across key freeways. Crash modification factors were calculated for specific sites as well as on an aggregate level, utilizing crash history data collected from a series of HOV lanes Florida Express lanes' locations. This analysis employed crash records from both before and after the implementation of the Express Lanes.

• Empirical Bayes (Before and After study)

For the analysis, data were collected from 132 untreated reference segments (HOV lanes), using crash records from three years prior to any intervention or the available data to develop Safety Performance Functions (SPFs) for these lanes. These models predict future crashes within the HOV lanes or project outcomes for the Express lanes system, assuming no changes were made.

• Comparison Group and Empirical Bayes Results

The study employed an Observational Before and After approach, complemented by Empirical Bayes techniques. For certain segments featuring HOV-Lanes in Florida, both 'simple' and 'Full' SPFs were developed.

• Developing Safety Performance Functions for all lanes

The results in Table 3 considering the introduction of Express lanes, underscores traffic volume and speed's significant impact on crash rates. Log (AADT) analysis revealed a direct correlation with crash frequencies across all categories, suggesting higher traffic volume increases crash likelihood. Particularly, these results illustrated how increased posted speed correspond with augmented crash occurrences, challenging the conventional belief that higher posted speed inherently led to more severe crashes. The analysis extended to specific crash types like rear-end, bad weather (BWC), and property damage only (PDO) incidents. Intercept values indicated baseline crash rates, while dispersion metrics and AIC scores gauged variation and model accuracy, respectively. Notably, rear-end crashes were significantly influenced by traffic volume. In contrast, BWC and PDO focused exclusively on volume's effects, excluding posted speed considerations, demonstrating nuanced impact on safety outcomes based on traffic dynamics.

This detailed evaluation of all lanes provides crucial insights into the dynamics influencing crash frequency, underscoring the importance of managing traffic volume and posted speed to enhance roadway safety.

• Crash Modification Factors for all lanes

Table 4 presents Crash Modification Factors (CMFs) for all lanes (both Express and General Purpose (GP) lanes together), comparing the effects of interventions using two methodologies: Before-After methods with Comparison Group (CG) and Empirical Bayes (EB). It categorizes crashes by type and severity, offering CMFs alongside their Standard Errors (SE).

For "All types" of crashes, CMFs are close to 1 (0.99 for CG and 0.96 for EB), suggesting minimal change in crash frequency postintervention. Lane Change Crashes (LCC) show a notable reduction in CMFs, particularly with EB analysis (0.91), indicating a decrease in these crash types. LC crashes also exhibit a reduction in CMFs, indicating some positive safety impact from the interventions. Rear-End crashes, interestingly, show a slight increase in likelihood according to the EB method (CMF of 1.12), possibly indicating areas needing further safety enhancements. Bad Weather Crashes (BWC) demonstrate reduced CMFs, suggesting an overall improvement in safety for these categories. Property Damage Only (PDO) crashes, uniquely, show an increase in CMFs (1.10 for CG and 1.13 for EB), hinting at increased occurrence post-intervention. This analysis provides insightful data on the effectiveness of safety interventions across different crash types within the whole freeway segment (both Express and GP lanes), highlighting areas of success and opportunities for further safety improvements.

Considering the standard errors and confidence levels, it can be concluded that most Crash Modification Factors (CMFs) will cross 1, indicating no significant difference in the overall safety of implementing express lanes. This suggests that while certain crash types may show reductions or increases in frequency post-intervention, the overall safety impact of express lanes remains neutral.

Table 4 CMFs for the whole segment (both Express and GP lanes).

Comparison Group Empirical Bayes
Crash Type Severity Levels CMF SE CMF SE
All types All 0.99 0.09 0.96 0.10
LCC1 All 0.95 0.06 0.91 0.09
LC2 All 0.90 0.07 0.91 0.06
Rear End All 0.98 0.11 1.12 0.08
BWC3 All 0.94 0.08 0.91 0.06
PDO4 Non 1.10 0.09 1.13 0.11

1 LCC=Lane changes crashes. 2 LC=Lost control. 3 BWC=Bad weather crashes. 4 PDO=Property Damage only.

Table 3 Safety Performance Functions for the whole segment.

Crash type
Level of injury		 Intercept Log (AADT) Posted speed Dispersion AIC
Est P-Value Est P-Value Est P-Value Est
All types All − 8.3355 0.0611 0.8775 0.0411 0.0455 0.0021 0.7850 1373
LCC1 All − 5.8522 0.0065 0.8121 0.0072 0.0324 <0.0001 0.1545 1133
LC2 All − 4.1142 0.0861 0.6275 0.0612 0.0495 0.0515 0.7113 1175
Rear End All − 3.5425 0.1012 0.5223 0.0420 − − 0.6933 1288
BWC3 All − 5.1142 0.0861 0.6275 0.0612 − − 0.7113 1175
PDO4 Non − 11.4284 0.0451 0.906 0.0501 − − 0.4552 977

1 LCC=Lane changes crashes. 2 LC=Lost control. 3 BWC Bad weather crashes. 4 PDO Property Damage only.

• Developing Safety Performance Functions for Express-Lanes only

Table 5 shows Safety Performance Functions for Express-Lanes focus on different crash types, showing both positive and negative correlations between crash occurrences and factors like traffic volume (Log (AADT)) and posted speed.

Higher Log (AADT) and posted speed are linked to increased crash probabilities, suggesting that traffic density and higher speeds elevate risks. Significant correlations between Log (AADT) and Lane Change Crashes (LCC) imply that traffic volume greatly influences crash likelihood, especially for LCC where speed also plays a crucial role. The Rear-End, Bad Weather Crashes (BWC) and PDO relate positively to traffic volume. This comprehensive analysis underlines the importance of managing traffic flow and speed to enhance road safety in Express-Lane areas, providing valuable insights for traffic safety strategies aimed at minimizing crashes and improving overall road safety within these specific lanes.

• Crash Modification Factors for the Express Lanes only

Table 6 presents the crash modification factors specific to Express lanes, using both Before-After methods with Comparison Group (CG) and Empirical Bayes (EB). methodologies to evaluate the impact of Express lanes on various types of crashes and their severity levels. Across all crash types, the CMFs are less than "1", indicating a reduction in crash occurrences post-E-Lane implementation. For instance, "All types" of crashes have CMFs of 0.73 (CG) and 0.71 (EB), suggesting a significant decrease in overall crash frequency with E-Lane usage. Lane Change Crashes (LCC) show even more substantial reductions, with CMFs of 0.63 (CG) and 0.59 (EB), reflecting a notable safety improvement in lane-changing behaviors within Express lanes.

The LC category exhibits CMFs of 0.71 (CG) and 0.74 (EB), consistent with the trend of reduced crash likelihood. Rear-End and Bad Weather Crashes (BWC) both demonstrate significant decreases in crashes, with Rear-End crashes showing CMFs of 0.68 (CG) and 0.64 (EB) and BWC3 presenting even lower CMFs of 0.56 (CG) and 0.60 (EB), highlighting enhanced safety during adverse weather conditions in Express lanes. Lastly, Property Damage Only (PDO) crashes have CMFs of 0.67 (CG) and 0.65 (EB), indicating a reduction in minor crashes that primarily result in property damage. The consistent decrease across all categories underscores the effectiveness of Express lanes in enhancing roadway safety by mitigating crash occurrences.

• Developing Safety Performance Functions for General Purpose Lanes only

Table 7 presents Safety Performance Functions specifically for General Purpose Lanes, focusing on various crash types. It indicates a strong correlation between crash occurrence and traffic volume, denoted by the logarithm of Annual Average Daily Traffic (Log (AADT)), across all crash categories. High Log (AADT) values consistently predict increased crash rates. These results focusing on Log (AADT) impact and dispersion

Table 6 CMFs for the express lanes only.

Comparison group Empirical bayes
Crash type Severity levels CMF SE CMF SE
All types All 0.73 0.11 0.71 0.12
LCC1 All 0.63 0.08 0.59 0.07
LC2 All 0.71 0.07 0.74 0.07
Rear End All 0.68 0.13 0.64 0.10
BWC3 All 0.56 0.08 0.60 0.09
PDO4 Non 0.67 0.09 0.65 0.11

1 LCC=Lane changes crashes. 2 LC=Lost control. 3 BWC Bad weather crashes. 4 PDO Property Damage only.

measures, which illustrate variability in data.

• Crash Modification Factors for the General-Purpose Lanes only

Table 8 displays Crash Modification Factors (CMFs) exclusively for General Purpose Lanes, comparing before-and-after scenarios with both Control Group (CG) and Empirical Bayes (EB) methods.

It lists CMFs across various crash types, indicating changes in crash likelihood post-intervention. Notably, CMFs above 1 suggest increased crash likelihood, particularly for all types, LCC, Rear End, and PDO crashes, indicating an elevation in crash occurrences. Conversely, LC and BWC crashes show CMFs below 1 with the EB method, suggesting a slight reduction in crash rates.

• Findings Using the Cross Sectional Method: Entire Segment (Including Express lanes and GPL)

This analysis assessed the impact of Express lanes on traffic safety along multiple locations such as Florida's Interstate 95 (including sections converted before 2012) and other locations completed around 2018. By ensuring consistency in traffic characteristics post-COVID-19, this method bolstered the accuracy of infrastructure studies. Given that some I-95 E-Lane segments predated the 2012 overhaul of FDOT's reporting system, the Cross-Sectional method was vital for estimating Crash Modification Factors (CMFs) using post-implementation data and examining crash data over years post Express Lane implementation (2013–2015) and (2019–2022) to develop Safety Performance Functions (SPFs) and derive Crash Modification Factors (CMFs). This approach filled data gaps for segments converted to Express lanes before 2012′s FDOT reporting system update and post-COVID-19, ensuring a comprehensive analysis of E-Lane safety implications, shedding light on E-Lane efficacy in enhancing roadway safety.

Table 9 shows Cross-Sectional method's CMFs for all lanes (both Express and General-Purpose lanes) assess the impact on various crash types and all severity levels. Notably, for all crash types, the CMF stands at 1.02, indicating a slight increase in crash frequency post-intervention.

Lane Change Crashes (LCC) and Lost Control (LC) exhibit CMFs below 1 (0.91 and 0.89 respectively), suggesting a possible reduction in

Table 5 SPFs for Express-Lanes only.

SPFs for Express-Lanes only
Crash Type
Injury levels		 Intercept Log (AADT) Posted speed Dispersion AIC
Est P-Value Est P-Value Est P-Value Estimate
All types All − 3.4211 0.0451 0.8521 0.0312 0.0511 <0.0622 0.2448 1315
LCC1 All − 2.7852 0.0465 0.3388 <0.0001 0.0721 <0.0424 0.4256 857
LC2 All − 4.2143 0.0431 0.5985 0.0321 0.0654 0.0725 0.8254 1254
Rear End All − 5.2458 0.0463 0.4156 0.0406 − − 0.7531 1168
BWC3 All − 6.1142 0.0364 0.5315 0.0512 − − 0.6551 1234
PDO4 Non − 8.3316 0.0402 1.3254 0.0102 − − 0.4132 1055

1 LCC=Lane changes crashes. 2 LC=Lost control. 3 BWC Bad weather crashes. 4 PDO Property Damage only.

Table 7 SPFs for general purpose lanes (Free-Lanes) only.

SPFs for General Purpose Lanes only
Crash Type Injury levels Intercept Log (AADT) Dispersion AIC
Est P-Value Est P-Value Estimate
All types All − 9.1425 <0.0001 1.2656 <0.0001 0.6298 1254
LCC1 All − 7.4658 0.0406 2.2547 0.0305 2.0332 746
LC2 All − 11.1492 0.0203 1.4685 <0.0001 3.3125 1211
Rear End All − 10.7325 0.0021 1.9529 <0.0001 2.7895 698
BWC3 All − 9.4526 0.0506 0.4358 0.0309 0.8641 810
PDO4 Non − 12.2569 0.0215 0.4582 0.0485 0.3221 756

1 LCC=Lane changes crashes. 2 LC=Lost control. 3 BWC Bad weather crashes. 4 PDO Property Damage only.

Table 8 Crash modification factors for general purpose lanes only.

Before-after with CG Before-after with EB
Crash Type Severity Levels CMF SE CMF SE
All types All 1.08 0.08 1.10 0.11
LCC1 All 1.19 0.10 1.21 0.06
LC2 All 0.90 0.09 0.89 0.06
Rear End All 1.22 0.11 1.23 0.08
BWC3 All 0.93 0.06 0.89 0.04
PDO4 All 1.22 0.09 1.26 0.07

1 LCC=Lane changes crashes. 2 LC=Lost control. 3 BWC Bad weather crashes. 4 PDO Property Damage only.

Table 9 The CMFs results from the C-S methods for Total and PDO crash data only.

| Cross-Sectional method | | | | | | | |------------------------|-----------------|------|------|--|--|--| | Crash Type | Severity Levels | CMF | SE | | | | | All types | All | 1.02 | 0.07 | | | | | LCC1 | All | 0.91 | 0.06 | | | | | LC2 | All | 0.89 | 0.07 | | | | | Rear End | All | 1.15 | 0.11 | | | | | BWC3 | All | 0.94 | 0.08 | | | | | PDO4 | Non | 1.14 | 0.07 | | | |

1 LCC=Lane changes crashes. 2 LC=Lost control. 3 BWC Bad weather crashes. 4 PDO Property Damage only.

these specific crash types at the negative side of the standard error. Conversely, rear-end crashes show an increased likelihood with a CMF of 1.15. Bad Weather Crashes (BWC) and Property Damage Only (PDO) also display a slight variance in crash probabilities with CMFs of 0.94 and 1.14 respectively, illustrating the nuanced impact of roadway interventions on safety across different crash categories.

4.1. Summary

In the extensive analysis, the safety impact of Express Lanes and General-Purpose Lanes in Florida were thoroughly evaluated using Control Group, Empirical Bayes, and Cross-Sectional methods. This study meticulously analyzed crash data, focusing on the transition of HOV lanes to Express Lanes. Results revealed that implementing Express Lanes did not compromise overall road safety. Safety improvements were noted within Express Lanes, with significant reductions in crash rates, contrasting with a slight increase in General Purpose Lane crashes. This underscores the potential of Express Lanes to enhance traffic flow and safety, suggesting the need for ongoing research to further understand their impact across various road sections when all projects are completed and the roads/traffic become stable.

5. Conclusion

The strategy of express lanes implementation aims to alleviate traffic congestion and finance transportation infrastructure. These lanes provide a faster travel option for drivers, bypassing congested General-Purpose Lanes (GPL). However, the safety impact of Express Lanes has been minimally explored. The Crash Modification Factors Clearinghouse, a key resource, lists only two relevant case studies, including one by our team. This study examines the safety effects of converting HOV lanes to Express Lanes or introducing new ones, and for the first time, assesses their impact on both Express and GPL individually.

The results of our comprehensive analysis demonstrate that converting HOV lanes to Express Lanes or introducing new ones does not compromise overall freeway safety. In fact, safety within Express lanes improves, as evidenced by a decrease in crash occurrences and Crash Modification Factors (CMFs) for Express lanes, which are below "1″ across all crash categories. This underscores the effectiveness of Express lanes in enhancing roadway safety. Conversely, incidents in General-Purpose Lanes (GPL) have increased, indicating a shift of crashes to these lanes, thus making Express lanes relatively safer. This highlights the importance of continued research into the safety benefits of express lanes and calls for further studies to refine traffic management strategies, aiming at enhancing travel efficiency while ensuring traffic safety, especially for the GP-Lanes, amid the expansion of express lanes. This study could serve as a crucial reference for assessing the safety of express lanes, significantly influencing the development of traffic management strategies. By focusing on enhancing travel times and ensuring safety on freeways, it aims to set a foundation for future improvements in traffic safety and efficiency.

CRediT authorship contribution statement

Muamer Abuzwidah: Writing – review & editing, Writing – original draft, Visualization, Supervision, Software, Resources, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Mohamed Abdel-Aty: Validation, Project administration, Data curation.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

The authors do not have permission to share data.

Acknowledgments

The authors extend their gratitude to the Florida Department of Transportation (FDOT) for providing crucial data for this study and their commitment to improving traffic safety. It's important to note that the insights and conclusions presented are exclusively those of the authors.

References