Assessment of sidewalk/bicycle-lane gaps with safety and developing statewide pedestrian crash rates.

Author(s)
Radwan, E. Abou-Senna, H. Mohamed, A. Navarro, A. Minaei, N. Wu, J. & Gonzalez, L.
Year
Abstract

Pedestrian and bicycle safety have become more common for governmental agencies to address and prioritize for strategic planning and construction. Public safety is a focal point for our decision makers to emphasize during the planning, design, and construction phases of projects. The prioritization of these projects provides more challenges for the decision makers to identify. The Florida Department of Transportation (FDOT) — District Five has utilized Geographic Information Systems (GIS) to identify sidewalk gaps within its nine-county region. While the gaps have been identified, there is a need to understand the relationship between the sidewalk gaps or bicycle-lane gaps and the safety along the agency maintained facilities. In addition, identifying the relationship between safety and the gaps provides for the ability to statistically link the gaps to prioritizing for construction based on safety and available funding. On the other hand, with the increased emphasis on the multimodal transportation, to date, there are no clear or uniform standards for a method to measure pedestrian incidents against a statewide average. However, the FDOT has developed vehicular crash rates that resulted in the hypothesis that geometrics and traffic characteristics influence vehicle incidents and corresponding crash rates. Therefore, research is needed to identify and evaluate crash rates for pedestrians that would result in the ability to identify locations with pedestrian crash rates higher than the statewide average. It is important to address the pedestrian-vehicular conflict as the State of Florida was reported in the “Dangerous by Design” report as having the highest four pedestrian incident locations in the country. Therefore, this research has two objectives: (1) Developing a safety prioritization tool that would assist governmental agencies in the prioritization of sidewalk gaps and/or bicycle-lane gaps based on a balanced approach between safety needs, socioeconomic evaluation, operational constraints, and fiscal assessment. (2) The development of a statewide average for pedestrian crash rates that would give transportation planners and engineers a barometer indicating how the regions (or FDOT Districts) compare against other areas, thus evaluating locations that are operating above statewide averages and therefore emphasizing the need for mitigations to be implemented. Pedestrian sidewalks and bicycle lanes in Florida are not continuous, and there is a concern among planners and engineers in the FDOT that these gaps constitute discontinuity of flow and are potentially posing threats to pedestrian and bicycle safety. Before these agencies attempt to develop a prioritization program to decide on which gaps need to be addressed, it was logical to carry out an analysis that investigates the correlation between safety and sidewalk/bicycle-lane gaps. The research team assembled a wide array of Geographic Information Systems (GIS) layers associated with the sidewalk/bicycle lane gaps; Roadway Characteristics Inventory (RCI), and the safety data from several agencies within District Five, including FDOT, MetroPlan Orlando, Orange, Seminole, Volusia, and Osceola Counties. It was found that every jurisdiction and agency uses their own independent GIS data that are not homogeneous and do not easily interact. This made it difficult to locate and analyze crash data along with roadway location and feature data using spatial analysis. As a result, the University of Central Florida (UCF) research team utilized the Florida Unified Basemap Repository (UBR). The UBR is a unified approach to GIS data management to develop a standard, comprehensive transportation network that could be used throughout the state, shared across jurisdictional boundaries through multi-agency involvement and coordination. Another source related to pedestrian and bicycle safety data was utilized: the Signal Four Analytics database. A 0.25-mile buffer radius, which is within walking distance from the intersections, was selected in order to differentiate between intersection-related pedestrian crashes and roadway segmentrelated crashes, and a spatial join command between the crashes and the intersection’s buffer was performed. Furthermore, intersections were separated based on the crash type (pedestrian or bicycle crash). The pedestrian dataset included minimum and maximum number of crashes within the buffer area at intersections as well as the overall average number of crashes per intersection. The sidewalk and bike lane gap layers were overlaid on the crash layer and a 50- foot buffer was created to identify the crashes along the roadway segments with no sidewalk or bike lane. Due to the fact that area population contributes to pedestrian activity within an area, it was imperative to include the population as a factor representing the intensity of pedestrian activity. The crash data was then overlaid on the population layer and a 0.5-mile radius was determined as a reasonable walking distance surrounding the crash location for the population parameters to be considered within. Based on the analysis and modeling results, a negative binomial (NB) regression model was developed. It was found that the absence of sidewalk along roadway segments is one of the main factors that has a significant impact on the expected number of pedestrian crashes at a specific location. Other factors included average annual daily traffic (AADT) volumes, roadway category (ROADCAT), specifically along urban two-way divided arterials with four to six lanes, and the average population within a 0.5-mile radius surrounding the crash location. The analysis also showed that the expectation of a pedestrian crash along roadways with no sidewalk is three times greater than the expectation of a crash with the presence of a sidewalk. The tool developed to prioritize gaps was crucial to the remaining research tasks. The developed tool takes into account the above-mentioned parameters as well as other pedestrian-related activity variables and proximity to generators using land use, income, and auto ownership data. The prioritization method was based on a multi-criterion ordinal ranking of the parameters of five main modules, using a scoring system that combines all criteria weights then aggregates them into a single indicator. The five main modules comprise roadway and traffic data, socioeconomic data, land use data, transit, and crash data. The need for roadway segment safety improvement was ranked according to its roadway pedestrian safety indicator (RPSI) threshold and categorized into five categories. The Sidewalk/Bike-Lane Gaps Safety Prioritization Tool (SBLPT) has the capability to generate sidewalk/bike-lane gaps map that can be viewed in Google Earth ®. The generated map is color-coded based on five prioritization ranks, where red, purple, orange, yellow, and green colors indicate urgent, high, medium, low, and no safety concerns priorities, respectively. On the other hand, the pedestrian crash rate methodology required the identification of an accurate yet practical exposure measure. Pedestrian exposure is one of the crucial factors needed in the analysis of pedestrian safety. Although there are numerous studies that attempted to identify pedestrian exposure factors, to date, there is no commonly accepted methodology to measure pedestrian exposure. In order to arrive at the correct exposure measure, specific data has to be available or collected, which is often a challenge, either due to its unavailability or the cost involved in collecting it. Researchers often use population density as a substitute for pedestrian exposure in pedestrian crash analyses because of its availability or the low cost to obtain it. However, it is not considered an accurate measure because it doesn’t account for the probability of pedestrians appearing on the road. The proposed approach focuses on the level of pedestrian activity and the potential conflict between pedestrians and motor vehicles expressed as the distance of walking while at risk of being involved in a motor vehicle accident. The majority of the pedestrian-vehicle crashes occur while crossing the street whereas an insignificant percentage of crashes occur while walking along the street. Therefore, it is believed that the pedestrian miles crossed (PMC) parameter is representative factor of the risks associated with pedestrian exposure. Furthermore, since the exposure measure should directly reflect the amount of walking in areas shared with vehicles, the vehicular traffic volume is another significant factor that should be included in the equation. Based on that, the three main significant parameters that were used in the exposure measure calculations were: pedestrian traffic, distance crossed and vehicular traffic. The proposed methodology is considered detailed and practical, and it provides a broad depiction of the main factors that directly contribute to pedestrian crashes. The pedestrian crash rate methodology involved two types of analyses: roadways and intersections. Statewide RCI data was collected from the GIS layers and classified according to their functional classification, area type, and number of lanes, as well as their average annual daily traffic (AADT) and total lengths. Intersection data was classified according to several parameters, including their total entering traffic, total crossing distance, total daily pedestrians, and type of control. Furthermore, the total statewide number of crashes over the five-year period (June 2009 to May 2014) was identified from the Signal Four Analytics database for all the roadways and intersections. In addition to the intersection GIS data, sample intersection data was collected based on a pilot study within the nine counties of FDOT District Five. The purpose of the pilot study was to identify the critical pedestrian safety locations within the district and correlate this with safety locations based on statewide averages. The main parameters used in calculating pedestrian the crash rates along the different roadway categories, which combine the functional classification, number of lanes, and area type were the AADT and the total length of the roadway category. Conversely, the main parameters used for computing the pedestrian crash rates for the different intersection classifications were the daily pedestrian volumes, distance crossed, and the AADT in addition to the number of pedestrian crashes either along the studied roadways or at intersections. Although the socioeconomic conditions of the geographic area may provide higher accuracy for estimating pedestrian volume than the population, there was not a specific data collection source that could produce data at these intersection locations. It should be noted that the analysis also could not identify a correlation between the number of crashes and the average daily pedestrians at the intersections due to the high variability of the data. The pilot studies conducted for the roadways and intersections revealed several critical safety locations within District Five when compared to the developed statewide average rates. This conclusion requires further investigation to identify main causes and emphasize the type of mitigations that can be implemented. It is recommended that an annual statewide pedestrian count program be initiated for the intersections and roadways, which would increase the sample size and assist in validating the assumptions used in this study. The purpose and duration of data collection are essential pieces of information for determining the appropriate technology. Active or passive infrared sensors are common practice for counting pedestrians, and they may also be used to collect combined counts of bicyclists and pedestrians. (Author/publisher)

Publication

Library number
20160829 ST [electronic version only]
Source

Tampa, FL, University of South Florida, Department of Civil, Environmental & Construction Engineering, 2016, XVIII + 184 p. + 6 app., 75 ref.; FDOT BDV24-977-07

Our collection

This publication is one of our other publications, and part of our extensive collection of road safety literature, that also includes the SWOV publications.