Integrating Trip and Roadway Characteristics in Managing Safety at Traffic Analysis Zones.

20 vitally important in transportation planning and have been emphasized in several macro-level studies during the past few years. Macro-level analyses of crash prediction models have been attempted for census block groups; traffic analysis zones (TAZs); census tracts; and counties considering various demographic, socioeconomic, road, and travel characteristics (1-6). The study in this paper was based on 1,349 TAZs of four counties in the state of Florida. This study aims to propose several safety management alternatives for the TAZs by investigating the association between crash frequencies and various types of trip productions and attractions in combination with the road characteristics of the TAZs. In addition, at the transportation planning stage if TAZs are to be defined on the basis of travel demand models, the results of this study point to different approaches when predicting crashes at the zonal level. It is speculated that using trip types to predict crash frequencies will help in understanding the safety consequences at an early stage of transportation planning. The study examined four response variables: total crashes, severe (fatal and severe injury) crashes, peak-hour crashes, and pedestrianand bicycle-related crashes per TAZ. Peak periods were estimated on the basis of temporal variations of the aggregated hourly crash frequencies and their association with different trip-and road-related covariates. Also, as both state and nonstate road characteristics within a TAZ were considered in the study, pedestrian-and bicycle-related crashes were of special interest as most of these categories of crashes occur on moderate-to low-speed roadways and streets. LITERATURE REVIEW Macro-level crashes have been investigated at various spatial aggregations. The crash models developed in these studies have incorporated different categories of variables in predicting crash occurrences. Amoros and Laumon (1) compared traffic safety in several counties in France taking different road types and socioeconomic characteristics into account. Aguero-Valverde and Jovanis (2) investigated crash risk for Pennsylvania counties with respect to sociodemographics, weather conditions, transportation infrastructure, and amount of travel. Noland and Oh (3) examined the association of various road network infrastructure and some demographic and socioeconomic variables with crashes in the counties of Illinois. Amoros and Laumon (1) found significant interaction between county and road type. Other positively associated road-related significant factors in the above-mentioned studies included road mileage and road density (2) and number of lanes (3). Among demographic variables, AgueroValverde and Jovanis (2) found that counties with a higher percentage Integrating Trip and Roadway Characteristics to Manage Safety in Traffic Analysis Zones Mohamed Abdel-Aty, Chowdhury Siddiqui, Helai Huang, and Xuesong Wang A transportation network is a conglomeration of road-traffic-environment modules and features multicategories of interdependent factors. This mix makes the management of safety in traffic analysis zones (TAZs) explicitly challenging. This study investigated the association between crash frequencies and various types of trip productions and attractions in combination with the road characteristics of 1,349 TAZs of four counties in the state of Florida. Crash safety management of these TAZs is emphasized through prioritizing them by examining the effects of trip and roadway factors on the aggregated crash frequencies. Models were developed separately for total crashes, severe crashes (fatal and severe injury crashes), total crashes during peak hours, and pedestrian-and bicycle-related crashes on the basis of various groups of estimators. It was found that the total crash model and the peak-hour crash model were best estimated by total trip productions and total trip attractions. The severe crash model was best fit by trip-related variables only, and the pedestrian-and bicycle-related crash model was best fit by road-related variables only. The results from this study pave the way for better safety management and the incorporation of safety measures in travel and network planning. The nature and extent of roadway safety vary widely depending on roadway type and facility, driver characteristics, land use pattern, and various other factors. Considerable research has been conducted to reduce the occurrence of crashes that take millions of lives and cause immeasurable human suffering each year throughout the world. As a crash is associated with a complex interaction of various factors, micro-level crash analysis (e.g., road-specific crash analysis, crash-specific safety analysis, event-specific analysis) can lead to better insight about factors that contribute to crashes. Also, by knowing the factors associated with crash occurrence one can enhance safety-conscious transportation planning and management. A transportation network is a conglomeration of various sets of road-traffic-environment modules and features by multicategories of interdependent factors. This imposes a challenge in macro-level and aggregate analyses of crashes. Aggregate level considerations ar