Investigating the effects of monthly weather variations on Connecticut freeway crashes from 2011 to 2015.

摘要: Abstract Introduction The objective of this research is to investigate the effects monthly weather conditions on traffic crash experience freeways, considering interactions between weather, volumes, and roadway conditions. Methods: Data from state Connecticut 2011to 2015 were used. Random parameters negative binomial models with first-order, autoregressive covariance estimated for representative types freeway crashes (front-to-rear, sideswipe-same-direction, fixed-object), most severe (i.e., fatal injury crashes), non-injury property-damage-only crashes). Results: Major findings are that variations in geometry, explain much crashes. Time exist panel data all Taking into account effect improves model prediction results. When raw measures highly correlated, using dimension reduction techniques helps extract more interpretable factors. By interaction condition variables, additional found. In general, lower temperature, heavy fog days, decreased precipitation, wind speed, higher narrower inside shoulder found be associated area type outside width change dramatically as number through lanes changes. Practical applications: could help researchers general readers gain a better understanding other factors give engineers practical guidelines improving safety.