Many studies have quantified the effects of traffic and geometric factors on the expected number of road crashes. However, crash prediction models that include also rainfall and hazardous points such as junctions or tunnels have rarely been developed. In addition, most research has paid more attention to two-lane roads rather than to multi-lane roads. Finally, as far as the authors are aware, few researchers have investigated the relationships in Italy between crashes occurring on multilane roads and the combined impact of all variables mentioned above. Thus, in this paper prediction models for estimating traffic crashes on Italian multilane roads as a function of infrastructure geometric characteristics, pavement surface conditions (wet or dry), and hazardous points (junctions or tunnels) were set up. Accident data were observed on a specific four-lane median-divided Italian motorway during an 8-year monitoring period extending between 1999 and 2006. Negative Binomial Distribution, applied separately to tangents and curves, was used to model the random variation of the number of crashes. Model parameters were estimated by Maximum Likelihood Method, and the Generalised Likelihood Ratio Test was applied to detect the significant variables to be included in the model equation. Goodness-of-fit was measured by means of both the explained fraction of total variation and the explained fraction of systematic variation. The candidate set of explanatory variables was: length (L), curvature (1/R) and the presence of point hazards such as junctions (J) or tunnels (T). Separate prediction models for total and severe crashes only were proposed. For curves it is found that the most significant variables are L, 1/R and J, whereas for tangents they are L and T. The effect of rain precipitation, examined on the basis of hourly rainfall data and assumptions about drying time, shows that with a wet pavement significant increases in the number of crashes are expected. In particular, rain considerably increases the number of accidents on curves than on tangents.
THE ASSOCIATION OF RAINFALL AND GEOMETRIC CHARACTERISTICS ON TRAFFIC CRASHES
CALIENDO, Ciro;GUIDA, Maurizio;PARISI, ALESSANDRA
2007-01-01
Abstract
Many studies have quantified the effects of traffic and geometric factors on the expected number of road crashes. However, crash prediction models that include also rainfall and hazardous points such as junctions or tunnels have rarely been developed. In addition, most research has paid more attention to two-lane roads rather than to multi-lane roads. Finally, as far as the authors are aware, few researchers have investigated the relationships in Italy between crashes occurring on multilane roads and the combined impact of all variables mentioned above. Thus, in this paper prediction models for estimating traffic crashes on Italian multilane roads as a function of infrastructure geometric characteristics, pavement surface conditions (wet or dry), and hazardous points (junctions or tunnels) were set up. Accident data were observed on a specific four-lane median-divided Italian motorway during an 8-year monitoring period extending between 1999 and 2006. Negative Binomial Distribution, applied separately to tangents and curves, was used to model the random variation of the number of crashes. Model parameters were estimated by Maximum Likelihood Method, and the Generalised Likelihood Ratio Test was applied to detect the significant variables to be included in the model equation. Goodness-of-fit was measured by means of both the explained fraction of total variation and the explained fraction of systematic variation. The candidate set of explanatory variables was: length (L), curvature (1/R) and the presence of point hazards such as junctions (J) or tunnels (T). Separate prediction models for total and severe crashes only were proposed. For curves it is found that the most significant variables are L, 1/R and J, whereas for tangents they are L and T. The effect of rain precipitation, examined on the basis of hourly rainfall data and assumptions about drying time, shows that with a wet pavement significant increases in the number of crashes are expected. In particular, rain considerably increases the number of accidents on curves than on tangents.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
