This paper describes and presents the main results of a recent investigation on the effects of substandard forward visibility upon the safety of urban dual carriageways. The objectives of the study were to: (a) identify sites where sight distance is very substandard either in the horizontal sense (walls/fences) or in the vertical sense (crests); and (b) to ascertain whether any statistical significant difference can be derived between the accident records down stream of the sites and the route generally. Detailed sight distance and speed measurements were carried out on a total of 13 sections of the road. Eventually 63 sites were identified on just 7 sections of road, but 3 being on the A102, A102(M), A40(M), A406, A4, and M4. A total of 19 control sites were identified on the 7 sectors and personal injury accidents were plotted for both controls and sites. A total of 592 accidents were plotted on approximately 28 kilometres of carriageway. Other data collected for each site included variables such as: horizontal curvature; lane width; numbers of junctions; details of safety barriers and skidding resistance. Statistical analysis was carried out to ascertain the relationship of each variable with accident rate. Modelling was also performed to select the group of variables best able to explain the accident rates. Both an additive and an exponential model was used. The results indicate that substantial visibility in combination with minor junctions and merges, is associated with higher accident rates. Particularly severe cases of substandard visibility such as linked or multiple sites may be associated with higher accident rates. Sites on A(M) and (M) roads had lower accident rates than sites on A roads.
Samenvatting