2003-09-25_PWETC_AgendaPacket - Laserfiche WebLink

Home

17 accidents is higher for the lighted section. (A larger night-day ratio indicates a cor- respondingly more hazardous night c~ndition.) Certain accident types and roadway ele- ments predominate as the percenLl!l!'es given at the bottom of Table 4 show. These data show that certain apparently dangerous locations actually had relatively few accidents. Such locations include the exit ramp, or gore position, and the ramps themselves. Type 4 accidents at these locations accounted for only 1 percent of the accidents on the lighted section and 2 percent on the unlighted section. Primary accident problems in the study sections involve rear-end collisions. The high concentration of such accidents at ramp entrances can be partially traced to lack of adequate acceleration lanes. The unlighted section had better designs in this respect. ..,. INTERCHANGE LIGHTING In several route sections, accidents were tabulated separately in the interchange areas (between extreme ends of exiting and entering ramp tapers) and areas between the interchanges. The percentage of night accidents occurring within each grouping was chosen as the best means of comparison. Summary data are given in Table 5. In most of the cases, a higher percentage of accidents occurred at night between the interchanges. This appears to be a characteristic that is not changed by lighting. The before-and- after sections show an even more pronounced difference after lighting was installed. A check was made of 5 freeway sections having 4,000 accidents. From 29 to 79 percent of all the section accidents occurred between interchanges. The average for all 5 sec- tions was 50 percent. Data such as these should generate careful review of policies that favor interchange lighting over continuous lighting. LIGHT-POLE COLLISIONS Most of the lighted routes studied had relatively standard mounting heights of 28 to 33 ft. Thus, the better illuminated routes had closer pole spacing. One 4-lane section with a measured HFC average of 0.50 had 35 poles/mile, and an adjacent section of the same 4-iane route with a 1.1 HFC had 66 poles/mile. The lower lighted route had 739 total accidents, 8 of which (1.1 percent) involved light poles as a first significant object struck. The total exposure (poles/mile x MVM) was 5,250 MV, or 655 MV /pole accident. The higher lighted route had 620 accidents, 17 of which involved llght poles (2..7 per- cent). Thus, an increase of 90 percent in the number of poles/mile also numerically increased the number of pole accidents by 90 percent. The pole percentage of total accidents was 2';.; times as great. The exposure was 12,700 MV for the higher llghted route, which is also 2';.; times that of the lower llghted route. The route had 1 pole accident/745 MV. A third route checked had 8 lanes, 0.9-HFC illumination, and 80 poles/mile. This route had 698 accidents during the study period; 27 (3.9 percent) involved llght poles. The total exposure was 22,000 MV or 815 MV/po1e accident. On a vehicle mileage basis, 2 lower lighted routes experienced rates of 0.05 and 0.06 pole accidents/MVM. The 2 better lighted route rates were 0.09 and 0,10/MVM. Thus, doubling the number of poles increased the actual number of pole accidents and tended to double the accident mileage rate involving poles. EFFECT OF ILLUMINATION LEVEL The routes selected for study have a wide range in accident rates. These vary from a low of 0.39 to a high of 9.24/MVM during the day. At night, the range is 0.62 to 9.98/ MVM. The average overall rate (for the routes with volume data) is 3.25/MVM. From the standpoint of testing a broad range of different conditions of congestion, geometric design, climate, and metropolitan area size, the variety of route sections is highiy desirable. The magnitude of differences, however, clearly precludes any direct comparison of accident rates, and analysis of lighting effect required use of the night-day ratio of rates. During the early part of this research, it was still assumed that field measurements on a sampling basis would provide a narrow range of factors, which could be used to relate calculated and actual field lighting levels and uniformities. It was assumed that % '\ I !. I