Lane width and safety. Literature

Ezra Hauer

Lane width and safety. Literature

Ezra Hauer

visibility

…

description

29 pages

link

1 file

terms lane width, roadway width, carriageway or pavement width will be used 1 interchangeably. Earlier drafts of this papers were prepared in the course of a project for UMA 2 Engineering (for the new Canadian Geometric Design Guide) and for DELCAN (in ORSAM 98). "On rural two-lane roads capacities of 10 and 11-fr lanes, expressed as a percentage of 3 12-ft lane capacity are 77 and 88 percent, respectively.r four lane undivided highways these percentages are 89 and 95 . . ." (Green book, 1984, p. 360). Capacity is the reciprocal value of the average minimum headway. Thus, with wider lanes drivers choose shorter headways. 1.1 C:\work\PROJECTS\HSIS\IHSDM-Multilane\Literature Reviews\1. Lane width\lane width.wpd This is an unedited draft reflecting my personal opinions. Ezra Hauer Lane Width and Safety. 1 E. Hauer. Draft , March 7, 2000. 2 1. Introduction. The link between lane width and safety is woven of two principal strands. First, the wider the lanes the larger will be t...

Sign up for access to the world's latest research

checkGet notified about relevant papers

checkSave papers to use in your research

checkJoin the discussion with peers

checkTrack your impact

Figures (27)

Figure 1. Non-junction accidents for two- lane roads in Buckhinghampshire. and Traffic Review, Autumn 1962, pp. 15-20) which shows the relationship in Figure 1. The squares show the accident rates fo as been reported. Thus, carriageways that are r the four carriageway width categories in which data 6-17 ft wide had 2.3 accidents/MV M (MV M=Million ‘ehicle Miles) as shown by point A while carriageways 22 to 24 ft wide had an accident rate of 1.7 ccidents/MVM as shown by point B. It is arriageway will reduce the number of accide tempting to interpret this to mean that widening the nts.

The kind of relationship between accident frequency and ADT that is depicted in Figure 3 vas been found repeatedly when data was examined (see, e.g., Roy Jorgensen Associates, p.18 or Zegeer, Deen, et al. 1981 ). Other researchers (see, e.g.,Silyanov , Slatterly & Cleveland 1969 p. 312} report an upward curving relationship, implying that accident rate increases with ADT. Be it 4S it may, when the accident rate decreases orincreases with ADT and the feature of interest (here ane width) is associated with ADT, one may not come to conclusions about the effect of that feature yn safety, without separating what part of the change in the accident rate is due to the ADT and what yart is due to the feature of interest.

1955. The earliest before-after study of pavement widening is by Cope. The data are for 22 pavement

Table 4. Relative Accident Rate by Babkov (1975). 1978. An elaborate regression analysis was performed by Roy J orgensen Associates . Based on date from Maryland, New Y ork and Was tegories x 3 categories of horizon idth and shoulder width were the i thors comments that: “. . . nts were common ationship between accide ines do not eff idth or across ADT leve resolved to esti Ca Ww au coeffici tal cu ndepe iS approach was nl ting that t hway geom how accident nd hig ively explai raight OU e Thus, abandoni icative model s” p.13. te a multi thors shown in Table 5. hington, initially 36 linear regression models were rvature x 3 categonies of sh ndent variables. The model te. Ch he model was not exp trics, if on adequa rates vary... e existed. Th und to ex aining th ese findin (with ng the lin in sign of et shoulder ear regression model oulder type. Terrai was fo anges gs rue phys for accident modification factors. The result fitted in 4 ADT n, pavemen’ plain little. The the regressi ica indicate t. and pavemen | fitting idea, t OF nat fl € is 1973, 1975. Silyanov and Babkov show graphs similar to Figure 1 and Figure 8 on which they compile data by many researchers from various countries (USSR, UK, Sweden, Germany, Hungary USA). They all form a band as in Figure 1, showing declining accident rates with roadway width. As far as I can tell, the method of all these studies is the same as th accident rates were computed for road sections having the same roadway width. The associatiot between ADT and roadways width has not been accounted for. at discussed in Section 2. That is This means that the results are confounded and one cannot say whatis due to roadway width and what is due to all the other factor which vary with ADT. The noted decline in rates may be cau associated with traffic flow. Drawing a bold curve through Babkov (1975) gives the numbers in Table 4. sed by the many factors that an the diverse results by many researchers,

Table 5. Accident Modification Factors for rural two-lane highways based on combined and smoothed accident rates in Maryland and Washington (p.18). The authors note that the increase from th e expectation” but also emphasize that they decided to join the two righ effect of conservatism this decision and the >23 ft category “. . of has the questions category to expectatio th f this, (1978) resu AMF (modi ns, they are cons behaviour and geometric stat mai. iste ndai fied). Were the u “ in estim nt with ntains th rd.”. The questi ted in the oft-quoted AMFs f e 21-22 ft category to tl tw sin met e the i . may have been ano it is consisten tmost column ating the geo at whi ith th to on ric € ncrease malous the general h he >23 ft g e width ca in accident in terms of ypothesi onable reasoning in Roy Jorgensen and Associates rom their Table 13 and reproduced in row 2 of Table 6 as nmodified data used the result would be that in Table 7. s of an in roup is “inco tegonies sayin ffects on safety”. McLean nsistent with e research findings” (p. 20.) In spite g that “This ( p.192) rates from conventional teraction bet the 21-22 ft engineering ween driver

f one may ass raits except substantive accide ncrease in ru Thus, 18' pave men nt reductions in ‘run ts may be expected ume that in the regions of overlapping ADT the compared for lane width, the results indicate that increased lane wid road’ and ‘opposite direction’ acci n-off-the-road accidents is by a factor of 1.12 per foot decrease -off-the to have .12*=1.57 time more run- off-the- road sections have similar th is associated with very dents. The average in pavement width. road accidents than 2' wide pavements. The average increase in ‘opposite- direction’ accidents is by a factor of 1.21 per oot of decrease in pavement width. T he pavement width groupings in this repo rt do not allow one 0 judge whether accident frequency increases as pavements width grows beyond 23 or 24 feet. 1980, 1981. Zegeer et al. merged data for about 17,000 accidents (in Kentucky) with highway section data for 25,670 km of road. The authors focussed on run-off-the-road and opposite- direction accidents (apparently excluding rear-end, intersection and driveway related accidents as unrelated to the issue of lane or shoulder width). The large majority of roads (70%) had no shoulders. The main results are shown in Figures 10 and 11. results are shown in Figures 10 and 11.

Table 8. Proportions of accidents by type. The authors note that: “Accident rates for oth run-off-the-road and opposite- direction acci er accidents generally increased as lane width inc indicating that the only accidents which would be expected to decrease with lane widening w dents.” (1980, p.4). Of course, if one assumes that lane widening reduces some accident types, one must be willing to accept that other accident types are increased by lane widening (perhaps due to in the earlier studies. creased speed). It is the joint effect on all acciden reased, ere the t types which ought to be of interest. Since this joint effectis n ot estimated here, one may only conclude that it is lesser than what has been estimated in this study and may be in line with the results obtained in

The authors also show how accident rates vary with shoulder and lane width (Figure 12)

Table 9. Number of accident and control sites by carriageway width. Thus, while 121 accident sites had a carriageway width wider than 6 m, only 95 control sites did. This association persisted even when sites on curves were eliminated.

he emerging picture is far from clear. Th ; due to limitations of sample size. A tte whole, the authors conclude that “Surface width h ident rate on rural, two-lane farm to m f the total) and that “surface widening can reduce sin e erraticjumps al arket roads with re explainable pa rtl mpts to fit a curve to the data met as no demonstrab ADT’s up to 1,50 idents are about 67% of the total in this data). gle vehicle accide y by the imprecision that with limited success. O1 e effect on multi-vehicle 0" (which are about 33% nt rate ...” (single vehicle

considered ‘relevant’. The authors state that: “Rates for other accident types were found not to b significantly related to lane or shoulder width.” (p.163) (“Acci As in Zegeer et al. (1980, 1981), dent rates for other accidents gene only accidents which would be expected to oppos ite-direction accidents.” 1980, p.4). 1 and only 39% of the cross-section data b more only single-vehicle and opposite- direction accidents we . This seems contrary to the earlier quo rally increased as lane width increased, indicating that th decrease with lane widening were the run-ff-the-road an re e € C ase. If the remaining 37%-61% accidents tend to becom numerous as lane width increases, t hey should not be omitted from the analysis. é The ‘relevant’ accidents are 63% of the primary data base é

Table 9. ‘Related’ Accidents/MV M. One may expect that the true relationship between safety and lane width is smooth and gradual. Thus, e.g., it would be most perplexing if in Illinois 9 ft lane width was associated with a third of the accidents for 8 ft lanes but twice as many as for 10 ft. lanes. The accident rate jumps in Figure 21 and in Table 9 make it difficult to distinguish between what is signal and whatis noise. In general, jumps of this kind may be an indication of:

1994 b. U freeways. following t the projec Based on he implem at the expense of sho ts were implemen occasional review of y 10.5 fe several p work entatio ulder an et) whi. ted ...”(p. 126). T e maintaining sho rojects materialized several years after claim that “accident migration is not a problem hey also note th thors note tha rbanik discusses the experience gained in narrowing lanes and shoulders on url reported in (McCasland and Urbanik & Bonilla) they maintain n of 24 projects (in which capacity was increased by adding a new d lane width) “... most sites experienced decreased accident rates after t: “. . . higher accident rates had on well designed projects”. at “narrowing of lanes to 11 feet ulders did not change accident rates.” Based on in California the au anes were narrowed and left shoulders were removed . . .”. They also that ane the not 1994 b. Urbanik discusses the experience gained in narrowing lanes and shoulders on urban

Table 10. Regression coefficients for Lane Width and Pavement Width, Mid-block crashes. Whether lane width was included amongst the regressor variables has been decided in the stepwise regression by the Akaike information criterion. Thus, the non-inclusion of lane width in certain models merely indicates that it was deemed statistically insufficiently important. The functional form used forces the conclusion that crash frequency decreases with lane width no matter how wide the ane. The implied reduction in crash frequence per foot increase in width (e.g., 39% on four-lane urban freeways) seems quite excessive. Taw nwnwmn nnnnn LAA Rnaath nm weit then lawmncrdAth anta nna bhatt wri atawmiann nwarnkhnn) Taw bkeistn linnn

The assumption is that only run-off-road and opposite direction accidents are affected by lane widening. The proportion of such accidents depends on ADT. If the information in Figure 22 is to be applied to total accidents an appropriate correction needs to be applied. Thus, e.g., if half of the accidents are of the run- off-road and opposite direction accidents type, the AMF in figure 22 needs to be halved. 1997, In Septe ‘eaching conse safety Design ffect on sing inearly up to the limiting valu yanel thought mber 1997 ag e-vehicle and that the wideni ng o modification factors n roup of experts met in Washington for a few days with the purpose of nsus on the accident Model (IHSDM). T eeded as input for the Interactive Highway he results of this deliberation are shown in Figure 22. Thus, the opposite direction accidents is minimal for ADT <400 and increases es shown. Because of the results obtained by Zegeer et al. (1994) the F 9 ft lanes to 10 ft is undesirable.

Table 12. AMFs for ADT=10,000 1998c. Vogt and Bared used data from Minnesota (704.5 miles, 1694 accidents in three years) and Washington (535 miles,1706 accidents in three years) to estimate models for rural two-lane roads. The models account for ADT, roadway width, Roadside Hazard rating, driveway density , average horizontal curvature, average vertical curvature and average grade. The resulting regression

Paul Jovanis

2009

The Federal Highway Administration (FHWA) organized a pooled fund study of 26 States to evaluate low-cost safety strategies as part of its strategic highway safety effort. The goal of this study is to evaluate the safety effectiveness of various lane-shoulder width configurations for fixed total paved widths as a countermeasure for roadway departure crashes. Where possible, crash modification factors (CMF) are provided for specific lane-shoulder configurations. The cost of this treatment is essentially zero because it involves only the location of pavement markings. A matched case-control analysis was applied to geometric, traffic, and crash data for road segments in Pennsylvania and Washington. In general, wider pavement widths 9.75-10.97 m (32-36 ft) are associated with fewer crashes than narrower paved widths 7.92-9.14 m (26-30 ft). For specific lane-shoulder configurations, there is a general safety benefit associated with wider lanes and narrower shoulders for a fixed pavement width. For 7.92-to 9.75-m (26-to 32-ft) total paved widths, a 3.66-m (12-ft) lane provides the optimal safety benefit; the CMF ranges from 0.94 to 0.97, indicating a 3-6 percent crash reduction for 3.66-m (12-ft) lanes compared with 3.05-m (10-ft) lanes. For a 10.36-m (34-ft) total paved width, 3.35-m (11-ft) lanes provide the optimal safety benefit; CMF was 0.78 compared with the 3.05-m (10-ft) baseline. For a 10.97-m (36-ft) total paved width, both 3.35-m and 3.66-m (11-ft and 12-ft) lanes provide the optimal safety benefit; CMF was 0.95 compared with the 3.05-m (10-ft) baseline. Based on the estimated safety effectiveness of this strategy, specific laneshoulder configurations have the potential to reduce crashes cost effectively on rural, two-lane, undivided roads. However, limited sample sizes make it difficult to identify statistically significant differences between certain lane-shoulder configurations within a total paved width.

Log In

Lane width and safety. Literature

Sign up for access to the world's latest research

Figures (27)

Related papers

Related papers