' ' ’ . 0" -.-...fi. ' ‘ ‘ ~ - ‘..... r _ . . . _ . . .u - 0-. - - .. . .vs \ . . . .v ‘. ” ~."" I " ’ o 0 —. .. . - .‘ A . . ‘ o . ‘ o _ _ ‘ . ' . ‘ . . ‘ . , . » ‘ . . I c . ‘ . . . . . . I. ‘ . ~ . . . - ' . . - ~ . . , . I ‘ I .. . ., _ . . . . - > -' . _ . . . r ' ' ‘ ‘ . . . . . V ‘ .. _ ‘ A ' . . . ~ . . . . , _ _ . . ‘ . . . , - . _ ‘ . ‘ . - '. , . - - ‘ . . ~ . ~ 0 . _ . . - . . A . ' ' ‘ o . . . . . ‘. _ _ ‘ . . . _ . . I . . . , ‘ . . ‘ . ' .‘ I . I ‘ .-. . . I I - . . . I ' ‘ - ‘ . u . - A . . . _ . ‘ i: . . ‘ - ._ I .‘ . . I . ' ' ‘ . . - I I ' . . A V ' . . ' . .'. . . . . ' ' ' . V . . , ., _ _ . > o . . ‘ . . ‘ . . .V . . . - . - . . . -.. - . .. 2 , _ ‘ - ‘ ... ' . - . . . . . h. . . , . _ v . . _ ~ . . . . . . . , . ' . u _ . . . ‘ . . . I . - . ' V . , . . I . - . o .. . . . . - o .‘ . ’ . . . I . ... . o u . . . . . . I . o . . - .. I . . o . . V . . . v ' I ' . ~ ' . . ‘ ' . , ‘ r 'u‘ . I . . . '. - . . V ' . ' . ' ' ' ' . . I . y - ~ ' . a . , . ' ' '-. . ' i t ‘ '. . - . . . . .. - ~ . O . . l ' . a ' . ‘ .‘ . _ . . h h ' ' . c . . . , . D . I . . . . , - ‘ .. h ‘ . O ' ’ u . V , . . . . . . . . . . ‘ . . . . . . . -. i - . . . . v . . ~ . '.‘ . .. _ ' .' . ' . , ‘ I ‘ ' . ’ - v . , , ~ I ’ ' l . . . . . o ' ; ‘ . _ . . . . ' y i. . . .. . - . . .‘.. _ ‘. ' I. . . — ., - . ' . ‘ l I n . . . . . . _ . ' v . . . a. . . , .. . . . ' . . ‘ ‘ u . . . . I‘ . . . - . . . . , . ‘ . ‘ , h . . ‘ ' o . .- .-. . . ' _ ' - ‘ . - o . - - .\. . - '- V . i . a ‘ , . _. _ ‘ D I ' -. rm LIBRARY "'ichigan State ; Univefsity PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE 2/05 cz/éTfiClDateDueJndd-pds __ PlanBPaper Thornton, Iatthev. APPROPRIATE FREEWAY SPEED ZONES: A CASE STUDY OF SAFETY AND COMPLIANCE MATTHEW THORNTON A Plan B Paper Urban Planning 889 MAY 23 , 1994 a" ' .. ‘ "~ "(A , 3" human} ‘ 'QN BOX to remove this c* “n or before data TABLE OF CONTENTS INTRODUCTION CURRENT CONDITIONS ISSUES FOR ESTABLISHING SPEED ZONES CASE STUDY CONCLUSION 14 INTRODUCTION The issue of traffic control and the appropriate use of traffic control devices is one of the primary concerns of traffic engineers. Of particular concern is the control of speeds and the determination of realistic speed zones. Appropriate speed zones are important for efficient traffic flow, highway safety, law enforcement as well as other factors. The following study was concerned with speed zones for urban freeways and evaluating their safety, appropriateness, and effectiveness. This research was concerned with three primary issues. First was an evaluation of safety issues associated with freeway speed zones, particularly 55 miles per hour (mph) zones against 65 mph ‘zones. This was accomplished through a literature review of studies in this area and an analysis of accident occurrence on several freeway segments in Michigan. Secondly was compliance with freeway speed zones. Similarly, this was accomplished through a literature review and a case study on the same freeway segments in Michigan. Thirdly, this study attempted to determine if more realistic and prudent guidelines for establishing urban freeway speed zones should be developed. In addition, to safety and compliance issues the following discussion will begin by address the current conditions of urban freeway speed zones in the United States (U .S.) and how they came into existence. Primary issues and concerns associated with the establishment of speed zones will also be included. CURRENT CONDITIONS The most recent change of significance in freeway speed zones came in 1987, when congress allowed states to increase speed limits on rural interstates and some US routes from 55 mph to 65 mph. The urbanized area boundary was used as to delineate urban freeway segments from rural segments. This boundary was established by the Census Bureau, and is defined as any central place with a population over 50,000 and its surrounding area which maintains a minimum population density of 1000 persons per square mile (5). Speed studies nor engineering studies of any type were used to create this rule for spwd zones. As a result many freeway segments designated as urban, and posted at 55 mph because they were within the urbanized area, exhibited rural characteristics (i.e. wide medians, widely separated ramps and interchanges, and few weave areas). Public officials and engineers alike felt that this was not appropriate, and speed zone changes in urban fringe areas, despite the fact that they were within the urbanized boundary, would benefit traffic flow, improve speed compliance, and would not compromise highway safety. As a result, in 1990 the US. Department of Transportation (U SDOT) allowed states flexibility in setting 65 mph freeway speed zones. Freeway segments that only passed through the periphery of an urbanized area (i.e. had no more than two interchanges within the urbanized area) were eligible for a speed limit increase to 65 mph. In addition, on freeways that passed directly through urban areas, states were allowed to move the 65 mph speed zone into the urbanized area, up to their first interchange. Despite this numerous freeway segments in urban fringe areas remained posted at 55 mph, although they display characteristics of rural freeways. That is, speed zone changes from 65 mph to 55 mph take place, although there are no changes in the geometric design of the freeway. This is directly the result of using only the urbanized area boundary to designate the speed zones. There are other factors beyond population and population density that need to be considered when establishing speed zones if they are to be appropriate and realistic. ISSUES FOR ESTABLISHING SPEED ZONFS There are numerous issues and conditions that need to be considered when setting speed limits. Certain issues are of more importance when dealing with freeways. These issues include traffic volumes, geometries, compliance, and safety (3). Traffic Volumes There are several issues related to traffic volume which need to be addressed when considering speed limit changes on freeways. First, is the effect volumes have on constraining speed. That is, at what point do volumes effect free flow conditions and force drivers to decrease their speeds. This is important because high volume are frequently cited as a reason for lowering spwd limits. Secondly it is important to be aware of when and to what degree peak hour volumes occur. Related to this it is also important to be aware of how speed changes effect the level of service. Thirdly one must also be aware of merging volumes and traffic densities, and how increased speed may effect driver’s ability to merge and make lane changes. Finally, the carrying capacity of the segment must also be considered. At high volumes the carrying capacity may be grater at lower speeds due to decreased head-ways. Geometric Design In addition to traffic volumes roadway geometries are also important in determining appropriate speed limits. Sight distance, lane, shoulder, and median width, entrance and exit ramp density, design speed, and depressed or elevated conditions all play a role in determining speed limits. Geometries are of particular importance for this study as road design often remain consistent across urbanized area boundaries. Speed Compliance and Driver Behavior Both geometries and volumes are important issues for setting speed zones, but are usually secondary to typical driver behavior, although driver behavior is a function of volume and design. It is widely agreed among transportation professionals that speed limits should be determined by what the typical driver feels as safe and appropriate. This is refereed to as the ”85th percentile speed". This is the speed at or below which 85 percent of traffic is traveling. It has been found that this speed accommodates the safe and prudent driver and that lowering or increasing the posted speed has little effect on the 85th percentile speed. In addition, it has also been shown raising the speed to this level caused no increase in accidents (4 pp.5-6). Speed limits that are determined by the 85th percentile are favored as they are the most realistic, and in turn, decrease compliance problems, decrease speed variation, and lead to better traffic flow (2). Since speed limits on freeways are not determined by 85th percentile speeds nor considerations of geometries or volumes, but rather urbanized area boundaries certain segments of freeway have unrealistic speed limits. The primary concern related to these speed limits and the final issue important to determining speed limits, is safety. Safety Issues Although, it was stated that realistic speeds determined by the 85th percentile speed are appropriate and do not effect safety it is widely felt in the political arena that constraining speeds on freeways will increase safety. This is primarily due to both the effects of lowering the speed limit on freeways to 55 mph in 1974 and the increase to 65 mph on rural freeways in 1987. It is widely believed that the first change drastically increased safety and the latter change had a converse effect. Despite numerous studies that attempt to support this, the true effect of these speed changes is not completely understood. In addition, if drivers are not complying with the posted spwd limit, it would indicate that factors other than speed are effecting accident involvement. As stated and indicated from other studies increased speed will not necessarily lead to higher accident rates and speed limit reductions will not necessarily effect driver behavior (10 pp.45-49). Indeed, accidents and accident related fatalities decreased after the interstate speed limit was lowered to 55 mph, but this was a continuation of an existing trend and coincided with accident decreases on freeway segments that did not experience speed changes. The increase in traffic safety on freeways can be explained by improved medical services, better automobile design, and improved road design (1 pp. 323-329). It should also be noted that although the speed limit remained at 55 mph on freeways for 13 years this spwd limit was not widely complied with during a large portion of that time. Again, it is not valid to argue that speed limit decreases reduced accidents if the posted speeds were not complied with. When the speed limit on rural freeways was increased to 65 mph in 1987 there was speculation that this would compromise safety. Numerous states that changed their speed limits also conducted accident studies. Many states found an increase in accidents on freeway segments effected by the speed change. Despite this other states found either a decrease or no significant change in accidents (2). The basic conclusion from this and the 1974 experience is that many factors influence highway safety, including excessive speeds, but speeds that are reasonable do not significantly affect safety. In order to provide more insight and determine if better guidelines for speed zoning on freeways should be established, particularly in urban fringe areas, a case study was undertaken. CASE STUDY Since freeway speed zones are established in a nontraditional fashion and are potentially inappropriate and unrealistic, an evaluation of freeway accidents and speed compliance was performed. The intention was to determine if and how speed zones may effect accident rates and driver compliance. The primary concern was to help identify freeway segments that have inappropriate speed zones and establish guidelines for determining better freeway speed zones. It was hypothesized that freeway segments in urban fringe areas would have the highest potential for being zoned inappropriately. For this study, fringe area was defined as a freeway segment within an urbanized area and therefore zoned at 55 mph, but had rural design characteristics. Since conditions in theses type of zones was important to the study they were evaluated separately from 55 mph urban zones. Medians width, lane and shoulder widths, access point spacing, presence of weave areas, and at grade versus depresses or elevated designs were the criterion used to distinguish fringe from urban segments. That is, 55 mph freeway segments with wide medians, wide lanes and shoulders, widely spaces access ramps, few weave areas, and at grade design were designated as fringe zones. The remaining 55 mph zones with narrow medians, closely spaces ramps, extended weave areas, and possibly elevated or depressed designs were designated as urban segments. Segments zones at 65 mph were designated as rural. These 65 mph rural zones were compared with both 55 mph fringe and urban zones. Most study locations consisted of all three types of segments. That is, many of the different study segments were adjacent to each other. Study Location The study took place entirely within the state of Michigan. Full access control Interstate and US. Route sections of freeway were selected through a stratified randomly process. Study sections were chosen from urban, fringe, and rural freeway sections in and around the major urban areas of Michigan. These included Detroit, Lansing, Jackson, Flint, Saginaw, Ann Arbor, Grand Rapids, Kalamazoo, Battle Creek, and Benton Harbor. The sample was stratified by urban area in order to ensure an adequate number of fringe and urban segments which were evenly spread among the different urbanized areas. Since the concern of the study was with the appropriateness of urbanized area boundaries as speed zones delineators, remote rural sections of freeway were not desired in the study sample. None the less, rural freeway sections made up the largest part of the sample data while urban sections made up the smallest portion of the sample. There were 27 rural section totaling to 255 miles, 17 fringe sections equalling 64 miles, and 10 urban sections totaling to 42. Data Collection and Manipulation These fifty four sections were selected from in an around the nine major cities in Michigan. Accident records and speed compliance data for the selected sections was then acquired from the Michigan Department of Transportation (MDOT). The accident data was the complete automated record from the Michigan UD-10 accident form, for the year 1991. The speed compliance data was collected in the summer of 1993 through the use of loop detectors and b included the average, range, variance, pace, standard deviation, and 85th percentile speed. Speed data was collected for single 24 hour periods and was only available for 21 of the selected sections, but they were evenly distributed among the three analysis zones and across the nine urban areas. In addition to analyzing the accident data as a whole, analysis was also performed on different variations of the data controlling for contributing circumstances and environmental effects. For some of the analysis non-speed related accidents were removed, these included backing, parking, pedestrian, animal, bicycle, and other non-freeway type accidents. Lighting and pavement conditions were also controlled for. That is, day dry, day wet, night dry, and night wet accident were separated for analysis. The evaluation process showed similar findings for controlled data under all conditions. Therefore, the finding reported in the following sections are for the accident data with out controlling for accident type or environmental conditions. Finally, it was felt that it would be meaningful to control for accident severity, fatal, injury and property damage accidents were separated during analysis. Accident Findings Evaluations of the accident data revealed several facts about freeway accidents and the effect of speed on safety. First accident type varied across the three different analysis zones. Four particular type of accidents contrasted greatly between the different zones. Rearend accidents were much more likely to occur in urban areas. Of all the accident in urban 55 mph zones 64 percent were rearend as opposed to only 26 percent of rural segment accidents being rearend. In the fringe areas 49 percent of the accidents were of the rearend type. This is not surprising as rearend accidents would be expected in high volume areas with narrower shoulders and less recovery area. On the other hand Animal related accidents were more likely to occur in rural areas. Of all accidents in rural areas 17 percent involved animals while only 0.6 percent of the accidents in urban areas were related to animals. This likely not to be a spwd related effect though, as exposure to animals in rural areas is much greater than would be expected in an urban area. Finally, it was also found that both fixed object accidents and overturn accidents were more likely to occur in rural areas than in fringe areas, and more likely in fringe areas than in urban areas. Indeed, these type off accidents could be related to speed, but these type of accident were relatively rare compared to rearend and other types of accidents. Only 8 percent of rural accidents were overturn accidents. This is opposed to 3 percent in the urban zones. Another interesting note is that 36 percent of accidents in 65 mph zones occurred near street lights, opposed to 15 percent in urban areas. Interestingly, this did not correlate with accident involvement at interchanges as may be expected. Only 20 percent of the accidents in 65 mph zones occurred at interchanges. More importantly it was found that accident rates were higher in urban areas than in rural and fringe areas, see table one. The accident rates are the number of accidents per 100 million vehicle miles. This was calculated by multiplying the number of accidents in the zone for that year by 100 million and dividing that product by the product of the average daily traffic, the zone length, and 365 (the number of days in a year). The traffic counts used to calculate these rates was also provided by the Michigan Department of Transportation. In all categories (property damage, injury, fatal, and total) except for fatal accidents the accident rate was higher in urban zones than in the other two zones. Due to the fact that fatal accidents are relatively rare events these figures are the least reliable. That is, the sample size is not sufficient to draw conclusions about the variation in fatal accident rates across the different zones. In oneway analysis of variance test on the accident rates indicted that there was no 10 -. statistically significant difference between the fatal accident rates. Conversely statistical evaluations of the other accident rates found that there was a significant difference between the urban and rural accident rates. That is, significantly different at a 95 percent confidence level. TABLE ONE. ACCIDENT RATES, Accidents per 100 million vehicle miles :ACCIDENT CATEGORY 55 MPHURB URBAN 55 MPH FRINGE 65 MPH RURA PROPERTY DAMAGE INJURY 243 135 81 FATAL 1.6 3 . 6 1.2 TOTAL 908 621 337 =====._==__JE_== Because of this injury accidents provide a better indication of accident severity. When . evaluating highway safety accident severity is of equal or higher importance than simple accident rates. It is argued that speed may not have a great effect on accident involvement, but it does effect accident severity as higher speed impacts are more damaging. This is also in agreement with the earlier finding of the occurrence of different accident types--rearend accidents, although more plentiful are not typically severe accident. Despite this, the findings of this study do not support this. In addition, to having higher injury accident rates, urban areas also have a higher occurrence of injury accidents than do rural and fringe zones, see table two. Although it would not be 11 expected property damage accidents are much more prominent in rural areas than in urban areas. This and the accident rate findings indicate that conditions other than speed limits account for freeway accidents and accident involvement. TABLE TWO. ACCIDENT SEVERITY, Percent of Total for each Category by Zone PROPERTY DAMAGE INJURY 30% 23 % 22% ] FATAL 0.2% 0.4% 0.5% Speed Compliance Findings Even more critical than accident rates is speed compliance and zone effectiveness. If speed zones are not appropriate many drivers will not obey the spwd limit. This is not desirable as it creates law enforcement difficulties as well as higher variation of speeds and nonuniform speed distributions. Such speed distribution variations are not safe and have been shown to create hazardous conditions (9). Indeed, if speed limits are not complied with any benefit from slower speeds, even if they are not highly significant, will not be realized. This study found that in general freeway speed limits are not widely complied with, but compliance in the 55 mph fringe zones were much lower than in the other two zones, see table three. This provides support for several critical issues associated with establishing speed zones. 12 First this shows that drivers behave according to how comfortable they feel and drive at speeds that they feel are appropriate, which for the most part is independent of the posted speed. In urban areas where drivers are more confined by geometric design and more likely to encounter heavy volumes and congestion, speeds are slower than in 55 mph fringe zones which have more open road design and less congestion. This also indicates that freeway speed limits set closer to the 85th percentile speed may be more appropriate and will lead to better compliance. TABLE THREE. SPEED COMPLIANCE 65 MPH RURAL I AVERAGE SPEED 59 MPH 63 MPH 67 MPH 85TH PERCENTILE 65 MPH 69 MPH 74 MPH SPEED PERCENT EXCEEDING 62% 89% 63% SPEED LIMIT PERCENT EXCEEDING 15% 25 % 9% SPEED LIMIT BY 10 MPH OR MORE E In light of these compliance findings, the accident data was analyzed in regard to actual spwd as opposed to posted spwd. Further analysis of variance tests with the use of average speed as a covariate, indicated that speed was not a significant factor in explaining the variation of 13 accident rates across the study zones. This was true for all accident rate categories fatal, injury, property damage and total accidents. CONCLUSION The results of the case study analysis combined with the non-traditional method for establishing freeway speed zones around urban areas indicate that better guidelines for determining these speed zones need to be established. This is not to suggest a universal 65 mph spwd for all freeway segments, but rather better evaluation if certain segments, particularly those through urban fringe areas. Such freeway segments posted at 55 mph, but have designs similar to preceding 65 mph zones should be considered for upgrades to 65 mph. The findings of this study suggest that this can be implemented without compromising safety and would improve compliance. Freeway segments in urban core areas which are depressed or elevated, have high ramp densities, narrow medians, and extended weave areas should remain at 55 mph. In addition, 65 mph zones out side the urbanized area that do not have rural design characteristics or are particularly hazardous could also be considered for a reduced speed zone. In general better guidelines and methods need to be established for determining freeway speed zones. These guidelines should consider geometric design, traffic volumes, zone length, road side development, compliance, and safety. These guidelines would lead to more appropriate speed zones which would have numerous benefits. First this would encourage better compliance which would decrease law enforcement needs and legitimize and create more effective speed zones. Secondly, this would create better flow characteristics which would lead to higher 14 carrying capacity, decreased travel times,and more uniform speed distributions. Better speed distributions and lower speed variation would also have the potential for also improving safety. 15 10. ll. 12. REFERENCES Deen, T. and Godwin, 8., 'Safety Benefits of the 55 mph Speed Limit," W W, July 1985, pp. 321-343. “catwalk Streff F mdSchulmll W W, University of Michigan Transportation Research Institute, December 1989. 'GuidelinesforDeterminingWheletheSSmphSpeedIimitCouldbeRaisecL' Institute W January 1987 PP 21-25 W Michigan Office of Highway Safety Planning. 18mins- WW, ”Area Classifieations' U.S. Department Commerce. -. , Michigan Department Sll El C I IS |' Eli ,Ii'l'ganD I IofT. |° ’1 . 1980. Nomsis Mafiia. Wards. Rush-MW. Clficaso 1986- Taylor, William, ”Speed Zoning a Theory and its Proot‘ W, January 1965. pp. 17-21. Ullman, G. andDudek,C., "Effects ofReducing SpeedLimitsinRapidly Developing Urban Fringe Arm“ MW. Transvomfion Web Board, Washington D. C. 1987. Calrk. W. and Hosldng. P.. W John Wiley and Sons. New York, 1986. May. Aldof. Wants. Prentice Hall. Englewood Cliffs, 1990- 31293 02656 9008