ML20210E714
Text
.
THE TEXAS A&M UNIVERSITY SYSTEM TEXAS TRANSPORTATION INSTITUTE COLLEGE STATON TEMAS 77343-3135 TRANSPORT OPERATCNS PROGRAM (409) 945-153b February 1, 1984 Gail Temple US Nuclear Regulatory Commission 1450 Maria Lane Suite 210 Walnut Creek, CA. 94596 -
Dear Ms. Temple:
I '..
ener as Thefollowingwilldescribethefe.g.alconceptofevacuationtimeestimates7&d-ey relate to baseline scenarios normal and adverse weather) and also- ' ij@.
how a variety of other factors are considered in the event of an accident. ;] "
Since it is not cossible to know the specific conditions at the time of an accident, a number of' key factors are considered in-several baseline scenarih4 as reference points for decisionmakers. This approach is. appropriate as explained below.
+
The number of scenarios that could be considered is potentially very large I if many possbile events are included in an analysis along with the various possible combfnations of factors. Analysis of the very large number of scenarios possible would likely be more confusing than illuminating. In search for the "right" scenario at the time of an emergency, the decisionmaker would likely find that the actual
. conditions differed from those assumed. He or she would then have to make judge-ments based on the available data. A more productive approach is to limit the number of scenarios such that decisionmakers know the. sensitivity of the roadway network to variations in the number of evacuees orichanges in capacity.
A couple of comments are appropriate concerning adverse conditions. The ad-verse weather scenario is not intended as a worst case scenario. Neither the fog nor rain scenario include visibility so low that one can not see far enough to drive. Under. severe adverse conditions, the decisionmaker would have to add addi-tional time (e.g. an estimate of when the fog would clear) to the adverse weather scenario.
t
-In regards to road capacities, available research considering snow, rain-and i
fog suggests that an approximate 20 percent reduction in speed and capacit.y is '
l appropriate for a range of adverse weather conditions under which roadways are still passable. This conclusion is based on data from Houston (rain), Chicago (rain and snow) and Elmira, NY (fog). While not an extensive database, the data does support informed traffic engineering experience and judgement. I might add i
that perceptions concerning massive congestion under adverse weather is consistent l with a 20 percent reduction in capacity. A lane of traffic flowing near capacity at 1800_ vehicles per hour could result in 2 miles of congestion and about 15 minutes j
e603280124 860123 PDR U
FOIA J'
- T $ PORTA 1CN RESEARCH ANO DEvn0PMENT r
}
i of additional delay if capacity were reduced 20 percent. Under such conditions, speeds could be as low as 8 mph.
Concerning the interpretation of the 20% increase in time for adverse weather conditions, the normal conditions should not be considered as uncongested conditions.
.That is to say, congestion and delay will exist even under normal conditions as defined in the evacuation time estimate. The analogy is, therefore, not similiar to driving under normal conditions without congestion one day and then one day being delayed by fog or rain induced congestion.
- In summary, the evacuation time estimates are technically sound and based on the best available data. Obviously, the estimates can not resolve the uncer-tainties that exist in any endeavor.
Sincerely,
- Tom Urbanik II Program Manager TU:jh 9
o e
e e
7
.e e a
1 NATION AL COOPERATIVE HIGHWAY RESEARCH PROGRAM REPORT HIGHWAY FOG WARREN C. MOCMONO AND MENNETH PERCHONOK CORNELL AERON AUTICAL LABOR ATORY BUFFALO. NEW YORK RESEARCH SPONSOREO 8Y THE AMERICAN ASSOCIATION OF STATE HsGHWAY OFFICIALS IN COCPERATION WITH THf SUREAU OF PUBLIC ROAOS 9
$U5 JECT CLASStFICATIONS:
'RANSPORTAftON ADM'NISTRATION .~.
HIGHWAY OESIGN M AINTENANCE. GENERAL HIGHWAY SAFETY TRAFFIC CONTROL AND OPERAt6ONS URBAN TRANSPORTATION ADMINISTRATION o e.
HIGHWAY RESEARCH BOARD DIVISION OF ENGINEERING N ATION AL RESEARCH COUNCIL 1970
! NATIONAL ACADEMY OF SCIENCES-N ATION AL. ACADEMY OF ENOINEERING
r 4
HIGHWAY FOG
SUMMARY
A review of the literature shows th'at fog has had the following efTects: (1) a slight reduction in accident frequency, (2) an increase in the likelihood that,an accident will result in a fatality, and (3) an increase in the !!Lelihood that accidents will involve either a single vehicle or more than three schicles.
Traf5c measurements made during this project indicate that: (1) speeds were slightly lower in fog, (2) the probability of overdriving one's visual range was greatly increased, and (3) laterallocation and vehicle interactions were not affected by fog. It is concluded that drivers exercise more caution in fog, but that the increase in overdrising probably explains the increased severity of accidents.
Field tests have demonstrated that visibility in dense fog can be improved by se,eding with practical amounts of carefully sized hygroscopic material. Additional.
- tudies s are needed to refine seeding procedures and to determine the scope- of application for highway fog abatement. Other concepts (e.g., vegetation barriers to influence the movement of shallow fog, monolayers to inhibit evaporation from water reservoirs, use of helicopters to mix drier air with fog) hase limited application and may be tailored to speci6c types of highway fog. These concepts are discussed in this report.
Previously suggested vehicle guidance peedures were studied. It was determined that specially designed, lights mounted near the road surface, producing an area of illumination directed about 110* to the direction of traffic flow can be used to effectifely provide illumination for night driving in fog.
A vehicle guidance system involving the use of polarized headlamps was evaluated in field experiments and judged impractical as an aid to drivers in fog.
' Measurements of the effect of vehicle lighting on visibility showed that rear lighting systems can be improved to allow better detection of vehicles in fog.
CHAPTER oNE INTRODUCTION AND RESEARCH APPROACH s
The occurrence of dense fog on highways a an obvious 3. Determine the effects of day and night fog levels on hazard capable of impairing the safe and efficiert operation driver performance and traffic operations.
of motor vehicles. Attempts have been made to prevent or 4. Explore the feasibility of warm and cold fog abate-abate fog and to improve visibility and guidance through ment and of vehicular guidance systems under highway fog, but to date no satisfactory solution to this problem has conditions.
been found. 5. Test fog abatement metho's o and guidance systems The project has had the fo!!owing objectives: under highway conditions.
- 1. Review past and current research of warm and cold 6. Suggest ways and means of obtaining maximum ef-fog as it affects highway operation. fectiveness of systems to combat reduced visibility due to -
- 2. Prepare a state-of-the-art summary of fog abatement gog, procedures, guidance systems, measures of visibility, and effects of fog on tra#ic operations. This report summarizes the research findings.
t-
w ,
2 CHAPTFM TWO RNDINGS LITERA1URE REVIEW AND STATE-OF.THE ART Guidance Systems Designed to Reduce the Fog Hazard .
SUM M ARY A review was made of a wide variety of vehicle guidance A re iew of pcrtinent literature was made to assess the Systems propmed for use during corditions of reduced siss bih. ty. In general, procedures using artificial light ng ha7e state of the art of fog abatement, methods of'schicle guidance in fog, and past and current research related to been most effective m improvmg a drner's perception an thz effect of fog on highway operations, in preparing the f g. Visibility m both day and mght fog has been shown to summary, consideration was given to several other areas be dramatically improved when artificial lights are em-directly related to highway fog. For example, supplemen- pl yed as (1) lane delineators and (2) specially constructed 1:ry discussions of the physics of fog formation, visibility street lamps which have been modified to produce a narrow in fog, and a fog climatology for the United States are pre-beam of illumination. De use of street lamps having a sented in the summary with the intent of providing a "*rr w beam spread, with the beam perpendicular to the driver's Ime of sight,' was studied by Pritchard and Black-rather broad back Eround to the researc'her interested in the well (1959) and Spencer (1961). Field studies by Spencer hi hway fog problem. The completed summary is pre-on a section of the New Jersey Turnpike demonstrated that sented in Appendix A of this report. (A listing of all refer- ,
lights placed at low elevations alongside the roadway were ences used in preparing the state.of.the. art review is in-quite effective in irnproving traffic operations m dense fog.
cluded as Appendix H.) Important findings of the resiew Tc5dtS Scre reported m Europe from a lighting sys-Era presented in the following.
tem which involves cylindrical lanterns placed about.3 f ENect of Fog on Accidents and Traffic Flow above ground level and producing a fan. shaped beam di-rected at about 110* to a driver's line of sight. The low He efect of fog on accidents has received only minor at- lighting system is said to have significantly improved vehicle
' IIntion in the past in terms of empirical research and is not guidance in dense fog (Illum. Eng.,1967). ..
well unt'erstood; this is due primarily to the difficulties as. In contrast, tests involving reflectorized pavement mark-sociated with the collection of valid data describing traffic ings (beaded lane delineators and edge striping) resulted in exposure to fog and nonfog conditions. On the basis of only slight increases in visibility and had virtually no erect studies of individual accidents, it can be said that fog is on traffic flow during daytime fog conditions. Interviews c:pable of inducing accident involvement. On the other with a sample of drivers revealed that at night, however, hand, results of an investigation in metropolitan Melbourne the reflectorized treatment alone provided as much informa-(Foldvary and Ashton,1962) showed that fog significantly tion as full overhead lighting.
reduced the probability of casualty accidents. These find. A guidance system involving polarized fog lamps (Na-ings suggest that fog has mixed efects in the sense that it than,1957) was evaluated by the research agency in the
, ia a hazard with which drivers attempt to cope, but not all laboratory and in field experiments. De concept was cre successful in doing so. shown to be impractical for use on highways. Additional In a sample of accidents on California freeways (John- discussion of the field tests appears in " Vehicle Guidance son,1965), the fatality rate in fog was essentially twice as Investigation"in this chapter.
high as in nonfog conditions. In another California sam- Very recently, a gated laser technique has been advanced pie [ Reduced Vitibility (Fog) Study (1967)],* it was
- by Laser Diode Laboratories (Prod. Eng.,1969), for use f und that accidents occurring in fog had an increased ,
during periods of poor visibility. The hand. held system em-probability of involving either a single vehicle or more than ploys a pulsed gallium arsenide laser that emits in the in-
- three vehicles. Based on these studies it is reasonable to frared and receives reflected light from objects on a photo.
conclude that fog acts to modify the nature of acciden*s. ' cathode image converter tube. At present, the system is Data obtained on selected sites on. California freeways said to have a range of only 300 ft, although work is con-and expressways in both day and night conditions indicate tinuing to develop a capability for greater range.
that the efects of fog on traffic flow are not large. Mean Research also has been conducted in the area of traffic .
speeds were reduced by only 5 to 8 mph; speed variation control during periods of fog. In one study (RVFS,1967), ,
tiso decreased. In addition, there was a decrease in the speed limit signs were cKectively used to reduce speeds of t number of very short and very long headways. De de- vehicles by 5 to 10 mph over the reduction attributable to !
fog alone; the techmque was eNective only as long as the creased speeds and increased uniformity attest to the fact posted speeds were greater than 40 mph. The presence of that drivers recognize fog as a hazard requiring some
. . . poh,ce patrol cars has been found to be eNective m reducing increased caution m driving.
- the speed of drivers in fog (RFFS,1967). ne results revealed that speed reductions averaging 3 to 6 mph could
- Herearter referred to as JtVrJ. l B
3 be obtained by using mosing or parked patrol cars along INVESTIG4 TION OF THE EFFECTE OF FOG ON TRAFFIC the roadway. In night fog, however, onl y parked patrol The changes in trafhe associated with reduced sisibility due cars had an effect on average vehicle speed.
to fo are not easily predicted. The difficulty arises pri.
Fog Abatement Systems marily from the ability of drivers to compensate for im-paired sision. Specifically, reduced visibility tends to de-As part of the literature review, an examination was made crease the driver's ability to coordinate schicular control of warm and cold fog modification concepts to determine with ensironmental demands. On the other hand, the driser whether, and to what extent, previously suegested abate- can be expected to exercise more caution; this iracrened care ment systems could be used on highways to allesiate the may or may not result in osert changes in vehicle move-fog probtem. Of those considered, four concepts were ment. For some measures of vehicle dynamics and their, judged more promising than the others. They are: relationships with the environment, the two effects can be expected to essentially cancel one another so that no
- 1. Fog seeding.-Briefly, this concept insohes dissemi-changes occur. For example, drivers m fog may desire to nating prescribed amounts of carefully sized hygroscopic rnamtam unif rm traf5c flow and yet be unable to see well materials into fog to cause a favorable redistribution of drop en ugh to do so. Although such cancelling effects need not sizes, which resuits in improved visibilit'y. At present, a seeding tec;inique, which was developed at the research *I* *Y5 ***"'., th,5 Potential interaction is capable of pro-
%cmg situatmns m which predictions of change, and even agency, is being used at some airports in the United States
- "* "5 I 8re Simply not available.
for warm fog modification. The technique is described in Before proceed. "E.th a discussion of the observed ef-mg wi
" Fog Abatement Investigation" in this chapter. Results of the CAL NASA fog seeding tests are presented in Ap- I* ' II 8 " I'* ** 'I 5" * **'I IIhe experimental.
design is given. Data were collected on s.ix different days pendix B. The concept may be useful for general highway n a f ur-lane, divided, rural highway, near Dmira. N.Y.
application. provided that reliable and easily maintined The analyses were performed primarily ,m the context of a equipment can be installed along the roadaay. Equally tw -way lay ut m which any particular variable could be important in the efficient use of this technique wi!! be the studied as a function of visual distance for a given time of development of improved methods of disseminating the ,
day. The visual distance was taken as the maximum di.tance sized hygroscopic nuclei into the fog.
fr m which the rear of a vehicle could be seen. (The ra-
- 2. Helicopters.-Recent experiments conducted by the Air Force (AFCRL) have shown that it is possible to dis- tionak {or choosing this measure of visual distance and the
- ethod of its determination are discussed m Appendix F.)
sipate 1,000-ft-thick stratus clouds by using very large heli-V.isual distances for purposes of analysis were grouped into copters to mix the saturated air with drier air from aloft 100-ft intervals. Time of day, m intervals of I hr, was (OAR Res. Review,1968). The technique can be used chosen as a control variable, so that the effects of changing equally we!I for dissipati. ; certain types of radiation fogs that are capped by warmer drier air. The system requires yisual distance could be studied with,m each such time interval. Selectmg time of day as a control variable mmi-that the humidity of the drier air be 90% or less, because mized contammation of the results due to changes m driver mixing nearly saturated air with the fog can cause the cloud Population and changes m traffic volume.
to become more dense. It is possible that, for highway Data describmg volumes, sample sizes, and the variation application, effective dispersal of fog could be achieved by f volumes w,i thin time mtervals are given in Appendix F.
seeding from the air using helicopters to disseminate hygro-scopic materials. The benefits of both techniques might . Only those variables having essentia!!y monotonic rela-ti nships with visual distance are discussed m this section. ,
then be realized Other detailed analyses, which include the study of lateral
- 3. Forest stands and vegetation barriers.-The proper ,
Pl acement, use of passing lane, time headway, vehicle spac.
use of vegetation to block movement of shallow fog from higher elevations to low-lyir.g surrounding areas appears to ing, a measure of th timg withm which speed or lane must be changed to avoid colhston, and the variance of all be-be applicable to many highway fog situations. Becaus'e no havioral measures, are treated in Appendix G. It is im-maintenance or operating costs are involved, implementa-tion of the concept would be inexpensive. Airborne surveys Portant to recognize, however, that although no syste,matic changes were found between fog density and the variables could be conducted to determine regions where vegetation -
discussed in Appendix G, the findings are germane to the barriers might be used to influence fog movement.
Program objectives. In Chapter Three the sigmficance of
- 4. Supercooled fog dispersal.-The dissipation of super.
the findings is evaluated.
cooled fog is an operational reality at many airports. For those areas of the United States where the incidence of speed supercooled fog is sufficiently high (e.g., Ncrthwest states), .
standard cold fog seeding techniques can be employed to The effect of fog on dr.iver speeds is plotted m. Figure 1.
cause dissipation over heavily traveled sections of highway. The data show that for the 7 AM (7 to 8 AM) and 8 AM (8 to 9 AM) periods there was a general tendency for speeds
~Ihe same problems that limit the hygroscopic nuclei seed-to decrease as visual distance decreased. Notice, however, ing technique also hamper etlicient use of this concept (i.e.,
that the speed reductions are not large, being m the neigh-installation, maintenance and operation of the equipment).
borhood of 4 to 5 mph. It can be seen that reduction m At the end of this chapter recommendations are made speed starts to occur well before visual distances are as low for use of fog abatement techniques on highways. as 500 ft. According to the AASHO Geometric Design
~
~
4 ,
S . t (1965), the stopping sig1-t distance for 65 mph on a dry in this analysis the probability of high speeds was investi.
road is 489 ft, allowing 2.5 see for driver perception and gated. There was some difficulty in selecting that speed reaction. (The road surface was dry during all data collec. which was to be used as the threshold of risk-taking. If an tions made in the program.) Thus, drivers begin to reduce extreme value had been used, the numbers of drisers ex.
speeds in response in reductions in visual distances which ceeding that value would have been low and relizbility are greater than stopping sight distances, would hase become questionable. If a value wiJch was not extreme had been used, the element of risk could have been Speed also was studied for a subsample of vehicles lost and the results would have tended to simply reflect the (which will be referred to as isolated vehicles). This sub.
sample included only those vehicles for which the previous curves obtained by plotting means. Therefore, the thr*cshold schide, in the same lane, was estimated to be at least value was chosen so as to minimize the characteristics of M mile away. These isolated vehicles, then, were virtually either extreme.
unconstrained by traffic with regard to speed; they also Figure 3 shows the probability that speed exceeds 60 mph were provided with only minimal guidance by traffic ahead. for the various visual distance intervah. It can be seen that Figure 2 shows the mean speeds of the isolated vehicles the likelihood of high speed decreased as visual distance as a function of visual distance. Although the. numbers of dropped. In the 200. to 300-ft range there is a total of observations are not high, the mean speed of the vehicles 12 drivers exceeding 60 mph. Although this frequency is follows essentially the same trends observed for those of the low, it is important to point to this evidence that some cimplete sample. The main difference is that the isolated drivers were willing to tolerate the inherent risks (ie., even vehicles averaged approximately 3 mph higher speeds. if most of the drivers had chosen to drive slowly, some A second type of analysis was ap;illed to study poten. did not),
tially hazardous conditions as a function of visual distance. Stopping and Stopping Sight Distances One measure of safety of traffic flow is the probability.that the speed and visual range relationship is such that a driver could not stop his vehicle before striking a nonmoving object in the road. This occurs when the distance traveled
~ ~ ~ '
during the time required to observe, perceive, make a de.
~~~
cision, react, and decelerate 'to a stop is greater than the
-u- ---
!. visual range. This act of overdriving is relevant only to j . . . .
those situations in which the driver might come across an
"= - object with essentially no velocity in the direction of traffic flow. This is normally not the situation; however, it is
- - - probably the mechanism whereby foggy weather accidents involving few vehicles are extended to accidents involving I
= = .= = = = **= "= many vehicles.
mim mima t""I , , , ,
Because the time required prior to overt response can be
....... estimated only roughly it is difficult to determine which formulation of stopping sight distance to use. For that Figure 1. Mean speed as a funulon of visual distance.
matter, simple response time itself is not available in terms of a single precise number. Furthermore, one might well expect drivers to be more attentive in fog than in clear weather conditions, thus rendering clear weather times invalid. In view of these problems, two measures of the e .
u - .... ,,, .
~~~~
-k E o . . . . .
[ .. . .
w - - . . _ _ .
m m me we = = = = ' * = * ' = , n .. . .
mm uum, gu n, nem mwa tunt t ..a.
.......... ---. g Figure 2. Mean speed as a function of visual distance for Figure 3. Probability that speed is greater than 60 mph.
isolated vehicles.
i .
V
, 3
,. .:e.
V -
. 5 TABLEI PROBABILITY OF OVERDRIVINO AS A FUNCTION OF VISUAL DISTANCE raosannuTv
- 4LL VEHICLES ISOLATED YEHICL ES I Tutt 201- 301- 401-' 201- 301- 401-ITEM -
(4M) 300Fr 400rr 500 rr 300rr 400rr 500 rr S10 7-8 0.05 0.01 0.00 0.14 600 0.00 8 0.25 0.02 - 0.32 0.04 -
525 7-8 0.82 0.50 0.00 1.00 0.80 0.00 8-9 0.88 0.39 - 1.00 0.58 -
No.of observations 7-8 141 155 202 7 10 24 8-9 185 230 0 28 26 0 time, and hence distance, required to stop have been used TABLE 2 in the analysis of fog data. All of the stopping distance PROBABILITY THAT A LEAD VEHICLE -
formulas are based on informition given in the AASHO WAS WITHIN .VISl; AL DISTANCE Geometric Design (1965). The analyses determined the raoanatuTv probability of overdrising, with stopping distance based on (1) 2.5 see for ** perception" and reaction time, plus brak. [,y,$c,g,7) 7,g ,, ,,9 , . ,
ing distance (labeled S25), and .(2) 1.0 tee for reaction 201-300 0.45 0.31 time plus braking distance (labeled S10). 0.50 0.30 301-400 Table I gives several relevant factors. As expected, the -
probability of overdriving increased as visual distance de- _
creased. . Second, because it has been shown that speeds 601-700 - 0.53 were higher for isolated vehicles, it is not surprising that 701-800 - 0.52 801-900 0.73 -
the probability of overdriving tends to be greater for iso. 000 0.7 lated vehicles. The most important fact revealed by this -
'g ,
analysis'is that the probability of overdriving in the low visible distance intervals is substantial. It is seen that one- 73,,, , i ,,, n ,,, ,, ,, ,,,,,,, ,,,,,i oi,,,,,, ,,,,, ,, y mii,,
fourth of the drivers are overdriving their visual ranges when in dense fogs.
A separate analysis was performed to measure the mean difference between visual distance and stopping sight dis-tance for isolated vehicles. For most fog :evels mean visual range exceeded mean stopping sight distance. In only two vehi.les. Reductions in speeds were observed even when cells was this not true. The mean of S25 minus visual dis-visual distance exceeded 500 ft. ~
tance was -87 ft in the visual' distance interval of 201 to 2. The probability of overdriving increases sharply as - '
300 ft, and -45 ft for the 301- to 400.ft interval (mean visual distance decreases. This probability is higher for -
stcpping sight distance exceeded mean v,sual distance only i
isolated vehicles.
for the 7 4w data). . 3. The likelihood of driving without a lead vehicle in '
For purely descriptive reasons the probability that spac- sight reached 55 to 69% for the denser fogs.
ing was less than visual distance was computed for each 4. No ' meaningful relationships were found between visual distance range. This probability is equivalent to the visual distance and (I) use of the passing lane, (2)s lateral probability that the lead vehicle could be seen. Results are position. (3) probability of short headways, (4) vehicle given in Table 2. . The data show that as visual distance spacing, and (5) collision course time (see Appendix G for dimimshes, the likelihood of dnymg without being aware definition).
of, or getting cues from, a lead vehicle becomes substantial.
It should be mentioned that although the results pre-Summary of Findings for Effects of Fog on Traffle sented in this section are based only on daytime data, the night and dawn data were similarly analyzed, in spite. of The following summary of findings is based on the analyses their small sample sizes, with the result that virtually no described in this section and the results discussed in contradictions with the daylight findings cccurred.
Appendix G. ,
- 1. Speeds in daylight fog tend to decrease by approxi- j FOG ABATEMENT INVESTIGATION mately 4.5 mph as visual distance decreases from more than 1,000 ft to approximately 250 ft. This result holds for j The principal objectives of the fog abatement study have th(. complete sample as well as for a subsample of isolated been twofold:
- t-22 - .
stitId that their first ef.!ision was with Ge vehicle ahead, THE EFFECT OF FOG ON TRAFFIC ;
% 22, or 61%, said this schicle was in sis'.: before the inci-dent occurred and did not " suddenly a pear." His sug. The only empirical sto*Jy of the c6ects of fog on traffic it w known to the smarchers is the RVFS (1967). In it, gests, perhaps, that o:e major problem is associated not -
the eHects of fog on speed and headway were studied on
? with seeing the sehicle ahead, but with be*ng unable to see thead cf him and, thus, being unable :o anticipate his freeways and esprenways in the Sacramento area. The :
data showe.d that, in general, fogs (all levels-day and -
actions. Another proMem might be in de inability of a mght) tended to reduce mean speeds by 5 to 8 mph. ,
driver to detect the me e subtic actions cf the lead vehicle.
The report also stated that there was no general e6cet of
' Miller also presented results cross-i.-feming estimated .
visuil range and speed prior to impact .~he researchers of fog level on speed variation. Howeser, the data on which i
-- this report used those speeds to determine required stopping this conclusion was based included observations on roads distincts with the aid of the AASHO Geometric Design where experimental speed limit signs were in use. Because these signs could well have influenced vehicle speeds, and, (1965); perception time, one of the fa:: ors in total stop.
ping distance. s-as set to zero on the assu. .ption that drivers hence, speed distributions, it was decided to resiew only -
those data collected in places where the normal California in fig tend to pay close attention. Of tl e 37 vehicles that ,
could be judged. 25, or 68%, had speeds requiring stopping maximum of 65 mph was in eNect.
' dist nces that exceeded the visual range. In this context, Table A-3 gives the standard deviation of speeds as a howIvIr, Mi::er observed, on a nonfog night on the same function of visibility. Each column in the table represents road, that speed and vehicle spacing was such that headways a separate data collection from a diNerent combination of were also ins :.*. icient for emergency stopping. Again, per. . site traffic volume lesel, and/or day. night condition. The haps this ind%tes the iraportance of the driver being able diferences between columns are not important, because -
ta see beyond the vehicle immediately'in front of him. It attention is focused within columns.
. is interenin; nat during the nonfog obse vations there was, It can be seen that in all but two columns (I and 7) in fact. a 13 schicle accident. Therefore, it cannot be. there is an obvious trend for decreased speed variation with Ldetrrmined ether the nonfog speed ar.d spacing observa. decreased v.isibility. With regard to the data in column 1, tions w ere r: rnal or were more representative of the un- a line fit to the four data points would reveal a slight trend usually unsaf s :ondition. in the same direction. Ignoring the strength of the trend for each column, the probability, assuming independence of visibility and speed variation, of observing eight or more trends in the mentioned direction is 0.02. This suggests a rejection of the hypothesis of "no trend" in favor of a hypothesis of " decreased variation in speeds as visibility
~
TABLE A.3 decreases."
STANCARC DEVIATION OF SPEEDS With regard to headways, the RFFS tends to show a
- AS A F~JSC lON OF VISIBILITY relatisely smaller number of short headways in foggy condi-r aNoaan osv:4T3ON, av DATA COLLECTION tions (RVFS, Table 13). It should be noted, however, that
- inasmuch as speeds are reduced in fog it is possible that the 2 3 4 5 6 7 8 9 relative frequency of short vehicle spacings is not materially f 0-10< 7.1 3.5 7.2 changed.
On the other hand, if one compares the headway data 4.s 3.9 5.1 5.2 in Figures 17 and 18 of the RVFS the approximate data in 200-SC: %. 9 5.2 7.7 '
i<L400-56*. 6.5 Table A-4 can be obtained.
500-76*. 4.0 6.7 6.4 Now, for example, considering the 85th percLntile head.
day at 2,400 vph, one sees that in fog 15% of the drivers 5 m. 6 6.1 7.4 had headways over 6.3 sec. However, in clear weather 700+ 5.4 5.6 5.8 6.5 7.6 15% of the headways were over 7.6 see; thus, for this condition, well over 15% had headways greater than 6.3 .
sec. It can be concluded that there were more long head- .
ways _in clear weather than in foggy weather. '. Dis state-
~
ment is shown to hold for all three volume conditions and
- -TABLE ^" at both the'67th and 85th percentiles. Thus, at least at the HEADWAYC, IN SECONDS high volume levels, these data show fewer long headways in fog; this is in agreement with a hypothesis that in' fog usAnwAYS (SEC) there is a tendency for drivers to shorten headways to 857H reacsNTits maintain visual contact with the vehicle in front.
! 677x enacsNTits It should be pointed out that, givert a Exed volume, the votuur. vr= d fog CLEAR FoO CLEAR
'II,200 4.3 5.0 3.9 '> 12>.1' events. As the number of short headways increases, so i 1,300 ' 3.3 3.s 6.7 B.8 ".'.W, must the number of long headways. Hus, it may be that,
.2,400 3.1 3.4 6.3 7.6 n,el*in fog, tailgating is unacceptable to the driver; as a result,
. 2.Y.lo
. 23 schicles sprnd out more, grouping becomes less evident, The eNectiveness of this type of experimental treatment and very Ices headways decrease in number. Or, it may for guidance in fog was tested in the RVFS. Ref!cetorized be that, in f:g. vehicle spacing exceeding visual range be- paint was applied to edge and gore stripes on the entrai.ce _ '
comes unacceptable; as a result, spacing becomes more (bine paint) and exit (yellow paint) ramps of a closcrieaf even, and.sbrt headways decrease in number. The true interchange. Trafvic movement (encroachments on or mechanism . educing the numbers of large and small head. across the gore and edge striping) was obsersed both in ways therefe e could not be determined from these data. fog and cicar weather before and after application cf the
- Changes 1 vehicle lateral placement, as a result of fog, experimental treatment. It was hypothesized that, if guid. ,
', were also 1:sestigated. .Two sections of a state highway ance was improsed by the r,cflective paint, there would bc (curve and nraight) were selected for the tests. Observa. less encroachment on the edge and gore stripes. Because
.(
tions revea! 3 that for the curve test segment, the propor. of a lack of fog, and the observation t!)st vehicles tend to ,
tion of veh :les crossing the center line during fog (day cross the edge stripes to reduce the sharpness of the on and night-ali densities) was greater than during clear and oK ramp curves, the 6ndings were inconclusive, weather. H: wever, these results were not considered sig- Warner (1958) recommended changes in the type as i nincant (P =,0.05). Furthermore,-there was no change in color of various highway markers (signs, Iane and edge encroachme:t on the pavement edge (all conditions tested). delineators, etc.). He claimed that primary emphasis must . r be placed on maximizing the contrast between the highway GtJtDANCE SYSTEMS DESIGNED TO ELIMINATE marker and the roadside environment. Even though a fog THE FOG KAZARD . , environment was not considered specifically, some of his suggestions might prove applicable and should be tested.
A variety of methods designed to aid the driver during For example, the use of two-tone redectorized guide posts periods of reduced visibility has been proposed and experi. might prove to be an eNective guidance system in both day mentally tes:ed.- Beaton and Rooney (1966) evaluated the and night fog conditions. Forbes, Pain, Fry, and Joyce use of dine:ent types of raised reflective lane delineators (1967) confirmed the influence of environment on sign (beaded ar d nonbeaded buttons and wedges, raised reflec- visibility: subjects observed sets of signs of varying bright-tors enclosed in acrylic plastic) for a variety of weather . ness against diNerent backgrounds. The subject's task was conditions (day-rain, dry; night-rain, dry). They found to Pi ck the sign, from each set viewed, that was recognized that, at night, reRex reflector and bended pavement markers first. The results revealed that signs highest in brightness provided the best delineation. However, during the day were seen most frequently against a night and day hill '
these markers proved unacceptable because of their tend. (daylight with a hill covering a large portion of the picture) ency to blend with the pavement. background. Conversely, signs lowest in brightness were
' Die usefulness of the bended type of lane delineator for seen most frequently against a day-snow background.
guidance in fog was investigated in the RFFS. The marker Pavement edge striping is another guidance technique was tested on a California state highway during daylight that has received considerable attention. Basile (1961) !
hours only. Judgments by trained observers indicated that designed a study to determine the eNect of the use of this '
the improvement in visibility (all fog densities) was negli- technique on accident rate. Weather was not included in gible. However, Beaton and Rooney showed thst the the experimental design. Basile found that edge ' striping beaded marker oNered the pentest visual improvement had no eNect on the number of accidents occurring on the '
during the night. Therefore, before any conclusions can be oPen road. However, at intersections and driveways, the drawn concerning the usefulness of this type of marker for guidance in fog, nighttime tests should be conducted. accident ing was present.ratesThese wereresults significantly reduced were confirmed when by Music The use of renectorized paint on both the pavement and (1961). The eNect of edge striping'in fog was studied in roadside signs is another technique that has been proposed the RVFS. Edge striping was installed on two test sterions to increase the driver's visual range. Huber (1961) used (curve and straight) of a state highway. The lateral place-
- this technique to increase visibility and provide informatica ment of a sample of vehicles was used as a meabre of the at a freeway interchange. He painted the on-r' amps and eNectiveness of the striping. It was hypothesized that, if acceleration Ianes a reflective yellow and the exit signs, oN- the striping was eNective, drivers would be able t6 maintain
} ramps, and deceleration lanes a reSective blue. By varying their vehicles near the center of their lane, thus reducing ,
the illumination from existing overhe'ad streetlights he was the probability of conniet with vehicles in the adjacent lane I . able to study the efects of five levels of day and night *or collisions with obstacles alongside the road. Observa.
visibility. His observations were made only during clear tions were made in various fog densities and cicar weather and dry weather and indicated that none of the experimental both before and after edge striping. The results indicated test conditions had any efect on the flow of traffic (vehicle that edge striping had no eNe~ct on vehicle lateral placement headway, speed or placement) through the interchange. (all test conditions). However, interviews with a sample Interviews with a sample of drivers revealed that, at night, of drivers revealed that the edge striping gave them a feel.
the reflectorization treatment alone provided as much in. . ing of increased safety, especially during dense fog condi-formation as full overhead lighting. Similar studies (Roth tions.
and DeRose,1966; Darrell and Dunnette,1960; Fitzpatrick, Guidance systems using artificiallighting have proved to 1960) using reflectorized paint as a means of illumination be eNective in combatting the hazardous eNects of fog.
have revealed essentially the same results. Finch (1961) proposed the use of small lights as lant and