Skip to main content

Full text of "All-red signal intervals in San Francisco : a progress report"

See other formats


SAN FRANCISCO PUBLIC LIBRARY 



3 1223 06957 0050 

letMiny uivisiuii 

nty of ^an Francisco 

OWN, JR., MAYOR 

], EXECUTIVE DIRECTOR 




5/5 a' Intervals in San Francisco 




cisco Public Library 

taient Information Center 
incisco Public Library 
'kin Street, 5th Floor 
ncisco, CA 94102 

51ENCE BOOK 

taken from the Library 



A Progress Report 



DOCUMENTS DEPT. 

SAN FRANCISCO 
PUBLIC LIBRARY 



^Department of Parking and Traffic 
May 1, 2001 



FAX (415) 554-2352 



25 Van Ness Ave., Suite 345 



San Francisco, CA 94102-4576 



Traffic EngiiietJiiny uivisiuii 

City and County of ^an Francisco 

WILLIE LEWIS BROWN, JR., MAYOR 

FRED M. HAMDUN. EXECUTIVE CIRECTOR 



All-Red Signal Intervals in San Francisco 

A Progress Report 



DOCUMENTS DEPT. 
MAY | 8 200! 

SAN FRANCISCO 
p UBLIC LIBRARY 



^Department of Parking and Traffic 
May 1, 2001 



41 5) 554-2300 FAX (41 5) 554-2352 



25 Van Ness Ave., Suite 345 



San Francisco. CA 94102-4576 



Introduction 

This is a status report about the implementation of all-red signal intervals at 
signalized intersections in San Francisco. All-red signal intervals are the brief, 
simultaneous display of red lights in all directions. They typically last from 1 to 2 
seconds and are intended to provide some time for vehicles to clear the 
intersection before the onset of the next vehicular or pedestrian movement. As 
such, their primary purpose is to act as an additional safety measure at 
signalized intersections. 

To date, San Francisco Traffic Engineers have implemented all-red signal 
intervals at 235 of its 1067 signalized intersections. 

This report will discuss the practice of using all-red intervals, provide a listing of 
those intersections with all-red intervals, and discuss the before-and-after safety 
records of those intersections provided with all-red intervals. 

The Practice of Using All-Red Intervals 

In 1980, The Federal Highway Administration published a report entitled A Study 
of Clearance Intervals. Flashing Operation, and Left-Turn Phasing at Traffic 
Signals . The City of San Francisco has generally used this document, as well as 
the California Traffic Manual , as guides for determining where all-red signal 
indications should be implemented. 

According to this first report, all-red intervals are recommended at intersections 
where the right-angle accident rate is greater than 0.8 accidents per million 
vehicles entering the intersection 1 . 

The Traffic Manual has this to say about all-red intervals, which they term "red 
clearance intervals": 

"Generally, red clearance intervals are not required. A red 
clearance interval may be used following the yellow change 
interval, at very wide intersections, offset intersections, or at other 
intersections where it is desirable to delay the green interval for 
opposing traffic. Normally, red clearance intervals range from 0.1 
second to 2.0 seconds." 2 

Some cities adopt a practice of using all-red intervals at all intersections. Such a 
policy may be adopted for the sake of consistency or to simply avoid the question 
about why certain intersections do not have all-red intervals. 



1 Barry Benioff and Thayer Rorabaugh, A Study of Clearance Intervals, Flashing Operation, and 
Left-Turn Phasing at Traffic Signals . FHWA-RD-78-46, May 1980, p. 22. 
Traffic Manual . California Department of Transportation, January 1992, p. 9-34. 



3 1223 06957 0050 



In San Francisco, there are several reasons why the Department of Parking and 
Traffic has taken a selective approach on the implementation of all-red signal 
intervals. Most importantly, the Department of Parking and Traffic believes that 
the power of all-red intervals as a safety measure depends upon their selective 
application. There is a concern that their overuse can dilute the warning power 
of the yellow interval as a cautionary signal. The fear is that if drivers know that 
an all-red interval follows the yellow interval, drivers would be more inclined to 
enter the intersection during the yellow interval, especially towards the later 
portion of the yellow interval. It is believed that this is riskier behavior, which may 
increase the chance of a traffic accident. 

There are also some capacity reduction consequences involved with the use of 
all-red signal intervals. All-red intervals are a period of time when no traffic 
movement is supposed to occur. The use of all-red intervals at a typical 
intersection in San Francisco can reduce the traffic-carrying capacity of an 
intersection from 2% to 6% 3 . At bottleneck intersections, this can be a serious 
reduction in traffic capacity, causing traffic back-ups in which vehicles are not 
able to clear an intersection within one signal cycle. Congestion itself also has 
safety consequences, with cars backed up across crosswalks, along with other 
environmental problems, such as air pollution and gasoline waste. 

All-Red Intervals in San Francisco 

Table 1 shows a year-by-year breakdown of all the intersections in San 
Francisco with all-red intervals. The first to be given an all-red interval was the 
intersection of Columbus Avenue/Green Street/Stockton Street in 1953. Up until 
about 1985, very few intersections had all-red intervals. After that time, more 
and more intersections have been given all-red intervals, particularly when entire 
traffic signal systems have been re-timed. To date, a total of 235 intersections 
have all-red intervals in San Francisco. 

Accident Experience with All-Red Intervals 

The Department of Parking and Traffic analyzed its experience with all-red 
intervals in a report entitled "Evaluation of All Red Phase in Reduction of 
Collisions," dated May 12, 1997. That report lists the following findings and 
recommendations: 

> "A literature search revealed that all-red phases have been effective in 
reducing the number of right-angle accidents." 

> "A before-and-after study of intersections in San Francisco which received all- 
red in 1987 shows a statistically significant reduction in the number of injury 
right angle accidents at treated intersections compared to intersections 
without the all-red." 



3 Assuming a 60 second cycle length and a 1-2 second all-red interval after one or both traffic 
signal phases at a two-phase signal. 





55 2 t _ 

Miff!! 




1 



t 

I* 



8 § o 8 

o 5 ■ c 

11*111 «S J333352S253«!2^i2S2le««tt5SS2«!5^S2« 5? « 



5 s |If!ll||IIlil!l!ll 

^ 5 s N To VTo vi o> s- N < < m m m co m m co o i 



2 L «2I 



oil- *J ipiliili p lilii 



.5 cecals &f i § S « Io9:?.e f "I-s elsf i issSa? 

o 2f$i>! slllll^ll^i!ii5lill $1111*^*11 

O •EJ-SSl SCCCrc^})^ ■ S 3 cS E E 2 S o « E EE E J J 

'5 
c 
a 
l> 
u. 

c 

IB 
W 



re ^ -g 



S "5. 2 £ S S » 



* i ill* illll iillil II Mlii 



8 




*>i 4) fl) I) 41 fl) i. ^ 3 




* ■ c u 
cf " 




o 



5 I*. 3 r- 



C c 



il 



§>° 

ra $ 



5 Ss 2s§ 2 e f | s § g|§ 8 1 1 




> "No statistically significant reductions from all-reds were measured for non- 
right angle accidents or for accidents involving pedestrians." 

> "All-reds implemented in 1987 appeared to remain effective over a seven year 
period in reducing right angle accident totals when compared to intersections 
without the all-red." 

> "All-reds should be between one and two seconds for most applications." 

> "The results of this study suggest that, given the all-red's effectiveness in 
reducing right angle accidents, all-reds should continue to be implemented at 
intersections with a high rate or total number of right angle collisions." 

As a partial update to this earlier study, this report examines the accident 
experience of intersections receiving new all-red intervals using available data for 
the five-year period of 1 995 through 1 999. Three different before-and-after 
comparisons were made: one for all-red intervals implemented in 1996, in 1997, 
and in 1998, respectively. 

Table 2 shows the before-and-after right-angle accident results for the ten 
intersections receiving all-red intervals in 1996. The "before" results reflect data 
from 1995. The "after" results show data in each of the years 1997, 1998, and 
1999, as well as an "after" average. 

Table 2. 1996 All-Red Installations 



Before and After Resu 


ts-Rig 


ht-Angle Collisions 


Intersection 


1995 | 1997 
(Before) | (After) 


1998 
(After) 


1999 
(After) 


'97-'99 
(Avg) 


Change 


l^/Howard 


12 


9 


9 





6 


-6 


4 m /Mission 


10 


7 


5 


8 


7 


-3 


1 ^/Harrison 


2 


3 


6 


6 


5 


+3 


Beach/Stockton 




















Embarcadero/Howard 








3 





1 


+1 


Embarcadero/Jefferson/Powell 




















Embarcadero/Mission 




















Embarcadero/Stockton/Pier 39 




















Mission/Otis/S. Van Ness 


2 


4 


1 





2 





Mission/Silver 


3 





1 


1 


1 


-2 



According to these results, three intersections show a reduction in the number of 
right-angle accidents. The intersection of 1 st and Howard Street has experienced 
a remarkable reduction. Two intersections show an increase in the number of 
right-angle accidents. Five intersections show no change. These results seem 
to support the findings from the DPT and FHWA studies cited previously: all-red 
intervals show a demonstrable safety benefit when used at intersections with a 
high number of right-angle accidents. For those intersections without a clear 
right-angle accident problem, the use of all-red intervals does not show a 
conclusive safety benefit. 



Table 3 shows similar results for 1997 all-red intersections. 



Table 3. 1997 All-Red Installations 



Before and Af 


ter Results-Righl 


-Angle Collisions 


iniersccxion 


1 QQ^ 
fRpfnrp^ 


1996 
(Before) 


'95-'96 
(Avg) 


(After) 


1 QQQ 

(After) 


'Oft 'QQ 
BO" 33 

(Avert 


v->nange 


2" a /King 


1 

1 


3 


2 


n 




w 


n 


_2 


3"VKing 




w 





11 


n 

V 


ft 


+ fi 


4 ,h /King 




| 1 


4 


1 


3 




5"7King 











1 

1 


1 


1 

1 


+1 


5 m /Mission 


4 


10 


7 


7 

f 


s 

w 




_1 
i 


^/Mission 


4 


2 


3 


3 


2 


3 


n 

V 


1Cf/Mission 


4 


5 


5 


4 


6 


5 


w 


22^/30. Van Ness 


2 


8 


5 


1 


1 


1 

1 


_4 


Alemany/Congdon/Justin 


2 


2 


2 


4 


2 


3 


■ 1 


Alemany/Lyell 


1 





1 


1 


o 


1 

1 


o 


Bay/Fillmore/Cervantes 





1 


1 


o 


o 


o 


_1 


Beach/Grant/Embarcadero | 








1 

1 


1 


1 

1 


+1 


Brannan/Embarcadero 


2 





2 


n 

V 


1 


1 


1 


Broadway/Franklin 


1 


4 


3 


1 


4 


3 





Broadway/Van Ness 


6 


8 


7 


7 


2 


5 


_2 


Bryant/Embarcadero 





1 


1 


o 


1 


1 


o 


BusnA/an Ness 


2 


| 2 


3 


6 


5 


+3 


CaliforniaA/an Ness 


2 


4 I 3 


3 


4 


4 


■ 1 


Cesar Chavez/Mission 


6 


4 


5 


4 


2 


3 


_2 


Cesar Chavez/Valencia 


1 


2 


2 


2 


o 


1 




Chestnut/Columbus/Taylor 











o 


o 


o 


o 


ClayA/an Ness 


1 


2 


2 


2 


o 


•j 


_1 


EddyA/an Ness 


2 


1 


2 


1 


4 


3 


+1 


ElIisA/an Ness 


3 


2 


3 


1 


o 


1 


-2 


Embarcadero/Folsom 











1 


o 


1 


+1 

* 1 


Em barcadero/Towrtsend 


1 





1 


1 


o 


1 


o 


FellA/an Ness 


5 


7 


6 


6 


5 


6 


o 


Foote/Guttenberg/Mission 


1 





1 


o 


o 


o 


-1 


Franklin/Lombard 


7 


9 


8 


5 


5 


5 


-3 


Franklin/Sutter 


1 


6 


4 


o 


2 


1 


-3 


GearyA/an Ness 


2 


1 


2 


3 


3 


3 


+1 


Golden GateA/an Ness 


2 


4 


3 


1 


1 


1 


-2 


GroveA/an Ness 


1 


3 


2 


3 


2 


3 


+1 


HayesA/an Ness 


2 


4 


3 


7 


7 


7 


+4 


McAllisterA/an Ness 





2 


1 


1 


1 


1 





O'FarrellA/an Ness 


5 


2 


4 


1 


6 


4 





Pine A/an Ness 


6 


1 


4 


5 


7 


6 


+2 


PostA/an Ness 





2 


1 


1 


1 


1 





SacramentoA/an Ness 


3 


2 


3 











-3 


Stanyan/Turk 


1 


2 


2 


3 


2 


3 


+1 


SutterA/an Ness 


2 


2 


2 


2 


3 


3 


+1 


TurkA/an Ness 


4 


4 


4 


4 


1 


3 


-1 



For the 1997 intersections, 16 intersections experienced a decrease in right- 
angle accidents, 11 intersections showed no change, and 15 intersections 



experienced an increase. These findings would suggest that there are other 
significant factors besides the absence or the presence of an all-red phase that 
may lead to right-angle accidents. Consistent with the previous findings, 
however, we do see a demonstrable benefit at some of the intersections with a 
higher number of right-angle accidents. See, for instance, the intersections of 
Broadway/Van Ness and Franklin/Lombard. 

Table 4 shows the results for 1998 all-red intersections. 



Table 4. 1998 All-Red Installations 
Before and After Results-Right-Angle Collisions 



Intersection 


1995 


1996 


1997 


'95-'97 


1999 


Change 




(Before) 


(Before) 


(Before) 


(Avg) 


(After) 




Beach/Jones 


1 





1 


1 





-1 


Beach/Taylor 














1 


+1 


Embarcadero/Washington 





3 


3 


2 


1 


-1 


Jefferson/Jones 




















Jefferson/Mason 




















Jefferson/Taylor 




















Randall/San Jose 





1 


2 


1 





-1 



Here we see no remarkable changes. This may be because of the low incidence 
of right-angle accidents before the implementation of the all-red intervals. This 
again lends credence to the suggested principle of using all-red intervals where 
the right-angle accident experience is relatively high. 

Conclusion 

The City of San Francisco has implemented all-red signal intervals at about 25% 
of its intersections. The Federal Highway Administration recommends their use 
at intersections experiencing a high incidence of right-angle accidents. San 
Francisco's experience with all-red intervals seems to support this advice. 
Where right-angle accident rates have been high, the use of all-red signal 
intervals does tend to show a demonstrable safety improvement. Where the 
right-angle accident rate is low to begin with, the safety benefits are not 
conclusive. The Department of Parking and Traffic must continue to explore 
other ways to improve intersection safety.