Evidence on performance

Impacts of 20 mile per hour zones in London

The London School of Hygiene & Tropical Medicine have undertaken two studies of the impacts on collisions and casualties road of 20 mile per hour zones in London. One report focuses on the overall impacts and the other on the impacts on inequalities in the city (Grundy et al. 2008a; Grundy et al. 2008b). The studies were promoted by perceived need for robust evidence on the impacts of the 20mph zones which: ‘in London [have] increased year on year since they were first introduced in 1990/91, to a total 399 zones by 2007/08, with some Boroughs far more enthusiastic about adoption than others’ (2008a, p.5). The methodology of the studies is described in detail in the reports, and took account of characteristics of each 20mph zone and the impacts of the zone over time.

The studies analysis found “a 42% reduction (95% CI 36%, 48%) in all casualties within 20 mph zones compared with outside areas, adjusting for an annual background decline in casualties of 1.7% on all roads in London. The largest effects of 20 mph zones were found for all casualties aged 0 -15 killed or seriously injured (KSI) and for car occupants. A reduction was evident for all outcomes examined. In areas adjacent to 20 mph zones, reductions compared with outside areas were evident for most outcomes, except for those killed”  (2008a, p.6).

“The effects of 20 mph declined over time, although those implemented in the most recent years (2000-2006) still had an effect of reducing all casualties by 23% (95% CI 15%, 30%) within the 20 mph zone, and 3% (95% CI -1%, 7%) in adjacent areas,” (2008a, p.7). The study indicated “some evidence that 20 mph zones are more effective in reducing KSI casualties in less deprived areas compared to more deprived areas” (2008a, p.7).

Objective	Scale of contribution	Comment
	?	-
	?	-
	?	-
		The London study found “some evidence that 20 mph zones are more effective in reducing KSI casualties in less deprived areas compared to more deprived areas” (2008a, p.7).  This should be considered alongside matters of whether more deprived areas have higher casuality rates and so begin from a position of greater disadvantage. Further it should be considered alongside evidence on inequalities in rates of KSI faced by users of different travel modes, and potential benefits to those groups of lower speeds.
		-
		-
		The London 20mph studies found positive benefit-cost ratios for introduction in zones with over 0.7 casualties per year per km.

	= Weakest possible positive contribution,		= strongest possible positive contribution
	= Weakest possible negative contribution		= strongest possible negative contribution

= No contribution

Meta analysis of case studies

The Handbook of Road Safety Measures (Elvik & Vaa 2003), provides a comprehensive survey of road safety measures, includes meta-analysis of a large number of studies on the effects of different speed limitation and enforcement measures as well as road markings. The meta-analysis assigns statistical weights to studies by sample size and sorts them by design quality and thus gives the most systematic overview of impacts, especially on traffic safety. The evidence of performance described is mainly based on this information. For environmental impacts the TØI Environmental Handbook (Kolbenstvedt et al 1999) is used as well.

Context
The studies included are from several countries; e.g. Australia, Canada, Denmark, Finland, Germany, Great Britain, Norway, Sweden, Switzerland, and USA.

Impacts on demand
The measures in this area are not primarily intended to affect transport demand, and most impact studies concern effects on speed and accidents. Indirect impacts on demand will have to be derived from knowledge about speed impacts.

Reductions in speed limits and transition from unrestricted speed to speed limits have been introduced in a number of countries. On average a reduction of the legal speed limit with 10 km per hour will result in a speed reduction of 3 km per hour. Though many drivers exceed the speed limits, implementing lower speeds increase the time taken to travel and transport goods. By reducing speed from for example, 80 to 70 km per hour for a 60 km trip, travel time increases from 45 minutes to around 51 minutes.

Physical speed-reducing devices reduce speed. This can induce individual vehicle delays and can deter traffic, especially heavy vehicles. It has not been shown that these effects always occur. On a typical access road with a length of up to 0.5 kilometres, a reduction in speed from 35 km per hour to 25 km per hour will lead to a delay of a maximum of 20 seconds per car. It is not known whether humps create problems for winter maintenance of roads.

It has also been shown that traffic volume goes down on roads where humps are constructed (e.g. Webster & Mackie 1996). On average, the reduction in traffic is around 25% (-33%; -14%). This indicates that the actual roads had a certain amount of through traffic before the humps were constructed.

Stationary and automatic camera speed enforcement affects the speed level and gives an average speed reduction of around 2 km per hour. The halo effects in time and space show that the reductions can be maintained from 2 days up to 10 weeks after increases in enforcement procedures end. The distance-halo effect which has been demonstrated, varies between around 1 km and 22 km from the point where the stationary enforcement took place (Ragnøy 2002). It is not known to what extent drivers compensate with higher speeds outside areas which they suspect are monitored manually or automatically.

Speed camera can to some extent give raise to "kangaroo driving". This may interfere with the flow of traffic, but the effect of this measure on demand and travel choices is not sufficiently known.

Figure 2: Longitudinal speed profile E6 Hedmark. Speed limit 90km/h. Relative average speed before and after speed cameras in km/h. (Adjusted for changes in comparison sites). Change in average speed in km/h. Copyright ã TOI

Drivers who have had their driving licence withdrawn due to speed enforcement, will have their individual mobility reduced for as long as the driving licence is withheld.

Effects of road marking on speed vary, and the results differ. Both decreased and increased speed are found.

Impacts on Supply
Speed limitation and enforcement does not occupy road space, i.e. capacity.

Lower speeds may have an effect in that the average speed on the actual road is reduced, thus increasing travel time for the individual, assuming that there are no capacity problems. Lower speed can also induce congestion when capacity is limited. However, reducing average speed can also lead to less dispersion and a more even level of speed, so that the flow of traffic eventually improves and the road capacity can be better utilised. This is especially important when traffic is heavy.

Road marking aimed at reserving certain parts of the road for certain traffic groups will alter the supply between different road users.

Other Impacts – Traffic safety
The objective of speed limitation and enforcement is to reduce traffic accidents. Meta-analysis of some hundreds of studies, clearly show that this objective is achieved (Elvik &Vaa 2003). If we look at speed limitation it may be concluded that a majority of the results show reductions in the number of traffic accidents. Very often, these reductions are also statistically significant. Furthermore, these measures are the most cost-effective traffic safety measures of all (Elvik 1999 and 2000). On the contrary roadmarking measures have no statistical significant effect on accidents.

Taking all types of accidents and levels of severity together, stationary speed enforcement is associated with a 2% reduction in accidents, while speed cameras give a reduction of 19 %, both statistically significant, see table 3. As to stationary speed enforcement the effect is largest for fatal accidents, which are reduced by 14%. Speed cameras appear to have a greater effect in densely populated areas (28% reduction) than in sparsely populated areas (4% reduction). The area of effect is limited to the road where speed cameras are installed.

Table 3: Best estimate and confidence interval for accidents of stationary and automatic speed enforcement. Percentage change in the number of accidents. (Source: Elvik & Vaa 2003)

Table 4 shows that humps reduce the number of injury accidents, with a given amount of traffic, by around 50 %. Rumble strips and speed zones also have a significant positive effect, around 30% on accidents. The majority of results come from simple before- and after- studies, which have not controlled for regression-to-the-mean in the number of accidents. On the other hand, a number of studies have measured changes in both the amount of traffic and speed levels in roads where the measures have been introduced.

Table 4: Effects on accidents of speed-reducing devices. Percentage change in the number of accidents. (Source: Elvik & Vaa 2003)

The main impression from the meta-analysis in the Handbook of Road Safety Measures (Elvik & Vaa 2003) is that the majority of road marking measures appear to have relatively little effect on the number of accidents. Changes in the number of accidents are in many cases not greater than + / - 5% and are, as a rule, not statistically significant. The exception to this rule are profiled edge lines, which appear to reduce the number of driving off the road accidents by around 30%, and distance markers on motorways, which reduce the number of accidents by more than 50%. The idea of distance markers is to help car drivers maintain an adequate distance from those in front. A combination of several road marking measures appears to have a more favourable effect on the number of accidents than individual road marking measures, see table 5.

Table 5: Effect on accidents of different effective road marking measures. Percentage change in the number of accidents. (Source: Elvik &

Vaa 2003)

Explanations for these results are little known. However, a number of studies have shown that different types of road markings can lead to higher speeds, see the paragraph on the effect on mobility.

Other Impacts – Environment
The environmental effects of road traffic depends, among other things, on the amount of traffic, the speed, the variation in speed, the composition of the traffic, the road alignment and the road surroundings. A significant change in environmental effects can be achieved by changing these conditions. Measures that improve the quality of traffic flow, i.e. which reduce queuing problems and lead to more even speeds, normally reduce the environmental problems along a road. The same is true of measures that reduce the amount of traffic.

Speed reductions however, have both positive and negative impacts on the environment.

Measures that reduce speed will in general have a favourable effect on the noise level. Specific measures like Rumble strips can however increase the noise level by 2-6 dBA. The increase in noise levels will be lowest for paving stones and highest for grooves in the road surface.

The global effects are related to energy consumption and greenhouse gases. Speed reduction at high speeds will reduce energy consumption and CO2, but at lower speeds the effect is opposite. For light traffic the energy consumption per km is high when starting, and decreases up to speeds of 40 km per hour. At speeds of 70 –80 km per hour the wind resistance will again increase energy consumption. Heavy vehicles have the same pattern, but the energy consumption will increase already form about 50 km per hour.

Local environmental impacts depend most strongly of the car’s age, driving style, cold starts and are also to some extent dependent on local geography. Speed is a less important factor below speed of 120 km pr hour. The clearest positive impact on pollution is that lower speed will reduce recirculation of dust particles (Amundsen & Ragnøy 2002). For modern cars no evident positive effects of lower speed as such are found on other local pollutants. Catalyst cleaning of exhaust from modern gasoline fuelled cars is not dependent on speed (SSB & SFT 1999).

If increased congestion involves variable driving speeds, local emissions will increase. More even speeds will reduce emissions. Driving pattern and transient driving have stronger environmental effects than the speed as such. The impacts of speed limitations on the driving pattern is less known.