Pre-50 km/h Implementation Driver Speed Behaviours on Perth Metropolitan Local Roads

Transcription

1 1 Pre-50 km/h Implementation Driver Speed Behaviours on Perth Metropolitan Local Roads Tony Radalj Main Roads Western Australia Waterloo Crescent EAST PERTH 6004 Ph. (08) Fax: (08) KEY WORDS: speed survey, functional road hierarchy, speed compliance Abstract A speed survey was conducted over the period October to November 2000 to measure driver speed behaviours on Perth metropolitan local roads, which were expected to be affected by 50 km/h blanket speed implementation. The collected baseline data will be used for subsequent evaluation of driver speed compliance to the speed limit changes on local roads affected by the 50 km/h blanket speeds and 60 km/h signage of road sections that will maintain current speed limits. A sample of local roads stratified by population size was chosen across the 30 Local Government Authorities. The sample consisted of 139 roads, classified into four categories: Access Road, Local Distributor, District Distributor A and District Distributor B. Distributor A and Distributor B roads will generally preserve current speed limit of 60 km/h and be signed prior to the 50 km/h implementation. Traffic volume, speed and vehicle composition data was collected using vehicle classifiers. The results of the survey indicated that approximately 40% of drivers on Perth metropolitan local roads in a free flowing traffic environment exceed the speed limit of 60 km/h, ranging from 23% on Access Roads to 54% on District Distributor A roads. It was estimated that, on average, 8% of drivers exceed the speed limit by 10 or more km/h. The percentage varies between the road types, ranging from 4.4% on Access Roads to 11.2% on District Distributor A roads. Local roads drivers in the metropolitan area are more likely to exceed speed limits during the night hours than during the day hours. Higher mean speeds are observed on Fridays, Saturdays and Sundays than other days of the week. Similarly, tendency to exceed speed limit is more pronounced on these days than other days. The best compliance and smallest average speeds are observed on Tuesdays.Medium and long vehicle drivers tend to have similar speed behaviours as the drivers of short vehicles. If it is assumed that compliance will be achieved through a mixture of public education and enforcement after implementation of the 50 km/h on local roads, then expected changes in driver speed indices are as follows:

2 2 expected reduction in average speed for Access Roads of 2.3 km/h and Local Distributors of 3.9 km/h; expected reduction in the 85th percentiles of 5.3 km/h and and 7.6 km/h, respectively. It the road safety targets are achieved then expected changes in driver speed behaviours on these roads will result in reduction of fatal crashes and serious injury crashes, by 5 to 9 (20 to 37% ) and 250 and 345 (12 to 17%) per year, respectively. Introduction The introduction of 50-km/h speed limit legislation in Western Australia is part of a National campaign designed to blanket a 50-km/h speed limit over residential areas within Australia. The 50-km/hr speed limit legislation was passed through the Parliament of Western Australia in May 2001 by which all local roads on a state-wide basis within built-up areas would be blanketed to the speed limit of 50 km/h unless otherwise signed under regulation 11 (2) of the Road Traffic Code The current code states: A person shall not drive a vehicle in a built-up area, at a speed exceeding 60 km/h, except within a speed zone in which a higher speed is permitted. The introduction of 50-km/h speed limits in Western Australia would complement other road safety initiatives aimed to reduce number of fatalities and injuries resulting from traffic crashes on residential roads across the state. Although Australian states have taken different approaches in speed reductions, the objectives are similar, that is, to provide opportunities for safer local roads. South Australia has taken the lead with introduction of 40 km/h on roads in adjoining Adelaide suburbs, while Queensland introduced a comprehensive South East QLD area with 50 km/h speed limit on local streets in NSW adopted a policy of allowing Local Government Authorities (LGA) to decide when and if they would lower speed limits. Western Australian model is similar to the Victorian model, which gained an exemption from the Australian Road Rules to have a 50 km/h builtup area speed limit introduced across the state in January Speed zoning to 60 km/h in the metropolitan area is mainly based on the Main Roads Western Australia Functional Road Hierarchy (1999), where Primary Roads, District Distributor A and B will generally be signed to 60 km/h if not already speed zoned. Anomalies to the Functional Road Hierarchy classification model were negotiated with all of the councils. In rural areas and major regional centres, 50 km/h will apply to all built-up areas not speed zoned at 60 km/h or higher. Main Roads Western Australia (MRWA) will have responsibility of providing and maintaining signage on the roads. International and national research shows that 85% of pedestrians hit by a vehicle at 60 km/h roads are killed, while at 50 km/h the percentage of people killed is reduced to 45%. According to overseas studies, lowering the speed limit is expected to reduce the number of pedestrian fatalities at least by 25%. According to the New South Wales Roads and Traffic Authority s 50 km/h Urban Speed Limit Evaluation (Roads and Traffic Authority, 2000), by introducing 50-km/h urban speed limitation the number of casualty crashes caused by motor vehicle accidents is expected to be reduced by between 20% and 25%, and fatal crashes up to 44%. In addition, the total

3 3 number of traffic accidents is expected to decrease by 22% on local roads where the limit was introduced. A community opinion survey conducted in New South Wales, South Australia and Tasmania, in 1996, prior to introduction of 50 km/h limit on local streets (Austroads, 2000) showed that, in general, public would be supportive of reducing speed limit in residential areas. Both, respondents that would be affected by the change in speed limit and those chosen randomly from the population in the cities and towns in which the survey took place largely believed that 50 km/h on local roads would improve safety in some way. MRWA has maintained an interest in supporting the reduction of speed limits, particularly introduction of the blanket 50 km/h speed limit in urban areas, which includes all Access Roads and Local Distributor roads classified according to the Functional Road Hierarchy (FRH) agreed upon by individual Local Government Authority. However, unlike the NSW evaluation that was conducted to support the State Government s introduction of 50-km/h legislation, in Western Australia, MRWA has taken responsibility for evaluation of the program by measuring changes in drivers speed behaviours over time, based on the data collected before and after the legislation takes effect. Effects of drivers speed behaviours will be assessed against traffic safety measured in terms of number and severity of crashes accounted for by the changes in legal speed limits. Prior to introduction of 50 km/h speed limit on local roads within metropolitan area, all Local roads as well as Primary roads have had speed limits of 60 km/h, unless they were signed otherwise. All roads retaining the current speed limit of 60 km/h will be signed by the date of implementation of 50 km/h speed limit in metropolitan area, planned for November It is anticipated that most current 60 km/h roads such as Distributor A, Distributor B and Primary Roads will preserve current speed limits and will be signed with the speed limit of 60 km/h. The objectives of the survey were: to assess current driver speed behaviours on unsigned 60 km/h Local Roads, those that are expected to be blanketed to 50 km/h (Access and Local Distributor roads) and those that will preserve the current 60 km/h speed limit but signed before 50 km/h is introduced (most likely District Distributor A and District Distributor B roads) to collect the baseline speed data which will be used to assess changes in driver speed behaviours over time on 50 km/h and 60 km/h local roads networks, 6, 12 and 24 months after full implementation of 50 km/h speed limit legislation in WA to monitor driver compliance to legal speed limits on signed and unsigned road sections to assess needs for review of speed limits and to assign appropriate speed limits for the roads of similar functionalities

4 4 to provide baseline speed data for assessment of effects of change in speed behaviours on occurrence of crashes over time. MRWA is committed to conduct three speed surveys over the period of two years: six, twelve and twenty four months after the introduction of 50 km/h speed limit in the metropolitan area. Changes in driver speed behaviours will be assessed with respect to the baseline behaviours as well as in the behaviours observed at the three occasions, across all local road types and speed limits. Comparative evaluation will be conducted to monitor trends in acceptance of the legislation over the time period. METHODOLOGY Survey Design Given that Perth metropolitan area road network covers a vast area, ranging from the Local Governments of Rockingham to the south and Joondalup to the north, Stirling to the west and Mundaring to the east, substantial care was taken in the design of the survey such that data collected would be representative of types of local roads and driver population within the thirty metropolitan LGAs. It was assumed that there were no differences in distribution of proportions of drivers between the LGAs. Factors like traffic exposure and driver population within the LGAs were estimated on the basis of population size using the Australian Bureau of Statistics Statistical Local Areas of the 1996 census, represented in Figure 1, below. Since there was no population data available by LGA, Statistical Local Area population size was used in determining number of survey roads per locality. Subsequently, the number of survey roads by 37 Statistical Local Areas were merged by LGA, where required, into the final sets of survey sites distributed over the 30 LGAs (see Table 1).

5 5 Figure 1. Perth metropolitan Statistical Local Areas Where possible, proportional number of road types was chosen within each of the LGAs. Each of the LGAs was represented with at least one road, in most cases with at least one Access Road. Due to lack of types of functional hierarchy roads within each LGA and methodology used in determining number of total roads by LGA based on population size, not all LGAs were represented by all four types of roads in the survey. For the purpose of this baseline survey, roads were classified according to the Functional Road Hierarchy developed by MRWA in conjunction with Local Government Authorities in the Perth metropolitan area based on AUSTROADS Functional Classification System, having the following traffic volume characteristics: Road Type District Distributor A District Distributor B Local Distributor Access Road Traffic Volume Above 8000 vehicles/day Above 6000 vehicles/day Max 6000, average 3000 vehicles/day Max 3000, average 500 vehicles/day Sample Size Previous research studies involving surveys of the size of vehicles have been proved to be sensitive in detecting changes in the 95 th percentile speed as small as 2

6 6 km/h between two consecutive years (Cameron & Vulcan, 1998). For the purpose of future speed evaluation programs, the survey was designed such that the minimum sample size by each of the principal strata (road type and subsequent differences in speed limits, 50 km/h and 60 km/h)) was not less than vehicles. The sample size of this magnitude, when stratified by the strata was thought to be sufficient enough to detect significant changes in speed percentiles and average speeds. A sample of speeds taken on 60 km/h locations in the metropolitan area suggested that the minimum sample size of 1000 vehicles was required to detect changes of 1 km/h in mean speed. The pilot survey indicated that the minimum sample sizes required to detect the change in 95 th percentile of 1 km/h with 99% confidence level and error of 1 km/h, assuming speed variances of two surveys were the same would be approximately 2N, where N was estimated at 2000 vehicles. The survey was confined to the four types of road and traffic environment, namely, District Distributor A, District Distributor B, Local Distributor and Access Road. In order to achieve full coverage of the metropolitan area and every LGA, it was envisaged that the sample of approximately 140 sites would be sufficient in size to provide reliable representativeness of driver speed behaviour on general road locations of the local roads network within each of the LGAs and road types selected in the authorities. Distribution of number of sites proposed and surveyed in the LGAs is represented in Table 1, below. In the process of selecting estimated number of sites across entire local roads network, care was taken to achieve the most feasible representation of all driver speed behaviours over all strata. Out of the proposed 141 roads/road links, 139 were surveyed. Distribution of survey locations by road type was as follows: Functional Road Hierarchy Type Total District Distributor A 12 District Distributor B 14 Local Distributor 41 Access Road 72 Total 139 The number of surveyed sites by LGA varied between 1 and 18. Therefore, some LGAs were represented by all four road types, while others were represented by one, two or three types, depending on population size and availability of roads of a particular functionality.

7 7 Local Government Authority Population Number of Survey Sites (LGA Number) Count % Estimated Proposed Surveyed % Cambridge Town (128) 24, Claremont Town (115) 8, Cottesloe Town (116) 7, Mosman Park Town (121) 7, Nedlands City (122) 20, Peppermint Grove Shire (123) 1, Perth City Inner (124) 10, Subiaco City (127) 15, Vincent Town (130) 25, Total Central: 120, Bassendean Town (111) 13, Bayswater City (112) 56, Kalamunda Shire (102) 47, Mundaring Shire (106) 33, Swan Shire [City] (109) 87, Total East: 238, Stirling City (125) 174, Wanneroo City (110) 61, Joondalup City (131) 141, Total North: 377, Armadale City (101) 52, Belmont City (113) 27, Canning City (114) 71, Gosnells City (104) 103, Serpentine-Jarrahdale Shire (108) 9, South Perth City (126) 35, Victoria Park Town (129) 26, Total South-East: 325, Cockburn City (103) 57, East Fremantle Town (117) 6, Fremantle City (118) Kwinana Town (105) 19, Melville City (119) 89, Rockingham City (107) 58, Total South-West: 254, Metropolitan Grand Totals: 1,315, Table 1. Estimates of number of sites in the sample by Local Government Authority based on population

8 8 Speed Survey Site Selection Methodology Selection of a physical location on a road was based on the following five criteria: 1. Speed limit 60 km/h 2. The spot must be chosen at a traffic free flowing section of a road such that no factors within the section could cause changes in vehicular speeds other than the changes voluntarily induced by the drivers (eg. speed limit signs, major intersections, road works, etc.). 3. Whenever it is possible, the speed measuring equipment should be installed approximately at the mid-point of the section defined by two intersecting roads. 4. No installation should take place within a school zone. 5. Road features such as bridges, culverts, railway level crossings, and floodways should be avoided. Other Survey Characteristics and Criteria Survey Period For the purpose of survey reliability, the data collection took place during expected stable weather periods, namely, between October and beginning of December The data collected on rainy days was excluded due to possible effects of road condition on usual driver speed behaviours. Site Data Collection Time Period Minimum data collection period per site on all road types was 4 days, covering 96 consecutive hours of vehicle data. In order to control for the differences between days of the week, survey of the metropolitan local roads network was arranged in such a way that each day of the week would be equally represented in the sample. Start day and finish day varied from site to site such that each day of the week was substantially represented in order to control for possible differences between days of the week. In order to control for the differences between times of the day at each of the locations, the vehicle data collected in excess of the multiples of 24 hours was discarded in the analysis Equipment Used Vehicle data was collected by Main Roads Western Australia pre-qualified contractors who had a substantial experience in installation of the equipment for the purpose of collecting speed data. The equipment used was Microcom traffic classifiers, which were capable of recording a number of vehicle parameters utilised in the study, such as: time, date, vehicle speed, headway, number of axles, number of axle groups, wheelbase and vehicle type. Analysis It is anticipated that comparison of the pre 50-km/h speed limit legislation surveys and post 50-km/h speed limit legislation surveys will be analysed for roads where the traffic speed limit will be and will not be affected by a 50-km/h speed limit. In addition comparisons will be made between currently signed roads and newly signed roads.

9 9 The speed analysis will compare the pre and post 50-km/h speed limit legislation surveys looking at the mean speeds, average 85 th percentiles, and proportions of motor vehicles exceeding the speed limit by road type. Traffic speed comparisons will also be conducted between Local Government Authorities. Results from a comparison of driver behaviour between Local Government Authorities can be utilised for targeted 50-km/h local speed limit driver education and communication campaigns. This analysis will be in both the Local Government Authorities and public interest. Data and results may be accessed and made available to respective Local Government Authorities for their perusal. RESULTS The four-day survey of the metropolitan local road network recorded 1.68 million vehicles of which 1.28 million were considered to have travelled with free speeds. For the purpose of speed data analysis, a vehicle is considered to be travelling under free-flowing condition if its headway was four or more seconds (Australian Standard, 1999). Vehicles with lower headway values were excluded from the analysis. In addition, the survey identified presence of 7400 bicycles (0.44% of all vehicles) on the local road network. Due to special characteristics of bicycles, the data associated with these types of vehicles were excluded from all the analyses presented in this report. The analysis presented in the sections below is constrained to the data obtained from vehicle details recorded under free flowing traffic conditions, having a headway greater or equal to 4 seconds. Therefore, the sample size consisted of 1.28 million vehicles of which: were recorded on Access Roads, on Local Distributor Roads, on District Distributor B roads and on District Distributor A roads. If it is assumed that each driver of a vehicle passed the survey spot twice a day, making trips to and from a particular zone along the same route, and if compensation was made for those drivers who made multiple trips by those drivers who made single trips, then one can estimate that approximately driver speed behaviours were recorded in the survey. A more conservative estimate of the number of driver speed behaviours would be if it was assumed that, on average, each driver traversed the same path four times a day, for example, going to shops twice a day or dropping children off to schools and picking them up from schools, then the number of different driver speed behaviours recorded would be estimated at One can infer that the sample consisted of between and drivers, representing approximately 9 to 18 percent of metropolitan drivers. Analysis of the speed data suggested that under free-flowing condition, in total, 39% of vehicles exceed posted speed limits and 7.9% exceed 10 km/h or more. More detailed analyses indicated that driver speed behaviours (measured in terms of mean speeds, 85 th percentiles, percentage exceeding speed limit and percentage exceeding 10 km/h or more) varied between FRH road types, days of week, time of day and vehicle types. The results are presented in the sections below.

10 10 Differences between Road Types Speed Distributions As expected, due to road geometry and traffic conditions driver speed behaviours was found to largely depend on FRH type. Although speed distributions seemed to be quite similar in shape, apart from Local Distributor being too leptokurtic (peaked) when compared to other road speed distributions, the main parameters describing the distributions, mean speed and 85 th percentile, were associated with the road type (see Figure 1, below). Mean speeds ranged from 51.8 km/h on Access Roads to 59.8 km/h on District Distributor A roads. Similarity in functionality between District Distributor A and B was indicated by similar vehicle speed distributions Mean (LD) = Mean (DDB) = No. of vehicles Mean (AR) = Mean (DDA) = Access Road Figure 2. Speed distribution by road type, 60 km/h local roads, Perth metropolitan area, Nov 2000 The relationship between 85 th percentiles and FRH road types is similar to the relationship between mean speeds and the road types, ranging from 62.8 km/h for Access Roads and 68.4 km/h for Distributor A roads. NON-COMPLIANCE BY ROAD TYPE The speed data suggested strong relationship between FRH type and driver tendency to comply with speed limit. The best compliance with the speed limit was observed on Access Roads (77.3%), followed by Local Distributor (61.1%), Distributor B (49.7%) and Distributor A (46.7). Average compliance across the local 60 km/h speed limit road network was 61%. Similarly, driver non-compliance with 10 km/h or move above the speed limit was found to be strongly related to the road type, as shown in Figure Local Distributor Dist. Distributor B Dist. Distributor A Speed

11 Average % = 7.9 % exc. limit by >=10 km/h Access Road Local Distributor Distributor B Distributor A Road Type Figure 3. Vehicles exceeding limit by 10 km/h or more by road type on 60 km/h local roads The higher FRH is the higher speed non-compliance is observed in the driver population when travelling on local roads. The non-compliance with the speed limit ranged between 4.4% on Access Roads and 11.1% on Distributor A roads, with the average of 7.9% across the network of all local roads. Similarly, higher percentage of drivers travelling at higher speed groups above the speed limit were found at higher functional roads than on Access Roads, as illustrated in Figure 4. It seems that roads of high FRH classification by their nature with respect to traffic and road environment provide far more opportunities for drivers to exceed the speed limit than the roads classified as having a low functionality. % of vehicles Access Road Local Distributor Dist. Distributor B Dist. Distributor A Road Type Under speed limit < or more km/h Figure 4. Distribution of speeds by road type Percentage of drivers exceeding the speed limit, as indicated, by the speed groups in Figure 4, is strongly related to road functionality. For example, a driver when travelling on Distributor A or B roads is 1.6 times more likely to exceed 20 km/h above the speed limit than when on an Access Road.

12 12 Differences between Vehicle Types As expected, there was a relationship between vehicle type, average vehicle type speed and road type. Average speeds of Short Vehicles (SH) were higher than average speeds of Short Vehicles Towing (SVT) and Medium/Long Vehicles (MLV) across all road types (see Figure 5). Mean speeds for SH ranged from 52.0 km/h on Access roads to 59.9 km/h on Distributor A roads compared to SVT and MLV, ranging from 49.1 km/h to 56.8 km/h and 47.2 km/h to 57.8 km/h, respectively, on similar road types Mean = Mean = Mean = Mean = Speed (km/h) Access Road Local Dist. District Dist. B District Dist. A Road Type Short vehicle Short veh. towing Medium/Long Veh. 85th percentile Figure 5. Mean speeds by vehicle and road type Similar relationships were shown between percentage of vehicles exceeding speed limit as well as exceeding the speed limit by 10 km/h or more and vehicle type and road type. Percentage of vehicles of every type exceeding the speed limit was significantly less on Access roads than on the higher functional road hierarchy types, ranging from: 14.2% for SVT on Access roads to 40.4 on Distributor A; 22.9% for SV on Access roads to 54.0 on Distributor A roads; and 18.6% for MLV on Access roads to 43.0 on Distributor roads (ref. Table 2). Vehicle Road Type Total Distributor Type Access Road Local Distributor B Distributor A % exc. % exc. % exc. % exc. % exc. % exc. % exc. >=10 % exc. % exc. speed enf. speed >=10 speed >=10 speed km/h speed >=10 limit speed limit km/h limit km/h limit km/h limit km/h limit SV SVT MLV Total Table 2. Percentage distribution of vehicles exceeding speed limit and 10 km/h or more by vehicle and road type

13 13 Similar relationship was shown between road type, vehicle type and percentage of vehicles travelling at speeds 70 km/h or more. Percentage of SVT travelling at speeds 10 km/h or more above the speed limit ranged from 2.9% on Access roads to 6.5 and 7.1% on Distributor A and B, respectively. Slight variation in percentage of SV and MLV travelling at similar speeds was observed between road types. However, there were no differences between the two vehicle types in the overall percentage of vehicles travelling 10 km/h or more above the speed limit, estimated at 7.9%. Association between Temporal Factors and Driver Speed Behaviours Speed data suggests that there is an association between temporal factors, such as day of week and time of day, and driver speed behaviours on all local road types. DAY OF WEEK The distributions of mean speeds and percentages of vehicles exceeding 10 km/h or more above the speed limit by day of week, shown in Figure 6, indicate that there is an association between day of week and driver speed behaviours. The lowest average speed was observed on Tuesday, gradually increasing towards weekend, with the highest average speed recorded on Saturday. As it can be expected, mean speeds were related to the percentage of drivers travelling at speeds above the speed limit. The data suggests that drivers are more likely to exceed the speed limit by 10 km/h or more towards the end of week and weekend days, with the highest percentage and Saturday (9.3%), than at the beginning of the week, Tuesday being the day when the drivers are less likely to exceed the speed limit by 10 km/h or more than any other day of the week. Mean speed (km/h) Mean speed Mon Tue Wed Thu Fri Sat Sun % exc. limit by >=10 km/h Figure 6. Distribution of mean speed and percentage of vehicles exceeding speed limit by 10 km/h or more by day of week on 60 km/h local roads Similarly, driver compliance to the 60 km/h speed limit across all local road types seems to be the best on Tuesday (65%) and the worst on Saturday (57%). Time of Day Day % exc. limit by >=10 km/h

14 14 The vehicle speed data suggests that drivers tend to, on average, travel at speeds allowable by traffic and road conditions, which vary during the day. Drivers are more inclined to drive faster during night hours, when traffic volumes are low, than during daylight hours, which coincide with higher traffic volumes and larger traffic conflicts (see Figure 7). Mean speed (km/h) Mean speed % exc. limit by >=10 km/h Hour % exc. limit by >=10 km/h Figure 7. Mean speed and percentage of vehicles exceeding speed limit by 10 km/h ormore by time of day on 60 km/h metropolitan local roads network Relative relationships between time of day and average speeds were similar for every road type in the functional road hierarchy. However, as expected, observed hourly average speeds differed between roads, Distributor A and B roads experiencing higher speeds than Access or Local Distributor roads. Similar relationship that existed between average speed, road type and time of day was also observed with driver non-compliance to travel within 10 km/h above the speed limit during the day (see Figure 8). Drivers were more likely travel with speeds 10 or more km/h above the speed limit during the night than during the day. This tendency of non-compliance to travel at speeds less than 10 km/h above the speed limit was more pronounced on the higher than on the lower FRH roads. 30 % exc. limit by >=10 km/h Hour Access Road Local Distributor Dist. Distributor A Dist. Distributor B Figure 8. Percentage of vehicles exceeding limit by 10 km/h or more by time of day and road type

15 15 The relationships between time of day and driver speed behaviour on the local metropolitan roads network was found to be similar to the driver speed behaviours across all the network of roads, national, state and local roads, and all the speed limits ranging from 60 to 110 km/h, identified in the general speed survey conducted by MRWA in June and July 2000 (Radalj & Kidd, 2000). It seems that driver speed behaviours, measured in terms of average speeds and tendency to drive above speed limits is significantly related to road environment and traffic volume at a particular road type. Apart from differences in magnitudes of speeds at various road types and speed limits, the behaviours relative to the roads and speed limits are quite similar over times of day. Expected Driver Speed Behaviours After 50 km/h Implementation Local roads like Distributor A and Distributor B are expected to remain at 60 km/h. These roads will be signed with the 60 km/h speed limit signs. Most of Local Distributor roads as well as all Access Roads are expected to be blanketed to 50 km/h but unsigned. Consequently, it is expected that driver speed behaviours on Distributor A and B will not be significantly affected by the changes in speed limits on Access Roads or Local Distributor roads. However, significant changes are expected to occur in driver speed behaviours on the roads that will experience changes in speed limits. If it was assumed, and the case seems to be very likely to happen, that behaviours of those drivers who currently travel below 50 km/h will remain unaffected, and that the changes may only be observed among the drivers travelling above 50 km/h, then using the individual vehicle speed data it is possible to estimate what the driver population speeds would be when the 50 km/h is fully implemented. It is highly likely that drivers who currently travel above 60 km/h will be more affected than the drivers who currently travel at the speeds between 50 and 60 km/h. In order to estimate the overall effect on the population of drivers it is assumed that reduction of all drivers whose speeds are above 60 km/h would best estimate expected speeds of all drivers in the speed group above 50 km/h. Using this approach in estimating expected population speed parameters one may arrive at an expected speed distributions shown in Figure 9, below. No. of vehicles Mean (LD after) =53.15 Mean (AR after) = Access Road AR after 50 km impl Local Distributor LD after 50 km impl. Mean (LD before) = Mean (AR before) = Figure 9. Expected speed distributions for Access and Local Distributors after 50 km/h implementation Speed

16 16 It is expected that speed distributions of the two road classes will be more leptokurtic than the current distributions with the means of 49.5 and 53.1 km/h for Access and Local Distributor roads, respectively. Average expected speed reductions on Access Roads and Local Distributor roads are estimated at 2.3 km/h and 3.9 km/h, respectively. The shapes of the distributions are expected to remain leptokurtic for a substantial period of time due to driver adaptation to the new speed limit. In order to comply with the speeds less than 10 km/h above the speed limit of 50 km/h on Access Roads, it is expected that percentage of drivers travelling under the speed limit will be reduced from the current 77% to 41%, with the increase in percentage of drivers travelling up to 10 km/h above the speed limit of 50 km/h, from 18% to 54% (see Figure 10). Similar downward and upward changes in the two speed groups, respectively, but with higher magnitudes are expected on Local Distributor roads. If the current level of enforcement and community acceptance of 50 km/h in residential areas is similar to the current one with respect to 60 km/h, then percentage of drivers exceeding the speed limit by 10 km/h or more will remain unchanged, estimated at 4.4% on Access roads and 7.9% on Local Distributors. % of vehicles Before Before After After Access Road Local Distributor Access Road Local Distributor Road Type Under speed limit < or more km/h Figure 10. Comparison between current and expected vehicle speeds after 50 km/h implementation In terms of absolute speed reduction benefits compared to the current driver speed behaviours, post-50 km/h implementation is expected to result in increase in number of vehicles travelling up to 60 km/h on: Access roads by 18% (from 77% to 95%), Local Distributor roads by 31% (from 61% to 92%). Similarly, number of vehicles exceeding 10 km/h or more above the speed limit is expected to be reduced on: Access roads from 4.4% to 0.8%, Local Distributor roads from 7.9% to 1.2%. These and other expected changes are presented in the Figure 11.

17 17 % of vehicles th % = th % = th % = 58.8 Mean = Mean = Mean = th % = Mean = Before Before After After Access Road Local Distributor Access Road Local Distributor Road Type <= 60 km km Figure 11. Expected actual changes in vehicle speeds compared to the current speeds Expected increase in number of drivers travelling with speeds up to 60 km/h on roads that will be affected by change in speed limit is estimated at 25.4% above the current level of 68%. Therefore, percentage of vehicles travelling below 60 km/h is expected to be approximately 93%. Likewise, the number of vehicles currently travelling at speeds of 70 km/h or more and the number travelling with speeds less than 10 km/h above the speed limit are expected to be 6 (reduced from 6.4% to 1.1 %) and 4.9 (reduced from 25.3% to 5.2%) times less, respectively. These estimates of expected driver speed behaviours should possibly be targets to be achieved through primarily public education, enforcement, if necessary, or combination of the two approaches. It is expected that 5 out of 6 drivers currently travel above 60 km/h will change their habits to drive below the current speed limit. Applying these estimates onto the fiveyear, , crashes for which speed was a contributing factor, one can conservatively expect reduction of 5 to 9 fatal crashes per year (excluding those for which speed was an unknown factor), representing 20 37% of all fatal crashes on Access and Local Distributor roads. Similarly, currently speed related serious injury crashes are expected to be reduced by 250 to 345, representing 12-17% of all serious injury crashes on the roads (excluding crashes with unknown speed factor). CONCLUSIONS The results of this survey suggest that driver speed behaviours in a free flowing traffic environment on the metropolitan 60 km/h local road network are largely associated with a number of factors such as: functional road hierarchy, vehicle type, day of week and time of day. The analysis of the study has shown association between functional road hierarchy and driver speed behaviours. The higher the functional road classification is the less compliance is observed. Non-compliance with the speed less than 10 km/h above the speed limit ranged between 4.4% for Access roads and 11.2% for Distributor A roads. Average non-compliance rate to the speed less than 70 km/h across the entire 60 km/h local roads network was estimated at

18 18 7.9%. Seventy seven percent of vehicles on Access Roads travel below the speed limit. This rate is reduced to 46% for District Distributor A roads. Drivers of medium/long vehicles were more likely to travel at speeds greater or equal to 70 km/h than short vehicles (9% vs. 8%). Drivers are more likely to exceed the speed limit and, on average, drive faster during the night hours than during the daytime hours. Higher average speeds are observed on Fridays, Saturdays and Sundays than other days of the week. Similarly, tendency to exceed speed limit is more pronounced on these days than other days. The best compliance and smallest average speeds are observed on Tuesdays. If it is assumed that the current 60 km/h compliance will be achieved through public education or enforcement or combination of the two after implementation of the 50 km/h on local roads, then expected changes in driver speed indices are as follows: (1) Expected reduction in average speed is for: Access Roads km/h (from to 49.51) Local Distributors km/h (from to 53.15). (2) Expected reduction in 85th percentile is for: Access Roads km/h (from 62.8 to 57.5) Local Distributors km/h (from 66.4 to 58.8). Changes in driver speed behaviours are expected to have a significant effect on likelihood of occurrence of speed related crashes, and very likely other types of crashes due to overall reduced speeds on roads affected by speed limit changes. Solely based on proportion of vehicles/drivers who currently travel above the speed limit or exceeding the speed limit by 10 km/h or more and who are expected to drive below current 60 km/h speed limit or exceeding the 50 km/h by 10 km/h or more, it is expected that, on average, fatal crashes will be reduced by 5 to 9 crashes per year (20 to 37% of all fatal crashes). Similarly, reduction in number of serious injury crashes is estimate to range between 250 and 345 crashes, accounting for 12 to 17% of all serious injury crashes on Access and Local Distributor roads. RECOMMENDATIONS It is recommended that the findings of this study on a baseline driver speed behaviours on the 60 km/h local roads in the Perth metropolitan area be utilised in evaluation of effects of 50 km/h speed limit implementation program and assessment of changes in driver speed behaviours over time. In addition, the results of the study in conjunction with crash information for the local road network should be used in future cost/benefit evaluation studies on traffic safety associated with the program implementation. Furthermore, the findings of this study on driver speed behaviours on 60 km/h local roads should be used as a component in formulating traffic safety and enforcement strategies, taking into consideration driver behavioural characteristics associated with road and traffic environments, vehicle type and temporal factors.

19 19 Due to road type associated driver speed behaviours, the findings of the study may be used for speed limit road classification purposes. REFERENCES AUSTROADS (2000). 50 km/h speed limit on local streets-community opinions and anticipated effects. Report AP-R161/00. Sydney, NSW. AUSTROADS. Cameron M.H., and Vulcan, A.P. (1998.) Evaluation review of the Supplementary Road Safety Package and its outcomes during the first two years. Report to Land Transport Safety Authority, New Zealand, July Radalj T. and Kidd B. (2000). Driver speed compliance in Western Australia. Handbook & Proceedings Road Safety Research, Policing & Education Conference, pp Brisbane, Queensland, November MRWA (1999). Metropolitan Functional Road Hierarchy. ISBN Perth, WA. August Standards Australia (1999). Manual of uniform traffic control. Part4: Speed controls. AS Homebush, NSW, 1999.