Issue 1
2020 has been a year like no other for road travel in Australia. The onset of the COVID-19 pandemic at the start of the year led governments to impose restrictions on the movement of people in all Australian states and territories by March. With people allowed to leave home only for specified essential purposes, it quickly became obvious to the casual observer that vehicular traffic flows were greatly reduced across much of the network, and congestion had all but disappeared. By August, as restrictions were gradually being eased across much of Australia, a second and much larger wave of infections led to the imposition of even tighter restrictions on travel in Victoria.
With vehicular travel greatly reduced, there were hopes for a reduction in crashes, deaths and injuries on the roads. Reduced road travel implies fewer opportunities for road users to be involved crashes. So was the reduction in motor vehicle use accompanied by a corresponding reduction in road trauma? The Australia and New Zealand Driverless Vehicle Initiative (ADVI) commissioned the Australian Road Research Board (ARRB) to investigate changes in road travel and road trauma during the lockdown period, and the links between them. This article provides a brief account of the findings.
Australian jurisdictions began to apply restrictions on movement during March 2020, so any impact of the pandemic on traffic volumes and crashes was expected to begin in that month. The investigation was conducted in September 2020, so information about travel and crashes was available only to the end of August 2020 at the latest. The COVID period was therefore defined as the period from March to August 2020.
To allow for the influence of seasonal factors, such as weather and school holidays, it was important that travel and crashes during the COVID period be compared with travel and crashes during the same months of earlier years. The pre-COVID period was therefore defined as the months of March to August in the years 2017 to 2019 when data were available for these years, or the months of March to August 2019 when data were not available for the earlier years.
Ideally, data on driving exposure would be used to convert crash counts to crash rates per unit exposure (crashes per million vehicle-kilometres travelled).
However, due to the recency of the COVID period, estimates of driving exposure (such as those generated by the Australian Bureau of Statistics [ABS] Survey of Motor Vehicle Use) were not available. It was therefore necessary to make use of surrogate indicators of vehicular travel, including fuel sales, vehicle detection counts and responses to an online travel survey.
Vehicle detection counts recorded at signalised intersections in SA and Victoria by the SCATS signal control system are not counts of passing vehicles, since each vehicle may have been detected more than once at the same intersection. Nevertheless, changes in the detection counts are indicative of changes in traffic volumes on urban major and minor arterials (where traffic signals are generally located).
Australian petroleum sales data were obtained from the Department of Industry, Science, Energy and Resources (2020). The data include sales of regular, premium and ethanol-blended petrol. Diesel sales were excluded because available data did not separate automotive diesel fuel from marine and industrial diesel fuel and biodiesel blends. In any event, 73% of registered vehicles in Australia are fuelled by petrol and only 26% by diesel, with 1% using other fuels (ABS 2020).
Table 1 summarises the decrease in petrol sales during March–July 2020 compared with the same period averaged across 2017–2019. (Data for August 2020 were not yet available.) Nationally, the reduction in petrol sales from March to July 2020 was 21%, with the biggest reductions occurring in Victoria and Tasmania. If data had been available for August 2020, when Stage 4 restrictions applied in Victoria, a larger decrease in petrol sales in Victoria may have been recorded.
At ARRB’s request, South Australia’s Department for Infrastructure and Transport (DIT) undertook an analysis of SCATS vehicle detection counts from permanent traffic counting stations at signalised intersections in Adelaide and rural South Australia. The Department’s analysis concluded that the maximum impact of the pandemic on traffic volumes was seen in early April 2020, when traffic counts were down 30–35% compared with the corresponding period in 2019. By mid-September 2020, traffic volumes had fully recovered to the same level as in 2019.
During the COVID period (March–August 2020), vehicle detections recorded by SCATS were reduced by 27% relative to the corresponding period in 2019.
The ABS divides Victoria into Greater Melbourne and the Rest of Victoria, with the majority of signalised intersections being located in Greater Melbourne. The drop in vehicle detections during the COVID period was larger in Greater Melbourne (28%) than in the Rest of Victoria (19%).
The first wave of a household travel survey by the University of Sydney (Beck & Hensher 2020), conducted in early April at the height of the first wave of COVID-19 around Australia, investigated the number of trips undertaken, rather than the distance travelled. All 1073 respondents were aged 18 or more and three-quarters were from NSW, ACT, Victoria and Queensland.
Respondents reported that the number of trips (all travel modes) made by households per week in early April was 54% lower than in early March (before the major impact of the pandemic). The number of car trips was down 53%. Active transport (walking and cycling) trips were also down in absolute terms but up as a percentage of all trips. Food shopping trips were down less than trips for other purposes, and hence comprised a higher proportion of all trips.
Data on fatal crashes during the COVID period and earlier years were obtained from the Australian Road Deaths Database, maintained by the Bureau of Infrastructure, Transport and Regional Economics (BITRE 2020). Fatal crashes during the COVID period (March–August 2020) were compared with the same months averaged across the previous three years (2017–2019).
Table 2 shows that the fatal crash count for the whole of Australia was down 10% during the COVID period. Table 3 shows that fatalities across Australia were down 8% for the same period. Two-tailed binomial tests revealed that the decrease in fatal crashes was statistically significant (p=0.044), but the decrease in fatalities was not (p=0.102). Every jurisdiction recorded decreases except Queensland, where there was an 11% increase in fatal crashes. Fatalities decreased for drivers, motorcycle riders, passengers and pedestrians during the COVID period, but increased for pedal cyclists.
The only state or territory to experience an increase in fatal crashes during the COVID period was Queensland (from an average of 111 in 2017–2019 to 123 in 2020). Comparison of Queensland fatal crashes and fatalities from 2017–2019 with those in 2020 revealed that:
ABS remoteness codes were not available in the Australian Road Deaths Database for crashes in 2017. Comparison of remoteness codes for fatal crashes in 2018–2019 with those in 2020 revealed that the increase in fatal crashes in Queensland was confined to Major Cities of Australia (i.e. Brisbane, Gold Coast, Sunshine Coast and surrounds) (+32%), with little change in the remainder of Queensland (–1%).
It is worth remembering that the numbers of crashes and fatalities are small, so many of the changes seen during the COVID period may not be statistically significant.
Table 4 compares changes in SCATS vehicle detections and fatal crash counts for 2019 and 2020, the two years for which vehicle detection counts were available. For Victoria as a whole, the reduction in fatal crashes (18%) was less than the reduction in SCATS counts (27%).
However, the outcomes differed markedly between Melbourne and the remainder of the state:
The Australian Road Deaths Database classifies fatal crashes as single vehicle (including pedestrian crashes) or multi-vehicle. Despite the reduced proportion of Victorian fatal crashes occurring in Greater Melbourne during the COVID period, the proportions of Victorian fatal crashes classified as single vehicle (57%) and multi-vehicle (43%) did not differ between the pre-COVID period (March–August 2017–2019) and the COVID period (March–August 2020).
The only road user group to experience an increase in fatalities nationally during the COVID months was pedal cyclists (from an average of 21 in 2017–2019 to 27 in 2020). Anecdotal accounts suggest there was a substantial increase in cycling during the COVID period, which is likely to have contributed to the increase in cyclist fatalities. The distribution of cycling across roads versus off-road paths is not known.
Examination of cyclist fatalities in the BITRE data reveals that:
Proxy data for exposure, including petrol sales across Australia, South Australian traffic volume data, vehicles detected at signalised intersections in Victoria and driving trips reported in online surveys all indicate that driving exposure was reduced between 20% and 30% during the COVID period.
Based on data available at the time of the study, it appears the decline in fatal crashes during the COVID period was substantially smaller than the decline in motor vehicle use.
The increased number of deaths of pedal cyclists at a time of reduced fatalities among other road user groups warrants further investigation. Possible contributing factors include an increase in on-road cycling, particularly among inexperienced cyclists, possibly resulting from concerns about risk of infection when using public transport.
Measures likely to yield further reductions in fatal crashes include:
A key limitation of the study is that complete data on serious injury crashes during the COVID period
(March to August 2020) will not be available until 2021 at the earliest. Analyses for this investigation were therefore limited to fatal crashes. In addition, exposure data (vehicle-kilometres travelled) were not available, necessitating the use of proxy measures, such as petrol sales, vehicle counts and self-reported number of trips taken.
The authors gratefully acknowledge the contributions of Will Hore-Lacy and Kenneth Lewis, both of ARRB, who procured the Victorian SCATS data and the fuel sales data respectively and assisted with data analysis.
Australian Bureau of Statistics [ABS] 2020, Motor vehiclecensus, Australia, https://www.abs.gov.au/statistics/ industry/tourism-and-transport/motor-vehicle-census-australia/latest-release, viewed 29/10/2020.
Beck M and Hensher D 2020, Insights into the impact of COVID-19 on household travel, working, activities and shopping in Australia – the early days under restrictions, working paper ITLS-WP-20-09, Institute of Transport and Logistic Studies, University of Sydney Business School, Sydney, Australia.
Bureau of Infrastructure, Transport and Regional Economics [BITRE]2020, Australian road deaths database, https://www.bitre.gov.au/statistics/safety/ fatal_road_crash_database, downloaded 18/9/2020.
Dept of Industry, Science, Energy and Resources 2020, Australian petroleum statistics 2020, https://www. energy.gov.au/publications/australian-petroleum-statistics-2020.
