< Previous | Contents | Next >

SCOTTISH TRANSPORT STATISTICS No 23: 2004 Edition

Chapter 6 ROAD TRAFFIC

1. Introduction

1.1 This chapter provides information about road traffic, such as the total volume of traffic by type of road, by type of vehicle, and by council area. It also provides some figures on traffic flows at selected points on the road network, some statistics on delays and congestion, and information about petrol and diesel deliveries in Scotland and some atmospheric pollutants.

1.2 The main changes in this edition are the inclusion of three new tables, on:

• Traffic on trunk roads and local authority roads, by Council area, 1993 to 2003 (Table 6.5);
• Traffic on trunk roads: average time lost per vehicle-kilometre on monitored roads (Table 6.8);
• Car drivers' journeys: whether delayed by congestion and, if so, how much time was lost (Table 6.9);

In addition, Table 6.10 has been expanded to show the percentage breakdown of petrol and diesel delivered in Scotland, and some more monitoring sites have been added to table 6.11.

1.3 The Department for Transport (DfT) has revised slightly its estimates of the volume of traffic on major roads (M and A roads), and on all roads, in Scotland, for each year back to 1993. Therefore, the figures which were published previously for 1993 onwards have been superseded by DfT's new estimates.

2. Main Points

2.1 The estimated total volume of traffic on Scotland's roads in 2003 was about 42 billion (thousand million) vehicle kilometres: about 1% more than the revised estimate for 2002.

2.2 When using the traffic estimates, it must be remembered that they indicate only the broad level of traffic, and may be unable to provide reliable information about year-to-year changes. This is because (as is explained in the "Sources" section) they are based on information from a very small cross-section of the roads in Scotland: "12 hours in one day" traffic counts taken at several hundred sites per year (over 750 per year from 2003; roughly 500 per year before then) and data from automatic traffic counters at about two dozen sites in Scotland (which are combined with data from automatic counters at similar sites in England and Wales). This limits their precision. "Scottish Transport Statistics no. 21 - 2002 edition" discussed how the extent of the revisions to the estimates, following more information becoming available and improvements in methodology, indicated the uncertainty in the estimates. In 2003, DfT introduced a new estimation method, which revised greatly the previously-published estimates for the total volume of traffic on minor roads in Scotland.

2.3 The total volume of traffic on major roads (Motorways and A roads) in 2003 was estimated to be 27.7 billion vehicle-kilometres. Traffic on Motorways totalled an estimated 5.8 billion vehicle kilometres (14% of all traffic). This was less than the estimated 9.8 billion vehicle kilometres on trunk A roads (23% of the total), and the 12.1 billion on non-trunk A roads (29%). Most of the traffic on A roads was on roads in rural areas, which accounted for 16.5 billion out of the A roads total of 21.9 billion vehicle kilometres. (Table 6.1)

2.4 Minor roads (B, C and unclassified roads) accounted for the remaining 34% of traffic in 2003: an estimated 14.4 billion vehicle kilometres, most of which was on unclassified roads (8.2 billion). Most minor road traffic (8.1 billion vehicle-kilometres in 2003) is on roads in urban areas. (Table 6.1)

2.5 The total volume of traffic on major roads (Motorways and A roads) in 2003 was 1% more than in the previous year, due to 2% growth in traffic on trunk A roads, a 1% increase on Motorways and a 1% rise in traffic on non-trunk A roads. There was also an increase in traffic on rural A roads of 2% while traffic on urban A roads fell very slightly. Minor road traffic was also 1% higher than in 2002. (Table 6.1)

2.6 When looking at the figures for earlier years, it must be remembered that the volume of traffic in 2000 was affected by the fuel protests in September. DfT's estimates for Scotland, like those for GB as a whole, show a slight fall in the total volume of traffic on major roads, and a slight rise in the total volume of traffic on minor roads, between 1999 and 2000.

2.7 The DfT estimates suggest that the volume of traffic on major roads in Scotland has increased in every year since 1993 (which is the first year for which DfT has produced revised estimates) apart from 2000, which was affected by the fuel protests. The estimated total volume of traffic on major roads in Scotland in 2003 was 20% higher than the estimate for 1993. Motorway traffic was estimated to have increased by 34% since 1993 - representing more rapid growth than the rises, over the same period, in traffic on trunk A roads (21%) and non-trunk A roads (13%). (Table 6.1)

2.8 It is estimated that traffic on minor roads rose by 15% in the period between 1993 and 2002. As a result, the overall total volume of traffic on all roads in Scotland in 2003 was estimated to be 18% higher than in 1993. (Table 6.1)

2.9 Cars account for nearly four-fifths (79%) of the total volume of traffic on the roads (i.e. of the total for major roads and minor roads combined), and light goods vehicles for 12%. Heavy goods vehicles contribute under 6%. (Table 6.2)

2.10 Since 1993, the volume of car traffic has increased on both major and minor roads, as has the volume of light goods vehicle traffic. Heavy goods vehicle traffic rose less rapidly. (Table 6.3)

2.11 Over a fifth of motorway traffic was within the City of Glasgow, and Highland was the Council area with the greatest total volume of trunk A road traffic (measured in vehicle kilometres). Other Council areas with large volumes of traffic on major roads (over a thousand million vehicle-kilometres, in total) were Aberdeenshire, Dumfries & Galloway, Edinburgh, Fife, North Lanarkshire, Perth & Kinross, South Lanarkshire and West Lothian. The areas with at least a thousand million vehicle-kilometres of traffic on minor roads (B, C and unclassified roads taken together) were Aberdeenshire, Edinburgh, Fife, Glasgow and North Lanarkshire (Table 6.4)

2.12. The monthly average daily traffic flows recorded at a selection of Automated Traffic Classifier (ATC) sites are given in Table 6.6. Unfortunately problems with the equipment at some sites meant that figures were not available for some months. The levels of traffic varied greatly. Taking the total of traffic in both directions, the average flows at the A80 and A720 sites were around 60-80,000 vehicles per day. In contrast, the average daily flow at the A835 Aultguish site was less than 2,400 vehicles in even its busiest month. Traffic levels can also vary considerably within the year: for example, the A9 Tomatin site in August averaged 10,814 vehicles per day — more than four times the average of 2,534 in January. (Table 6.6)

2.13. A range of information about traffic flows at selected points on the trunk road network is given in Table 6.7. Of the sites listed, the M8 Harthill was the busiest, with an annual average of over 51,105 vehicles per day in 2003. Its Monday-Friday average was 56,331 vehicles per day, and its Monday-Friday peak hourly flows were 4,833 vehicles in the morning and 4,343 vehicles in the evening. At the opposite end of the scale, a point on the A835 Aultguish averaged 1,515 vehicles per day over the year as a whole (2,326 in August: figures are also given for that month to show any effects of the tourist season), and its Monday-Friday peak hourly flows were 161 vehicles in the morning and 229 vehicles in the evening. Of the sites listed, the A75 Carsluith and the A1 Granthouse had the highest percentage of traffic accounted for by heavy goods vehicles: in 2003, 22% and 21% respectively of vehicles on these roads were HGVs (15% and 16% respectively on weekdays). (Table 6.7)

2.14 Table 6.8 provides estimates of the "time lost" by traffic (i.e. the additional time taken compared with "free flow speed" conditions) on the trunk road routes which the Executive is monitoring. The table shows the overall average delays (in seconds) encountered by vehicles travelling one kilometre on each route in some months in 2003. (These figures are provisional, and may be updated in due course. As the amount of data collected has been increasing, the quality of the estimates will have improved during 2003, and will improve further during 2004. Definitions and sources are described in sections 3.3 and 4.4.) The reasons for the delays vary, and include traffic congestion, roadworks, increases in traffic for particular events, and seasonal factors - for example, the unusually large values for the A96 in November and for part of the A90 in December may be due to, respectively, very bad weather and traffic management during the reconstruction of a roundabout. Routes which usually have fairly high overall average "time lost" values are the most likely to be affected by congestion. However, a route with low figures is not necessarily "congestion free": it could have occasional spells of severe congestion, which are not sufficiently frequent, prolonged or widespread to produce a high overall average "time lost" per vehicle per kilometre. In general, the time lost is only a few seconds per vehicle per kilometre, on average over the whole of a month. If one disregards a few unusually large values, like those mentioned earlier, average delays per vehicle per kilometre tend to be highest on the M8 St James Interchange / Baillieston and the Kincardine Bridge approaches, with averages of up to 12 seconds per vehicle-kilometre. As these are overall averages, considerably more time could be lost at certain times on particular days. In addition, it must be remembered that these are estimates of the average "time lost" per vehicle per kilometre. One would identify other routes as having greater delays if one used figures on other bases. For example, longer routes would be identified as the worst-affected if one used the total time lost by a vehicle travelling over the whole of the route (rather than "per kilometre" figures); and more heavily-trafficked routes would be identified as the worst-affected if one used the total delay for all vehicles (rather than "per vehicle" figures). A report by the Scottish Executive Traffic Controller (see section 5.4) will provide more detailed information about (e.g.) traffic levels, speed and congestion/delay on each of the routes which the Executive is monitoring. (Table 6.8)

2.15 Table 6.9 provides estimates of delays attributed to congestion on the roads, as reported by the drivers of cars. With effect from April 2003, the Scottish Household Survey collected information from adults who had driven a car on the previous day. For each car journey made, the interviewer asked "was this part of your trip delayed due to traffic congestion?" and, for those who said that it was, the interviewer asked "how much time do you think was lost due to traffic congestion?" No definition of "traffic congestion" was provided, so the respondents could interpret the term as they wished. Approximately 11% of journeys made as the driver of a car were said to be delayed due to traffic congestion. 4% of car drivers’ journeys were delayed by about 5 minutes, 3% by about 10 minutes and 3% by about 15 minutes or longer. 24% of car drivers’ journeys which started between 8:00 and 8:59 a.m. on a weekday suffered delays due to congestion, as did 19-21% of such journeys started between 7:00 and 7:59 a.m. and between 4:00 and 5:59 p.m. At weekends, only 6-7% of car drivers’ journeys suffered delays due to congestion. Few delays (2-3% of car drivers’ journeys) were experienced by people residing in "remote" small towns and "remote" rural areas, compared to 14% in large urban areas. (Table 6.9)

2.16. The Department of Trade and Industry (DTI) publishes figures for the overall total amounts of petrol and diesel delivered in Scotland. These figures should be used with caution since they are based on company reports that may not distinguish properly between Scotland and the rest of the United Kingdom. Consequently they can provide a misleading indication of the amount of fuel actually consumed in Scotland. In particular, there are concerns that inconsistency in companies' reporting of countries may have caused the apparent discontinuity between 1999 and 2000. The DTI figures suggest that the total amount of petrol and diesel delivered in Scotland fell by almost a quarter between 1993 and 2003, from 2.7 million tonnes in 1993 to 2.1 million tonnes in 2003. However, roughly half of that drop occurred between the figures for 1999 (2.6 million tonnes) and 2000 (2.3 million tonne), which may be due to inconsistency in reporting. Despite these doubts, it is clear that there have been large changes in the types of fuel delivered in Scotland. In 2003, petrol accounted for just under half of all the reported deliveries, compared with over three-fifths in 1993. The decline in petrol's share of the total is due to deliveries of leaded petrol falling from 27% of the total in 1993 to just 1% in 2003; over the same period, unleaded petrol's share rose from 35% to 47%. Diesel's share of the total increased from 37% in 1993 to 51% in 2003. This was due to retail diesel rising from 11% of the total in 1993 to 27% in 2003; commercial diesel's share was 25-26% in both years, having fluctuated in the range 22-28% throughout the period. (Table 6.10)

2.17 At the selected monitoring sites, carbon monoxide concentrations were below the level of the air quality strategy objective (see section 3.4.2) for December 2003 in every year from 1993 to 2003, and the lead concentrations were below the value set as the objective for December 2008. However, nitrogen dioxide concentrations at monitoring sites in both Edinburgh and Glasgow were above the level set as an objective for December 2005 in every year from 1993 to 2003. The air quality strategy objective for ground level ozone states that by the end of 2005 the maximum daily concentrations should not exceed 100 m g/m³ on more than ten days per year. While ozone concentrations at the selected monitoring sites have fluctuated over the years, the target value was exceeded on more than ten days a year at the Strath Vaich site in most of the years from 1993 to 2003. Eskdalemuir was above the target in several years between 1993 and 1999, met the target between 2000 and 2002, and exceeded it again in 2003. However, Edinburgh met the target in every year from 1993 to 2003. Annual mean particulate concentrations in Edinburgh and Glasgow centres were below the December 2004 objective level in all the years from 1993 to 2003 for which figures are available, but exceeded the level set for December 2010. (Table 6.11)

3. Notes and Definitions

3.1 The traffic estimates produced by the Department for Transport

3.1.1 The methods that have been used to estimate the volume of traffic on major roads (Motorways and A roads) in Scotland have changed over the years. Section 4.1 describes the method which the Department for Transport (DfT) used to produce the estimates for 1993 onwards, and section 4.2 explains how the figures for 1992 and earlier years were calculated. Estimates of the volume of traffic on minor roads (B roads, C roads and unclassified roads) in Scotland that are suitable for publication are only available from 1993. Section 4.3 describes the methods used.

3.1.2 Please note that the DfT traffic estimates provide only a rough indication of the likely volume of traffic on the roads in each local authority area, and that the DfT traffic estimates for individual Council areas are not National Statistics. DfT provides the estimates that it produces for individual local authority areas as being the best that it can produce from the limited amount of data available to it - rough indications of the likely volumes of traffic on roads in each Council area, for use with caution as no better estimates are available. After considering the description of the methods of estimation in Section 4, one should appreciate why:

• it is not possible for DfT to quantify the possible "margins of error" around the estimates for individual local authority areas;
• they are not classed as National Statistics;
• more detailed breakdowns of the estimates for individual Council areas are not published.

3.1.3 DfT's methodology for estimating traffic volumes distinguishes between Motorways, "urban roads" (i.e. roads, other than Motorways, which are in urban areas) and "rural roads" (i.e. roads, other than Motorways, which are in rural areas). For the purposes of the DfT traffic estimates, the general rule is that an "urban" road is a road (other than a Motorway) that lies within the boundaries of an urban area which had a population of 10,000 or more in 2001 (DfT identified such areas using the Population Census boundaries for settlements); a "rural" road is one located in an area with a smaller population. However, there are exceptions. DfT adjusted the "urban/rural" classification of stretches of major road which are on the outskirts of urban areas, in some cases where it was not possible to break them at a junction with another major or minor road. For example, a stretch of road which is part of a trunk road bypass will usually be classified by DfT as "rural" (even the part of it which runs through an "urban" area) whereas a relatively short road between two urban areas that are close to each other will normally be classified by DfT as "urban" (even the stretch which is in a "rural" area). DfT's view is that the effects of such adjustments on the overall traffic estimates are likely to be small.

3.1.4 DfT's "urban / rural" classification of roads differs from the "built-up" / "non-built-up" classification of roads, which was used for the traffic estimates which DfT produced in 2002 and earlier years. The "built-up" / "non-built-up" classification was based on speed limits, with roads with a speed limit of 40 mph or less being classed as "built-up"; those with a higher speed limit being "non-built-up". For example, a dual carriageway with a 50 mph speed limit in an urban area is counted as an "urban" road on the basis of its location, but as a "non-built-up road" on the basis of its speed limit. In contrast, a road with a 40 mph speed limit in a small town (population under 10,000) is classed as a "rural" road on the basis of its location, but as a "built-up" road on the basis of its speed limit. While most roads in urban areas have speed limits of 40 mph or less (so are "built-up"), there are many roads in small towns and villages in rural areas which also have speed limits of 40 mph or less (so are also "built-up"). Therefore, "urban / rural" traffic figures are not comparable to "built-up / non-built-up" traffic figures: the two could differ noticeably for some local authority areas. It will not be possible to quantify this, because each set of DfT's estimates were produced using only one of the two classifications, so there is no table which cross-tabulates the traffic estimates by both "urban / rural" and "built-up / non-built-up". Another point is that urban boundaries tend to change slowly over time, whilst there has been a trend for more roads in rural areas to be assigned speed limits of 40 mph or less. So, a time series for traffic on "urban" roads may show a different trend from a time series for "built-up" roads.

3.1.5 On 1st April 1996, local government was reorganised, and the 32 present Councils replaced the former Regions, Districts and Island Areas. At the same time, changes were made to the trunk road network: some former non-trunk roads became trunk roads, and some former trunk roads ceased to be trunk roads. Section 4.3 of the "2002" edition described how this affected the traffic estimates produced by DfT's previous methodology, and caused discontinuities in the series of figures for traffic volumes on major roads. DfT's traffic estimates are no longer affected by such discontinuities, because they count major roads on the basis of their "trunk road status" at a recent date, rather than on the basis of their trunk road status in the year in question. As a result, there is no discontinuity in the figures between 1995 and 1996. The new estimation method which DfT introduced in 2003 also removed some other discontinuities from the figures (again, details of these were given in previous editions).

3.2 Traffic flows at selected sites

3.2.1 The average daily traffic flows at Automated Traffic Classifier Sites are "total past the point" figures: traffic is counted in both directions. The estimated traffic flows are based on 7-day averages which include both weekdays and weekends. On occasion, the ATCS counters are not in operation for enough of the month to provide a reliable estimate: in these cases, ".." is used to indicate that no estimate is available.

3.3 Traffic on specific trunk road routes: average time lost

3.3.1. Estimates of the "time lost" by traffic on particular routes are produced by the Scottish Executive Trunk Roads Network Management Division. The figures are estimates of the additional time taken compared with the time that would have been taken had the vehicles been travelling in "Free Flow Speed" conditions. The reasons for the delays may vary from month to month and from route to route, and include traffic congestion, roadworks, increases in traffic for particular events and seasonal factors (such as tourist traffic and the effects of bad weather). For example, particularly bad weather in one month may lead to a very high "time lost" figure for that month for a route for which the "time lost" figures are normally very low. A route which has major roadworks which last for a period of two or three months may have greatly increased "time lost" figures for those months. Routes which have fairly high "time lost" values throughout the year are the ones which are the most likely to be affected by congestion. However, because the figures are overall estimates, it does not follow that a route which has low figures is necessarily "congestion free": it could suffer from occasional spells of severe congestion, which are not sufficiently frequent or prolonged to produce a high overall average.

3.3.2 The Free Flow Speed for a stretch of road generally represents the speed that is seen outwith periods of high traffic flow and other known events on the road network (e.g. traffic management for roadworks etc). The early hours of the morning are generally excluded, as they often have a higher than usual percentage of heavy goods vehicles, which usually travel at speeds lower than the overall free flow speed. The Free Flow Speed for each stretch of a particular route is derived from information about the actual speeds of vehicles travelling on that road. The additional travel time at a particular time on a particular day is then calculated from the average speed of vehicles using that stretch of road then and from its Free Flow Speed. For example, suppose that, on a stretch of road of length 1 kilometre, the average speed of vehicles (in a particular 15 minute period of one day) was 60 kilometres per hour, and that the Free Flow Speed for that stretch of road was 100 kph. The additional travel time per vehicle in that period would be calculated thus:

• average time taken to travel 1 km at 60 kph = 1 minute
• time taken to travel 1 km at Free Flow Speed of 100 kph = 0.6 minutes
• so, additional travel time per vehicle = 0.4 minutes

If 300 vehicles went through in that period, the total additional time would be 300 x 0.4 = 120 minutes. (NB: for the purpose of these calculations, vehicles whose average speed is above Free Flow Speed are treated as if they were travelling at Free Flow Speed, so that their "reduced" travel time does not "offset" any of the additional travel time that is incurred at other times of day or on other days.)

3.3.3 Such figures can be aggregated to produce a number of additional travel time values, such as the average time lost per vehicle-kilometre for a route for a month. This represents the average delay encountered by a vehicle travelling one kilometre on that route. As it is an overall average for the month as a whole, it could conceal considerable day-to-day and/or hour-to-hour variation - for example, a stretch of road which has only one or two periods with very long delays due to congestion (perhaps when there is a lot of traffic to events such as football matches), and traffic travelling (on average) at or above Free Flow Speeds at all other times, will have a low overall average "time lost".

3.3.4 The average time lost per vehicle-kilometre is only one of a number of possible measures of the delays that are due to traffic congestion and other factors. A report by the Scottish Executive Traffic Controller (see section 5.4) will provide information on a range of such measures, and give more detailed information about (e.g.) the levels of traffic, speed and congestion/delay on each of the routes which the Executive is monitoring.

3.4 Pollutants

3.4.1 The atmospheric pollutants listed in Table 6.11 have been selected because they are considered to be a threat to human health, and transport is understood to be a significant contributor to emissions of the pollutants. In February 2003, the Scottish Executive, UK Government and the other devolved administrations published an Addendum to the Air Quality Strategy containing air quality objectives for eight pollutants (benzene, carbon monoxide, lead, nitrogen dioxide, ozone, particulates (PM10), sulphur dioxide and 1,3-butadiene). The objectives are policy targets expressed as a maximum ambient concentration to be achieved, either without exception or with a permitted number of exceedences, within a specified timescale. The table below sets out the agreed air quality objectives for the first six of these (the ones to which transport is understood to contribute significantly).

3.4.2 Carbon monoxide produced from road transport, particularly petrol-fuelled vehicles, accounted for 62% of UK emissions in 2001. Emissions have been declining and this is attributed to the installation of catalytic converters in vehicles. Additives in petrol have been the primary source of atmospheric lead. Significant reductions in urban lead levels have occurred since the introduction of unleaded fuel in 1986. The general sale of leaded petrol was banned in the UK from 1 January 2000. All combustion processes in air produce oxides of nitrogen (NO_x). Road transport accounts for about half of all NO_x emissions in the UK; this contribution is greatly increased in urban areas. The greatest source of PM₁₀ (particulate pollution less than 10m m in diameter) is combustion. In particular, this includes fuel from road transport, which accounts for around 20% of UK emissions of PM₁₀ . Ozone is not emitted directly from any man-made source in significant quantities but arises from the chemical reaction of other pollutants such as NOx and volatile organic compounds produced by road transport, industrial processes and solvent uses. Benzene is a recognised genotoxic human carcinogen and in the UK the main atmospheric source is the combustion and distribution of petrol. Road transport accounted for 36% of UK benzene emissions in 2001, compared with around 65% prior to 2000, due to a reduction in the benzene content of petrol.

AIR QUALITY OBJECTIVES FOR SCOTLAND

4. Sources

4.1. The method of estimating major road traffic volumes for 1993 onwards

4.1.1. Estimates of traffic volumes on major roads (Motorways and A roads) in Scotland by road type, vehicle type, and area within Scotland are produced by DfT in conjunction with the Scottish Executive Trunk Roads Network Management Division (SE TRNMD).

4.1.2. The method of estimation has two main stages. First, traffic flows (which represent the numbers of vehicles flowing past particular points in a specified period) are estimated for approximately 1,900 individual road "links" on Motorway and A roads in Scotland. (A link is normally a section of road between two major intersections). The estimates of the traffic flows on these road "links" are then combined with information about the lengths of the links, to derive total traffic volume estimates (measured in millions of vehicle kilometres) for major roads by road type, vehicle type and Council area. The type of a road is determined by its class (Motorway or A road), by whether or not it is a trunk road (trunk roads are those roads for whose upkeep Scottish Ministers are responsible), and by whether it is in an "urban" area or a "rural" area (see Section 3.1). The steps involved in each of these stages are described in subsequent paragraphs.

4.1.3. The estimates of traffic flows for the individual major road "links" for each year are derived by a methodology which involves the use of two different types of traffic counts: "link" and "core":

i. The road "link" traffic counts are taken manually, for 12 hours in one day, on a rotating basis (on average about once every four years), at each of the approximately 1,900 road links covering nearly all of the major road network in Scotland. These counts take place in "neutral weeks" during late March, April, May, June, September and October (the aim is to avoid counting, for example, during school holidays, and so to obtain counts which are representative of the level of traffic on each link). Traditionally, roughly one sixth of all the road links on the major road network were counted each year in Scotland, but it is now about one in four (compared with around 40% in England and Wales). At one time, the aim was to count each Scottish site once every six years. However, in 1999, the counting schedule was changed in order to improve the accuracy of the estimates: now, the more important links in Scotland should be counted more often, and the less important should be counted less often. Up to and including 2002, about 300 or so counts were taken each year. However, following a study of possible ways of improving the road traffic estimates for Scotland, SE increased the number of counts to about 460 or so per year. These "12 hours in one day" counts must be scaled up to estimate the total flow of traffic for the year as a whole, and in order to reflect changes in traffic levels in the years after each count was taken. The "core" counters provide the information that is used in the scaling.

ii. The "core" counters are automatic traffic classifiers, which are located at selected sites on major roads through Great Britain. These operate, on the whole, continuously: 24 hours per day, throughout the year, and provide information about traffic flows classified by category of vehicle according to their length and number of axles. The locations of the "core" counters, taken together, cover a good cross-section of types of road. There are around 150 "core" sites on major roads (including motorways) in Great Britain, of which about 25 are in Scotland.

4.1.4. For the purpose of combining the data from the manual counts and the automatic counters, DfT allocates each road link, and each "core" counter, to one of 22 groupings of road type. These were based on a detailed analysis of the results from all the individual automatic counter sites, and take into account traffic flow levels, (GB) regional groupings, and the road's "category", which is a combination of its class (e.g. Motorway, A road, etc) and its urban/rural classification. The groupings range from lightly-trafficked roads in holiday areas, such as Devon and Cornwall, to major roads in Central London. There are no groupings which consist solely of Scottish roads, because there are not enough "core" counters on roads in Scotland which are in the same category, and have similar levels of traffic flow, to form any separate Scottish groupings.

4.1.5. The estimated traffic flows for each major road "link" for the latest year are then derived by a series of calculations of which the following provides only a broad outline. The "core" traffic counters are used to derive two sets of factors, which are then applied to each of the 1,900 "link" counts:

a. "Expansion Factors" for road type and vehicle type are used to scale the single day 12 hour "link" counts to provide estimated traffic flows for the whole year in which the counts were taken.

b. "Growth Factors" for each road and vehicle type are used to scale estimated traffic flows in the previous year forward to the latest year, for those links which were not counted in the latest year.

4.1.6. DfT estimates the total traffic volume (in vehicle-kilometres) on each major road link by multiplying together the estimated traffic flow for the link and the length of the link. DfT obtains the length of each major road link, and identifies the Council(s) in which it is located, using a Geographic Information System (GIS). When a link lies completely within the area of one Council, its estimated traffic volume is counted wholly against that Council. In a case where a link crosses a boundary between Councils, it is split (for the purposes of the calculations) at the boundary into two separate links. Similar calculations are performed for each new link: the length within the relevant local authority (which DfT obtains from the GIS) is multiplied by the average traffic flow calculated for the original link (regardless of the Council area in which the traffic count was taken - because the original link was a section of road between major intersections, the traffic flow should not vary much along its length).

4.1.7. DfT compared its estimates for some motorway and trunk road links with the information that was available from the "volumetric" automatic traffic counters which are operated on motorway and trunk road links by TRNMD, its equivalent in Wales, and the Highways Agency in England. In general, there was a much closer correlation between the two sets of data than for the estimates which DfT had made in 2002 and earlier years. DfT noted that its estimates were slightly lower, and thought that there might be a number of reasons for this (e.g. the manual counters might miss some vehicles, the fact that the DfT "core" counters cannot be positioned on the most congested roads, etc). DfT therefore adjusted its expansion factors in order to eliminate the apparent slight bias in its overall estimates. DfT did not attempt to make its estimate for each individual link agree exactly with the total from any "volumetric" counter on that link because, for example, the volumetric counters on some links did not provide information for the whole of the year.

4.1.8. These calculations produce estimates of traffic volumes for each road link (or part of a road link) which is within the area of each Council. The estimated traffic volume for each Council is then obtained by adding up the estimates for the relevant links (or parts of links), and the estimates for Scotland as a whole are then produced by adding up the estimates for each Council. As indicated earlier, DfT produced the figures for trunk roads by counting each major road link on the basis of its "trunk road status" at a recent date.

4.1.9 DfT's estimates of the total volume of traffic on major roads in each local authority area are based on "12 hours in one day" manual counts at an average of under 10 (up to 2002: under 15 for 2003 onwards) sites on major roads per Council per year - so they are clearly not based on much data. And, because the manual traffic counts are taken on a "rotating census" basis, there may be several years between successive counts at a particular site: in which time, there could be large changes in the volume of traffic there. The estimates therefore provide only a broad indication of the likely volume of traffic on major roads in each Council area. DfT notes that there could be some large percentage errors in its traffic estimates for the major roads in some local authority areas. Therefore, DfT's estimates for individual Council areas are not classed as National Statistics.

4.2. The method of estimating major road traffic volumes for 1992 and earlier years,

4.2.1. The method that was used to produce the estimates for 1992 and earlier years differed significantly, in several respects, from the current method.

4.2.2. Estimates for 1992 and earlier years were produced by the then Department of Transport (DoT) alone. There were significant differences in the kinds of data that were available for use. DoT did not have GIS-based information about the lengths and locations of individual major road links. Instead, it used information about the total length of roads of each type in each of the nine former Scottish Regions, and the three Island Areas, which was obtained from the road lengths returns (see Chapter 5; the lengths of Motorway slip roads were excluded from the calculations). In addition, because automatic counters had not then been introduced, the scaling factors were calculated from manual "core" traffic counts at about 130 fixed sites throughout GB (including about 20 in Scotland). These manual "core" counts were taken on three days in each month of the year (a weekday, a Saturday and a Sunday) for 16 hours each day.

4.2.3. The calculations were performed for each road type, for each Region (and Island Area). DoT first calculated the average traffic flow for each road type and area for the latest year by weighting the estimated traffic flow for each individual road link of that road type in that area (calculated as described above) by the total length of the link (as supplied to DoT by the then Scottish Office National Roads Directorate). Not having the GIS-based information required to split links which crossed boundaries, DoT counted each link as being in the Region which included the location at which the link's traffic count was taken. Therefore, each link contributed to the estimated average traffic flow for only one Region.

4.2.4. DoT then estimated the total traffic volume (vehicle kilometres) for each type of road in an area by multiplying the estimated average traffic flow for the road type and area (calculated as described above) by the total length of roads of that type in that area (as had been reported in the road length returns). The figures for the total road lengths for each area took proper account of links which crossed boundaries, because the people making the returns had to include only the length of each link that was within an area in the calculation of the total road length for that area. Therefore, the figures for an area's total road lengths could cover a somewhat different road network from that used to estimate its average flows (remember that the latter were calculated using data for only those road links for which the locations of their traffic counts were within the area).

4.2.5. It follows that old method of estimation was likely to be less precise than that used to produce the revised estimates. For example, suppose that there were only two major road links in a particular Region: a short low-flow link whose traffic count was taken at a point within the Region, and a long high-flow link, which crosses the boundary into another local authority, whose traffic count was taken at a point in the other area. Using the old method of estimation, the average traffic flow for the Region would be calculated using only the data for the low-flow link, and then multiplied by the total road length for the Region (including the length of the part of the high flow link that was within its boundaries). The total traffic volume for the Region would therefore be under-estimated: the method could not take account of the high traffic flow on the long link, because its traffic count took place in another local authority.

4.2.6. The estimates produced using the previous methods were also affected by a number of discontinuities, which were caused by changes in local government and trunk road organisation, changes in the availability of data and changes in methodology over the past ten or so years. Some of these discontinuities have been referred to earlier, and others are described in the previous edition. The introduction of DfT's revised method of estimation has removed all the discontinuities that previously affected the estimates for 1993 and subsequent years.

4.2.7. The earliest year for which there are estimates of the total volume of traffic on major roads in Scotland is 1983.

4.3. Method used to estimate traffic on minor roads for 1993 onwards

4.3.1 Estimates of traffic volumes on minor roads (B roads, C roads and unclassified roads) in Scotland by road type and vehicle type are produced by DfT in conjunction with the SE TRNMD.

4.3.2. The method used differs from that used for the major roads, because far fewer data are available for minor roads: up to and including 2002, only 200 or so "12 hours on one day" manual traffic counts per year were taken at Scottish minor road sites. In each of the years up to 1997, a fresh sample of sites was picked by, in effect, taking a series of random points on a map, looking within a circle with a specified radius around each point, and identifying which (if any) minor road was nearest to the selected point. The number of other minor roads within the circle was used, at a later stage, when the results were grossed-up to produce the overall traffic estimates. This method of sampling was suitable for the production of results for GB as a whole, but not for Scotland: the kinds of minor roads in the Scottish sample could vary greatly from one year to the next, and, as a result, the Scottish component of the GB estimates was not sufficiently reliable to be published in its own right.

4.3.3. Over the years, a list of all the minor road sites that had been chosen in this way built up, and became the basis for selecting a "panel" sample of minor road sites to be counted in 1998 and later years. Taking the counts at the same sites each year should produce a better estimate of the year to year percentage change in the volume of traffic on minor roads. The sample was picked from a list of all the sites at which traffic counts had been taken between 1992 and 1997. Disproportionate stratified sampling was used, with a higher sampling fraction for roads which had had a greater volume of traffic, as this should produce more accurate results than a simple random sample of minor road sites. Sites with average flows of less than 200 vehicles per day were excluded altogether. Some of the sites chosen for the panel for 1998 were found to be unsuitable, and were replaced by substitute sites in the panel for 1999. There was little change in the composition of the panel of sites until 2003, when, following a study of possible ways of improving the traffic estimates for Scotland, SE increased the number of minor road traffic counts in Scotland to about 320 or so per year.

4.3.4. As with the major road traffic counts, the minor road "12 hour" traffic counts must be "expanded" to estimate the flows for a whole day, and a whole year. This is done using expansion factors calculated from information recorded by a set of "core" automatic traffic classifiers located on a sample of roughly 40 minor roads across GB, of which about 5 are in Scotland.

4.3.5. The data from the GB-wide "core" automatic traffic classifiers were used to calculate growth and expansion factors for minor roads outwith London (with separate sets of factors for "urban" and "rural" roads of each class). There are too few "core" classifiers in Scotland for there to be any separate Scottish groupings.

4.3.6. The number of manual counts per year at minor road sites across Scotland represent an average per local authority area per year of only 6-7 (up to 2002) and only 10 (2003 onwards) - clearly, too few to be the basis for reliable estimates of minor road traffic for individual local authority areas calculated solely from the data collected in each year. DfT had therefore to estimate the volume of traffic on minor roads in individual local authority areas in other ways. DfT started by producing estimates of the volume of traffic on minor roads in each local authority area in 1999 (as that is the new base year for its panel of minor road manual traffic count sites). The "information base" for these estimates was widened to include manual counts taken in other years by "uprating" them to 1999 using the growth factors produced from the "core" counters. DfT used different methods for "B" roads and for other minor roads ("C" roads and unclassified roads).

4.3.7. "B" roads: DfT looked at the location and traffic levels of all the "B" road manual traffic count sites, including ones counted in the past that were not included in the panel sample, identified gaps in coverage and initiated extra counts where necessary. Using its knowledge of the variation in B road traffic by type of location, and the length of B roads in each area, DfT produced estimates of B road traffic for each local authority area.

4.3.8. "C" and unclassified roads: Estimating traffic on other minor roads was more difficult, and had to be done in another way. First, DfT estimated the average levels of traffic flow on each type of these roads across GB (e.g. "urban C roads", etc), using the information from the minor road manual counts and "core" counters. Second, DfT compared the average levels of traffic flow on the non-trunk A roads in each local authority area with the GB average traffic flows for such roads. Third, DfT made the assumption that an area which has non-trunk A road flows that are above the GB averages will also have minor road flows that are proportionately greater than the corresponding GB averages, and that an area whose non-trunk A road flows are below the GB averages will have proportionately lower flows on its minor roads. DfT then estimated the flows for each type of minor road in a local authority by applying to the GB average flows for each type of minor road the relevant ratios (of its non-trunk A road flows to the corresponding GB averages). The resulting estimates were multiplied by the length of minor road of each type in that local authority to give the estimated minor road traffic volumes for the area. This produced what DfT considered to be sensible results for many local authorities. However, there were some areas for which DfT felt the results were odd in relation to those for nearby areas or similar areas. For these local authority areas, DfT undertook a more detailed study. This involved looking at the minor road traffic count data for different parts of the local authority, deriving a "traffic intensity" value for each part, and comparing the results with the "traffic intensities" of other local authorities for which DfT was confident about the minor road traffic estimates, in order to produce what DfT considered to be more credible estimates for some parts of the local authority. The resulting estimates were then added together to produce totals for the local authority as a whole, and the results for all the local authorities in Scotland were then added together to produce minor road totals for each area and for Scotland as a whole.

4.3.9. DfT used its estimates for 1999 as the basis for the estimates for earlier years and for later years. The minor road traffic volumes for the years prior to 1999 were estimated by applying year to year change factors, which were calculated from the information produced by the panel survey. The estimates for later years were calculated using the estimated overall percentage changes in traffic flow levels (obtained from the results of the panel survey) and information about changes in the length of the minor road network.

4.3.10. Given the assumptions that DfT has to make, and the fact that its estimates of the total volume of traffic on minor roads in each local authority area are based on "12 hours in one day" manual counts at an average of 6-7 (up to 2002: about 10 for 2003 onwards) sites on minor roads per Council per year, it is clear that these estimates can only provide a broad indication of the likely volume of traffic on minor roads in each local authority area. That is why figures for individual minor road types are not published for local authority area: only the total volume of minor road traffic for each area appears in Table 6.4, with no breakdown by type of minor road within local authority. DfT notes that there could be some large percentage errors in its traffic estimates for the minor roads in some local authority areas. Therefore, DfT's estimates for individual Council areas are not classed as National Statistics.

4.3.11. 1993 is the first year for which there are estimates of the volume of traffic on minor roads for individual local authority areas, and also is the first year for which there are estimates for Scotland as a whole. There are no reliable estimates of the total volume of minor road traffic in Scotland for 1992 or any earlier year.

4.4 Average time lost by traffic on specific trunk road routes

4.4.1 The Scottish Executive Trunk Roads Network Management Division produces the estimates of the average time lost by traffic on specific trunk road routes. The routes for which the estimates are produced are those sections of the trunk road network which presently experience congestion, or which are thought likely to experience congestion over the coming years, and which are therefore covered by TRNMD's congestion monitoring work.

4.4.2 Contractors working for TRNMD produce the estimates from two sources of data about the speeds of traffic on those sections of the trunk road network: automatic traffic counters and so-called "floating vehicle" surveys.

4.4.3 The automatic traffic counters use sensors which are buried under the surface of the road. They run continuously, and record the numbers of vehicles passing each site, and the speeds at which they travel. The counters collect large amounts of data, which are then aggregated and stored as overall figures for 15-minute periods. The number of counters used has been increasing: while some of the counters already on the monitored routes were collecting the kind of data needed, others had to be upgraded, and some new counters had to be installed, in order to achieve the required level of coverage of the monitored routes. Inevitably, this took time - for example, because roadworks are required to install a new counter. As a result, the quality of the estimates will have improved during 2003, and will improve further during 2004. Once all the new and upgraded counters are operating, data will be available from automatic traffic counters at over 300 locations on the monitored routes, with information collected about the speed of traffic in both directions at each location.

4.4.4 The speed data for each section of road covered by a particular monitoring site are validated and calibrated using what are called "floating vehicle" surveys. In these, vehicles drive the routes at speeds which are representative of the traffic flow in which they are travelling (by balancing the numbers of vehicles that they overtake and which overtake them) and record their speeds and times taken along the route. A particular stretch of road is surveyed several times, on different days and at different times of the day, in order to obtain a representative range of results. The surveys also provide some information which is unavailable from the automatic traffic counters, such as the time which is taken by traffic queuing at junctions.

4.4.5 The contractors produce the estimates by combining the information from the two sources, using a specially-developed methodology and considerable computer processing of the data. The figures for 2003 are provisional, because they may be revised following improvements to the methodology and further data becoming available. The contractors scrutinised the results, and checked that some apparent anomalies could be the result of known unusual circumstances rather than problems with the data or the computer programs. For example:

• the high values for the A96 for November - very bad weather in that month;
• the high value for the A90 Inchture eastbound towards Dundee for December - traffic management during the reconstruction of a roundabout near a new supermarket;
• the low values for some Glasgow routes in July - reduction in traffic in the holidays; and
• the high values for the A725 northbound in November and December - major roadworks.

4.4.6 A more detailed description of the method of producing these estimates appears in a report by the Scottish Executive Traffic Controller - see Section 5.4.

4.5 Scottish Household Survey

4.5.1 Information about the Scottish Household Survey is given in Chapter 12.

4.6 Pollutants and air quality objectives

4.6.1 The information on pollutants is taken from the Scottish Executive online publication "Scottish Environment Statistics Online". Some of the data are additionally published in the Scottish Executive National Statistics publication "Key Scottish Environment Statistics". The air quality objectives are taken from "The Air Quality Strategy for England, Scotland, Wales and Northern Ireland: Addendum".

5. Further Information

5.1 Further information on GB road traffic statistics can be found in the annual DfT publications "Road Traffic Statistics" and "Transport Statistics Great Britain", and also in the former DETR's "Focus on Roads" publication.

5.2 For enquiries about DfT's methods of estimating road traffic, contact Mr John Garnsworthy of the Department for Transport (Tel: 020 7944 6396).

5.3 For further information on average daily traffic flows at selected Automated Traffic Classifier (ATC) Sites and on key routes on the road network contact Mr Stuart Hay of the Scottish Executive Trunk Roads Network Management Division (Tel: 0131 244 0458).

5.4 Further information about the time lost by traffic on trunk roads will be given in the Scottish Executive Traffic Controller's report on congestion on Scottish trunk roads, which is being prepared at the time of writing. Enquiries regarding the availability of this report, and for more detailed information about the basis of the figures, should be made to David Hamilton of the Scottish Executive Trunk Road Network Management Division (Tel: 0131 244 0447).

5.5 For further information about the Scottish Household Survey figures on drivers’ experience of congestion, contact Frank Dixon of the Scottish Executive Transport Statistics branch (tel: 0131 244 7254).

5.6 For further information on petrol and diesel deliveries contact Mr Ian Corrie of The Department of Trade and Industry, Energy Policy and Analysis Unit (Tel: 020 7215 2714).

5.7 For further information on pollutants see "Scottish Environment Statistics Online", found at: http://www.scotland.gov.uk/stats/envonline/menu0.asp, alternatively contact John Cameron of The Scottish Executive, Environment Statistics branch (Tel: 0131 244 0445).

< Previous | Contents | Next >