Do Sustrans figures add up?

How are the figures obtained?

Sustrans uses a mixture of automatic cycle counters and manual surveys to determine the total usage of the network. The sites at which this information is gathered are classified by type (eg urban, rural, etc) and the results are assumed to be typical of the parts of the National Cycle Network which are of that type. Changes in the system of classification can have a dramatic effect. In 2004 Sustrans decided to reclassify areas such as town parks and the fringes of urban areas as urban rather than rural. As a result of this and some other changes, the number of trips on the network in 2003 was revised from 126 million to 180 million!

Automatic Cycle Counters

Automatic counters would seem to be the ideal way of measuring NCN use, both in terms of absolute numbers and trends. There have to be enough of them, though, and they must be in places which are representative of the whole network and they need to be in place all year round. Automatic counters are owned by local authorities and there are about 300 of them throughout the country, at sites covering both NCN and non-NCN routes, traffic-free and road routes, and both urban and rural areas. In 2003 Sustrans took the data from the 153 counters that could provide two years' worth of figures (ie for both 2002 and 2003), but only some 40 or 50 of those were on NCN routes, roughly one for every 200 miles of the network, so the chance of these counters being in places which are typical of the wider network seems somewhat remote.

The Yeo Bridge at Barnstaple is fitted with an automatic cycle counter but how typical is this location of the wider National Cycle Network?

Unfortunately, as far as accuracy of the results in concerned, the counters aren't spread evenly across the land. As of 2004 tiny Rutland had six, the same as Glasgow. Lancashire (county) has 2 permanent automatic counters. Lancaster (city) has three that are moved between 12 sites, while Pendle and Burnley districts were about to install some. Surrey had none! Hampshire had 43 permanent cycle counters, but these were clustered around Winchester, Eastleigh, Gosport, North Baddesley and the New Forest and information was typically collected for only 2 or 3 months of the year. Of Devon's 49 counters, 31 were concentrated in Barnstaple, Exeter, and Newton Abbott. Essex was coy about how many it had and where they were.

Since the counters are primarily intended to show that councils are meeting their cycling targets, they're put in places that will show the most favourable results, not half way along a muddy footpath! It's also worth remembering that the purpose of these counters is to measure year-on-year changes in cycle usage in specific areas – they're not sited to assess typical cycle usage across the country.

Applying the counter results

Sustrans don't explain precisely how they calculate overall use of the network from so few counters and extrapolating 40 (probably non-typical) counts to cover the whole country requires a massive leap of faith! It appears that the site of each counter is classified according to its location – urban or rural, traffic-free or road, etc. Each section of the NCN is classified in a similar way and the results from the counters are applied (extrapolated) to those parts of the NCN with the same classification. Of course, the accuracy and reliability of this method depends on the sophistication of the classification system.

Both of these scenes are on the Devon Coast to Coast route of the NCN and only a few miles apart. Both could be classed as rural/traffic-free. But could a count from one realistically be applied to the other?

For example, a counter on the Tarka Trail between Barnstaple and Bideford, which forms part of the NCN's route 3 West Country Way, might be classed as rural/traffic-free. The section of the Tarka Trail further south, near Meeth, might also be classed as rural/traffic-free, but anyone who knows the route will appreciate that the one is far more heavily used than the other. To apply the the results from Barnstaple-Bideford to Meeth would give a totally misleading result. A very detailed classification system is needed to overcome this problem, with enough counters to fully cover each of the different categories of location. This, obviously, is not the case today.

In addition, the system cannot prevent double-counting, and the clustering of counters in urban areas is likely to make this a problem. A cyclist making one trip could pass two or more counters in the course of that journey, making that one trip seem as if it was two, three or more.

The impact of changes

In 2004 Sustrans decided to make some changes to the classification of parts of the network – town parks and urban fringe areas were reclassified from rural to urban, urban paths having a much higher usage than rural ones. The result of this and some other changes was to restate the calculated usage of the network in 2003 from 126 to 180 million trips, an increase of nearly 50 per cent. By 2005 this change had added around 70 million trips a year onto the claimed usage of the NCN. Clearly, getting the classification of the parts of the network right has a big impact on the results.

Manual surveys

The results from automatic cycle counters are supported by manual counts and surveys. A paper published in 2003* described how the manual counts and surveys were carried out. In 2001 22 monitoring points used in the survey. It counted 15,118 cyclists over the 4 days of the exercise – an average of 172 at each survey point per day – and 1,464 of those cyclists were interviewed – that's 17 per point per day. Such short-term counts cannot tell us very much about the total usage of the network so I have to question why they were carried out but the surveys can tell us more about the people who use the NCN and the purposes for which they make their journeys.

Half of the four survey days were weekend days, despite the fact that weekends only account for a much smaller proportion of the week. Two of the four days were during school holidays, which, again, account for a far smaller proportion of the year. And the survey dates were in late summer and early autumn, which hardly reflects the year's weather patterns. All of these factors will tend to sway the results.

The number of surveys varies each year. Between 2000 and 2002 it was an average of 24 a year, in 2003 it was 36, bu then in 2004 it leapt to 80. In 2005 50 surveys were carried out, but this is still only one for each 200 miles of the network.

The Millennium Coastal Park near Llanelli, once industrial waste land but now converted into a wonderful recreational facility, but hardly typical of the National Cycle Network. An NCN user survey was carried out here in 2001.

The choice of survey points in 2001 was interesting. They included:

• Strathyre, near Stirling, is on a long-standing railway path where there is a picnic area, car park, etc.
• Bowling Harbour is on the equally well-established cycle path between Glasgow and Dumbarton (on the way to Loch Lomond)
• Wansbeck Estuary is on a cycle path beside a new dual carriageway bridge over the River Wansbeck near Blyth in the north east
  – the alternative route, if you don't want to ride on the main road, is a 4-mile diversion to the next bridge upriver
• Elswick Riverside is in the centre of Newcastle
• Darlington to Stockton Railway is on a railway path linking Stockton-on-Tees and Darlington
  – the alternative route is on the A66 or the A67
• Eston and Normanby: a good cycle path through parkland in an otherwise heavily built-up area near Middlesbrough
  it's part of NCN route 1 and the White Rose Middlesbrough-York-Hull cycle route
• Blackburn, Oakenhurst Road is a minor road route out of Blackburn to the Witton Country Park
• Rishton Railway Path, Lancashire, is railway path between Blackburn and the Burnley area
  – it links to Leeds and Liverpool Canal towpath
• Spen Valley Greenway is 8-mile route opened in 2000, a railway path running near the River Spen between the towns of Cleckheaton, Dewsbury and Heckmondwike, eventually linking to Bradford.
  – the route runs through densely populated urban areas with long distance moor-land views, it passes a wild life reserve and a rolling golf course.
• Rhyl Promenade is a 10-mile route along the sea-front
  – How many people ride along it not realising it's part of the NCN and how many already used it before it became legal to do so?
• Stoke Greenway: running along the canal towpath through the centre of the built-up area
  – the same applies as for Rhyl.
• The Millennium Coastal Park – a largely recreational facility developed out of industrial wasteland along the shore of the estuary near Llanelli
• Fremington: between Barnstaple and Bideford on the Tarka Trail railway path, primarily recreational.

Few of these routes seem typical of the average stretch of the NCN. Many are medium length stretches of railway (and similar) paths in built-up areas or holiday spots which are are bound to prove popular, but which are unlikely to attract typical users of the average section of the network.

The importance of accuracy

It is, of course, vitally important that the figures about the usage of the National Cycle Network are accurate. They are a measure, perhaps the only one, of whether the Sustrans concept is effective in delivering the benefits it claims. It informs funding partners – government, local authorities, the lottery bodies, etc – and is used to justify further funding decisions. Successful figures help to assure Sustrans the further funding it needs in the future.

*The UK National Cycle Network: an assessment of the benefits of a sustainable transport infrastructure. Cope A et al. World Transport Policy & Practice, Vol 9 no 1, 2003