Selecting TRICS sites for a Transport Assessment

TRICS site selection decides the trip rate, and the trip rate decides almost everything downstream: junction capacity results, the need for mitigation, the scale of any highway works and, often, whether the scheme is recommended for approval. Two competent transport planners working from the same database and the same scheme can produce rates differing by half, purely through filtering choices. That is why highway authorities audit the selection rather than the arithmetic.

The defensible position is a filter set you can explain line by line, with every exclusion justified on grounds that have nothing to do with the number it produced.

Why selection matters more than the calculation

The calculation is trivial. You take the average rate per unit across the selected sites and multiply by the quantum proposed. The judgement sits entirely in which surveys enter the average.

A residential rate drawn from car-dependent edge of town sites in the East Midlands will be higher than one drawn from suburban sites with good bus provision. Neither is wrong. One of them is wrong for your site.

Start from the site you are assessing and write down its characteristics before you open the database. Land use, unit count or floorspace, location type, settlement size, parking ratio, walking distance to a centre and public transport provision. Filter towards that description rather than browsing and then rationalising.

Filtering by land use, region and location

Four filters do most of the work in a Transport Assessment.

Land use category and sub-category come first, and the sub-category is where people go wrong. Privately owned flats and affordable housing generate differently. A retail warehouse and a discount food store sit in different sub-categories for good reason.

Region and area type come next. Greater London behaves differently from everywhere else and should normally be excluded from a non-London assessment, and included alone for a London one. Beyond that, be careful about excluding regions too aggressively, because a tight regional filter can leave you with two surveys.

Location type, described in terms such as town centre, edge of town centre, suburban area, neighbourhood centre and free standing, is the filter that best reflects accessibility. Match it honestly. A site on the ragged edge of a market town is not a suburban area.

Settlement size, expressed through population ranges, is the fourth. A 15,000 population market town and a 400,000 population city produce different travel behaviour for the same land use.

Parking, sample size and the single comparator problem

Parking provision is the filter most often skipped and the one that changes vehicle rates most directly. Where the database records parking spaces, filter to a range around your proposed ratio. A scheme at 0.5 spaces per dwelling assessed on surveys of sites with 2.0 spaces per dwelling overstates its own traffic, which sounds conservative until the highway authority asks why you have justified a car park you do not need.

On sample size, practitioners generally look for at least five comparable surveys and highway authorities often press for more. The reason is stability. With three sites, one atypical survey moves the mean by a third.

The single comparator problem is worse than a small sample. If one survey has a much larger unit count than the others, it can dominate a weighted rate entirely. Check the spread before you accept the output, and report the range and the standard deviation alongside the mean. A report that gives only an average invites the question of what it concealed.

Where the filtered set is too small, say so and widen one filter deliberately, explaining which and why. Widening the region is usually safer than widening the location type.

Survey period and day selection

Use surveys taken on neutral weekdays, and exclude weekend surveys from a weekday peak assessment. That sounds obvious and it is routinely got wrong in mixed sets.

Survey age needs a stated cut-off. Travel behaviour has shifted, and surveys taken during 2020 and 2021 are not representative of normal conditions. Set a date range, state it in the report, and apply it consistently rather than case by case.

The peak hours themselves should come from your own baseline counts at the site, not from an assumption. Derive the network peak from the junction surveys, then extract the corresponding TRICS hour. Assessing a residential scheme at the network peak of 08:00 to 09:00 when the site's own generation peaks at 07:45 is a mismatch a highway officer will find.

Multi-modal rates and mode share

Vehicle rates alone are no longer enough for most authorities. The database holds multi-modal surveys recording pedestrians, cyclists, public transport users and vehicle occupancy, and these support the person-trip approach that sits behind a credible Travel Plan.

Two cautions. The multi-modal sample is smaller than the vehicle sample, so a filter set that works for vehicles may return almost nothing for mode share. And mode share is more sensitive to location than vehicle rates are, so a loosely filtered multi-modal set is less reliable than a loosely filtered vehicle set.

Where the sample will not support a mode share derived from the database, take the vehicle rate from TRICS and the mode share from census travel to work data for the relevant area, and say plainly which figure came from where.

Justifying the selection in the report

Include the full database output as an appendix, not a summary table. The output records the filter criteria, the sites selected, the sites excluded and the resulting rates, which is what makes the work reproducible.

In the body of the report, set out six things: the site characteristics you filtered towards, each filter applied with its range, the number of surveys returned, any sites manually excluded and the reason for each, the resulting rates with the range around the mean, and the sensitivity of the assessment to the choice.

Manual exclusions deserve particular care. Excluding a survey because the count contains an obvious anomaly, or because the site has an attribute the database does not capture such as an adjacent railway station, is legitimate and should be stated. Excluding a survey because it produced a high rate is not, and an auditor comparing your excluded list against the rates will see it immediately.

How a highway authority audits the choice

Most highway authorities re-run the filter set. That takes them ten minutes, and it is why the appendix matters.

They look for four things. A filter set that reproduces your stated sample. Exclusions with reasons that are independent of the outcome. A location type and parking range consistent with the application drawings. And consistency between the rates used and any earlier submission on the same site, including anything in a Transport Statement submitted at pre-application stage.

Agree the filter set at scoping. A short note proposing the land use sub-category, region, location type, settlement size band, parking range and date range, sent before you run the assessment, converts a contested chapter into an agreed one. Where the authority proposes different criteria, run both and present the comparison rather than arguing.

Parking demand is a separate exercise again. A Parking Survey using an overnight occupancy methodology answers a different question from the trip rate, and the two should not be conflated in the same chapter.

An AI-assisted draft can structure the transport chapter, set out the policy tests and prompt for each input the assessment needs, flagging missing site data rather than inventing it. PlanScope provides AI-assisted draft reports for planning applications. Reports should be reviewed before submission. We do not guarantee planning approval. Councils usually expect a transport assessment to be prepared by a suitably qualified transport planner, and professional review is strongly advised.

Browse all 66 report types at PlanScope to scope the rest of the evidence base alongside the transport work.