Trip distribution using travel to work data

Trip distribution decides which junctions see your development traffic, so a weak distribution will sink an otherwise sound Transport Assessment. Census travel to work flows are the standard evidence base for residential distribution in England, and using them well means knowing what the 2021 data can and cannot tell you. The short version: the 2021 Census was taken in March 2021, when homeworking was at its peak, so its journey-to-work flows understate commuting and distort mode share in ways that a highway authority will question.

Trip generation gives you a number. Distribution gives that number a direction, and the direction is what the capacity results depend on.

Why distribution drives the junction results

A scheme generating 60 two-way peak hour vehicle trips is harmless or fatal depending on where they go. Push 70 per cent of them right at a signalised junction already running at 95 per cent degree of saturation and you have a refusal reason. Spread them four ways and nothing moves.

That sensitivity is why highway authority officers read the distribution chapter before the generation chapter. They are looking for one thing: whether the split was derived from data or assumed to be convenient.

The failure mode is familiar. A consultant adopts a 50/50 split either side of the access, offers no source, and the officer reruns the junction with 80/20 towards the town centre. The scheme then fails on a basis nobody tested, and the applicant pays for a second round of modelling.

Census travel to work data as a source

Origin-destination data from the Census records where usual residents live and where they work, down to output area and workplace zone geography. It is published by the Office for National Statistics and remains the only consistent national source of commuting flows with that resolution.

For a residential scheme the method is direct. Take the flows out of the output areas that best represent your site, aggregate the destinations to meaningful zones, convert to percentages by car driver mode, then route each zone through the network to see which arm of the access it uses.

Two cautions matter more than anything else in the chapter.

The first is the 2021 distortion already mentioned. Many practitioners still base distribution on the 2011 Census, or present both and compare them, because 2011 reflects a normal working week. Say which you have used and why, in the text, rather than in a footnote.

The second is that travel to work data covers commuting only. It tells you nothing about the school run, the supermarket trip or a Saturday peak, and for anything other than a residential or employment scheme it is the wrong dataset. A retail or leisure proposal needs survey-based distribution from comparable sites, or household survey data where it exists locally.

Gravity models and when they are used

Where census flows are too coarse, or where the land use has no commuting pattern to borrow, a gravity model distributes trips in proportion to the attraction of each destination and inversely to the cost of reaching it.

In practice that means a doubly constrained model, balanced iteratively until row and column totals match the generation and attraction estimates. It is defensible work when the inputs are defensible. The attraction term is usually floorspace, jobs or population, and the deterrence term is generalised cost or journey time from a highway model.

But a gravity model built on invented attraction weights is worse than a census-based split, because it looks sophisticated while hiding the assumption. If you use one, publish the attraction and deterrence terms, the convergence criterion and the number of iterations. An officer who cannot reproduce the output will not accept it.

Local knowledge adjustments

Raw census percentages always need some adjustment, and this is where judgement enters the method.

The adjustments that tend to be accepted are a correction where the output area geography straddles a river or railway with no crossing, a reallocation where a new link or closure has changed routeing since the census, a seasonal correction for a coastal or tourist location, and a tenure correction where the scheme is substantially different from the surrounding housing stock. Each needs a sentence of justification and ideally a map.

The adjustments that get rejected are the ones that move trips away from the constrained junction with no stated reason. Officers have seen that often enough to look for it.

Assigning trips to the network

Assignment converts destination zones into turning movements. On a small scheme that is a manual exercise with a routeing table, comparing journey times by the two or three realistic routes from the access to each zone and splitting where they are close.

Record the route choice logic. A table showing each zone, its percentage, the chosen route and the junctions it passes through lets a reviewer check your work in ten minutes, and that single table resolves more objections than a longer narrative.

Where the network is congested the assumption of shortest-time routeing breaks down, because drivers use rat runs. Residents' groups will tell you exactly which ones, and ignoring that is how a transport chapter loses credibility at committee. The geometry side of those local routes is covered in our note on visibility splays and the stopping sight distance calculation.

Sensitivity testing the distribution

Any distribution is an estimate, so test it. The standard approach is to rerun the critical junction with a distribution weighted towards the worst arm, typically by shifting ten or twenty percentage points, and report whether the conclusion changes.

If the junction still operates within capacity under the pessimistic case, the chapter is effectively argument-proof and the officer can agree it without further work. If it fails, better to know before submission than at the consultation stage, because the mitigation design then has time to catch up.

Sensitivity testing also protects against the single-source problem. A 2011-based distribution tested against a 2021-based one, with both reported, shows the authority that the recommendation does not depend on which census you preferred.

Agreeing the method before modelling

The Planning Practice Guidance on travel plans and transport assessments expects the scope of assessment work to be agreed with the local highway authority, and distribution method belongs in that scoping discussion rather than in the submitted report.

Put four things in the scoping note: the census year and geography used for distribution, the zones and their percentages, the assignment routes, and the sensitivity tests proposed. Ask for written agreement. An officer who has signed off the method in advance will rarely reopen it, and the same discipline applied to a Transport Statement on a smaller scheme saves a round of correspondence.

Where the scheme also triggers public transport contributions, the distribution evidence feeds that case directly, as set out in bus infrastructure contributions and their triggers.

One honest limitation: no distribution method predicts individual driver behaviour, and the usual defence is transparency about inputs rather than a claim of accuracy.

PlanScope provides AI-assisted draft reports for planning applications. Reports should be reviewed before submission. We do not guarantee planning approval. For specialist technical reports, qualified professional input may be required.