Junction capacity modelling outputs an officer can read

Junction capacity modelling outputs are read by officers who are not modellers, so the test of a good set of results is whether a case officer can see, in one table, which arm fails, in which scenario, and by how much. Most of the arguments in a Transport Assessment about junction capacity modelling outputs are really arguments about presentation. The numbers are often agreed. What is in dispute is which scenario they belong to and what they mean.

What capacity modelling is actually testing

Capacity modelling takes the turning flows at a junction and asks whether the junction geometry can process them within the modelled hour. It does not test safety, it does not test severance, and it says nothing about whether a scheme is acceptable in planning terms.

The output is a measure of how hard the junction is working, plus a queue and a delay. A junction operating within capacity can still be refused on road safety or on the quality of the pedestrian crossing. A junction shown as over capacity can still be acceptable where the increase caused by the development is a small fraction of the problem.

Keep that distinction visible in the text. Officers who have been handed a 60-page modelling appendix with no narrative tend to default to the highway authority's view, whatever it is.

Priority junctions, roundabouts and signals

Three families of method, and they produce different metrics, which is the first source of confusion in a mixed junction assessment.

Priority junctions and roundabouts are modelled with empirical capacity relationships derived from observed behaviour, with capacity falling as the conflicting flow rises. The headline metric is the ratio of flow to capacity.

Signal junctions are modelled from saturation flows, cycle times and green splits. The headline metric is the degree of saturation for each approach lane, with practical reserve capacity expressing the spare or deficit for the junction as a whole.

Those thresholds are conventions rather than policy. Some highway authorities apply 0.85, others look at 0.9 with queue lengths, and a few will accept more in an urban centre where the alternative is widening a street nobody wants widened. Ask which convention applies locally before the modelling is run.

Reading saturation, flow to capacity and queues

The ratio of flow to capacity is a decimal. At 0.85 the arm is working at its practical limit and queues begin to grow non-linearly. At 1.0 demand equals theoretical capacity, queues lengthen through the hour and the model's queue output becomes unreliable.

Degree of saturation works the same way as a percentage. Ninety per cent is the usual practical limit. Above 100 per cent the approach is oversaturated and the residual queue carries into the next hour, which a single-hour model does not represent.

Queues are reported in passenger car units, not vehicles, and usually as a mean maximum queue rather than the worst observed. One passenger car unit is about 5.75 m of road, so a mean maximum queue of 12 units is roughly 70 m. Converting queues to metres in the text, and saying where 70 m of queue reaches, is the single most useful thing a transport consultant can do for a case officer. If it blocks the access to the school, say so.

Delay is reported in seconds per vehicle and is the metric most often ignored. A signal junction running at 88 per cent saturation with 90 seconds of delay per vehicle on one arm is unpleasant in a way the saturation figure hides.

Scenario labelling and the do minimum case

Scenario naming is where assessments go wrong, and fixing it costs nothing.

1. Base year, validated against observed counts and queues.
2. Future year do minimum, which is the base grown to the assessment year with committed development included and without your scheme.
3. Future year do something, which adds your development.
4. Future year do something with mitigation, where mitigation is proposed.

The do minimum case decides the argument. If it already shows the junction over capacity, your scheme is adding to an existing problem and the question becomes whether the increment is material. If it shows the junction comfortable, the development owns the whole impact.

Growth to the assessment year normally comes from the national trip end model datasets, and committed development comes from an agreed list of consented schemes. Both should be stated and both are negotiable, so agree them in writing at scoping. A disputed committed development list is the most common cause of two sets of modelling that cannot be compared.

Picking the hour and the period

The weekday morning and evening network peaks are the default, usually taken as the busiest hour within 07:00 to 10:00 and 16:00 to 19:00. Use the surveyed peak hour rather than an assumed one.

Where the development's own peak falls elsewhere, model that too. A food store peaking on Saturday afternoon, a school peaking at 08:30 for 20 minutes, a stadium peaking after an event. Modelling a single weekday hour for a use like that is an invitation for the highway authority to ask for more work later, and our note on school access, drop-off and school streets shows how short, sharp peaks behave.

One honest limitation. A one-hour model cannot represent a queue that carries over from the previous hour, so where the preceding hour is also near capacity the results understate the problem. Say that in the text rather than waiting to be told.

Reporting results so they can be checked

Put a summary table in the main report as well as the appendix, with one row per arm and one column per scenario, showing the headline metric, the queue in units and in metres, and the delay. Flag the arms over the local threshold.

Give the model files or the full output in an appendix, state the software version, and list the inputs that were assumed rather than measured, such as saturation flows, cycle times and geometry taken off a drawing rather than surveyed. A modeller who tells you which numbers are assumptions is easier to trust than one who presents everything as measurement.

Where mitigation is proposed, show the junction with and without it, and be specific about what the mitigation requires. A signal upgrade or a widened entry on land outside your control needs a highway agreement under section 278 of the Highways Act 1980, and the bonding and commuted sum consequences belong in the appraisal. Servicing arrangements interact with the same geometry, which our note on servicing yards, standing and reversing covers.

Where results get challenged

Highway authorities challenge four things, in this order. Whether the base model was validated against observed queues as well as flows. Whether the committed development list is complete. Whether the peak hour is the right one. And whether the geometry in the model matches the drawing in the application, which is a frequent mismatch after a layout revision.

Answer all four in the text and most technical objections disappear. Where the junction fails outright and cannot be mitigated, the honest route is to say so and argue the planning balance, supported by the mode shift commitments in a Travel Plan, rather than to produce a fifth scenario with optimistic saturation flows.

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