Junction modelling outputs a highway officer will check
A junction model is only as good as its base year and its assumptions log. Here is what a highway officer checks first and where submitted models usually fail.
A highway officer reading a junction model looks at the base year before looking at anything else. If the model cannot reproduce the queues that form on the approach every weekday morning, the future year results are arithmetic without meaning. That single check disposes of more submitted models than any argument about growth rates, and it is the reason experienced transport planners spend most of their effort on validation rather than on the forecast.
Everything else in the assessment follows from whether the base is believed.
What junction modelling is meant to show
Modelling answers a narrow question: will the junction operate acceptably with the development in place, in the assessment year, at the times that matter.
It is not a prediction of what will happen. It is a comparison between scenarios built on the same assumptions, and its value lies in the difference between them rather than in the absolute numbers. A model showing a queue rising from 14 to 16 vehicles is telling you something useful even if neither figure is exactly right.
That framing should shape the presentation. Report the without-development and with-development cases side by side for each scenario, and state the difference. Officers are assessing the impact of the development, not the general health of the junction.
The tools differ by junction type. Priority junctions and roundabouts are modelled with the standard capacity packages based on empirical relationships between geometry and capacity. Signal controlled junctions are modelled with signal design software that optimises stage timings. Complex or interacting junctions may need microsimulation, which is more work and should be reserved for cases that need it.
Base year validation and observed queues
Build the base year from primary data. Classified turning counts at the junction, collected in a neutral month, on days representative of normal conditions, covering the peak periods relevant to the site.
Then validate. Compare modelled queues against observed queues recorded during the same surveys, and modelled delay against journey time observations where the junction is busy. Report the comparison rather than asserting that the model is validated.
Where the model will not reproduce observations, the usual causes are recognisable: geometry measured from an aerial image rather than on site, flare lengths taken as the full length of the nearside lane, pedestrian crossing demand omitted, blocking back from a downstream junction, or an inappropriate assumption about how drivers actually behave at a give-way line.
Say what you found and what you did. A model that has been adjusted, with the adjustment explained, is far more credible than one presented as a perfect fit.
Growth factors and committed development
Future year flows come from two sources, and both need a stated provenance.
Background growth, normally derived from the national trip end forecasts, applied for the correct area, the correct time period and the correct assessment year. State the version of the dataset, the geography used and the date you extracted it.
Committed development, meaning schemes with permission that are not yet built and not yet in the surveyed counts. Get the list from the local planning authority rather than assembling it yourself, agree it in writing, and set out for each scheme its size, its assumed trip generation and how the trips were assigned to the network.
Double counting is the classic error. Applying background growth and then adding a committed development that is already inside the growth forecast inflates the future case and makes your own scheme look worse than it is. Say explicitly which approach you have taken.
Reading capacity ratios and queue outputs
Two families of output do the work.
| Junction type | Capacity measure | Practical limit conventionally applied | Queue output |
|---|---|---|---|
| Priority junction | Ratio of flow to capacity | Around 0.85 | Mean maximum queue, in vehicles |
| Roundabout | Ratio of flow to capacity per arm | Around 0.85 | Mean maximum queue, per arm |
| Signal junction | Degree of saturation | Around 90 per cent | Mean maximum queue, per lane |
These limits are conventions, not statutory standards, and local highway authorities apply them with varying strictness. What they represent is the point at which small increases in flow produce large increases in delay, so an arm at 0.83 and an arm at 0.92 are in materially different conditions even though the flows differ by little.
Report queues in the same units as the observations you validated against, and give the peak period for each. A single figure for the morning peak hour hides the fifteen-minute period when the school run and the commuter peak coincide.
Where a result sits above the conventional limit, address it directly. Explain the magnitude of the development's contribution, describe what mitigation would achieve, and set the residual position against the policy test, which is whether the residual cumulative impact on the road network would be severe.
Sensitivity testing and why officers ask for it
Sensitivity tests exist because the central case rests on assumptions that could reasonably be different.
The tests most often requested are a higher growth scenario, an alternative trip distribution reflecting a different assumption about where residents travel, a scenario with no mitigation to show what the mitigation is doing, and occasionally a test with a large committed development assumed to be complete earlier than expected.
Run them and publish them. An assessment that includes an uncomfortable sensitivity test and explains why the conclusion still holds is more persuasive than one that presents a single flattering case. Officers who suspect a test has been run and withheld will ask for it, and the correspondence costs more time than the test would have.
Common errors in submitted models
The recurring faults are easy to check for and easy to avoid.
- Geometry entered from desktop measurement, particularly flare lengths and entry widths at roundabouts, which have a large effect on modelled capacity.
- Surveys taken in school holidays or during nearby roadworks.
- Pedestrian and cycle demand omitted at signal junctions, which understates the green time those movements need.
- Only the network peak hour assessed, missing a site-specific peak such as a school or nursery arrival period.
- Trip rates from a database extract with no filtering rationale given.
- Mitigation modelled without checking that it can be built within the highway boundary.
- No comparison of modelled and observed queues anywhere in the report.
That last one is the fastest way to lose a highway officer's confidence.
Presenting outputs alongside the assumptions
Put the assumptions where a reviewer can find them. An assumptions log at the front of the Transport Assessment, listing survey dates, growth source and version, committed development list, trip rate derivation, assessment years, peak periods and software versions, lets a consultee reproduce the work.
Then present the results consistently. Same table format for each scenario, same units, base and future cases adjacent, difference columns included. Put the full model outputs in an appendix and share the model files where the authority asks for them.
For smaller schemes, all of this scales down. A Transport Statement with a short capacity check and a clear description of the access is often the right level, and it should say why fuller modelling is not proportionate rather than leaving the question open.
Where mitigation is proposed, be clear about how it will be delivered and secured, whether by condition, by planning obligation or through a highway agreement. Demand management measures belong in a Travel Plan with monitoring and targets attached. And where construction traffic is itself a concern at the junction, a Construction Traffic Management Plan covering routing and delivery timing answers a question that would otherwise arrive late.
An illustrative example: a priority junction with a modelled ratio of flow to capacity rising from 0.71 to 0.76 on the minor arm in the evening peak, with the mean maximum queue moving from four vehicles to five, and observed queues in the base surveys of four to six vehicles. That is a defensible position, and the observed comparison is what makes it defensible.
This is general guidance, not legal advice. Check the position with your local planning authority or a qualified professional before relying on it. 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.
Frequently asked questions
What is a base year model for?
To show that the model reproduces conditions as they actually are. It is built from surveyed turning counts and junction geometry, then checked against observed queues and delays on the same days. If the model predicts free flow where a queue forms every morning, the geometry or the flows are wrong, and nothing built on top of it can be relied on.
What capacity threshold is used?
At priority junctions and roundabouts, a ratio of flow to capacity of around 0.85 is conventionally treated as the practical limit, beyond which queues grow disproportionately. At signal controlled junctions the equivalent measure is degree of saturation, with about 90 per cent treated as practical capacity. Both are conventions rather than statutory standards, and local authorities may apply their own.
When should traffic surveys be carried out?
In a neutral month, avoiding school holidays, August, the period around Christmas, and weeks with unusual local events or roadworks. Surveys taken in an atypical period get discounted, and repeating them costs both money and programme. Where existing counts are more than about three years old, expect to be asked for new ones.
What sensitivity tests are usually requested?
Higher growth than the central case, an alternative trip distribution, a scenario without any proposed mitigation, and sometimes a test with a nearby committed development built out sooner. The purpose is to show whether the conclusion holds up under reasonable variation, so present the tests even where the results are less comfortable.
Does a small development need junction modelling at all?
Usually not. Modelling is proportionate for schemes adding meaningful flows to a junction that is already close to capacity. For most minor applications a description of the access arrangements, visibility splays and parking is enough, and commissioning a model where none is needed adds cost without improving the case.
Sources and further reading
This article is general guidance about the planning system in England and, where stated, the rest of the UK. It is not legal or professional advice. Check your local planning authority’s validation requirements and take professional advice for your specific site.