One and two dimensional modelling: which is needed

Choose a one dimensional hydraulic model where flow stays in a defined channel and floodplain flow paths are simple, and a two dimensional model where water spreads across a floodplain, interacts with buildings and embankments, or takes routes a cross-section cannot describe. For a site specific flood risk assessment the practical question is narrower than the modelling theory suggests: can you answer the regulator's questions by extracting levels from an existing model, or does the site need a new build. 1d and 2d hydraulic modelling for planning is a proportionality decision, and it should be agreed before any fee is committed.

A great many site assessments need no new model at all.

What each approach represents

A one dimensional model represents the watercourse as a series of cross-sections with flow calculated along the channel line. It solves for water level and discharge at each section, and it assumes flow is along the direction of the channel, with floodplain storage represented either as part of the section or as a storage area linked to it.

A two dimensional model represents the ground as a grid or mesh and solves flow in both horizontal directions across it. Water finds its own route, so flow paths, velocities and the effect of walls, embankments and buildings emerge from the topography rather than being assumed.

Linked one and two dimensional models combine the two, keeping the channel in one dimension for efficiency and the floodplain in two dimensions where spreading matters. For most urban river sites that is the standard arrangement, and the term "2D model" in correspondence usually means a linked model.

Where a one dimensional model is sufficient

Four situations where one dimension answers the question properly.

A watercourse in a confined valley where the floodplain is narrow and flow is clearly along the channel. A culvert or structure assessment where the question is afflux at a specific point. A channel capacity check, for instance whether a proposed outfall or a modified channel section conveys the design flow. And a sensitivity test on an existing one dimensional model, for example to confirm the effect of a climate change allowance on peak level at the site.

In those cases a one dimensional model is not a compromise. It is the right tool, it is quicker to build and audit, and its results are easier for a reviewer to check. Over-specifying is a real cost: a two dimensional build on a confined upland reach can take weeks and deliver the same level at the site.

When floodplain spreading requires two dimensions

Two dimensions become necessary when the assumption of channel-aligned flow breaks down.

The indicators are a wide flat floodplain where water spreads laterally, a defended site where breach flow behaves as a spreading inundation rather than channel flow, an urban floodplain where streets, walls and buildings route water in ways a cross-section cannot capture, a site where velocity and hazard mapping are needed rather than level alone, and any case where the proposal itself changes the floodplain, such as a platform or compensatory storage scheme whose effect on flow routes has to be shown.

Hazard mapping is the clearest trigger. Where the assessment needs depth and velocity across the site to demonstrate safe access and escape, a one dimensional model cannot produce it, because velocity across a floodplain is not a one dimensional quantity.

Compensatory storage is the second. A level-for-level, volume-for-volume argument can be made from a one dimensional model in simple cases, and where the storage is remote from the loss or the floodplain is complex, the effect on third parties has to be shown in two dimensions.

Model build, calibration and sensitivity

Whichever approach is used, the same evidence makes the model credible, and its absence is what generates review queries.

Calibration against a recorded flood event is the strongest evidence a model can carry, and many sites have no suitable record, in which case say so and rely more heavily on sensitivity testing. Blockage scenarios matter at structures: a culvert partially blocked is a standard sensitivity test and reviewers ask for it.

Mesh resolution is where two dimensional models are commonly under-specified. A grid too coarse to resolve a road, a wall or a flood route gives a smooth and reassuring answer that is wrong at the scale that matters for a site layout.

Proportionality on small sites

For a small site the first question is whether modelling is needed at all, and often the answer is no.

The sequence to work through: does the Environment Agency product for the reach already give levels at the site, has a strategic flood risk assessment or a catchment model been produced for the area whose outputs can be requested, can the levels be interpolated defensibly between modelled sections, and does the proposal change anything the existing model assumed.

Where an existing model exists, requesting its outputs and interpolating with a short technical note is a fraction of the cost of a new build and is frequently accepted. Where the proposal alters the floodplain, or the existing model is too coarse at the site, a new or refined model is justified. Those two positions should be set out in writing and agreed rather than assumed in either direction.

The cost asymmetry is large enough to matter on a small scheme. A technical note extracting levels from an existing model is a modest piece of work. A new linked model with survey, hydrology, calibration and sensitivity testing is a different order of fee, and on a two-unit infill it can exceed the value the assessment adds.

Agreeing the approach with the regulator first

Write a modelling scoping note and send it to the Environment Agency, and to the lead local flood authority where surface water is the issue, before commissioning.

Put in it the site and the proposal, the sources of flood risk identified, what existing modelling is known to exist for the reach, the proposed approach with the reason, the return periods and climate change allowance proposed, the outputs to be produced, and the calibration and sensitivity tests intended. Ask for confirmation or comments.

What that achieves is twofold. Where the regulator holds a model you did not know about, you learn before spending anything. Where they expect two dimensions and you proposed one, you learn before the report is written rather than after it is submitted. Pre-application charges for this engagement are modest relative to the modelling fee.

Illustrative example: a 1.2 hectare infill site beside a main river in Flood Zone 2. The initial expectation is a new linked model. The scoping response confirms a recent catchment model covers the reach with sections 80m apart, and that extracted levels plus a short interpolation note will be accepted because the proposal includes no floodplain encroachment. The modelling fee becomes a technical note, and the Flood Risk Assessment is written in three weeks instead of three months.

Reporting conventions that avoid queries

Reviewers ask the same questions, so answer them in a fixed order and the review is short.

Lead with a model build note as an appendix: source data, schematisation, parameters, boundaries, calibration and sensitivity, with the software and version named. Present results as both tables and mapping, with levels given to a stated datum and the datum named. Give the modelled level at defined points on the site rather than a single site-wide figure. Show the proposed finished floor levels and site access levels against the modelled levels with the freeboard stated. And include the sensitivity results rather than describing them as acceptable.

Keep the hydraulic work consistent with the drainage design, because the two share levels and discharge rates, and a mismatch between the Drainage Strategy and the flood model is an easy query to raise. Where groundwater is part of the picture, the evidence approach in evidencing groundwater flooding in a flood risk assessment applies, and surface water mapping should not be overstated, as set out in reading the surface water flood risk map without overstating it.

Hydraulic modelling is work councils and the Environment Agency expect a suitably qualified modeller to carry out, and professional review is strongly advised. PlanScope provides AI-assisted draft reports for planning applications. Reports should be reviewed before submission. We do not guarantee planning approval.