HX HydroclimateX

LISFLOOD Overview · Community flood risk assessment

Flood risk, modelled to the community.

LISFLOOD-FP is a physically based, two-dimensional surface-flow model. The online application lets you draw a study area anywhere over the world. Using the Qixia District, Nanjing as an example, run a design rainfall event, and inspect the resulting flood hazard, exposure, depth and velocity layers in your browser.

Launch LISFLOOD Explore methodology ↓

Methodology

Rainfall in, hazard and risk out.

A design storm is built for the selected return period, the two-dimensional surface-flow model propagates it over the terrain, and the map layers are derived from the model's maximum depth and maximum velocity simulations.

Design rainfall

  • Chicago design-storm hyetographs for Qixia District, Nanjing
  • Return periods: 5, 10, 20, 50 and 100 years
  • 3-hour storm, 1-minute resolution

Hydraulic model

  • LISFLOOD-FP 8.0.3, acceleration (ACC) solver
  • 30 m grid cells, EPSG:32650 (UTM 50N)
  • Surface flow only — no sewer or channel routing
  • Each run: 12 h of simulation time per scenario

Map layers

What each layer means — and how it is computed.

The interactive map shows one layer at a time. Layers are rendered from the model's maximum-over-event rasters; continuous layers are stretched to the 2nd–98th percentile of their values.

Flood risk

Hazard × exposure.

Flood risk combines the physical flood hazard with the local population exposure in a single score. Each grid cell is assigned a hazard class (from the Flood hazard layer) and an exposure class, then looked up in a risk matrix:

  • Exposure class — WorldPop population counts within the study area are split at their 25th, 50th and 75th percentiles into four classes.
  • Hazard class — the Flood hazard rating, classified at 0.75, 1.25 and 2.5 (see below).
  • Risk matrix [[1,1,1,2],[1,2,2,3],[2,2,3,4],[2,3,4,4]] — indexed by hazard class then exposure class, producing Low, Moderate, High or Extreme.
  • Cells with maximum depth below 0.10 m are treated as no flood.
Flood hazard

The physical danger of the water.

The model computes the DEFRA flood-hazard rating at every time step and records its maximum over the event:

HR = d · (v + 1.5)

where d is water depth (m) and v is flow velocity (m/s). Deep and fast water scores higher than deep still water. The rating is classified:

  • Low — HR < 0.75
  • Moderate — 0.75 ≤ HR < 1.25
  • High — 1.25 ≤ HR < 2.5
  • Extreme — HR ≥ 2.5

Flood hazard describes the water. It is one input to Flood risk and is not a duplicate of it — risk additionally weighs how many people are exposed.

Maximum depth

How deep the water gets.

The maximum water depth (m) reached anywhere during the simulated storm, taken from the model's maximum-depth raster and shown only where depth reaches at least 0.10 m.

Maximum velocity

How fast the water moves.

The maximum flow-velocity magnitude (m/s) reached anywhere during the simulated storm. Fast-moving floodwater is far more dangerous than still water of the same depth — this is the velocity term inside the hazard rating.

Population

Who is exposed.

2025 population count per grid cell from WorldPop R2025A (constrained estimate for China), resampled to the model grid with conservative cell-summing so totals are preserved. This is the exposure input to Flood risk.

DEM

The terrain the water runs over.

Ground elevation in metres above sea level from Copernicus DEM GLO-30 (30 m). Elevation drives where water accumulates and how fast it flows.

Hazard vs risk — the short version.

Flood hazard asks "how dangerous is the water here?" It depends only on depth and velocity. Flood risk asks "how much danger is there to people here?" It combines the hazard with the population exposure — a remote, hazardous flood cell can still be low risk because nobody lives there. The two are related, not duplicative.

Data sources

Open data, reproducible pipeline.

Population

Basemap

  • OpenStreetMap contributors

Limitations

What the model does not capture.

This is a research demonstration. Sewer networks and engineered drainage are not represented — there is no underground drainage or artificial drainage capacity in the model, so results are not suitable for emergency response or engineering design. The 30 m grid resolves the floodplain but not street-scale or building-scale detail, and the design storms are synthetic rather than observed events. Always validate against local engineering guidance before use.

Run a flood scenario in your browser.

Choose a return period, draw a study area anywhere over Qixia District, run the simulation, and inspect Flood risk, Flood hazard, Maximum depth, Maximum velocity, Population and DEM layers.

Launch LISFLOOD →