Surveying the catchment…
🌧 Storm and structures
The science

Model. A 56 x 56 grid of 2 m cells (a synthetic 112 m catchment). Each cell keeps a water depth h; mass balance dh/dt = r − f − ∇·q with rain r and infiltration f = I · φ (capacity I times a permeability φ: 0.90 grass, 0.12 road, 0.06 roof/paving).

Overland flow. The flux between neighbouring cells is a diffusive-wave (Manning) law, q = (Δx / neff) h*5/3 √(|ΔS| / Δx), where ΔS is the water-surface (stage) difference, h* the depth above the higher bed and neff = n (1 + 0.38 (2 − φa − φb)) makes built-up cells rougher. Outgoing flux of a cell is scaled so depth never goes negative; the scaling is identical for all of its edges, so volume is conserved exactly.

Structures. The sluice is a broad-crested weir, Q = 0.60 · opening · W √(2g) (h − 0.085)3/2; closed means exactly zero flow. The pump removes min(available water, Qp dt) from its footprint disc.

Honest simplifications. Diffusive wave (inertia and backwater waves are dropped), no sewers or culverts, constant-rate infiltration (no Horton or Green-Ampt decay), uniform rain, a made-up terrain and a time scale of up to 240x. Parameters are illustrative; this is not a flood forecast and must not be used for safety decisions.

Independent checks (run on load with fresh model instances):

running…

Catchment readout

Simulated time-
Surface storage-
Peak depth-
Flooded area (> 3 cm)-
Rain / infiltrated-
Sluice / pump out-
Water-balance error-
rainoutflowstorage
-
Drag to orbit, pinch or scroll to zoom. The model is a classroom-scale illustration, not a forecast.
Made with 🟠 Blender · Three.js