Water cycle: evaporation, cloud, rain and runoff

Sea, wind, mountain and river driven by one water-budget model, in a Blender-modelled diorama.

Switch to 2D

Presets

Playback

Experiment

Atmosphere and ocean

Time

Camera

Layers

Quality

Model check

The science

What is modelled

A vertical slice of the water cycle: sunlight evaporates sea water, wind carries the vapour onshore and up a mountain, the vapour condenses into cloud if the mountain is higher than the lifting condensation level, cloud water falls as rain, and the runoff returns to the sea by a river. The 3D scene is driven by exactly the same water-budget model as the 2D page (same equations, constants and fixed time step), so numbers agree with it when the time scale is 1x.

Equations

  • Saturation vapour pressure (Magnus): es(T) = 6.112 exp(17.62 T / (243.12 + T)) hPa, about +6.5 % per kelvin (Clausius-Clapeyron).
  • Evaporation: latent flux LE = min(Rn, f(u) (es(Tsea) − ea)), with f = 0.6 + 0.25 u W/m²/hPa, air 1 K cooler than the sea and ea = RH es(Tsea − 1); then E = LE / (Lv ρw), Lv = 2.5×106 J/kg, so 1 mm/day needs about 29 W/m².
  • Lifting condensation level: LCL ≈ 125 m/K × (Tair − Tdew) (dry adiabatic cooling 9.8 K/km, dew point falling about 1.8 K/km).
  • Cloud forms when mountain height > LCL; the condensed fraction grows with the excess height and the wind; rain starts above 0.5 u of cloud water; ground water drains to the sea at 25 % per model-second.

What to try

  • “Cold coast, no cloud”: the condensation level is above the mountain, vapour leaves the slice, no rain.
  • “Tropical monsoon coast”: warm sea, humid air, strong evaporation and heavy rain.
  • “High mountain, snow cap”: the snow cap appears above 2200 m.
  • Drag the sea temperature up by 1 K and watch E and es in the first graph.

Honest simplifications

  • One 2D slice extruded in depth; no latent-heat feedback on the air, no ice or snow storage, fixed land storage. Vapour that leaves the slice without forming cloud goes to an “exported” reservoir so the budget stays closed.
  • Water amounts are in model units u (1 u = 1 mm over the sea slice) and the model runs 3x faster than real time; the absolute rain rates are a classroom scale, not a climate estimate.
  • The vertical axis is exaggerated: 3000 m is drawn about 4.6 units high against a 12 unit wide mountain. Vapour dots and rain streaks are decoration driven by the model state (spawn rates follow E and the cloud water), not simulated droplets. Trees, houses and the boat are scenery.
  • The ridge profile is the 2D page’s profile, tapered toward the front and back faces so the diorama reads as a hill; the cross-section face is the true profile.

Checks run in the page

  • Textbook es values, a hand calculation of 100 W/m² in mm/day, the Clausius-Clapeyron ratio, mass conservation of the full budget, a dew-point round trip by bisection, and an experiment that sweeps the mountain height to find where cloud and rain start (this one tests the implementation of the model’s own LCL rule, not nature).
Loading model...
Loading Blender model...
Evaporation vs sea temperature
Water in each reservoir
Condensation level vs sea temperature
Made with 🟠 Blender · Three.js