Read a rural English weather station the way an observer does: dry and wet bulb thermometers in a louvred screen, a cup anemometer and vane, a rain gauge and a barometer. See what the instruments say, what the real weather was, and why they differ.
Click the anemometer, the Stevenson screen or the rain gauge in the 3D view.
Educational model with a synthetic weather generator. Instrument constants marked [assumption] are illustrative, not manufacturer data. Not for official observations. The station is drawn larger than life and the mast is 3 m (a standard anemometer mast is 10 m).
The science
Vapour pressure. Magnus: es(T) = 6.112 exp(17.62 T / (243.12 + T)) hPa (WMO-form coefficients, over water). Dew point inverts it; RH = 100 e / es(T).
Wet and dry bulb (psychrometer). e = es(Tw) − A p (T − Tw) with A = 6.6×10−4(1 + 0.00115 Tw) K−1 for an aspirated instrument. The page solves it by bisection for Tw from the true humidity, and the observer inverts it again. RH uncertainty is propagated from ±0.15 K reading uncertainty (0.1 K graduation plus parallax) in each bulb: near saturation a small wet-bulb error moves RH a lot.
Thermometer lag. A bulb of heat capacity C and surface conductance h follows dT/dt = (Tair − T)/τ, τ = C/(hA). [assumption] τ = 90 s / (1 + 0.6√u) in the screen (ventilation by wind u, m/s), 40 s for the bare bulb.
Radiation error and why the Stevenson screen exists. A bare bulb absorbs sunlight: ΔT = α S / (π h), h = 5 + 8√u W m−2 K−1, α = 0.3 [assumption], so in strong sun and light wind it reads several kelvin too warm. The white double-louvre screen blocks direct sun but lets air through; its remaining error is modelled as a small fraction of that [assumption: about 0.3 K at 1000 W/m² and 1 m/s with the door shut]. Opening the door lets sunlight reach the bulbs again (6× larger error).
Cup anemometer. The rotor follows the wind with a distance constant L: dUc/dt = (U − Uc) U / L, i.e. a time lag τ = L/U. Real cup wheels accelerate faster than they decelerate (Lup < Ldown; here 2.5 m and 5 m [assumption]), so in gusty wind the mean reading is too high ("over-speeding"). The page demonstrates it numerically; spin speed ω ≈ 3.3 Uc rad/s. Beaufort force is from the standard thresholds.
Rain gauge. A funnel of area Af (127 mm diameter) drains to a measuring glass of area Af/10, so the water column is 10× the rainfall depth: 2 mm of rain is 25.3 mL, a 20 mm column. Wind blows drops over the rim: catch ratio 1/(1 + 0.04 u) at gauge height (u about 0.55 of mast wind) [assumption, illustrative]; real gauges sit in pits or shields for this reason. The water in the glass is drawn 8× taller than its true column. Daily totals run from 09:00.
Pressure. Station pressure is reduced to sea level with the hypsometric formula p0 = p (1 − 0.0065 h / (T + 0.0065 h + 273.15))−5.257, h = 120 m. The 3 h tendency gives the classic forecast: falling fast means strong wind or rain, rising fast means clearing and gusty.
Honest simplifications. A synthetic, smooth day (deterministic sum-of-sines gusts), no ground heat flux, no dew or frost, no hail or snow, no lightning, no instrument drift, no site-exposure effects other than wind; the weather scenarios are teaching sketches, not forecasts. Sources by type: WMO CIMO Guide (instruments and observing methods), standard meteorology and psychrometry texts, Magnus-type formulae.
Models built with Blender (screen, instruments, wall, cottage, trees, terrain) and rendered with Three.js.