Atmospheric River 3D: Coast Range, Orographic Rain, Snow Line and Flood Risk

Setting up the lab bench...

Presets

Live readouts

Graphs

Along the plume centre-line: IVT, terrain and rain

Event time series: IVT, peak rain and the three rivers

Rain shadow: fraction of IVT that crosses a ridge

Clausius-Clapeyron: IVT against warming

What is happening now

Starting.

Numerical checks

running...

How to use

Press a preset and watch the map: the cyan band is the plume, the heat colours are rain. Then (1) raise the ridge: the lee side dries out (graph 3 shows the fraction that gets through); (2) raise the freezing level and watch the snow layer vanish while the river bars climb; (3) set drift to 0 so the plume stalls over one catchment; (4) add +3 K of warming (graph 4). Hover over the map for local values. Drag to orbit, wheel to zoom. Space pauses, R resets.

The 3D page uses a Blender-built coast range with ocean shelf, forested slopes, snow caps, valley, river, town and dam; clouds, rain and snow are drawn from the same model.

The science

An atmospheric river (AR) is a long, narrow corridor of strong water-vapour transport in the lower atmosphere, typically 400 km wide and 2000 km long, carrying as much water as the Amazon's mouth. Its strength is measured by the vertically integrated vapour transport:

IVT = (1/g) ∫ q v dp (kg m⁻¹ s⁻¹, from the surface to 300 hPa)

with specific humidity q, wind speed v and pressure p. Values above about 250 kg/m/s count as an AR; above 750 they are strong and above 1250 exceptional (the real AR scale also uses duration, here only peak IVT is used). Atmospheric rivers deliver a large share of the winter precipitation of western coasts and are behind most floods there.

Orographic rain budget. When the plume meets a ridge the air is forced up, cools and condenses its vapour. Following a streamline with unit direction e, the vapour flux is depleted in proportion to the rate of ascent:

d(IVT)/ds = −ε IVT max(0, e·∇h) / H_v ⇒ IVT_down = IVT_up exp(−ε Δh_up / H_v), P = − d(IVT)/ds

H_v ≈ 2.5 km is the vapour scale height and ε the condensation efficiency. This gives the classic pattern: rain concentrated on the windward slope, a rain shadow behind the crest (flux already lost, no ascent) and, for a straight crossing of a ridge of height h_max, a transmitted fraction exp(−ε h_max/H_v) (graph 3). The solution is exact cell by cell, so the water budget closes: rain removed + flux leaving = flux entering (checked below).

Warming. Saturation vapour pressure grows about 6.8% per kelvin (Clausius-Clapeyron, Magnus formula e_s = 6.1094 exp(17.625 T/(T+243.04)) hPa), so a warmer source region puts more vapour into the plume: IVT ∝ exp(0.068 ΔT), and the freezing level rises about 150 m per kelvin, turning snow into rain on the slopes (graph 4). Rain that falls on snow and a high snow line is what turns an AR into a flood.

River response. Three catchments on the windward slope (south, middle, north) respond to the mean rain on them through a linear reservoir, dQ/dt = (R − Q)/τ with τ = 6 h, so a drifting plume gives each river a lagged peak.

Honest simplifications. Streamlines are straight, the wind is prescribed (no dynamics), there is no moisture recycling, evaporation, melting, infiltration or real river routing, the condensation efficiency is a single number and the plume profile is Gaussian. It is a teaching model of the moisture budget, not a forecast. The mountains are an artificial coast range, not a specific place.