MY SIMULATOR
- ELA - m melt today - mm FPS 0
Oblique Top Terminus
White = snow cover on that elevation band today, blue-grey = bare ice, brown tint = dust. Red-white stake = equilibrium line altitude (where yearly snow gain equals melt). Blue beads = meltwater. Vertical scale is stretched (slider).
WebGL is not available here — the numbers, charts and numerical checks on the right still work.
Glacier Albedo Lab: Surface Energy Balance, Dust Feedback, Degree Days and the Equilibrium Line
Snow reflects 80-90% of sunlight, bare glacier ice only 30-40%, dusty ice even less. As the melting season strips away the snow, the surface darkens, absorbs more energy and melts faster: the albedo feedback. This lab runs a daily surface energy balance on 17 elevation bands of a valley glacier for a whole year, finds the equilibrium line altitude (ELA) and compares it with a simple degree-day model.
Educational point model with typical parameters. Fixed glacier geometry (no ice flow or thinning), daily means, clear/cloud mixing by a single cloud fraction, no cold-content or refreezing, turbulent fluxes with constant bulk coefficients, precipitation rising 4% per 100 m. Values resemble a mid-latitude Alpine glacier but are not calibrated to any real one.
Pause Reset Quality: High
Scenario
Time and place
Climate
Surface
Live readout (band at the chosen elevation)
Daily energy fluxes Surface albedo through the year Yearly mass balance against elevation Experiment 1: warming or darkening?
How far does the equilibrium line climb for +1 K, +2 K, dust and a cloudy year, with everything else as set above?
Run scenarios
Experiment 2: how much does ice albedo matter?
Run albedo sweep
Experiment 3: calibrate the degree-day factor
The factor that makes a degree-day model reproduce the energy-balance melt (all bands, whole year), and what it does to the ELA.
Fit factor
Independent numerical checks running...
Also run in Node: node src/check.js.
Controls Drag / one finger: orbit; wheel / pinch: zoom. Camera buttons: Oblique, Top, Terminus. The date slider scrubs through the hydrological year (1 October to 30 September); Play runs it. The elevation slider chooses the band shown in the charts and readout. The science
Energy balance. At a melting surface (0 °C) the energy available for melt is Q = (1 − α) S + LWin − LWout + H + LE (W/m²). Melt is M = Q · 86400 / Lf kg/m² per day with Lf = 334 kJ/kg, so 100 W/m² melts about 26 mm w.e. per day.
Shortwave. S is the daily-mean top-of-atmosphere insolation (closed form with declination and sunset hour angle, checked against numerical integration) times a clear-sky transmissivity (0.75 + 2×10-5 z) and a cloud factor 1 − 0.75 N3.4 . Longwave : incoming εa σT4 with εclear = 9.2×10-6 T2 blended to 0.98 under cloud; outgoing from the surface at 0 °C (315.7 W/m² for ε = 1). Turbulent : sensible H = ρ cp Ch u (Ta − Ts ) and latent LE = ρ L Ce u (qa − qs ) with C = 0.002 and 0.0015 (typical).
Albedo and feedback. Fresh snow αsnow decays toward 0.62 with a 6-day e-folding time after the last snowfall; below 30 mm of snow the ice shows through; once snow is gone the surface is ice (αice about 0.4). Dust lowers all three by a fraction of the dust amount (a simple linear scaling, not a measured relation). With "albedo feedback" off the albedo is held at 0.75.
Degree days. The classic alternative is M = DDF · max(T, 0) with DDF about 3-5 mm K-1 d-1 for snow and 6-9 for ice. It uses air temperature alone, so it cannot see dust or cloud: that is what Experiments 1 and 3 show.
ELA. For each 100 m band the yearly balance is snowfall (when T < 1 °C) minus melt. The ELA is where it crosses zero (linear interpolation); above it the glacier gains mass, below it loses mass.
Honest simplifications. No ice dynamics, no debris cover, no meltwater refreezing, no aspect or shading, one cloud fraction for the whole year, no snow densification, constant bulk coefficients. The dust relation and all coefficients are typical or illustrative values.
3D valley and glacier modelled in Blender, baked PBR textures; rendered with Three.js