ScaleLAB

Analogue experiments scaled to planetary flow laws

1 · Lab setup

More settings
My lab materials

Start here: the lab is fixed, nature is fitted to it. The anchor layer holds your base fluid; its viscosity sets the time scale and it is never changed or moved by the app. Material menus show only your lab materials. A closed lock on a value means nothing in the app will change it.

2 · Box and bodies

Coordinates are in cm from the centre of the floor, z upward. Layers are stacked from the floor with the model thicknesses from the scaling. A body placed by depth below the surface keeps its place in the upper layers when a deeper layer changes thickness; one placed by height above the floor stays fixed in the box. Bodies are saved with the setup.

3 · Layers, top to bottom

Each layer's flow law is evaluated at the reference strain rate times that layer's own factor; leave the factor at 1 unless the whole layer deforms at a different rate. Thickness takes the literature range: min / reference / max. The reference value has the highlighted border. Each layer also has a model thickness in cm under "In the lab": it is the reference thickness seen through the length scale, and you can type in either one.

4 · Nature (prototype)

Model geometry

Material recipe

Volumes use the thicknesses selected above and the container footprint, minus the parts of bodies inside each layer. Add your own margin for losses.

Model design

Density contrasts

Nature fixes only the viscosity ratios between layers. Autofit fits only the unlocked layers and only from your lab materials: the stock fluid within 10 %, a blend of two stock fluids by the log-viscosity rule (weight fractions, approximate: measure the blend), or an own mix with the viscosity to reach. The anchor layer, locked values, densities and thicknesses are never changed, and the anchor never moves. A layer that would fall outside your working range is held at the nearest end of the range and flagged with its mismatch.

Buoyancy-driven flow scales with density differences. "Needed" is the natural contrast times the density scale behind the stress ratio. Set a body's natural density in "Box and bodies" to include it here.

Fit nature to the lab

Shows where every ductile layer can be built with your viscosities while the anchor stays as it is. Click a cell to use that gradient and thickness.

Brittle layers and deformation

Model strength is Mohr–Coulomb from the friction angle, cohesion and pore space you enter per brittle layer; nature is cohesion + β ρ g z (1 − λ) from the Nature panel. Similar faulting needs the same friction angle and a model strength that, taken to nature with the stress ratio, equals the natural one.

Bodies and behaviour over time

Lab

Body height over time

Lab

Material flow curves

Lab

Body temperature over time

Lab

Body viscosity over time

First-order estimates. Body motion is Stokes flow of a viscous drop of equal volume in an unbounded fluid, using your material's viscosity where one is chosen and the required viscosity otherwise; walls, the free surface, neighbouring layers and shape change are ignored. A body that starts warmer or colder than the lab relaxes to lab temperature by conduction; meanwhile its density follows its thermal expansivity and its viscosity its activation energy, so a warm plume rises fast at first and stalls as it cools. Heating of the surrounding layer is ignored. For a body that comes to rest, the time shown is when it has covered 90 % of the way. In the flow-curve panel the vertical dashed line is the model strain rate; solid curves are your materials and dashed curves of the same colour are the natural layers scaled to the lab. Where a pair crosses, the layer is matched; where the slopes differ, the match holds only near that strain rate. The shear-thinning runaway time uses the form of Houseman & Molnar (1997) with your C and Z₀/h, and is unchecked against their paper.

Prototype viscosity

Envelope = spread of the layer value over the 8 end-member cases (all thicknesses min or max × gradient min or max × strain rate min or max).

Profiles

layer value used for scalingyour materialbrittle
Nature

Viscosity with depth

Lab

Required viscosity with model depth

Nature

Strength with depth

Lab

Scaled strength with model depth

Nature

Temperature with depth

Flow laws

Viscosity against temperature

Nature and lab panels share the same vertical geometry: lab depth is natural depth times the length ratio. In strength panels the solid line is viscous stress and the dashed line brittle strength; the weaker one holds. In the last panel the thick part of each curve is the temperature range that layer spans. Every panel redraws as you type.

Export to numerical codes

The LaMEM and ASPECT files are templates that have not been run. MILET and MILEVA have no dedicated exporter yet: the neutral JSON carries every number they need, in SI units.

Flow-law library

Imported from the creep-law database of GeoParams.jl (version and commit are recorded in data/flowlaws.json), plus two laws entered from memory and marked "verify against source". A in MPa⁻ⁿ µmᵖ s⁻¹, E in kJ/mol, V in cm³/mol. "Experiment" is the geometry each law was measured in and sets its conversion to stress and strain-rate invariants. Hover a row for the database's note on that law, including known discrepancies with the source paper. On the published site this table is read from data/flowlaws.json, which the CMS edits.

Lab materials library

"Datasheet" and "measured" entries carry their source; "indicative" ones are starting values from memory. Measure your own batches either way. Viscosity follows η = η₀ (ε̇/ε̇₀)^(1/n − 1) exp[Eₐ/R (1/T − 1/T₀)], so n above 1 is shear-thinning and n below 1 shear-thickening. Picking a material in a layer copies these values there, where they can be changed.

Validation

Every test

Runs each time the page opens. "Published value" reproduces flow-law numbers printed in the source. "Published experiment" runs published analogue experiments through the app: the scaling factors of two drip studies, the descent of their seeded drip (pass within 30 %), and the rise speed of heater-fed plumes taken as a sphere of the heater's radius (pass within 25 %). "Lab window" asks whether a law predicts a believable laboratory stress at laboratory temperature and strain rate: it catches unit errors, not a factor-of-few error in A. The remaining groups test the code against closed-form results.

Method

Flow law. ε̇ = A σⁿ d⁻ᵖ Wʳ exp(−(E + PV)/RT) for each mechanism. Dislocation creep, diffusion creep and low-temperature plasticity act in parallel: at each depth the stress is solved so that their strain rates sum to the imposed strain rate, and η = σ_II / 2ε̇_II.

Low-temperature plasticity. ε̇ = A σⁿ exp[−(E/RT)(1 − (σ/σ_P)ᵖ)^q], with the barrier set to zero above the Peierls stress σ_P. The exponential form leaves a finite rate at zero stress, so each law is used only above a minimum stress (200 MPa unless you change it) and is switched off below it. Where the mechanism would be faster than the imposed strain rate at that minimum, the stress sits at the minimum.

Lab-data conversion. Each law carries the geometry it was measured in. Axial-compression laws are converted to second invariants with ε̇_II = (3^((n+1)/2) / 2) A σ_IIⁿ; simple-shear laws with ε̇_II = 2^(n−1) A σ_IIⁿ.

Geotherm. T(z) = min(T_s + G z, T_p + G_a z). Pressure is lithostatic from the layer densities.

Brittle strength. Δσ = C₀ + β ρ g z (1 − λ), compared with the viscous differential stress taken as 2σ_II. Where brittle is weaker, the layer is counted as brittle and left out of the viscosity average.

Scaling. With starred ratios model/nature: σ* = ρ* g* L* (or Δρ* g* L* when the stress scale is taken from a density contrast), η* = σ* t*, v* = L*/t*, ε̇* = 1/t*. The time ratio follows from the anchor: t* = η*/σ* with η* the anchor material's viscosity over the anchor layer's natural viscosity, so one lab hour is 3600/t* seconds of nature and the model strain rate is the natural rate divided by t*. Because natural viscosity depends on strain rate, the time scale changes when the reference strain rate does. Required model viscosity is η_nature · η*; a model cohesion scales with σ*.

Limits. A Newtonian analogue matches a power-law layer only at the reference strain rate. Grain-boundary sliding, melt, strain hardening and grain-size evolution are not included. Published low-temperature plasticity laws disagree widely, so compare more than one.

Paste a setup

Paste the text you got from "Copy setup".

Add a flow law

A in MPa⁻ⁿ µmᵖ s⁻¹, E in kJ/mol, V in cm³/mol. Saved with this setup in your browser. To share it with everyone, use "Copy library JSON" or the CMS.