About this lab
Five topics, one thread: flowing water strips material from one place, carries it away and puts it down in another. On one river the upper reach cuts while the lower reach fills, and in a bend one bank retreats as the other advances.
Erosion strips material from bed and banks, transport moves material already in the flow, and deposition is material settling out or staying put. All three can occur on the same reach; the difference is the local net change: a lower bed is erosion, a higher bed is deposition.
The flume topic is at true size: 80 cm × 30 cm with 5 cm of sand. The four outdoor topics use original teaching terrain, not any real place, and the cutaway walls stretch heights 3 to 5 times, as each topic’s description states. Time is model time: thousands of years, years or seconds. It is not the erosion rate of a real river, and it cannot be inferred from playback speed.
Five findings
Five findings
- Erosion, transport and deposition: the larger the stream power Ω = ρgQS (in W/m), the stronger the transport potential, but sand moves only above the bed material’s threshold. A 5° slope gives a deeper maximum cut than 2°; at the same flow, coarse sand has a higher threshold and barely moves at low flow: faster water does not mean sediment is carried off at once. The sand removed equals the outlet fan plus the ridges deposited along the groove.
- V-shaped valleys and waterfalls: as the bed cuts down, the valley sides adjust under gravity and retreat, and the cross-section changes from a shallow trough to a V. Record the fall of the valley floor and the retreat of the valley sides separately. Where hard rock caps soft rock, the soft rock is undercut, the overhanging hard rock collapses, the waterfall retreats upstream and leaves a gorge downstream.
- Alluvial fans and deltas: where a river leaves a mountain front, the slope drops abruptly and the flow spreads out, so transport capacity falls and sediment is laid down as a fan whose radius grows with the cube root of time, coarse at the apex and fine at the edge; the active channel on a fan switches. Only a river entering still water builds a delta; when base level rises, the shoreline retreats and so does the place where new sediment lands. Names depend on position, not just on shape.
- Meanders and oxbow lakes: secondary flow in a bend erodes the outer bank and deposits on the inner bank; with constant channel width the outer bank retreat equals the inner bank advance. As the bend grows its neck narrows until a flood cuts through, the two ends of the old channel are plugged with sediment and an oxbow lake remains, which then slowly fills. This has nothing to do with the Coriolis force.
- Floodplains and terraces: a floodplain is a surface that today’s floods can cover, and a terrace is a higher, older surface. Each fall of base level makes the river cut down again and leaves the old floodplain hanging on both sides, so the number of terrace levels equals the number of falls. Tectonic uplift and climate change can also make terraces, so no single cause explains them all.
Model and sources
Model and sources
- The flume model is a teaching model, not a fluid erosion simulation: incision depth d = k(Ω − Ωc)⁺(1 − e^(−t/τ)). The entrainment threshold Ωc is 0.012 W/m for fine sand and 0.045 W/m for coarse sand, k = 0.32 m per (W/m) and τ = 45 s; all are teaching values. The groove cross-section follows (1 − u²)², the ridges beside the groove take 12% of the sand removed, and the rest builds a conical fan at the outlet. The table gives the values this model produces at 120 s, not measurements. Tracer grains carry no mass, so grain counts cannot give a sediment rate.
- V-shaped valley: 1 mm/yr = 1 m/kyr; incision depth d = U(t − τ(1 − e^(−t/τ))), with τ = 40 kyr divided by the flow multiplier; the valley-side angle tends from 10° to 32° with a time constant of 70 kyr. The waterfall retreats at 8 m/kyr × √flow × (10 m ÷ cap thickness), and a block falls for every 4 cap thicknesses of retreat. These are teaching values that illustrate the process.
- Alluvial fan and delta: deposited volume = sediment supply × time, so fan radius R ∝ t^(1/3); the active channel switches every 10 kyr to the azimuth with the least accumulated deposit. The delta front radius is solved numerically from the volume. Raising the water level 2 cm in the flume stands for a 4 m rise of model base level, a schematic scale. Supply and slopes are teaching values.
- Meander: every 2 years each point of the centreline moves toward the outer bank in proportion to the upstream-weighted curvature, and the channel is fixed at 60 m wide. A cutoff occurs when two parts of the centreline come closer than 2.2 channel widths (a neck narrower than 1.2 channel widths between banks). The migration coefficient and the starting perturbation are teaching values, and the cutoff time is not a forecast for any real river.
- Terraces: bankfull depth is 3.2 m and base level falls by 4.0 m and 4.5 m. The silt left by a flood is 0.12 times the depth above bankfull, drawn 8 times thicker on the cutaway wall. Small, medium and large floods peak 2.4 m, 4.0 m and 5.6 m above the bed.
- Stream power · Wikipedia ↗ The definition and units of Ω = ρgQS; checked 2026-11-17
- Meander · Wikipedia ↗ Secondary flow, outer-bank erosion and inner-bank deposition, cutoffs, and Einstein’s Coriolis hypothesis; checked 2026-11-17
- Oxbow lake · Wikipedia ↗ A neck cut through and the old channel sealed into a lake; checked 2026-11-17
- Alluvial fan · Wikipedia ↗ Where fans form, avulsion and the role of base level; checked 2026-11-17
- River terrace · Wikipedia ↗ Base-level change making terraces; checked 2026-11-17