KINEMATIC WAVE vs DYNAMIC WAVE
One equation, two solvers: what each keeps, what each drops, and when it matters.
Symbol key
The seven-panel breakdown
1 · The St. Venant pair
∂Q/∂t + ∂(Q²/A)/∂x + gA·∂y/∂x + gA(Sf−S0) = 0
2 · What kinematic wave keeps
Q = (K/n)·A·R2/3·S01/2
3 · The dropped terms, at a glance
✕ Convective acceleration
✕ Pressure-gradient force
✓ Bed slope balanced by friction
Show full equation
→ Sf = S0
4 · Which way information travels
cdyn = V ± √(gA/T) — free-surface only
5 · Where the shortcut holds
6 · The loop it can't enter
7 · Dynamic wave, contrast
The practical difference: a backwater profile
Kinematic wave's normal-depth line can't see the elevated tailwater. Dynamic wave's profile curves up to meet it.
Try it: toggle the momentum terms
Uncheck a term to drop it from the equation and see which solver you land on.
Scenario explorer
Pick a condition and see how each solver responds.
Kinematic wave
Dynamic wave
Routing-method advisor
Check anything that applies to your network.
Evidence: compare your own runs
Run the same model once with KINWAVE and once with DYNWAVE, then drop both .rpt files here. They're parsed locally in your browser — nothing is uploaded.
| Output | Kinematic Wave | Dynamic Wave |
|---|---|---|
| Routing method detected | — | — |
| Peak conduit flow | — | — |
| Maximum node depth | — | — |
| Max hours surcharged | — | — |
| Total flooding volume | — | — |
| Flow routing continuity error | — | — |
| Total elapsed time | — | — |
What SWMM won't tell you
A converged run is not the same as a correct one. A continuity error under 10% is necessary, not sufficient — a Dynamic Wave run can converge cleanly and still report implausible depths if the network itself is wrong. And once a network has been vetted against a Dynamic Wave run, switching to Kinematic Wave for long continuous-simulation or calibration work is a legitimate, and much faster, engineering choice — not a shortcut.