Incoprea hypothesis tests · volume II
Stormwater retention fails above a threshold in impervious fraction, because infiltration capacity is lost faster than storage is added.
Retention should collapse over a narrow band of imperviousness; the crossing should be set by impervious fraction rather than by rainfall depth.
Falsified by: Failure tracking rainfall intensity would remove imperviousness control.
SCS curve-number method. Pervious curve number 75, impervious 98, composite by area. Runoff Q = (P - 0.2S)^2/(P + 0.8S) with S = 25400/CN - 254 mm. Retention is 1 - Q/P. Impervious fraction swept from 0 to 0.90 at design storms of 25, 50 and 100 mm.
code/tests/stormwater-impervious-threshold.py from the recorded run.| Retention loss per 0.1 impervious | 5.8 points |
| Largest 5-point drop | 5% |
| 50%-retention crossing | 25 mm: 83% · 50 mm: 60% · 100 mm: 21% |
| Crossing spread across storms | 62% |
Retention falls smoothly, not in a cliff. At a 50 mm storm it loses 6 percentage points per 0.1 of imperviousness through the middle of the range. No 5-point band of imperviousness loses more than 5 points of retention, against the near-100 a step would. The 50%-retention crossing moves from 83% impervious at 25 mm to 21% at 100 mm, a spread of 62 points — so failure tracks storm depth as much as imperviousness.