GPU particles advect down the Ottawa River loop through Lima — its Muskingum–Cunge routed design-storm hydrograph, over the real NHDPlus river centerlines. Particle
density and speed encode each reach's routed storm-peak flow; reaches whose receiving water
can't dilute its permitted effluent at the design low flow — a deficit — thin and read
oxblood. Motion is deliberately restrained.
Modeled screening flow — [inference]
The particle magnitudes are the Tier-0 Muskingum–Cunge routed hydrograph —
a screening model over cited channel geometry, not a gauged flow record — so the flow field
is [inference], never [verified]. The river
centerlines themselves are real NHDPlus geometry; the deficit shading is the cited low-flow
assimilative screen. It shows where the model routes water, not a measured discharge.
Routed per-reach attenuation
Each reach's routed inflow vs. outflow storm peak — where a channel took a bite out of the
peak (attenuation) and delayed it (lag). Every figure is [inference] — a
Tier-0 screening route, most-attenuating reach first.
Reach
Length (ft)
Inflow peak (cfs)
Outflow peak (cfs)
Attenuation
Lag (hr)
Pike Run (headwater)
38,000
6,405
6,201
3.2%
0.8
Dug Run (headwater)
21,000
2,917
2,859
2.0%
0.6
Ottawa River upstream of Lima
82,000
13,479
13,348
1.0%
1.9
Dug Run -> Ottawa River
6,000
2,859
2,832
0.9%
0.2
Pike Run -> Ottawa River
4,000
6,201
6,175
0.4%
0.1
Ottawa River at Lima (assimilative reach / USGS 04187100)
20,000
14,311
14,283
0.2%
0.5
Lima supply abstraction reach (cooling draw — unbuffered bound)