Upper GuadalupeJul 16–22, 2026

July 2026 Guadalupe flood

Peak flow: 200,000cfs
Canyon Lake rise: 21.4ft
Stations recorded: 6
NWS forecasts saved: 8

The story of this flood: The night the Guadalupe filled Canyon Lake

The wave, station by station

Downstream order, headwaters first. Each station’s highest reading of the event and the moment it hit, which is how you see the wave move. A gauge marked over-toppedwas pushed past what it can measure, its rating curve ran out of river. That’s a reading we don’t have, not a place the flood skipped.

Thu 8:00 AM

Central time · drag to scrub

Hunt

7,947cfs

Kerrville

26,400cfs

Comfort

200,000cfscresting

FM 474 (Bergheim)

2,980cfs

Spring Branch

6,260cfs

Downstream order, headwaters at the top. The dot is each gauge’s flow at the moment shown; a row turns red as its crest passes. Watch a downstream gauge sit calm while the wave is still upstream. That lag is the whole point, and the reason a quiet gauge near you is never the all-clear. Real 15-minute USGS readings; values between samples are a straight line, the crest is exact.

Peak numbers, station by station

Station	Peak flow	Stage	Crested
Hunt	40,400	20.4 ft	Jul 16, 3:35 AM
Kerrville	49,800	17.0 ft	Jul 16, 4:50 AM
Center Pointgauge over-topped	Stopped rating the flow, no valid crest
Comfort	200,000	37.1 ft	Jul 16, 8:00 AM
FM 474 (Bergheim)	196,000	57.8 ft	Jul 16, 2:30 PM
Spring Branch	69,900	36.8 ft	Jul 17, 12:30 AM

What the forecast knew, and when

The Weather Service publishes only its current forecast for Canyon Lake’s inflow, and each re-issue erases the last. We saved every one. Here is what the NWS expected the inflow to crest at, issuance by issuance, against what the river actually delivered.

First NWS forecast crest

40.1 ft

The river actually crested

36.8 ft

first call ran 3.4 ft high, then walked down each issuance

the river (observed) NWS forecast (selected) earlier / later issuances

NWS issued 8 inflow forecasts during the event. Tap one to trace it:

This basin is thin soil over limestone: it sheds a storm almost instantly, and a crest can build and pass faster than a forecast cycle can track it. The first issuance, published while the wave was still upstream in the dark, reached highest; each later one settled toward what the gauges were actually recording. Forecast stages are the captured NWS/NWPS issuances for SRGT2; the solid line is the observed record. Provisional, subject to revision.

The data, and how it was captured

Every 15 minutes for the length of this event we recorded the USGS instantaneous readings at each main-stem gauge, the Canyon Lake level from the National Weather Service, and every forecast the Weather Service issued for lake inflow. That last one matters: NWPS publishes only the current forecast, so each re-issue erases the one before it. This archive is where those survive.

Nothing here is modelled or smoothed. It’s the raw official record, kept.

Sources: USGS instantaneous values · NOAA/NWS National Water Prediction Service. Readings are provisional and subject to USGS revision. FloatTX informs and is not the go/no-go authority.