LateBoof The Oracle Oracle HowTo Support + Contact

STIKINE — GRAND CANYON · window opens when Telegraph Creek drops under 550 cms

OPEN since Jul 30 · 295 and holding under 550
Odds it's still runnable on Nov 01: 98%. Rain re-closes this window in 4 of 10 years — the sky gets a vote. Fly-in windows: watch the rain, not the calendar.
295 cms · ↓ falling under 300 — too low, done for the year gauge data through Oct 02 ↗

This year against the last 10

550300MayJunJulAugSepOct2005001,0002,000
P95–max
P90–95
P75–90
P25–75 · half of years
P10–25
P5–10
min–P5
this year
median
550/300 window
draw a year:
Today's 295 cms sits at the 20th percentile for Oct 02. Hover or touch for any date's numbers; tap a year below to draw it. Where the black line drops inside the green dashes, the canyon opens.

Next 10 days in the basin

basin headwaters · issued Oct 3, Open-Meteo
Sat
3
***
3°-3°
avg 2°/-4°
3 cm snow
Sun
4
***
4°-1°
avg 2°/-4°
9 cm snow
Mon
5
***
3°-1°
avg 2°/-4°
6 cm snow
Tue
6
***
3°-2°
avg 2°/-4°
2 cm snow
Wed
7
***
1°-4°
avg 1°/-4°
1 cm snow
Thu
8
***
1°-4°
avg 1°/-4°
dry
Fri
9
***
0°-8°
avg 1°/-5°
dry
Sat
10
0°-10°
avg 0°/-5°
dry
Sun
11
3°-8°
avg 0°/-5°
dry
Mon
12
***
4°-2°
avg -0°/-6°
dry
High/low at the basin headwaters (upper Stikine + Iskut, the points the model uses). The faint avg is the 1991–2020 normal for the date. Rain runs straight off; snow banks in the alpine — so a cold, snowy storm barely moves the river while a warm, rainy one lifts it fast.

Next 10 days at Dease Lake

Dease Lake (town, staging) · issued Oct 3, Open-Meteo
Sat
3
10°2°
avg 8°/0°
1.6 mm rain
Sun
4
12°5°
avg 7°/0°
3.1 mm rain
Mon
5
10°4°
avg 7°/-0°
0.5 mm rain
Tue
6
10°2°
avg 7°/-0°
0.5 mm rain
Wed
7
9°1°
avg 6°/-0°
dry
Thu
8
7°0°
avg 6°/-1°
dry
Fri
9
6°-2°
avg 6°/-1°
dry
Sat
10
6°-4°
avg 5°/-1°
dry
Sun
11
10°-2°
avg 5°/-1°
dry
Mon
12
10°2°
avg 5°/-2°
dry
The town most crews stage from — warmer and lower than the headwaters above. This is what to pack for, not a flow signal: Dease Lake drains north to the Liard, over the divide from the Stikine.

Odds the window is open, day by day

25%50%75%90%window open (7+ days under 550)opened, then rain reclosed itOct 15
From here on it's pure reclose risk — historical years that opened near our date, matched to today's status. Orange: inside a 7+ day window; blue: reclosed. As of Oct 02.

The watch log

  • Jul 30 window opened — gauge crossed under 550
  • Jul 26 opens ~Aug 05 ±8 d
    • gauge near the line (568 cms) — model lands near Aug 05
  • Jul 20 opens ~Aug 11 ±11 d
    • gauge still above the line (713 cms) — model lands near Aug 11
  • Jul 19 opens ~Aug 07 ±11 d
    • gauge still above the line (737 cms) — model lands near Aug 07
  • Jul 16 opens ~Aug 03 ±11 d
    • gauge still above the line (693 cms) — model lands near Aug 03
  • Jul 15 opens ~Aug 10 ±14 d
    • gauge still above the line (699 cms) — model lands near Aug 10
  • Jul 08 opens ~Aug 08 ±17 d
    • gauge running high (896 cms) — model lands near Aug 08
  • Jul 04 opens ~Aug 16 ±16 d
    • gauge running high (941 cms), but a cold, dry July — model lands near Aug 16
  • Jul 03 opens ~Aug 12 ±16 d
    • cool but unusually dry July
    • dry wins: less rain means a faster drop
  • Apr 01 odds only
    • season outlook opened
    • this canyon runs on summer glacier-melt and rain, so the first dated call lands Jul 1 with real melt data in hand

This season on the gauge

IN Jul 30AugAug 15SepSep 15550300cms (Telegraph Creek)
Flags mark every 550 crossing after the peak. Rain re-closes this window in 4 of 10 years — the opening is forecastable; whether it stays open depends on the fall rain.

When each season first dropped under 550

Jul 1Aug 1Sep 1Oct 1
Every dot a season’s first opening under 550, 1954–2026 · orange = it faded out (dropped low, no reclose) · blue = rain pushed it back over 550 at least once. Hover for the year.

Snow in the bank — not in the model

Forrest Kerr (high) 11 yrs from 2015 · peaked 1927 mm May 19 · now 1180 mm, 272% of the Jul 9 normal
AprMayJunJulAug1927SWE mm · shaded = middle-half of 2015– · orange = this year · dashed = median · hover for numbers
Tumeka Creek (mid) 32 yrs from 1989 · peaked 639 mm Apr 21 · melted out (normal for the date)
AprMayJunJulAug652SWE mm · shaded = middle-half of 1989– · orange = this year · dashed = median · hover for numbers
Kinaskan Lake (valley) 32 yrs from 1990 · peaked 487 mm Apr 12 · melted out (normal for the date)
AprMayJunJulAug479SWE mm · shaded = middle-half of 1990– · orange = this year · dashed = median · hover for numbers
Snow-water at three elevations, each against its own history over the melt season. The high station (Forrest Kerr) holds into summer and feeds the late river; the valley is bare by June. We checked and it does not cleanly predict the opening date, so it is context, not a model input — but a deep, slow-melting high snowpack tends to hold the river up longer.

What goes into this call — and where to check it yourself

Flow
Stikine at Telegraph Creek (Water Survey of Canada 08CE001), 71 years of record + live · BC River Forecast Centre
Weather
Basin temp & rain (ERA5 via Open-Meteo) · 16-day forecast models · seasonal ensemble out to October
History
Every recorded season's post-opening flow, spliced into the odds curve — the 4-in-10 rain reclose risk is included in the curve
Snow
Not in the call — 70 years of backtesting show snowpack adds no predictive skill to the 550 opening date. Live telemetry: Forrest Kerr high elevation (Iskut headwaters) holds 39 mm of snow-water; Kinaskan Lake and Tumeka Creek are melted out (BC automated snow stations, live).
These are forecasts, not gauge readings. Check the live gauge before you commit to travel. Seen a run come in? Let us know — real dates improve next year's forecast.