How this site works

This site tracks real daily fish counts recorded at sonar and weir counting stations across Alaska, then uses years of that history to project how a season is likely to unfold.

During the season, each river's page shows the actual daily counts recorded so far alongside a projected count for the same days, based on how previous seasons typically played out by that same date. This lets you compare this year's pace against history in real time.

Before a season starts, the site builds a full day-by-day projection for the upcoming year using historical timing patterns alone, since there's no current-season data yet to go on.

Finding the best settings: a projection can be built several different ways — which years to draw from, how to average them, whether to exclude unusually large or small years, and more. For many rivers, the site has already tested 504 different combinations of these settings against 15 years of real, already-known seasons to find the combination that would have predicted each one most accurately, and uses that automatically. Rivers that don't yet have enough years of history on record show a notice that generic default settings are in use instead.

You can also adjust every setting yourself — years of history to use, averaging method, outlier exclusion, uncertainty ranges, trend adjustment, and date alignment — to see how the projection responds.

A note on accuracy: these are historical-timing-based statistical estimates, not guaranteed forecasts. Treat them as a rough planning aid rather than a promise of what you'll see on the water.

A salmon wearing glasses and holding a tally clipboard, counting a school of salmon swimming past underwater.

Fish Count Projections

Historical run data and in-season projections for salmon counting locations. Pick a dataset below to see its counts.

37 rivers, 126 options

Fish Creek

Daily sonar counts and an in-season run projection based on historical timing.

Season
2026
7 days of data
Count to date
119
through Aug 2, 2026
Projected season total
303
18 prior seasons used
% of projection reached
39.3%
avg. historical pace: 39.2% by this date
Projection method: for each prior season, we compute what fraction of that season's total run had passed by August 2, average those fractions, then divide this season's count-to-date by that average. It's a simple historical-timing estimate, not a statistical forecast — treat it as a rough in-season indicator.
2027 Projected Daily Counts
Each day's projection averages only years sharing 2027's position in sockeye's roughly 4-year dominant brood cycle (2023). This is computed across 1 prior season (2023–2023), applied to those seasons' average total of 695. It's a rough forecast for a season with no data yet — not a statistical model.
Days are aligned by time since season start rather than calendar date, so historically early or late seasons line up by day-of-season.
DateProjected Daily CountProjected Cumulative
Aug 31, 2027 3 695
Aug 27, 2027 2 692
Aug 26, 2027 9 690
Aug 25, 2027 4 681
Aug 24, 2027 5 677
Aug 23, 2027 12 672
Aug 22, 2027 7 660
Aug 21, 2027 12 653
Aug 20, 2027 2 641
Aug 19, 2027 1 639
Aug 18, 2027 6 638
Aug 17, 2027 12 632
Aug 16, 2027 36 620
Aug 15, 2027 23 584
Aug 14, 2027 17 561
Aug 13, 2027 11 544
Aug 12, 2027 11 533
Aug 11, 2027 14 522
Aug 10, 2027 3 508
Aug 9, 2027 24 505
Aug 8, 2027 23 481
Aug 7, 2027 13 458
Aug 6, 2027 20 445
Aug 5, 2027 20 425
Aug 3, 2027 44 405
Aug 1, 2027 76 361
Jul 31, 2027 83 285
Jul 29, 2027 79 202
Jul 26, 2027 52 123
Jul 25, 2027 54 71
Jul 24, 2027 17 17

Highlighted cells are the top 5 projected daily counts.

2026 Daily Counts
Showing 7 actual days of counts. Not enough historical data to build a projection for comparison.
Date Actual Daily Count Projected Daily Count Actual Cumulative Projected Cumulative
Aug 2, 2026 96 119
Aug 1, 2026 0 23
Jul 31, 2026 6 23
Jul 30, 2026 10 17
Jul 29, 2026 3 7
Jul 28, 2026 2 4
Jul 27, 2026 2 2

Highlighted cells are the top 5 actual and top 5 projected daily counts.

Backtest a Prior Year
This uses only years strictly before 2025 as the reference pool (19 available, 1996–2024) — 2025 itself is never used to predict itself, so this is a genuine out-of-sample test of the projection method.
Each day's projection averages only years sharing 2025's position in sockeye's roughly 4-year dominant brood cycle (2021). Computed across 1 prior season (2021–2021), applied to those seasons' average total of 5.
1 outlier year excluded (2023).
Days are aligned by time since season start (day 1 = Jul 17).
Date Actual Daily Count Projected Daily Count Actual Cumulative Projected Cumulative
Sep 2, 2025 0 1,056
Sep 1, 2025 0 1,056
Aug 31, 2025 2 1,056
Aug 30, 2025 2 1,054
Aug 29, 2025 6 1,052
Aug 28, 2025 9 1,046
Aug 26, 2025 6 1,037
Aug 24, 2025 5 1,031
Aug 23, 2025 5 1,026
Aug 22, 2025 5 1,021
Aug 21, 2025 5 1,016
Aug 20, 2025 14 1,011
Aug 19, 2025 53 997
Aug 18, 2025 45 944
Aug 17, 2025 16 899
Aug 16, 2025 17 883
Aug 15, 2025 16 866
Aug 14, 2025 90 850
Aug 13, 2025 56 760
Aug 11, 2025 41 704
Aug 10, 2025 58 663
Aug 9, 2025 0 605
Aug 8, 2025 20 605
Aug 7, 2025 19 585
Aug 6, 2025 88 566
Aug 5, 2025 24 478
Aug 4, 2025 99 454
Aug 2, 2025 44 355
Jul 31, 2025 123 311
Jul 30, 2025 127 188
Jul 28, 2025 47 61
Jul 25, 2025 0 0 14 5
Jul 24, 2025 0 5
Jul 23, 2025 1 14
Jul 22, 2025 1 13
Jul 21, 2025 4 12
Jul 20, 2025 1 0 8 5
Jul 19, 2025 1 0 7 5
Jul 18, 2025 2 6
Jul 17, 2025 4 5 4 5

Highlighted cells are the top 5 actual and top 5 projected daily counts for 2025.