Recorded shad and sockeye passage overlap showed no clear mean change at Bonneville Dam from 1962–1993 to 1994–2025

We tested a change in the seasonal overlap of recorded adult American shad and sockeye salmon passage at Bonneville Dam. We used open daily counts for May–August 1962–2025. Each species and each eligible year had equal weight. We set blank species cells to zero and excluded years with negative counts. Mean seven-day overlap was 0.535 in 30 early years and 0.509 in 24 late years. The change was −0.027, with a 95% four-year block resampling interval of −0.120 to +0.076. Count-size, smoothing, window and negative-count checks did not establish a mean change. A strict blank-cell check reversed the point estimate but also left it uncertain. An added check compared mean annual overlap with overlap of mean seasonal curves. Averaging curves first gave a larger decline, −0.108. The difference between these measures fell by 0.082 (0.038 to 0.106). A direct triangle-inequality argument explains why averaging curves can increase overlap. The package has 192 unchanged source files from three US Army Corps dams, source hashes, code, results, figures and offline repeat instructions. Missing values, incomplete counting, and change within each period limit inference. The intervals do not include all observation error. These results do not establish equal overlap, competition, or harm to salmon.

seasonal overlap · American shad · sockeye salmon · Columbia River · phenology · open data · reproducibility