Twenty-one Maizuru fish species showed no clear mean seasonal timing change between 2002–2012 and 2013–2023

We tested seasonal timing change at one Maizuru Bay fish survey area in Japan. Open data gave 528 twice-monthly surveys in 2002–2023. The earlier period selected 21 regularly detected seasonal species. For each species, we compared the circular phase of detected visits in 2002–2012 and 2013–2023. The equal-species mean change was −0.015 half-month slots. Negative values mean earlier timing. Shared three-year blocks gave a descriptive range of −1.158 to +0.745 slots when undefined phase changes were bounded by one half-year in either direction. In 602 of 9,999 resamples, at least one species had no defined phase. This range is an envelope of percentile ranges, not a confidence interval with proved coverage. Counts, log counts, equal annual count totals, other block lengths and other early selection rules also left the mean change unresolved. Three selected species had at most nine late detections. A later support check kept 13 species and also gave a range that included zero. The records did not establish a shared advance, delay or constant timing. The result concerns recorded visits at one site and does not identify a climate cause. The package preserves exact inputs, provenance, code, all results, figures, resample indices and independent checks. A separate-directory repeat reproduced all result and figure files.

fish · seasonal timing · circular statistics · biodiversity · Maizuru Bay · visual survey · undefined phase · open data