Full-bloom dates were absent only in southern stations in 4,933 selected Japan cherry records from 1953 to 2024

We tested whether winter warmth was linked to a missing full-bloom date after a recorded first flowering in Japan. A new raw-data selection gave 4,933 station-years at 76 stations from 1953 to 2024. All 43 missing full-bloom dates were in southern stations. Northern samples had no missing outcomes, so their missing-date slopes had no useful estimated uncertainty. Both southern missing-date slope intervals included zero. With each missing interval in 0-40 days, the early southern winter interval slope could range from -1.752 to +2.641 days per C; the later range was +0.217 to +1.034. These are deterministic sensitivity ranges, not confidence intervals. These bounds apply to fixed records and assume full bloom occurred. Missing dates do not identify failed flowering. Selected long records, unrecorded first flowering, rare missing outcomes, unknown visits and trees, and coarse monthly weather limit inference. We do not test pollution, daily extremes, or causal climate-change effects. The package includes code, source hashes, parsed source rows, the weather panel, all model and bound results, independent checks, and a figure.

phenology · cherry · winter temperature · missing data · selection bias · partial identification