Local temperature adjustment gave no clear reduction in annual cherry bloom agreement across three site pairs in 1951-2025

We tested whether local seasonal temperature reduced shared annual changes in cherry peak bloom dates at Kyoto, Liestal and Washington. We used 225 site-years and separate tests in 1951-1987 and 1988-2025. Site models removed linear year trends, then added January-February and March temperature. The six correlation reductions ranged from -0.194 to +0.425. All six intervals included zero after correction for six tests. For Liestal and Washington in 1988-2025, correlation fell from 0.486 to 0.061. Its reduction was 0.425, with a corrected interval of -0.045 to +0.793. This supports local temperature as a candidate explanation for the prior pattern, but does not establish it. Fixed sites and complete years reduce changes in coverage. Different species, bloom rules, unknown visit histories, coarse weather and short periods limit inference. These descriptive fits do not show a causal climate effect or measure pollution, extreme weather or species loss. Code, source hashes, derived evidence and offline checks are included.

phenology · cherry bloom · temperature · spatial synchrony · observation bias · null result