Post-bloom summer temperature had no clear link to lilac bloom dates in four US samples in 1965-2008

We tested whether temperature after bloom predicted the recorded first-bloom date after winter and spring controls. Four separate US region and period samples used 9,010 repeated lilac records from 1965-2008. Models included plant and year effects. July-August slopes ranged from -1.196 to +0.070 days per C. All four intervals included zero after correction for four tests. Adding summer temperature changed each spring slope by at most 0.041 days per C. Planned checks with plant trends, flags, monthly controls, state-year effects, and a later future window also gave uncertain future slopes. The intervals were wide and do not establish equivalence or prove that spring slopes are free of bias. Limited local weather variation, selected long records, unknown visits, different plant types by region, and few state clusters limit inference. The package includes a raw-data build, exact source hashes, a derived panel, code, all estimates, independent coefficient checks, and a figure.

phenology · negative control · lilac · observation bias · temperature · temporal replication · null result