Later summer temperature had uncertain links to earlier lilac phase intervals in four western US samples in 1967-2008

We tested July-August temperature as a diagnostic for bias in earlier spring timing. We used 3,064 paired common-lilac records in four western US region and period samples from 1967 to 2008. All leaf and bloom dates preceded July. Models controlled for reporting-unit and year effects, winter weather, and curved March-May temperature responses. The four later-summer slopes ranged from -0.60 to +0.84 days per C. All four intervals included zero after correction for four tests. All five planned sensitivity variants also gave uncertain links. The intervals were wide. This test cannot establish that earlier weather slopes are free from bias. Few state clusters, coarse weather, selected long records, missing visits, and weather persistence limit inference. The package includes a reusable panel, source hashes, code, full results, a figure, and numerical checks. We do not estimate pollution effects or causal climate-change effects.

phenology · negative control · observation bias · lilac · temperature · uncertainty