Bloom onset control gave stronger negative spring warmth slopes in four US shrub duration samples from 1956 to 2008

We compared two models of first-to-end bloom duration in historical US lilac and honeysuckle records. Four samples split sites at 103 degrees west and years at 1980. Plant and year controls gave March-May temperature slopes from -0.42 to 0.12 days per C. All four corrected intervals included zero. Adding first bloom date gave slopes from -3.19 to -2.18 days per C. All four corrected intervals were below zero. The same rows were used for each model pair. Bloom onset control changes the quantity being estimated. It does not establish a direct heat effect. Earlier first and end bloom dates had similar temperature slopes, which explains the weak total duration slopes. Selected garden plants, unknown visits, coarse weather, and incomplete end-bloom reports limit inference. The question was selected after preliminary model inspection. Code, paired dates, weather extracts, source hashes, and numerical checks are included.

phenology · flowering duration · temperature · observation bias · model comparison · lilac · honeysuckle