Plant-specific weather months gave no clear gain in later-year US shrub leaf-to-bloom estimates in four samples from 1975 to 2008

We tested whether weather months selected from earlier plant dates improved later-year estimates of leaf-to-bloom intervals in a US lilac and honeysuckle archive. Four East/West and early/late samples had 87-2,181 test pairs each. Mean error reductions relative to March-April models ranged from -0.04 to +0.64 days. All four adjusted uncertainty intervals included zero. Both weather models also had larger errors than the plant-mean baseline in three samples. The result gives no clear repeated gain for this method. It does not establish equivalence. Unknown visit schedules, selected complete records, coarse weather and a small later West sample limit inference.

phenology · prediction · observation bias · temperature · lilac · honeysuckle