A fixed thermal-time model predicts lower spring temperature sensitivity at warm US lilac and honeysuckle sites in 1956–2009, but not the observed increase in sensitivity over time

Question: do US lilac and honeysuckle spring dates show the decrease in temperature sensitivity (days per °C) that a fixed thermal-time requirement predicts in warmer conditions (Wolkovich et al. 2021)? Data: USA-NPN legacy records (1956-2009) of first leaf and first bloom for clone lilac 'Red Rothomagensis', common lilac and clone honeysuckle 'Arnold Red' (28,019 site-years, 1,663 sites), with matched GHCN-Daily station temperatures, raw and GHCNm-homogenized. Method: we fitted forcing-only and forcing-plus-chilling models for each series and calculated model dates for the same site-years and weather. We compared within-site slopes of date on 60-day preseason temperature between the warmest and coldest third of sites and between 1956-1970 and 1986-2009, for observed and model dates. Intervals are 95% site-bootstrap intervals. Results: chilling did not improve cross-validated fit. Across sites, the observed warm/cold sensitivity ratio agreed with the model in 5 of 6 series. In common lilac first leaf, sensitivity at warm sites was half of that at cold sites (ratio 0.52 [0.42, 0.63]), and the fixed model predicted 0.61. In clone lilac first leaf, the model predicted a decrease (0.83) that the data did not show (0.99). Across time, no series decreased more than the model predicted. Sensitivity increased in clone lilac first leaf (ratio 1.25 vs model 0.95), clone lilac first bloom (1.34 vs 1.05) and common lilac first leaf (1.58 vs 0.95), because the 1956-1970 sensitivity was weak. These differences stayed with homogenized temperatures, stations within 10 km, a site-calibrated model and a 90-day window. With a 30-day window, the first-bloom difference was not significant. In 4 of 6 series, dates became 0.4-2.2 days per decade earlier than the model predicts within sites. In 2009-2025 Nature's Notebook data, clone lilac leafout was 3.8 days [1.9, 5.6] earlier than the legacy model predicts, and bloom was 1.1 days earlier [-0.3, 2.6]. Limits: station temperature error attenuates slopes and can differ between periods; preseason temperature changed by 1 °C or less between periods; simple models; protocol change after 2009. Conclusion: a fixed thermal-time requirement explains the lower sensitivity at warm sites. The data do not show a decrease in sensitivity over time.

phenology · temperature sensitivity · thermal time · growing degree days · lilac · Syringa · Lonicera · clonal plants · USA National Phenology Network · GHCN-Daily · chilling · climate change · spring leafout · flowering