Spring leaf dates gave no clear gain in 2016–2025 autumn predictions at 95 Swiss beech stations with observed weather
An earlier spring can come with earlier autumn leaf colouring. We tested whether this link improved later predictions in open MeteoSwiss records. At 95 European beech stations, we fitted linear models to 3,760 station-years in 1961–2015. Both models used station effects, a time trend and measured seasonal weather. One also used the spring leaf-unfolding date. The historical spring coefficient was +0.157 autumn days per spring day (95% interval +0.045 to +0.270). In 849 station-years in 2016–2025, prediction RMSE changed from 11.177 to 11.246 days when spring was added. The change was +0.069 days, with a conditional bootstrap interval of −0.106 to +0.268. The later spring coefficient was −0.007 (−0.077 to +0.063). Eight other tree species also gave no clear prediction gain. Planned model and panel checks gave the same conclusion for beech. The models use observed weather in each test year. This is a conditional prediction test, not a forecast made in spring or a causal test of tree physiology. The package preserves source data, provenance, code, predictions and figures.
phenology · European beech · leaf unfolding · leaf colouring · MeteoSwiss · temporal validation · conditional prediction · autumn
References
7- usesThe Swiss tree leaf-out elevation gradient stayed near 25 days per 1000 m in 2017–2025, about 3 days above the 1960–2016 decrease trend (four species, MeteoSwiss data) sci:sha256:afc41ed41413ca69d9e9ed2aefe6f5adbb89cd29b6d75953ef22fc44c0de72e4
We use the unchanged MeteoSwiss raw-data snapshot preserved with this study. We build a new paired spring and autumn panel.
- supportsEarlier budburst did not cause earlier autumn yellowing or leaf fall in the same ginkgo and mulberry trees at Japanese weather stations in 1953–2020 when autumn temperature is controlled sci:sha256:11f49f6ad83249631ee87728c5812f778dc56e0d8946e8e047891f31291f2783
We add a Swiss conditional prediction test to its evidence that the spring–autumn link depends on species and study conditions.
- doi:10.1073/pnas.1321727111
- doi:10.1111/gcb.12890
- doi:10.1126/science.abd8911
- doi:10.1111/gcb.16104
- doi:10.1111/gcb.17429
Cited by
2 publicationsPublic catalog entries on Sciscape.
- Mean date correlation among nine tree species was higher in spring than in autumn at 110 Swiss stations in 1996–2025
Prior analysis of the same source snapshot tested spring dates as autumn predictors within species. We test correlation between species within each season.
- Previous autumn colouring gave no clear gain in linear spring predictions at 88 Swiss beech stations in 2011–2025 with weather through February
It tested the reverse direction, from spring dates to same-year autumn prediction, with these Swiss source records.