Earlier 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

Question: does an earlier spring cause an earlier autumn in the same tree (spring-to-autumn carry-over)? We used Japan Meteorological Agency phenology records. At each station, the staff observe one reference tree per species. Ginkgo budburst, yellowing and leaf fall come from the same tree, and mulberry budburst and leaf fall also come from one tree. The data have 3,719 ginkgo station-years at 82 stations (1953–2020) and about 1,250 mulberry station-years at 36 stations. We used GHCN-Daily temperature from the same stations. Within each station, we regressed the autumn date on the budburst date. We controlled for the 60-day pre-event autumn temperature and for pre- and post-solstice temperature, and we also used year effects, station trends, monthly temperatures, instrumental variables and a negative control (cherry flowering date from a different tree). Without temperature control, an earlier budburst came with a later autumn (−0.17 to −0.22 autumn days per spring day). With control, the same-tree slopes were near zero: from −0.24 to +0.04 in 16 specifications per ginkgo outcome, with upper 95% limits of 0.11 or less. The pooled station slopes were −0.002 [−0.045, 0.041] for yellowing and −0.018 [−0.058, 0.023] for leaf fall. All ginkgo specifications reject a carry-over of 0.2 days per day or more (one-sided p < 0.0001). The mulberry slopes were also near zero, with wider intervals. The cherry date from the other tree gave equal or more negative slopes, so the small negative residual slopes agree with shared weather and not with tree physiology. Thus each day of earlier budburst added about one day to the leaf-on season, and the ginkgo season became longer by 2.4 ± 0.3 days per decade. In a secondary test, warmer May–June temperatures advanced the autumn events by 0.6–0.8 days per °C after detrending, but warmer July–August temperatures did not delay them. This result depends on the detrending model. Limits: one urban reference tree per station, observational data, and temperature as a confounder that we can control only partly.

phenology · leaf senescence · carry-over effect · Ginkgo biloba · Morus · Japan Meteorological Agency · growing season length · climate change · GHCN-Daily · summer solstice · negative control · autumn phenology