Leaf colour records gave no clear change in annual agreement in four fixed Japan networks in 1961-2020

We tested whether annual leaf-colour changes became more or less synchronized across fixed Japan station networks. We used 3,050 accepted ginkgo and maple records at 33 stations in 1961-2020. We compared the same station pairs in 1961-1990 and 1991-2020 after separate local trends and report-mark controls. Four mean correlation changes ranged from -0.020 to +0.091. All four 98.75 percent block intervals included zero. September-October temperature adjustment reduced correlation in all eight region, tree and period samples; 3 eight-test adjusted intervals were above zero among valid resamples. Bounds that include failed resamples included zero in all eight tests. September-only adjustment gave 0 clear reductions. Unknown visits, tree changes, five northern stations per tree type, and short time samples limit inference. These are selected-network associations. They do not identify a climate cause, pollution effects or species loss. The package gives source hashes, fixed-pair data, code, all draws, checks and repeat instructions.

autumn phenology · spatial synchrony · leaf colour · ginkgo · maple · temperature · observation bias · fixed station networks · null result