A 1961–2020 test found no clear difference in shared early and late ice break-up events at 17 Finnish lakes

Do unusually early lake ice break-up dates occur together more often than unusually late dates? We tested 1,020 observed dates at 17 Finnish lakes with complete records in 1961–2020. At each lake, we removed a linear date trend and selected the 12 earliest and 12 latest residual dates. Mean pair event agreement was 0.532 for early events and 0.546 for late events. The early-minus-late difference was −0.014, with a nominal 95% interval from −0.154 to +0.152 from 5,000 whole-year residual bootstrap draws in three-year blocks. Checks with other event fractions, trend rules, block lengths and a broader 41-lake panel also gave intervals that included zero. The test does not establish equal agreement: substantial differences in either direction remain possible. Inference applies to this fixed network and assumes approximately stationary residuals with short memory. These ice dates do not measure a species response or identify a climate cause. The package contains unchanged open source data, provenance, code, all estimates and draws, figures, and independent numerical checks.

lake ice · ice break-up · Finland · extreme events · spatial synchrony · null result · reproducible research