Mean ice break-up correlation fell by 0.165 from 1961–1990 to 1991–2020 in 41 Finnish lakes, but its 95% interval included zero

We tested change in ice break-up synchrony in a fixed panel of 41 Finnish lakes. Syke source records gave 2,410 observations in 1961–2020. We kept one site per lake and at least 27 dates per 30-year period. We removed a separate linear trend from each site in each period. Mean Pearson correlation across 820 fixed pairs was 0.697 in 1961–1990 and 0.532 in 1991–2020. The change was −0.165. A bootstrap that sampled whole regional years in three-year blocks gave a nominal 95% interval of −0.361 to +0.034. Thus, the test did not establish a decrease or equal synchrony. The interval is conditional on this lake network and approximately stationary residual variation within each period. The estimate was also negative with complete records from 17 lakes, with adjacent-year date changes, and in both distance groups. All planned change intervals included zero. Mean break-up date was 6.32 days earlier in the later period. Earlier dates therefore did not give evidence for increased spatial synchrony in this panel. The analysis does not identify a climate cause or measure species change. The package includes unchanged open source data, provenance, a local analysis plan, code, results, figures and independent numerical checks.

lake ice · Finland · spatial synchrony · ice break-up · time series · block bootstrap · ecosystem timing