Two lake reconstructions showed no clear autumn-to-winter evaporation compensation for 1950-2021 autumns

We tested whether high September-November evaporation preceded low January-March evaporation in Lakes Superior and Michigan-Huron. We used 72 seasonal pairs from two L2SWBM releases on the same 1950-2021 autumn years. After linear trend removal, v2023.2 correlations were 0.231 and -0.027. Conditional 97.5% five-year block intervals were -0.099 to 0.484 and -0.284 to 0.175. Both intervals include zero. The earlier release and all planned checks also gave intervals that include zero. Thus these reconstructed medians give no clear evidence for seasonal compensation. The result does not prove absence or identify a physical process. The intervals omit joint model uncertainty. Related releases are not independent replication. The package provides original code, aggregate results, all bootstrap statistics, a figure, an independent calculation check and exact external-data retrieval instructions. Source records remain outside the bundle.

Great Lakes · evaporation · seasonal compensation · water balance · null result · L2SWBM