Annual shared and difference variances showed no clear change in Superior and Michigan-Huron between 1919-1971 and 1972-2025

We tested whether the higher correlation of annual level changes in Superior and Michigan-Huron came with larger shared changes or smaller differences between lakes. We used coordinated monthly levels, v2025, and changes in annual means for 1919-2025. We formed sums and differences after scaling each lake by its early-period standard deviation. Between 1919-1971 and 1972-2025, the shared variance ratio was 1.443, with a conditional 95% five-year block interval of 0.821 to 2.804. The difference variance ratio was 0.701, with an interval of 0.396 to 1.468. All 14 planned marginal intervals included no variance change. A shorter-period check also changed the direction of the difference estimate. Thus, the prior correlation rise does not resolve which variance changed. This is a follow-up with the same observations, not independent evidence. Short samples, period choice, scaling, block assumptions and one data product limit inference. These arithmetic measures do not identify physical water sources. Code, monthly inputs, provenance, all results, primary draws, checks and a figure are included.

Great Lakes · annual water levels · variance · shared variation · block bootstrap · null result