Summer dry-pair excess shows no clear change across 344 US divisions between 1931-1960 and 1996-2025

We tested a change in shared dry summers across 344 climate divisions in the contiguous United States. We used fixed NOAA nClimDiv data and June-August totals. Dry means at or below each division's 1901-1930 20th percentile. Mean pair co-dry frequency was 6.224% in 1931-1960 and 4.389% in 1996-2025. The expected frequency from independent local dry rates fell from 5.196% to 3.782%. The excess above that reference fell by 0.421 percentage points. Its basic 95% three-year block interval was -1.010 to +0.116 points. Thus, the excess change is uncertain. The marginal component accounts for 77.1% of the observed pair-frequency decrease as an arithmetic result. All planned excess-change intervals include zero, and longer comparison periods change the point estimate sign. These results do not show a change in land area or hydrological drought. Short periods, fixed estimated thresholds, one retrospective product, and limited dependence assumptions restrict inference. The package contains exact source rows, code, all uncertainty draws, figures, and independent arithmetic checks.

summer precipitation · dry summers · spatial co-occurrence · nClimDiv · marginal frequency · block bootstrap · null result