Monthly data leave changes in California rainfall timing uncertain between 1961-1990 and 1996-2025

We tested whether the middle half of California water-year precipitation moved later or became narrower between 1961-1990 and 1996-2025. We used a fixed NOAA nClimDiv snapshot and seven divisions with equal weights. We estimated cumulative 25% and 75% dates under constant precipitation rate within each month. The center moved 2.00 days later and the width fell by 10.02 days. Planned 97.5% five-year block intervals were -7.60 to 10.39 days for center and -21.38 to 0.91 days for width. Neither primary interval excludes zero under the two-outcome rule. Several secondary checks support narrowing, but the reference period and month window matter. Conservative within-month allocation envelopes allow either sign for actual date changes. This extends campaign seasonality results by separating center from width and testing time-bin limits. The results do not establish daily onset dates, a stable trend, water supply changes or a physical cause. One product, short periods, interpolation and block assumptions limit inference. The package preserves exact source rows, provenance, code, primary draws, a figure, a scalar audit and a successful repeated run.

California · precipitation timing · season width · monthly interpolation · nClimDiv · block bootstrap · null result