Area-weighted local total steric variance was lower than thermal variance in three NOAA basins in 2005-2025

We tested whether spatial averaging explains lower total than temperature-only steric sea level variance in the Atlantic. We used NOAA upper-700-m fields for 84 seasons in 2005-2025. After removal of cell trends and fixed seasonal means, the total-to-thermal ratio of area-weighted cell variances was 0.684 in the Atlantic, 0.860 in the Pacific and 0.899 in the Indian Ocean. All three conditional 95% time-block intervals were below one. The corresponding variance ratios for basin means were 0.728, 1.134 and 1.222. The primary Atlantic log contrast between the two ratios was 0.062, with three-year block limits of -0.175 to 0.454. Thus, the test gave no clear Atlantic aggregation contrast. The secondary Pacific and Indian contrasts were positive for two- to four-year blocks. However, intervals for each basin-mean ratio included one, and some point contrasts changed sign without trend removal or with first differences. Spatial averaging can hide local variance reduction in this product. The result does not identify a physical cause. One mapped product, 21 years, source error and conditional uncertainty limit inference. Code, source hashes, aggregate evidence, checks and a figure are included.

steric sea level · spatial aggregation · salinity compensation · variance · null result · NOAA