The negative Pacific volume slope contrast remained under three SST reference periods in 1980-2024 history-adjusted fits

We tested whether the SST reference accounts for a negative warm-minus-cold slope contrast for six-month Pacific warm water volume change. We used fixed NOAA GODAS and ERSSTv5 records and 531 starts from April 1980 to June 2024. Fits controlled for starting volume, month, time and prior three-month changes. Changing from the source 1991-2020 reference to a rebuilt 1981-2010 reference changed the contrast from -1.171 to -1.117 in units of 10^14 cubic metres per degree C. The paired change was 0.054, with a conditional 95% three-year-block interval of -0.005 to 0.118. This is not evidence of equivalence. The negative contrast remained with a 1951-1980 reference, month-specific linear slopes and five fixed knots from -0.5 to +0.5 C. Five-year blocks gave the same qualitative result. These tested reference choices did not explain away the contrast. One reanalysis, few ENSO cycles, fixed climatologies and conditional uncertainty limit inference. The test does not establish a physical cause or forecast skill. Code, exact input hashes, aggregate estimates, coefficient draws, a figure and a separate numerical audit are included. External monthly source records are excluded.

ENSO · warm water volume · SST reference · recharge asymmetry · block bootstrap · sensitivity analysis