Stevenson, SamanthaDeser, ClaraCoats, SloanFalster, GeorginaKonecky, BrowenMaher, NicolaPfleger, Cali2026-04-112026-04-11PubMed:41790897WOS:001708103300009ORCID:/0000-0001-8567-7413/work/211161473ORCID:/0000-0003-3922-9833/work/211162339https://hdl.handle.net/1885/733808416The equatorial Pacific sea surface temperature (SST) zonal gradient has worldwide impacts and is expected to be highly sensitive to future climate change. However, biases in climate models call the reliability of future SST gradient projections into question. Here, we combine multiple climate model Large Ensembles to show that equatorial precipitation and cloud feedbacks have a controlling influence on the future Pacific SST gradient. An "SST gradient sensitivity" parameter is computed for each model, which shows that models with stronger historical equatorial precipitation have systematically higher sensitivities (more El Nino-like changes). This arises from the stronger negative SST-shortwave radiation feedback, which then creates a wind response that favors El Nino-like warming. Notably, when simulated historical deep convection is sufficiently strong, a "saturation" effect occurs that tends to inhibit this effect. These results imply that models likely underestimate future El Nino-like changes but that the "true" magnitude of changes may be predictable.S.S. was supported by NSF CAREER, OCE-2142953. C.D. was supported by the NSF National Center for Atmospheric Research, which is a major facility sponsored by the NSF under the Cooperative Agreement 1852977. N.M. was supported by the Australian Research Council Discovery Early Career Researcher Award DE230100315. G.F. was supported by the Australian Research Council Discovery Early Career Researcher Award DE250100071.11en©2026 The authorsEl-ninoEventsFrequencyProjectionsResponsesSst warming patternUncertainty2026-03-0610.1126/sciadv.aea8070105032622883