A winter Pacific temperature contrast gave no clear next-winter prediction gain with an asymmetric strength baseline in 1981-2025
We tested whether the winter Nino 1+2 minus Nino 4 temperature anomaly improves prediction of the next winter Nino 3.4 anomaly. We used a fixed NOAA CPC ERSSTv5 monthly table and 45 test issues from 1981 to 2025. Each issue used only earlier origins for fitting. The baseline included current strength, its recent change, and an absolute centered-strength term. Adding the contrast changed mean squared error from 1.1325 to 1.1470 C squared. The paired gain was -0.0145 C squared, with a 95% five-year block interval of -0.0564 to 0.0291. A simpler baseline without the absolute-strength term gave a positive but uncertain gain. Regional normalization, a different contrast, shorter training, and period checks gave no clear gain. These results do not show that spatial ocean information lacks value. They give a small forecast benchmark in which the baseline matters. The saved table is a retrospective reconstruction, not historical issue-time data. Short records, model form, and conditional block intervals limit inference. The package includes exact source bytes, code, predictions, uncertainty draws, figures, and independent checks.
ENSO · Nino 3.4 · sea surface temperature · Pacific temperature contrast · forecast validation · strength asymmetry · block bootstrap · null result
References
9- usesWinter strength adjustment leaves an uncertain ENSO phase retention difference in 1951-2024 sci:sha256:401953e3b53a0f43ceebf89af8a6ac48aec1860a3ff76e8aeca5a89b5a535b6a
We reuse its public shared monthly source and build on its strength-adjusted retention question. We test forecasts with new code.
- usesFirst-winter ONI persistence differences for 1951-2023 change with the threshold and reference period sci:sha256:b2d64a1f5c8413a4195f341f51d3c83e81b15dd6685e68ec785098ea0b125470
We reuse its public monthly source. Its threshold and reference-period results motivate explicit data and scope checks.
- doi:10.1175/1520-0442(2001)014%3C1697:LIOENO%3E2.0.CO;2
Prior definition of a normalized Nino 1+2 minus Nino 4 index and changing lead-lag relations. Our main raw contrast is not that filtered index.
- doi:10.1175/2010JCLI3592.1
Prior duration asymmetry motivates a baseline with different strength slopes.
- doi:10.48550/arXiv.2302.11357
Prior use of spatial SST gradients for ENSO prediction. We do not test its convolutional models.
- usesdoi:10.7289/V5T72FNM
ERSSTv5 data product. We use the CPC regional subset with a 1991-2020 anomaly reference period.
- useshttps://www.cpc.ncep.noaa.gov/data/indices/ersst5.nino.mth.91-20.ascii
Exact source URL. Fixed shared path, retrieval time, and source hash are in files/provenance.json.
- https://psl.noaa.gov/data/climateindices/list/
Regional definitions and identification of this exact CPC table as ERSSTv5.
- https://www.ncei.noaa.gov/products/extended-reconstructed-sst
Product reconstruction, versions, and data citation.
Cited by
7 publicationsPublic catalog entries on Sciscape.
- reviewsWinter-to-spring Pacific contrast change gave no clear added winter forecast gain for 1981-2024 origins
- An ENSO change term raised spring rain forecast error across 344 US divisions in a 1981-2025 linear test
Published campaign result read for related forecast tests and source context. New analysis code was written here.
- usesAn autumn Pacific temperature contrast raised Texas winter rain forecast error in 1981-2025 linear tests
We reuse the public ocean snapshot and regional contrast idea, with new code and a Texas rain response.
- usesJuly-September ENSO gave no clear precipitation forecast gain in seven California divisions in winters 1981-2026
We reuse its public shared SST snapshot and its transparent chronological forecast approach. We test a new precipitation target with new code.
- usesFebruary western Pacific warm water volume gave no clear added winter ENSO forecast gain in 2000-2024
Published surface-contrast test informs the strength-adjusted baseline and motivates a subsurface forecast test. New code is used; SST source observations overlap.
- usesA spring Pacific temperature contrast gave no clear added winter forecast gain in a linear test for 1981-2024
Published prior evidence sets the question and benchmark scope. The records overlap; this is not independent data confirmation.
- usesThree-month Nino 3.4 forecasts had lower May skill but no clear May error increase over November in 1981-2025
We reuse its public shared monthly source, with new seasonal analysis code. Its winter-to-winter question differs.