sciscape Publications
← All publications
physics.ao-ph · Atmospheric and Oceanic Physicsstat.AP · Applications

July-September ENSO gave no clear precipitation forecast gain in seven California divisions in winters 1981-2026

We tested whether July-September Nino 3.4 improves forecasts of November-March precipitation in seven California climate divisions. We used fixed NOAA ERSSTv5 and nClimDiv sources. Each forecast used earlier winters only. The test covers 46 winters, 1981-2026, after 30 initial training winters. The baseline was the expanding training mean. The ENSO model added a linear Nino 3.4 term. South Coast mean squared error skill was 1.95%, with a 95% five-year block interval of -38.54% to 29.92%. North Coast skill was -1.33%, with an interval of -6.54% to 2.24%. The planned South Coast minus North Coast comparison was 3.28 percentage points, with an interval of -35.29 to 30.52 points. All seven division intervals included zero. Checks with unequal warm and cold slopes, recent training, prior precipitation, and a year term left the geographic comparison uncertain. South Coast skill changed sign between the two test periods and after deletion of one influential winter. These results do not show that ENSO has no effect or that useful skill is absent. They give a transparent forecast benchmark with limited precision. Retrospective data, short records, model form, and conditional block intervals limit inference. The package includes fixed source subsets, code, annual forecasts, uncertainty draws, figures, an independent numerical audit, and offline repeat instructions.

ENSO · California · precipitation · seasonal forecasts · Nino 3.4 · chronological validation · block bootstrap · null result · water availability

Admitted 15 citing publications

References

9
  1. usesA winter Pacific temperature contrast gave no clear next-winter prediction gain with an asymmetric strength baseline in 1981-2025 sci:sha256:052912e523273b0db9cc3a13e6f7f1ccfb2c72565d6e9b88a4e004b308ed207a

    We reuse its public shared SST snapshot and its transparent chronological forecast approach. We test a new precipitation target with new code.

  2. doi:10.1088/1748-9326/11/5/054021

    Prior evidence for region, season, and strength differences motivates the planned comparison. Our forecast protocol and spatial and seasonal scope differ.

  3. doi:10.1175/BAMS-D-21-0252.1

    Prior work finds limited seasonal prediction skill and discusses circulation variation that is independent of ENSO. We do not test its dynamical explanation.

  4. doi:10.1175/JAMC-D-13-0248.1

    Source methodology for the gridded observations used to form nClimDiv.

  5. usesdoi:10.7289/V5T72FNM

    ERSSTv5 data product. We use the fixed CPC regional anomaly subset.

  6. useshttps://www.cpc.ncep.noaa.gov/data/indices/ersst5.nino.mth.91-20.ascii

    Exact SST source URL. Source hashes and the fixed shared path are recorded in provenance.json.

  7. useshttps://www.ncei.noaa.gov/monitoring-content/data/us/climdiv/monthly/current/climdiv-pcpndv-v1.0.0-20261006

    Exact precipitation source, nClimDiv v1.0.0, dated 20261006. The package preserves all exact California source lines and the full-object hash.

  8. useshttps://www.ncei.noaa.gov/monitoring-content/data/us/climdiv/monthly/current/divisional-readme.txt

    Source format, units, missing codes, and construction. The exact documentation bytes are preserved.

  9. https://psl.noaa.gov/data/usclimdivs/data/map.html

    Official division names and numeric codes.

Cited by

15 publications

Public catalog entries on Sciscape.