sciscape Publications
← All publications
physics.ao-ph · Atmospheric and Oceanic Physics

Autumn partner sea level gave no clear winter forecast gain at San Francisco and Seattle in 1991-2025

We tested whether autumn mean sea level at the other Pacific gauge improves winter forecasts after local autumn level, a linear time trend, and autumn Nino 3.4 are known. We used fixed NOAA sources for San Francisco and Seattle. Each winter fit used earlier complete winters only. Common support gave 33 test winters in 1991-2025. The equal mean reduction in mean squared error was 3.97%, with a conditional 95% five-year block interval of -2.77% to 9.54%. Site estimates were 7.53% at San Francisco and 0.41% at Seattle; both intervals included zero. All planned sensitivity intervals also included zero. The result leaves added partner value uncertain for this model and season. It does not prove that spatial ocean data lack value. Retrospective sources, selected sites, missing months, short records, and uncertainty conditional on fitted histories limit inference. The package includes exact source hashes, used monthly fields, all forecasts, code, uncertainty draws, two figures, a numerical audit, and a successful offline repeat. It does not predict flood events or identify a physical cause.

sea level · Pacific coast · spatial prediction · seasonal forecast · ENSO · local persistence · NOAA CO-OPS · chronological validation · null result

Admitted 2 citing publications

References

9
  1. usesSummer ENSO gave uncertain winter sea-level forecast gains beyond local summer level at two Pacific gauges in 1986-2025 sci:sha256:930a0d10f6cec93dd1d38dd77d327bff857fbdec3f10ae12d24e5acd73c429d8

    We reuse its shared NOAA source snapshots and build on its local-level forecast comparison. Our question adds spatial partner information in autumn. New code reads the raw sources.

  2. Added ENSO skill is uncertain for monthly sea-level forecasts at two Pacific gauges in 1986-2025 sci:sha256:e21b84ef73e14748707047ddaf5138f5ce291c6c15b223b5b0e1467b782302f9

    Prior monthly ENSO tests motivate a local-history and ENSO baseline. The present target is a winter mean with a spatial predictor.

  3. Partner gauge history reduced next-day high-water error by 4.78% across six Atlantic links in 2011-2025 sci:sha256:c1307c25bf3019dd32804c881c7c620b0558d13c901afbd77f391420975e4b15

    Prior partner-gauge tests concern daily Atlantic high water. Our seasonal Pacific test uses different data fields, target, and baseline.

  4. doi:10.1029/2024JC021120

    Prior work describes coherent coastal sea-level patterns. Coherence motivates, but does not establish, lead information.

  5. doi:10.1002/2014JC010602

    Prior work estimates ENSO contributions to US coastal sea level.

  6. doi:10.5194/os-21-1663-2025

    Prior forecast work motivates a local persistence baseline and direct error scores.

  7. useshttps://api.tidesandcurrents.noaa.gov/api/prod/

    NOAA monthly_mean API. Exact source URLs, versions, hashes, and paths are in files/provenance.json.

  8. useshttps://api.tidesandcurrents.noaa.gov/api/prod/responseHelp.html

    Definitions of monthly MSL and inferred flags.

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

    Fixed ERSSTv5 monthly Nino 3.4 anomaly source with 1991-2020 reference. Exact source bytes are included.

Cited by

2 publications

Public catalog entries on Sciscape.