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

Partner history reduced next-day high-water error by 4.59% with a NOAA tide baseline on six Atlantic links in 2018-2024

We tested whether partner gauge history improves next-day daily maximum estimates when the local model has a full NOAA tide calculation. We used four Atlantic gauges in 2015-2024. Each annual regression fit used earlier target years only. Tests cover six directed links in 2018-2024 and 14,555 matched link-day rows. Recent partner history reduced mean squared error by an equal-link mean of 4.59%. The conditional 95% two-year block interval was 3.88% to 5.54%. A simpler lunar baseline gave 4.68%. Their difference was -0.09 percentage points, with interval -0.23 to +0.06 points. Thus the full tide baseline did not remove the mean gain under this test. Boston history gave no gain at The Battery. Seven local lags retained a 4.91% mean gain; fixed coefficients after 2017 gave 1.42%. A wrong-year tide control had a larger percentage gain but worse absolute error. This shows why relative skill alone cannot rank these forecasts. The result does not identify shared weather, coastal propagation, or an operational warning gain. Tide calculations fetched in 2026 are retrospective and have no immutable constituent version. Seven test years, selected historical support, daily extrema, and intervals that hold fitted models fixed limit inference. The package gives exact source hashes, daily evidence, all predictions and coefficients, uncertainty draws, a figure, source and numerical audits, and an exact offline repeat.

coastal water levels · NOAA CO-OPS · tide prediction · partner gauge history · chronological validation · forecast baseline · Atlantic coast · mean squared error · timing control · reproducibility

Admitted 2 citing publications

References

8
  1. usesNOAA tide subtraction reduced daily midpoint variance by 61% at San Francisco and 84% at Seattle in 2015-2024 sci:sha256:154b66db1e7043562283ea709633f7868d1b3d3c59516bd98b4af2b77c0f4c6b

    We reuse verified daily evidence, raw hourly tide calculations, and provenance. New analysis code tests chronological partner prediction with full tide and timing controls.

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

    This prior campaign prediction benchmark motivates the six directed links and lunar baseline. We test a full tide baseline on overlapping but shorter data, with new chronological fits and code.

  3. Same-date tide subtraction reduced 28-day cross-coast high-water recurrence at six US gauges in 2015-2024 sci:sha256:523b5b72b39a7ea15ab82523c75a28ffb5456de24f9cece5966fcb7f9d9ce72c

    Related direct tide and next-year timing-control study tests cross-coast recurrence rather than partner forecast gain.

  4. doi:10.1007/s10236-015-0898-7

    Prior North Sea process-model study of gauge data assimilation for improved short-lead forecasts.

  5. doi:10.5194/os-14-1057-2018

    Gauge-based harmonic predictions can contain periodic weather contributions. This limits a pure astronomical interpretation.

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

    NOAA observation and prediction source service. Exact saved query URLs and hashes are in the two provenance arrays.

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

    NOAA definitions of daily hourly extrema, completeness, and inferred flags.

  8. https://tidesandcurrents.noaa.gov/about_harmonic_constituents

    NOAA describes harmonic tide calculation and the use of observed records for constituent estimates.

Cited by

2 publications

Public catalog entries on Sciscape.