. . . . . . "Replication of the satellite-era marine heatwave day trend using ESA SST CCI and XMHW" . . . . . "1. Sea surface temperature source. ESA SST CCI Analysis v3.0 rather than NOAA\n OI SST, and a 1 degree analysis grid rather than the original's 0.25 degree\n grid. This is the intended independence of the replication, not an\n incidental difference.\n\n2. Detection software. XMHW rather than the original authors' own\n implementation of the same published definition.\n\n3. Longitude sampling. The ARCO reader takes bands 60 native cells wide starting\n every 64, so 0.2 degrees of longitude is skipped every 3.2 degrees and the\n analysis grid holds 336 of a possible 360 one-degree columns - 93.3 percent\n coverage. Measured effect on the headline statistic: plus or minus 0.64 days.\n\n4. ENSO removal is fitted to deseasonalised anomalies. The original describes\n regressing daily sea surface temperatures onto the multivariate ENSO index.\n Applied literally to raw temperatures, 25 predictors spanning plus and minus\n twelve months can combine into an annual harmonic and absorb the seasonal\n cycle; in this implementation that accounted for 66 percent of the signal\n removed and more than doubled detected marine heatwave days, when removing\n ENSO should reduce them. Regressing the anomaly confines the fit to\n interannual variability. This is believed to be what the original intends\n rather than a departure from it.\n\n5. ENSO index version. The original multivariate ENSO index of Wolter and\n Timlin, which the original paper cites and which NOAA last updated in\n December 2018, rather than the maintained MEI.v2 successor, which uses a\n different variable set and base period.\n\n6. Annual sea surface temperature variance and skewness are computed on\n deseasonalised anomalies. At mid-latitudes the seasonal cycle is roughly 95\n percent of daily variance, so computing these on raw temperatures measures\n seasonal amplitude rather than the variability that governs threshold\n exceedance.\n\n7. The excess-trend significance test against a stochastic climate model was not\n performed, so the corresponding maps carry no significance hatching." . . "Daily sea surface temperature from ESA SST CCI Analysis v3.0 was streamed from\nthe Copernicus Marine Service ARCO store and area-averaged from its native 0.05\ndegree grid to a 1 degree analysis grid while streaming, covering 1 January 1982\nto 31 December 2016. Cells with continuous sea ice were excluded, following the\noriginal's ice-exclusion rule.\n\nMarine heatwaves were detected with XMHW, an xarray implementation of the\nHobday et al. (2016) definition, using the parameters the original states: 90th\npercentile threshold, minimum duration of five days, gaps shorter than three\ndays merged, an 11-day window for the percentile, 31-day smoothing of the\npercentile, and a 1983-2012 baseline climatology.\n\nMarine heatwave days were attributed to the calendar year in which they fall;\nevent duration and intensity were attributed to the year each event started, as\nthe original specifies. Per-cell annual statistics were aggregated to a global\nmean weighted by the cosine of latitude, and trends were estimated with the\nTheil-Sen estimator with 95 percent confidence intervals, as in the original.\n\nTo separate the secular trend from interannual variability, the analysis was\nrepeated on a sea surface temperature series with the ENSO signature removed:\ndaily anomalies at each cell were regressed onto the multivariate ENSO index\nwith monthly leads and lags to plus and minus one year, and the fitted ENSO\ncomponent subtracted. Detection on that series used the climatology and\nthreshold derived from the original, unmodified series, so that events remain\ndefined relative to real-world conditions." . "This study tests the claim's satellite-era component: the change in globally\naveraged annual marine heatwave days between 1982 and 2016, and the 1980s\nbaseline level against which that change is expressed. The two supporting\nmetrics the original reports alongside it — trends in marine heatwave frequency\nand duration — are also evaluated, since the claim states that the day-count\nincrease follows from them.\n\nOut of scope: every pre-satellite element of the original paper. Its\ncentury-scale results (1925-1954 versus 1987-2016), the proxy reconstruction\nfrom monthly gridded SST, and the six century-long in situ station records are\nnot tested here, because no independent daily global SST record exists before\n1981 that could test them. The abstract's most-quoted figure — a 54 percent\nincrease in marine heatwave days - belongs to that pre-satellite proxy analysis\nand is therefore not the claim under test.\n\nAlso out of scope: the excess-trend attribution of Figure 3a-c, which asks\nwhether marine heatwave trends exceed what mean sea surface temperature warming\nalone would produce. That test requires a Monte Carlo ensemble of synthetic\ndetections and was not performed." . . . "Anne Fouilloux" . "2026-08-15T16:05:02.065Z"^^ . . . . . "Replication of the satellite-era marine heatwave day trend using ESA SST CCI and XMHW" . . "RSA" . "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" . "ZOcytJK2xqFHehvKOu8rCPL9LdLwQqlYWb2mZ2S7TOYN0yRg4ZVXwYXY6qKynZmldls2NHZqf19DWvp0YIbDuOye8jP9XEx+SbIaAI21LBclosB0lFYC/IVDFmUaMHGv/xhezArpo4EaJJ1FkuHy7YYAcG6mLr0F1p9ZRbaNw0b6e3kTJegctRp1F37+7OiCemomn+cIzyF9tkpvPF4GZwcOu8/QFwA11xcRNp4eQOqnTRtea1qVKsisWufogpJzxSBzwPS7PvpteMSdy0cYfALbLIts0+msIgap81fPLYCvxz9MkLJpjApRrUz0iCwmIdEAkWUWWjAWdLHN6FmmEZ1xD8tVRoSBdk9cNkv20c2PjAbkpMZki1NVCu8MOjpInBQGLeFFlP1m2YPdynB6UbEuCBLLlQl49qKSlrTQQdguALv0qXrNAULw85nFA+b1PGknmLIssnu03Bm/RrMBrF1x2JWtAXzOB/Qz2cUTB44SvlyCcQ5nfQPPASH4bmpxLyXX1I2aMpy1jxSzawdFKhUaht73U2NxE+4vKICBXHoVpzZrehHZepmlqMzIuShBdMKpkY9gdWQhNMDUU/araDKuFFDaOT2P9UsZrIl24wzhdXTXG6X7DU4fzIKdJkjQviVkdPE6lRo8WWMsgffyScm87v4m7VTwLiPR1dh2Vzg=" . . .