Benchmark TICON stations against CHS APIs

[joeberkovitz]

A quick check of TICON predictions for several Canadian Hydrographic Services (CHS) tide stations shows some differences. The timings look fairly good (minutes off), the water levels not so much (0.3 meters off). This could be due to many issues, including differing TICON constituents or datum interpretation. Anyway, it seems worth doing by analogy with NOAA as the whole reason for ingesting TICON data is to get more coverage.

One example is the CHS tide station for Halifax, NS. The following code snippet:

      const station = nearestStation({lat:44.6700038,lon:-63.5825803});
      const start = new Date("2026-02-20T00:00:00Z");
      const end = new Date("2026-02-21T00:00:00Z");
      const { extremes: predictions } = station.getExtremesPrediction({start,end,timeFidelity: 60,datum: "MLLW"});

      console.log(predictions.map((e) => {return {t:e.time.toGMTString(),level:e.level}}))

yields these predictions:

{t: 'Fri, 20 Feb 2026 01:38:10 GMT', level: 1.5069309390205714}
{t: 'Fri, 20 Feb 2026 07:51:58 GMT', level: -0.06816336800737344}
{t: 'Fri, 20 Feb 2026 13:51:17 GMT', level: 1.4742727026708542}
{t: 'Fri, 20 Feb 2026 20:08:29 GMT', level: -0.13600777832477196}

The CHS REST API call https://api-sine.dfo-mpo.gc.ca/api/v1/stations/5cebf1df3d0f4a073c4bbcbb/data?time-series-code=wlp-hilo&from=2026-02-20T00%3A00%3A00Z&to=2026-02-21T00%3A00%3A00Z&resolution=ALL yields these results:

  {    "eventDate": "2026-02-20T01:36:00Z",  "value": 1.817,  },
  {    "eventDate": "2026-02-20T07:57:00Z",    "value": 0.31,  },
  {    "eventDate": "2026-02-20T13:46:00Z",   "value": 1.774,  },
  {    "eventDate": "2026-02-20T20:09:00Z",    "value": 0.181,  }

[bkeepers]

Yes, that is an excellent idea!