WebTide Northwest Atlantic Data (v0.7)

Source:

• Bedford Institute of Oceanography

Notes:

• Five constituents available: M2, K1, N2, S2 and O1
• More Details

Reference for model / data description:

• Dupont, F., C.G. Hannah, D.A. Greenberg, J.Y. Cherniawsky and C.E. Naimie. 2002. Modelling System for Tides. Can. Tech. Rep. Hydrogr. Ocean Sci. 221: vii + 72pp. (5.7M PDF format)

---

More Model Details:

The overall tidal prediction error was assessed by doing a comparison of one- year tidal predictions at almost all the tide gauge stations in the MEDS data base. The root-mean-square (rms) was computed for the one-year prediction at each station. Then these errors were averaged over different regions of interest after removing problematic stations.

The numbers on the left are based on time series computed using the 5 major constituents (5 vs 5). On the right, all of the tidal constituents were included for reconstructing the observed time series but the model time series still used the 5 major ones (5 vs all). Thus 5 vs all is a measure of expected model error when compared to the tide tables.

The Super Stations are 15 stations selected for their long records and stable tidal constituents.

Overall the results are very good. We note that even though the magnitude of the error is large in the Gulf of Maine and Bay of Fundy, the errors are respectable relative to the large M2 tide. However there is clearly a problem with the model at the northern end of the domain (Arctic).

The abstract from the report:
We assimilated Topex-Poseidon tidal observations in a finite element model of the northwest Atlantic. The modelling system consists of running a forward time-stepping non-linear model that predicts the tides and an inverse harmonic linear model that determines the elevation boundary condition at the open boundaries. In the coastal zone, we achieve an accuracy of 4-5 cm for M2 and 2-3 cm for the other constituents (outside of the Gulf of Maine and Bay of Fundy). On the shelf and offshore the system is capable of an accuracy of 2 to 3 cm per constituent. A goal of 10~cm accuracy along the coast was ascribed to this system based on a vision of what is required for future applications. The analysis indicates that the 10 cm prediction error is being achieved in Labrador, Newfoundland, the Atlantic coast of Nova Scotia, parts of the Gulf of St. Lawrence and the offshore and shelf regions. The prediction errors are generally of the order of 6~cm offshore, and between 5 to 10~cm along the coast. The prediction error increases in the Gulf of Maine and reachs 60~cm in the Bay of Fundy. The addition of SA and SSA to the predictive system improves the prediction by 0-2~cm.