Analysis of the uncertainties in tidal constants obtained from short tide gauge records and their value for tidal studies

We conduct a study to estimate uncertainties in tidal constants from M2, S2, N2, K1, O1, Q1 and related K2, P1, 2N2 constituents from 35-day tide gauge records in the northern Australia and Papua New Guinea regions. The motivation for this study stems from the availability of ∼300 short tide gauge records (most ∼ 30 days long) in these regions, but their accuracy for tidal studies is not clear. We simulate the 35-day uncertainties by dividing a selected set of 14 long tide gauge records (19-years where available) from the GESLA3 data set into consecutive 35-day sections. Amplitudes and phase lags computed from each long record are treated as the ‘true’ values, from which we compute and analyse inference information for the short records. Comparison of empirical amplitude ratios and phase lag differences with the relationships from the Equilibrium tide show significant differences in both amplitude and phase lag in some constituents and locations. We also compare inference information derived from the FES2022b ocean tide model, which suggests that such models could be used in this way in some instances. Empirical uncertainties in the 35-day records were no more than 0.045 m with maximum errors reaching 0.093 m. The largest 35-day errors appeared in the K1 constituent, mostly in the Torres Strait and northwest Australia. Empirical inference information showed improvement on the Equilibrium assumption for S2 and K1 reference constituents and related constituents K2, 2N2 and P1, demonstrating that the latter can be accurately derived from short records with accurate inference information.