Tidal signals in ocean bottom magnetic measurements of the Northwestern Pacific: observations versus predictions

Motional induction in the ocean by tides has long been observed by both land and satellite measurements of magnetic fields. Previous studies have reported major discrepancies between observations and numerical predictions of the tidal magnetic signals. This study aims to characterize the magnetic signature of tidal currents from six different ocean bottom electromagnetic stations over the North Pacific and to compare these observations with numerical predictions. We expect the deep ocean measurements provide a low-noise and high-signal environment to detect the weak tidal magnetic signals. The ocean bottom magnetometers were located in the Northwestern Pacific Ocean and provided vector magnetic data. The stations span the time periods from August 2001 to August 2002 and from October 2005 to November 2008. The long-term trend in the data was removed by subtracting a spline fit to the data. The spectral content of the magnetic data was obtained using two different methods: Welch's periodogram method and Thompson's multitaper method. For each station, the magnetic spectral amplitude increases with an increase in periods. There are clearly defined peaks near the major tidal modes, with the largest peaks occurring at S1, K1, and P1. The S2, M2, and N2 modes also all share a large peak in the spectrum. The results from the periodogram and multitaper analysis are consistent. The large peaks in both of the spectra show that the tidal modes are detectable and significant in the ocean-bottom magnetometer data. In order to separate the daytime ionospheric signals from the data, the analysis was limited to the nighttime (18-06 local time). To estimate the tidal amplitudes from the time series with gaps, the tidal harmonics were directly fit to the data in a least-square sense. In order to minimize the effect of outliers in the data, a robust fitting method was employed. The statistically significant amplitudes (determined by their p-values) of the largest ocean tidal signals (M2, N2, K1, O1) were identified. Using the TPX07.2 tidal model and a 3D electromagnetic induction code, we predicted the tidal magnetic signal on a 0.25 x 0.25 degree global grid. In general, the observed tidal signals were very consistent with the predictions. For station NWP (Toh et al., 2004), the difference between observed and predicted tidal amplitudes is within 11%. However, for the stations in the Philippine Sea (Baba et al., 2010 PEPI), the predicted and observed amplitudes differ in the ranges between 8-40%. The largest differences between the estimated tidal amplitudes and the predicted fields appear to occur in areas where the predicted field undergoes a steep change in amplitude, suggesting the model resolution in these areas may be the cause for the discrepancies.