The Empirical Determination of a Noise Function for NDBC Buoys with Strapped-Down Accelerometers

In most wave-measuring payloads deployed by NDBC, the displacement spectra of the ocean is obtained from the acceleration spectra, as calculated from the readings of a vertical heave sensor, which is fixed in the hull of the buoy. Due to being fixed (strapped-down, as distinct from vertically stabilized), the accelerometer is subjected to motions other than vertical, and this distorts its readings, as discussed in [1] and [2]. Loosely speaking, there is a "spreading around" of energy from one frequency to the next. This is not a problem at higher frequencies. What a given frequency loses to its neighbors, it pretty well gets back from them, and any difference is a small percentage of the total energy in a real signal But at low frequencies, where there is little or no real signal, the spurious energy acquired from the rest of the spectrum has been shown to be significant. It is especially significant since the power transfer function, which converts from acceleration to displacement spectra, multiplies by the reciprocal of frequency to the power four, and will, therefore, greatly magnify any false signal. Because of this, NDBC applies a correction function to eliminate low-frequency noise, but there has been an ongoing problem with periodic noise spikes in some buoys. This paper describes a correction function that is designed to eliminate these problems.