In Figure F3, raw P- and S-wave data recorded at receivers 1 and 4 and frequency spectra for receiver 1 (P-wave) and for receiver 5 (S-wave) are shown. The spectral plot indicates that most of recorded P-wave energy exists in the 10- to 13-kHz and 16- to 18-kHz frequency bands. Modeling indicates that the P-wave signal (the earliest arriving energy) is negligible above about 10 kHz, and late-arriving high-energy fluid and borehole modes generate these peaks. In these soft sediments, especially in the shallow interval, the formation P-wave velocity approaches the fluid velocity and these modes mix. This causes artifacts in the shipboard log because the software "switches" between picking the P-wave and the high-energy fluid modes. With enlarged hole conditions, the fluid arrivals are stronger and mode switching increases, generating a variable and unreliable log.
To remove this effect, filters were applied on the raw monopole data over a fixed frequency range from 5500 to 8500 Hz to isolate the P-wave energy. This bandpass filter was run three times for pass 1 and once for pass 2. In addition, the P-wave arrives earliest in the wavetrain and was separated from other later-arriving modes, such as the direct fluid and Stoneley arrivals, by selecting a variable time window length of 1450-1600 ms starting at 1400 ms for pass 1 and 1900 ms for pass 2. Testing shows that these parameter differences between pass 1 and pass 2 produce a negligible change in the results. A fixed slowness range from 80 to 280 ms/ft was used, and P-wave slowness is determined automatically by picking the first arrival time in the filtered data for both passes. The P-wave mode is also nondispersive at these frequencies (Paillet and White, 1982), and no subsequent corrections are required. Sufficient energy and high signal-to-noise in these data enable results to be obtained over the entire interval for both logging passes. Differences between the two passes are discussed below.