Calibration Scheme

The high quality of the Blake Outer Ridge data and the relatively deep water (more than 3 km) enable me to use a more statistical amplitude calibration: I use the seafloor reflection as a reference reflection to calibrate the amplitudes for the amplitude distorting effects relevant to the data. This approach does not attempt to compensate for all effects that diminish amplitudes individually. Rather, it corrects empirically by predicting the seafloor AVO and calibrating the data to the theoretical curve at all offsets. Amplitude correction by means of a reference reflector has been described by Chiburis (1984, 1987) and Ross and Beale 1992.

The calibration is done by assuming a functional form of the seafloor AVO response based on reasonable contrasts in P- and S-impedance at this interface. First, I normalize the data so that the seafloor peak amplitudes are equal to the near-offset seafloor amplitudes. I pick the seafloor amplitude at each offset [$\rm A(\rm x)$] and normalize the trace amplitudes [$\rm trace(t,x)$] at each offset by ${\rm A(0)\over A(x)} \:\rm trace(t,x)$. Then each trace is scaled to match the predicted seafloor AVO trend. This method will preserve the magnitude of the zero-offset seafloor amplitude variations, which are indicators of changing lateral lithology, but predetermine the shape of the seafloor AVO behavior.

By calibrating the amplitudes in this way, the described hydrophone attenuation will be automatically corrected. Given the thick water layer and small structural dips in the area, the propagation angles at the different reflectors are only about 2-3$^{^{\circ}}$ apart. Therefore, such an offset-depending calibration procedure will be valid for the reflectors beneath the BSR.

This offset dependent amplitude calibration, however, does not compensate for the effects of possible amplitude loss underneath the seafloor. Inelastic attenuation in the sediments and thin layering effects are therefore not included. In case of frequency dependent attenuation, higher frequencies will be attenuated faster than lower frequencies, thus yielding lower frequencies and amplitudes at the BSR in comparison with the seafloor. This would result in still undercorrected amplitudes at the BSR after calibrating with the seafloor AVO response. Frequency attenuation, however, will also have an effect on the hydrophone attenuation. Waves with less frequency content will be attenuated less. Thus, the BSR amplitudes would be affected less by the array attenuation than the seafloor. Consequently, correcting with the seafloor as a reference reflector would yield slightly overcorrected amplitudes at the center offsets where the array attenuation is strongest. Some of these effects might be visible in the data (see Figure ), but are treated as negligible in the later AVO analysis.