VELOCITY DISPERSION OF SEISMIC WAVES *

S ummary In well velocity surveys made to calibrate Sonic (CV) Logs the calibration survey uses frequencies around 50 Hz whereas the Sonic Logging tool uses frequencies around 20 kHz. There thus exists the possibility of making a direct measure of velocity dispersion. In any one survey the disturbing factors, both instrumental and operational, will often mask any dispersive effect that might exist. Consequently this paper reports on a statistical analysis of the velocity differences resulting from calibration surveys and Sonic logs. Only Borehole Compensated Sonic Logs were used. Four areas were investigated: the North Sea, Abu Dhabi, Libya and Alaska. After rejecting logs and calibration records which were obviously in error there remained 424000 feet (about 130 km) of usable log distributed throughout 66 wells. The four areas were analysed separately and in no case was the estimated dispersion significantly different from zero. However, the mean values did correlate with lithology from (− 0.17 ± 0.18)% for the essentially carbonate section in Abu Dhabi to (+ 0.45 ± 0.25)% for the sand‐shale section in Alaska, a positive sign meaning that the higher frequencies travelled faster. Except for Alaska the calibration surveys were made with a wall‐clamp geophone, and for these areas amplitude measurements were made. After suitable corrections estimates of the absorption parameter Q were obtained. These varied from 20 to 200 with mean values of 63 for Libya, 70 for Abu Dhabi and 88 for the North Sea (excluding the Tertiary). If, as is usually assumed, the absorption mechanism is linear and is described by a Q which is independent of frequency, then these values would necessarily imply dispersion of several percent. As instanced above no such dispersion was observed. It is possible that the expected dispersion was compensated for by invasion of the mud filtrate into the borehole walls, but it is more likely that the absorption mechanism was substantially non‐linear.