VERY HIGH RESOLUTION PSEUDO WELL LOGS OF VELOCITIES, PERCENTANISOTROPY, AND DENSITY FROM SEISMIC DATA

THE TRAVEL TIMES OF COHERENT EVENTS IN SEISMIC SURVEYS OF THE CRUST PRINCIPALLY GIVE INFORMATION ABOUT THE LOW SPATIALFREQUENCY VELOCITY FIELD; DETAILS OF THE SEISMIC SIGNATURES OF BACKSCATTERED EVENTS PRINCIPALLY GIVE INFORMATION ABOUT THE HIGH SPATIAL FREQUENCY OF THE VELOCITY AND DENSITY FIELDS (WHICH CAN BE REPRESENTED BY PSEUDO WELL LOGS). TRAVEL TIME INVERSION IS BEGINNING TO FIND PRACTICAL APPLICATION IN THE OIL INDUSTRY. HOWEVER, COMMERCIALLY USEFUL EXTRACTION OF PSEUDO WELL LOGS FROM MOST SEISMIC DATASETS IS SOME TIME OFF. THE KEY TO BOTH SOLUTIONS LIES IN A SUCCESSFUL COHERENT EVENT-PICKING PROCEDURE THAT CAN OPERATEACCURATELY AND STABLY IN THE PRESENCE OF BOTH RANDOM AND COHERENT NOISE. THE ACCURACY OF THE TRAVEL TIME INVERSION WILL INCREASE WITH THE NUMBER AND ACCURACY OF THE PICKS. PICKS MUST BE MADE BASED ON TARGET WAVELET ESTIMATES AT VARIOUS SOURCE-RECEIVER OFFSETS. THESE TARGET WAVELETS CAN BE INVERTED FOR THE PSEUDO WELL LOGS. PHASE I WILL DEVELOP A SEISMIC MODELING-AIDED, COHERENT EVENT DECOMPOSITION CLASS OF PICKING ALGORITHMS THAT SHOULD BE ACCURATE AND STABLE IN THE PRESENCE OF BOTH RANDOM NOISE AS WELL AS INTERFERENCE FROM OTHER COHERENT EVENTS. SIMULTANEOUSLY, ITSHOULD BE ABLE TO EXTRACT ACCURATE TARGET WAVELET ESTIMATES AT EACH OFFSET THAT HAVE A MAXIMIZED SIGNAL-TO-NOISE RATIO.