The study of generating prestack synthetic seismograms and its parallel method in frequency-slowness domain

Forward modeling is the basis of prestack elastic-wave inversion.The slowness method is adopted to compute the prestack synthetic seismograms in layered medium,and the seismic data gathers in time-space domain can be resulted from integral transformation for frequency and slowness separately.At one time,some means are put forward about computing noise which is generated in the course of slowness integral.Layers need to be fractionized to get high precision synthetic seismograms,whereas its computation complexity is beyond the capability.If Layers are partitioned roughly with accelerating the computing speed,much information of synthetic seismograms will be lost.Thus,partition principle of layers is brought forward.The method could compute a complete seismic response including mode conversions and multiple reflections.However,high precision synthetic seismograms cause to increase massive computation complexity and depress computation efficiency massively.Hence,parallel algorithm of full wave-field forward modeling is designed on the basis of slowness method,which adopts parallel technique of segmenting computation domain and work pool.And,its parallel algorithm is established so that the problem of elastic-wave forward modeling is solved more effectively,making the foundation for prestack inversion with prestack seismic data well for further study.