First‐break interpretation using generalized linear inversion

Publisher Website
Google Scholar
Abstract
ABSTKACI .The comput"Go" ofdetailed near-surface inform&j" using Ihe refracted arrivals on seismic shot records has been the subject of much research since the early days of the seismic method. This near-surfxe information is used 10 compute initial static vdues for the seismic data. In this paper. the redundancy found in CDP shooting is used toe"hancetheacc"racyofthesolu,io"andreducetheeffec, of pickingerrors. An initial rubswfxe model is input by the usev. normally consisting simply of a "umher offlat. consla"l-velocily layers. The model is the" itevafivcly updated. by usingagener~ alired linear inversion CGLI) algorirhm. in such a way "E t" reduce the difference between the observed breaks and those calculated from the model. These model breaks are c;,,cu,a,ed by ray-tracing. Two advantages of the GLI algorithm "vu previous appmache\ arc what the full redundancy of observed breaks reduces the sensitivity ofthe s"luti"n LO pickingerrors. and lhat Ihe final answer is constrained t" be re"wnably chxe to the input geological model. The paper is divided in10 four pwts. In Ihe first part. we describe the procedure concep\u+ I" the second. we look m"w closely al the theory ofthe algorithm. I" the third. model rewlfs are show" which demonstrate the convergence of rhe method through rewul iterations. Finally. Ihe fourth par! of thepaperrhowithe peti"rma"ceofthe merhodon sever,, 1-4 data euam,3e\. The presence of shallow, low-velocity weathering anomalies can cause serious problems in the processing of seismic data. This paper is concerned with compen- sating for these effects by estimating a set of static corrections using refracted xrivals. Two of the major effects that result from such near- surfxe conditions are the deteriot-ation ofthe stack due to misalignment within the CDP, and distortion in the structural times of deep retlectors. Both these effects result from time delays introduced by geological layers of varying velocity and thickness. If this time delay pattern is Fourier-analyred, it may be shown that com- ponents whose wavelength is shorter than the cable are

This publication has 2 references indexed in Scilit: