The Integrated Interpretation Technique Of Geophysical (Gravity, Magnetic And Seismic) Exploration Data

The integrated interpretation of gravity data, magnetic data and seismic data is animportant research direction in current integrated interpretation of geophysical data, which isan effective method to solve complicated geological problems. Gravity and magneticexploration can finish the measurement of region fast, and can obtain the distribution ofstructures. Seismic exploration is a usual exploration tool that can divide the fine structure oflayers, which has high vertical resolution, but the seismic exploration is insensitive to somespecial structures (dome, high-angle fault). Gravity and magnetic data have the functions thatincrease the resolution to high angle structure, identify seismic blind area (Mudstone fault,dome bottom), improve the simulation of seismic velocity, etc. So the integrated simulationof gravity data, magnetic data and seismic data can build more accurate geology-geophysicsmodel, and the precision of the source parameters and joint inversion method is the key to theaccuracy of the inversion results. I improve the parameter calculation and joint inversionmethod, and get more accurate results.In this research, the flow of the integrated inversion of gravity, magnetic and seismic dataas follows:(1) Applying high-precision event identification technique to interpret seismic data, andobtain the distribution of layers. Using improved edge detection methods to interpret gravityand magnetic data, and get the distribution of faults and the boundaries of stratigraphic marks.(2) Applying improved potential field interpretation method to compute parameters (depth,density and susceptibility) of the salt dome, layers and shallow sources, and then using theparameter information to build geology-geophysics model.(3) Using seismic stochastic inversion to compute the impedance, and then calculate thedensity distribution, and use the uncertainty analysis technique to get more accurate resultsdepending on the evaluation of the inversion results.(4) Applying the gravity, magnetic and gravity gradient data to compute the layers of thearea under the constraint of seismic and log data. We can build geology-geophysics modelaccording to the above information, and use the fast simulated annealing (FSA) to finish thejoint inversion of gravity, gravity gradient and seismic data, and get the finalgeology-geophysics model.Event recognition is a necessary task in the interpretation of seismic data, and can reflectthe distribution of layers clearly. In order to improve the resolution to seismic event, I presentenhanced mathematical morphology (EMM), which uses the combination of mathematicalmorphology and horizontal derivatives to identify seismic event. The test results show that themethod can identify the event effectively. Through the comparison with the other methods wecan see that the EMM method can display the seismic event more clearly, and is insensitive tonoise, and has high resolution to weak event. I also apply this method to real seismic data, andobtain the distribution of layers. Meanwhile, I found that the EMM method after a minor modification can be used to detect the edges of potential field data, which directly use the ratioof the erosion of horizontal derivative to the dilation of the horizontal derivative to finish thistask, and the maximum values of the results are corresponding to the edges of the sources. TheEMM method does not need computing the vertical derivatives of the data, so it will notincrease the effect of noise and environment. The EMM filter is tested on synthetic potentialfield data, and this method can display the edges of shallow and deep bodies. I apply the EMMmethod to real data, and this method can display the edges more clearly and preciselycompared to the other filters. But the edges recognized by the EMM method have a smalldiffuse, which is difficult to divide the edges correctly. In order to get more satisfactory results,I use the power transform and the exponential transform of EMM filter to make the edges ofthe sources clear. I demonstrate this method on synthetic and real potential field data, andenhance the contrast of the results effectively, and can display the edges more clearly.A key technique to build geology and geophysics model is to compute the propertyparameters. Seismic exploration is insensitive to salt dome, steep dip strata and small bodies,which is hard to get the accurate information about the depth, density and susceptibility of thesource, and we use gravity and magnetic anomalies to finish this task. Many of the existinginterpretation methods have the disadvantages of low precision and slow speed, which againstthe integration of multivariate data. I present several high-precision methods to compute theresults. I use the combinations of horizontal and vertical derivatives of gravity anomaly tocompute the depth, density and dip of the fault. The tests show that the parameters computedby the proposed method are close to true values. I apply the method to real gravity anomaly ofSichuan area, and the inversion results are consistent with the results of seismic data, and theresults show that the proposed method can help seismic data to compute the parameters offaults effectively. In this paper, I use analytic signal (AS) method and local wavenumber (LW)method to compute the parameters of magnetic data, and these two methods are insensitive tomagnetization direction, and the equations are simple, but the methods require the computationof third-order or higher-order derivatives, which enhance the effect of noise and lower thestability of the results. I improve the analytic signal and local wavenumber methods, and theimproved methods only require the computation of second-order derivatives, which canincrease the accuracy and stability of the results. The improved analytic signal method uses thecombinations of the horizontal and vertical derivatives of the analytic signal to compute thelocation and structural index of the magnetic source. The test shows that the improved analyticsignal method can successfully finish the inversion of magnetic source, and has strongadaptability. I apply this method to real magnetic anomaly of known dike, and the inversionresults are consistent with borehole log. The improved local wavenumber methods use thelinear combinations of the local wavenumbers at different locations and heights to compute thesource parameters. The tests show that the improved local wavenumber method is insensitiveto noise, and can get more accurate results, and the error between the inversion results and truevalues is less than5%. The practical application of the improved methods is tested on realmagnetic anomaly, and the inversion results are consistent with the results computed by theEuler deconvolution of analytic signal, so this method has good real application effect. In order to truly describe the density function of stratum, I use seismic stochasticinversion to compute the density information of the area. Stochastic inversion is thecombination geostatistics theory and inversion based on model, is also called geostatisticsinversion. Stochastic inversion combines the advantages of reservoir prediction method andseismic stochastic modeling, and fully integrates different data, and breaks the limitation ofseismic frequency bandwidth, can get high-precision impedance model, and can get theinformation about the porosity and gamma. Stochastic inversion can produce multi-realizations,and I use uncertainty analysis (US) technique to evaluate the inversion results, so theuncertainty analysis technique plays an important role in the inversion. I demonstrate thestochastic inversion on theoretical model, and the results show that the stochastic inversion iseffective in the inversion of impedance. I also apply it to real seismic data, and obtain theimpedance, and then use the uncertainty analysis technique to evaluate the results, and use themean, probability and variance to get more reliable data. Through the comparison between thestochastic inversion and determinate back analysis we can see that the stochastic inversion hashigher resolution, and obtain more subtle details. We can compute the density after obtain theimpedance of the seismic data.We should use the gravity, magnetic and gravity gradient data to compute the layers ofarea under the constraint of seismic and log data for the area that does not havethree-dimension seismic data. For the area that has three-dimension data, we directly use theevent identification technique to obtain the layers. Gravity gradient data is sensitive to shallowgeological body, detail features and change of the layers, but there are not detailed study thatuse the gradient data to invert density interface. First, I derive the relationship between densityinterface and gravity gradient, and give the flow that using the fast simulated annealing (FSA)to invert density interface depending on gravity gradient data. The theoretical tests show thatthe new method can finish the inversion of density interface, and has higher-resolution to thechange of the surface. Last, I apply the new method to interpret gravity gradient of ChinaSouth Sea, and obtain the seedbed terrain, and the inversion results are consistent with thedepth measured by LDEO boat. I use the information about the layers and property parametersto build the geology-geophysics model, and then accomplish the integrated inversion. I alsointroduce the gradient data to the integrated inversion, which can increase the resolution to thesurface. The computation speed of the fast simulated annealing (FSA) method is fast, and isgood for the interpretation of big-volume data, and I use the FSA method to finish the jointinversion of gravity, gravity gradient and seismic data. The theoretical example proves that theFSA method can finish the joint inversion of gravity, gravity gradient and seismic dataeffectively, and has higher resolution to details. Last, I apply the FSA method to process realgeophysical data, and the results reveal the distribution of layers, and provide more reliableinformation for the next exploration.