Study Of The2D Gravity Inversion For Basement Relief Of Rift Basin Using Total Variation Regularization Method

The structure of basement is very important in sedimentary basin researching.Rift basin is an important type among sedimentary basins. Rift basins, controlled byfaults on both sides or on one side, often extend very long along the strike in mostcases, so they can be approximated to two-dimensional bodies. We can get theinformation about the basement relief of rift basin by using2D gravity inversionmethod in the area with less seismic, drilling and well logging data.The basement relief of rift basin is a non-smooth interface, which cannot bedelineated by traditional gravity inversion method. Besides, the gravity observingsurface does not coincide with the upper interfaces of the sedimentary layers in mostsituations, which should be an up-and-down surface alone. Therefore, the basin modelshould be a dual interface model composed of an upper interface and its basement. Wehave developed a2D gravity inversion method to estimate basement relief of riftbasin using total variation (TV) regularization method and described the theory offorward and inverse problem in detail. We also deduced the calculating formula. Theinversion method can be directly used with gravity data observed on rough terrain.The density contrast between the sediments and the basement is assumed to be knownand decrease monotonically with depth according to the hyperbolic law. We introducethe total variation function as constraint to the sedimentary basin model in order toestimate the basement relief caused by faults. The result will be non-smooth, whichcan well accord with the basement relief of the rift basin. We also use a few knowndepth points to reduce the multiplicity of the inversion. We build an objective functioncomposed of data misfit function, known depth misfit function and TV function, thenwe use nonlinear conjugate gradients (NLCG) algorithm to minimize it.Tests conducted with synthetic data show that the inversion method can estimatebasement relief of rift basin based on a dual interface model with gravity dataobserved on rough terrain. Our understanding of model tests is mainly four aspects.Firstly, adaptive regularization method cannot be used to calculate the basement relief of rift basin. We need to give the value of the regularization parameter to start thecalculation. Secondly, the resolution of TV regularization method is related with thestructure of rift basin model. The computed results are satisfactory on basins exceptfor those basins whose depth is large and their width-to-depth ratio is less than4.Thirdly, The TV regularization method is stable, which is hardly affected by datameasurement errors. However, some gravity anomalies that caused by othergeological body rather than the basement should be removed, or the computed resultwill be an inveracious basement relief. Lastly, Tests conducted with compositionmodel show that TV regularization method can be used to calculate complicatedstructurual basement of rift basin and its performance is better than intertive method.We have applied TV regularization method to process the field data fromYuncheng Linfen Rift Basin. The results clearly delineate the characteristics of thebasement, which can provide important information for researching the structure ofsedimentary basin and exploration and evaluation of mineral resources.