Estimation Of Near-surface Density Based On Gravity Bayesian Analysis And Its Application In Yunnan Area

Bouguer gravity anomaly is calculated by terrain correction and Bouguer correction of Free-air gravity data.Near-surface density is an important parameter in these two corrections.The direct acquisition method of near-surface density is to collect rock samples and measure their density.However,the sampling conditions of samples in some areas?such as sediment-covered areas?are insufficient,which requires indirect estimation of near-surface density.In addition,when the study area is large,the near-surface density is often no longer constant.If the terrain correction and the Bouguer correction are calculated with constant density according to the conventional method,the obtained Bouguer gravity anomaly will have a non-negligible deviation,which will affect the subsequent processing and interpretation.This paper studies a method of estimation near-surface density based on gravity Bayesian analysis,which can be used to obtain the lateral density distribution and smooth Bouguer gravity anomaly.In this paper,the inversion objective function is constructed from the observed data of free air gravity anomaly and the parameters of near-surface density and Bouguer gravity anomaly model.At the same time,with the assumption that the Bouguer gravity anomaly is smoothly changing relative to the surficial topography,we fit the Bouguer gravity anomaly model by cubic B-spline function and the Bayesian optimal posterior estimation algorithm based on ABIC criterion is used to iteratively solve the inversion problem.Assuming that Bouguer gravity anomaly smoothly varies with surficial topography,this method gets Bouguer gravity anomaly fitted by cubic B-spline function.The ABIC method is used to automatically determine trade-off parameters between the data misfit and model smoothness.Theoretical 1D and 2D synthetic data tests verify the effectiveness of the method in this paper.The evaluation of the effect of parameters indicates that better reliable estimated results can be obtained with sufficient spline coefficients.What's more,the ABIC method can automatically increase the weight of prior smoothness constraints when the noise level is increased and decrease it when the model roughness is increased.The application results of two gravity profiles in the Yunnan area?Yunxian-Huidong and Simao-Qidian?show that the near-surface density along the profiles varies obviously,reaching 2.45².8g·cm^-3.The density of Precambrian and Paleozoic strata is relatively high,mainly 2.53².75g·cm^-3,while that of Mesozoic is low,which is2.45².73 g·cm^-3,and the near-surface density difference between both sides of the Honghe fault is relatively large,reaching 0.4 g·cm^-3.The application result of 2D area in Yunnan area indicates that the near-surface density of the study area is roughly distribution between 2.3 and 2.75 g·cm^-3,and the overall distribution characteristic is that the density values in the northern and eastern regions are slightly higher than those in the western and southern regions.The density values of the Precambrian are2.55².65 g·cm^-3,the density values of the Paleozoic are 2.5².65 g·cm^-3,some areas have density values above 2.7 g·cm^-3,and the Mesozoic density is widely distributed between 2.45 and 2.7 g·cm^-3.The results of near-surface density estimated in this paper are roughly consistent with the regional physical properties and geological characteristic of the surface,the estimated Bouguer gravity anomaly has good smoothness.In this paper,the method provides technical support for the estimation of regional near-surface density structure,and provides more reliable Bouguer gravity data for the study of the regional deep structure and tectonics.