Statistic Properties for Impedance Estimation of MT Data with Noise Using the Robust Method
For the magnetotelluric (MT) method, the weak signal obtained can be easily affected by various kinds of noise.How to obtain unbiased impedance estimations from measured data with complex noise is a focus of current research.To deal with the interference of various noises on the impedance estimation, the robust method is often used to obtain the unbiased impedance estimation.The essence of the robust method is that it can remove data outliers from the measured data and decrease the influence of large errors.Its optimality can be examined by posterior statistic tests on the residual.We added data with noises of different distributions and scales, and checked the statistics on the residual in terms of q-q and p-p plots.While the q-q plot emphasizes the tail of the data, the p-p plot emphasizes the distribution mode.For the synthetic data, it indicates that even with Gaussian noise, the p-p and q-q plots for the residual represent a strong bias from the Gaussian distribution.The sine transformed p-p plots show a better demonstration for Gaussian assumption tests.
robust method magnetotelluric impedance estimation q-q and p-p plots
Hang Chen Rongwen Guo Jianxin Liu
School of Geosciences and Info-Physics, Central South University, Changsha, Hunan, China Key Laboratory of Non-ferrous Resources and Geological Hazard Detection, Central South University,Ch
国际会议
杭州
英文
281-283
2018-06-10（万方平台首次上网日期，不代表论文的发表时间）