基于地质统计学的重力梯度全张量数据三维约束反演

With the development of high precision and high resolution full tensor gravity gradient (FTG) measurement technology, three-dimensional physical properties inversion of the tensor data become a research hotspot. However, three-dimensional inversion of gravity and magnetic is faced with the problems of serious non-uniqueness and poor resolution. Cokriging method based on geostatistics can combine with all kinds of prior information fully, so as to reduce the uncertainty of inversion. However, this method applies only to isolate source or multiple sources with the same size and shape, which restricts the practical application of this method. Therefore, we will propose a new cokriging method based on the latest space non-stationary theory of geostatistics, which can be applied to complicated geological conditions. Firstly, we will establish cokriging equation to invert single tensor component data; secondly, we will establish a joint cokriging equation for inverting multiple tensor components based on weighting with the direction, which can highlight the advantage of different components to improve the resolution and reliability of inverted results; thirdly, the cokriging equation with the sparse physical properties as constraint will be established to further improve the accuracy of inversion. This research work carried out in our project will provide a new method for exploration of mineral resources and detection of the fine structure in earth interior, and, furthermore will provide theoretical and technical support for the development of the gravity gradient measurement in China.

随着高精度、高分辨重力梯度全张量测量技术的发展，全张量数据三维物性反演成为研究热点。重磁物性反演存在多解性严重和分辨率差的主要问题。基于地质统计学的协同克里金方法在与各种先验信息结合、降低反演多解性方面具有巨大的潜能。然而，当前该方法仅适用于孤立场源或大小和形态一致的多场源情况，限制了该方法实际应用。项目拟引入地质统计学最新的空间非平稳理论，建立复杂地质条件下的协同克里金反演方法：首先，建立非平稳地质条件下张量单分量反演的协同克里金方程；在此基础上建立基于方向加权的重力梯度多分量联合反演的协同克里金方程，发挥不同分量各自的方向优势，提高反演的分辨率；最后，建立稀疏物性约束下的协同克里金反演方法，进一步提高反演的分辨率与可靠性。本项目最终建立一套基于地质统计学的重力梯度全张量数据约束反演方法，为矿产资源勘查和地球内部精细结构研究提供新的方法，并为我国重力梯度测量的开展提供理论与技术支撑。

随着高精度、高分辨重力梯度全张量测量技术的发展，全张量数据三维物性反演成为研究热点。重磁物性反演存在多解性严重和分辨率差的主要问题。基于地质统计学的协同克里金方法在与各种先验信息结合、降低反演多解性方面具有巨大的潜能。项目引入了地质统计学理论，建立了复杂地质条件下的协同克里金反演方法：首先，引入Marginalizing 理论实现了地质非平稳地质条件下重磁及梯度数据三维反演，该方法解决了地质统计学反演方法仅适用于孤立场源或大小和形态一致的多场源情况的限制问题；在此基础上研究了重力梯度多分量加权联合反演方法，发挥不同分量各自的方向优势，提高反演的分辨率；最后，实现了稀疏物性约束下的重力梯度全张量数据反演，进一步提高反演的分辨率与可靠性。本项目最终建立了一套基于地质统计学的重力梯度全张量数据约束反演方法，为矿产资源勘查和地球内部精细结构研究提供新的方法，并可为我国重力梯度测量提供理论与技术支撑。