Spatio-temporal analysis of the land subsidence in the UK using Independent Component Analysis

Abstract

Due to the causes of excessively underground mining, increased land surface loading and water extractions etc., land subsidence becomes a global serious problem of damaging buildings, roads and highways, flood-control structures etc. More and more new technologies such as GNSS are used to monitor the land subsidence. Analysis of the land subsidence monitoring data plays a more important role since it helps to interpret the current subsidence phenomena and predict the future subsidence scenarios. Principal Component Analysis (PCA) is used as an effective tool for spatio-temporal filtering and extracting the geophysical events from the GPS displacement series. However the modes obtained with PCA may be contaminated in cross model, which is a disadvantage to exact the signals of geophysical events. Independent Component Analysis (ICA) is a statistical method of Blind Source Separation (BSS) and can separate original signals from mixed observables. In this paper, the spatial and temporal features of land subsidence in the UK have been analyzed with the data of GPS vertical coordinate time series from the British Isles continuous GNSS Facility (BICF) stations using ICA method. In this method, vertical coordinate time series of 80 GPS stations were processed using ICA as input signals, and some independent temporal modes of vertical displacement without contamination can be extracted from these data. At the same time, the spatial response of these independent modes to each monitoring point was obtained from the mixing matrix in the process of ICA. The results showed that the land subsidence in the areas of UK can be considered consist of several independent subsidence components. And these subsidence components can be interpreted reasonably based on their features of spatial distribution.