MAPPING AND PREDICTION OF LAND SUBSIDENCE

Abstract

The exact location and timing of a land subsidence-related calamity cannot be foreseen with any degree of confidence. This is true for both progressive subsidence caused by fluid (groundwater, oil or gas) withdrawal and abrupt subsidence on surface caused by underground mine collapse. The best way to deal with these dangers is to mitigate them. In an ideal world, all regions prone to such risks would be well-known, and measures would be made to either prevent causing the problem if it is human-caused, or to avoid inhabiting such areas if they are prone to natural subsidence. Extensometers, levelling, hydrogeology modelling, and GPS are common methods for monitoring subsidence, although they require precise field data and are time-consuming. PSI (Persistent Scatterer Interferometry) is a strong radar-based remote sensing technique for measuring and monitoring surface displacements over time. The techniques of PSI using microwave sensors has made monitoring of earth deformation precisely more reliable and generates large number of Persistent Scatterer Candidates (PSC’s) or sampling points. Although, all the above techniques have their own merits and de-merits but mapping and prediction of susceptible subsidence prone zones is an issue. This paper deals with the application of the Exploratory Data Analysis (EDA), a powerful environment in Data Science provided a set of tools that made it easier to interpret PSI processed Big Data and also improved the capability of validation, management, visualization, and presentation of results with greater reliability.