Analysis of Modified SSA for Interpolation of Missing Values in Geo-Physical Data

Abstract: To learn about the natural geophysical process first we need regular observations of various parameters such as precipitation, cloud water, temperature, and soil moisture. For example, the spatiotemporal data set of soil moisture is collected National and Snow Ice data centre. This spatiotemporal data set size is{ 44 � 102 � 92}.In this dataset, there are several inherent gaps due to various reasons such as sensor error, failure of data recording devices, lack of coverage and so on. The goal of an interpolation process is to estimate the unknown values in the gaps based on knowledge from existing values in the spatio-temporal neighbourhood. Here we plan to study an efficient spatio-temporal interpolation scheme based on singular spectral analysis (SSA). It was complicated because multiple hydrologic processes are there in soil moisture interpolation, this method uses to drive the content the modes of variation in soil moisture pattern to estimate missing values. The reconstructed data includes the estimates of missing values. The main goal of this research is to know the efficiency of the SSA based spatio-temporal data filling method.

Keywords: NSIDC, AMSRE Data, Empirical Orthogonal Function (EOF), Singular Spectral Analysis (SSA).