tvf-EMD based time series analysis of $^{7}$Be of the CTBTO-IMS network

TL;DR

This paper introduces an adaptive time series analysis method using tvf-EMD to study $^{7}$Be activity variability from CTBTO-IMS data, effectively handling nonlinear and non-stationary signals for site characterization.

Contribution

It presents a novel application of tvf-EMD combined with denoising for analyzing $^{7}$Be data, enhancing detection of meaningful oscillatory modes without prior assumptions.

Findings

Identification of significant oscillatory modes in $^{7}$Be data.

Effective removal of trends and noise for clearer mode analysis.

Insights into site-specific activity variability patterns.

Abstract

A methodology of adaptive time series analysis based on Empirical Mode Decomposition (EMD) has been employed to investigate $^{7}$Be activity concentration variability, along with temperature. Analysed data were sampled at ground level by 28 different stations of the CTBTO-IMS network. The adaptive nature of the EMD algorithm allows it to deal with data that are both nonlinear and non-stationary, making no a priori assumptions on the expansion basis. Main purpose of the adopted methodology is to characterise the possible presence of a trend, occurrence of AM-FM modulation of relevant oscillatory modes, residuals distributions and outlier occurrence. Trend component is first estimated via simple EMD and removed. The recent time varying filter EMD (tvf-EMD) technique is then employed to extract local narrow band oscillatory modes from the data. To establish their relevance, a denoising step is then carried out, employing both the Hurst exponent as a thresholding parameter and further testing their statistical significance against white noise. The ones that pass the denoising step are considered to be meaningful oscillatory modes of the data, and their AM-FM modulation is investigated. Possible applications of the adopted methodology regarding site characterisation and suggestions for further research are given in the conclusions.