Investigating the Leptokurtotic Nature of Low‐Latitude GNSS Scintillations With a Multiscale Approach

Abstract

We investigate the statistical nature of Global Navigation Satellite System (GNSS) signal fluctuations through the multiscale analysis of a strong scintillation event recorded at the Ionospheric Observatory of the Luigi Broglio Space Center in Malindi, Kenya. By applying the Fast Iterative Filtering (FIF) technique, raw 50 Hz amplitude, phase and ionosphere‐free linear combination (IFLC) signals were decomposed into Intrinsic Mode Components (IMCs), enabling a rigorous examination of scale‐dependent properties. The analysis reveals that low‐latitude scintillation is inherently non‐Gaussian and intermittent, with statistical characteristics that evolve distinctively across frequency scales. Specifically, these modes exhibit pronounced leptokurtotic behavior, reflecting the combined effects of fine‐scale ionospheric turbulence and nonlinear signal propagation. In the higher‐frequency range, the probability density functions of the IMCs deviate significantly from a Gaussian distribution and are accurately modeled by Castaing‐family distributions. Strong intermittency begins above the Fresnel frequency and crucially extends into the high‐frequency range where standard spectral analysis suggests the signal is dominated by noise. We demonstrate that non‐Gaussian signatures persist in this regime, showing that meaningful physical information is preserved in the heavy tails of the distribution despite Gaussian dilution from thermal receiver noise. Similar behavior is observed in phase and IFLC signals, where excess kurtosis increases with frequency beyond the Fresnel scale. This multiscale approach demonstrates that advanced statistical models are necessary to characterize GNSS scintillation and provides a powerful framework for “superstatistical” modeling.