Figure 1: Impact of non-uniform spatial sampling – schematic
representation (not to scale). (Top) Spatially homogeneous ionosphere at
times t1 to tn. Increase in the
intensity of the color represents increasing concentration of TEC at a
constant rate. TEC is sampled along track A at IPPs (black dots marked
as A1 to An) placed at uniform interval
of time and space. Along the track A’ the TEC is sampled (green diamonds
marked as A’1 to A’n) at non-uniform
spatial intervals and uniform time intervals. Inter-IPP distances along
track A
(d1=d2=d3…=dn=d)
are uniform; but along the track A’
(d’1≠d’2≠d’3….≠d’n)
are non-uniform. (Bottom) Spatiotemporal gradient along the track A
(black) and A’ (green) plotted as a function of inter-IPP distances.
Considering the growing importance of ionospheric perturbations forced
from below the ionosphere and its multifaceted applications, the
accurate detection of ionospheric perturbations and robust determination
of its characteristics using GPS are essential. However, to the best of
our knowledge, there is no study available in the published literature
analyzing the aliases, artifacts, and methodology dependent errors in
computing the ionospheric perturbations using GPS. In this study, for
the first time, we analyze the aliases and artifacts present in the
tsunami and earthquake induced ionospheric perturbations obtained using
conventional methods, and show that adopting Spatio-Periodic Leveling
Algorithm (SPLA) proposed by Shimna and Vijayan (2020) is efficient in
obtaining aliasing and artifact free TIPs and CIPs irrespective of the
elevation angles. We also demonstrate that SPLA removes the necessity of
applying elevation cut-offs.
We use differential and residual methods to demonstrate the impact of
aliasing and artifacts for brevity, though the problem of aliasing and
artifacts are common to all the conventional methods including the ones
which directly filters the TIP (Manta et al., 2020) or CIP from the TEC
time series using a frequency filter. Further, we validate the
efficiency of SPLA by testing the algorithm under two theoretically
simulated scenarios (section 5), and using GPS observations carried out
during the 2004 Indian Ocean tsunami and 2015 Nepal-Gorkha earthquake
(section 6). Further, we quantify the improvement in characteristics
viz. Phase, frequency, and Signal-to-Noise Ratio (SNR) of the TIPs and
CIPs upon removing aliases and artifacts caused by the non-uniform
spatial sampling (section 7).