Impact of gaps in the asteroseismic characterization of pulsating stars. I. On the efficiency of pre-whitening

TL;DR

This study investigates how pre-whitening, a common technique in asteroseismology, can introduce biases in frequency analysis of pulsating stars with gaps, emphasizing the importance of appropriate interpolation methods.

Contribution

It demonstrates that pre-whitening may be inefficient for deconvolving spectral windows in { extdelta} Scuti stars and highlights the need for better interpolation techniques to preserve frequency content.

Findings

Pre-whitening can introduce biases in frequency analysis.

Interpolation methods affect the accuracy of asteroseismic characterization.

Proper interpolation minimizes spectral window effects.

Abstract

It is known that the observed distribution of frequencies in CoRoT and Kepler {\delta} Scuti stars has no parallelism with any theoretical model. Pre-whitening is a widespread technique in the analysis of time series with gaps from pulsating stars located in the classical instability strip such as {\delta} Scuti stars. However, some studies have pointed out that this technique might introduce biases in the results of the frequency analysis. This work aims at studying the biases that can result from pre-whitening in asteroseismology. The results will depend on the intrinsic range and distribution of frequencies of the stars. The periodic nature of the gaps in CoRoT observations, just in the range of the pulsational frequency content of the {\delta} Scuti stars, is shown to be crucial to determine their oscillation frequencies, the first step to perform asteroseismolgy of these objects. Hence, here we focus on the impact of pre-whitening on the asteroseismic characterization of {\delta} Scuti stars. We select a sample of 15 {\delta} Scuti stars observed by the CoRoT satellite, for which ultra-high quality photometric data have been obtained by its seismic channel. In order to study the impact on the asteroseismic characterization of {\delta} Scuti stars we perform the pre-whitening procedure on three datasets: gapped data, linearly interpolated data, and ARMA interpolated data. The different results obtained show that at least in some cases pre-whitening is not an efficient procedure for the deconvolution of the spectral window. therefore, in order to reduce the effect of the spectral window to the minimum it is necessary to interpolate with an algorithm that is aimed to preserve the original frequency content, and not only to perform a pre-whitening of the data.