An Improved Wiener Filtration Method for Constructing the Ensemble Pulsar Timescale

An improved Wiener filtration (WF) method for constructing the ensemble pulsar timescale (EPT) is proposed to solve the existing problems in the WF algorithm. The improvements consist of three parts: (i) adjusting the cross-power spectral density (PSD) and PSD estimated by the weighted average (WA) algorithm by the concept of the energy upper limit; (ii) combining the cross-PSD and WA PSD to form a different signal modulus for the filtration of residuals of different noise levels; and (iii) setting a weight for each residual by a more general algorithm based on the concept of effective power. We use this improved WF method to construct the EPT by both simulated data and observational data from the second data release of the International Pulsar Timing Array. The results from the simulated data indicate this improved WF can successfully suppress the noise level and reform the common signal. For observational data, this method also successfully detects the fluctuations of International Atomic Time (TAI) and local atomic time TA(NTSC). A stability analysis shows that an EPT will have the potential to reveal the instability of TAI on a scale longer than 7 yr. This improved WF method can also be used to detect and monitor the noise in pulsar timing residuals inversely, by using multiple atomic time standards, which can in turn improve the EPT.