Contamination of transient gravitational waves in LISA data by gaps and glitches

TL;DR

This paper develops a probabilistic framework to quantify how data artefacts (gaps and glitches) contaminate transient gravitational-wave signals in LISA data by modeling artefacts and transients as independent Poisson processes with dead-time-induced data loss. It introduces a Normal-approximation approach to estimate the distribution of the total contamination time $T_{\text{ctmn}}$ and self-contamination probabilities under multiple contamination rules (merged vs reset) and dead-time distributions (constant, uniform, exponential), validating the method against simulations. The authors provide analytical and numerical tools for each scenario, offering a practical figure of merit to assess instrument scenarios and set contamination-rate constraints, thereby enabling rapid mission-design evaluations. These results yield insights into how glitches and gaps can bias or mask transient GW signals and establish a framework applicable to LISA planning and similar experiments.

Abstract

We present a probabilistic framework to quantify the impact of artefacts (glitches and gaps) in LISA data on transient gravitational wave signals. By modeling both artefacts and transient signals as independent Poisson processes, and characterising the contaminating effect of an artefact by an associated dead time, we estimate the probability distribution of the total contamination time during the observation period using a Normal approximation under various contamination scenarios. Using the same approach, we also estimate the probability that a population of transient signals contaminates each other. We demonstrate the validity of the Normal approximation by comparing it to the numerical distribution obtained via simulations. Our approach provides a rapid means to assess the potential impact of glitches and gaps on LISA science, and can be used as a figure of merit to evaluate different instrumental scenarios.

Figures (8)

• Figure 1: Illustration of the scenarios described by the two different terms in \ref{['eq:recurrence_merged_scenario']}.
• Figure 2: Comparison of the PDF for the total contamination time $T_\text{ctmn}$ between the Normal approximation and the simulation in the constant dead time scenario. $T_\text{obs} = 365d$, $\lambda = 2.08\per d$, $\tau = 36s$.
• Figure 3: Comparison of the PDF for the length of contamination intervals $\Delta T_i$ (left) and the total contamination time $T_\text{ctmn}$ (right) between the analytical solution and the simulation for the uniform contamination population with the merged interval scenario. $T_\text{obs} = 365d$, $\lambda = 2.08\per d$, $\tau_{\text{max}} = 72s$. The analytical solution is truncated at $k=2$. In the right panel, the Normal approximation computed from the truncated analytical solution is shown in blue, while the Normal approximation computed with the empirical mean and variance of $\Delta T_i$ from the simulation is shown in dashed black.
• Figure 4: Comparison of the PDF for the length of contamination intervals $\Delta T_i$ (left) and the total contamination time $T_\text{ctmn}$ (right) between the analytical solution and the simulation for the exponential contamination population with the merged interval scenario. $T_\text{obs} = 365d$, $\lambda = 0.92\per d$, $\tau_{\text{mean}} = 9400s$. The analytical solution is truncated at $k=3$. In the right panel, the Normal approximation computed from the truncated analytical solution is shown in blue, while the Normal approximation computed with the empirical mean and variance of $\Delta T_i$ from the simulation is shown in dashed black.
• Figure 5: Comparison of the PDF for the length of contamination intervals $\Delta T_i$ (left) and the total contamination time $T_\text{ctmn}$ (right) between the analytical solution and the simulation for the uniform contamination population with the reset interval scenario. $T_\text{obs} = 365d$, $\lambda = 2.08\per d$, $\tau_{\text{max}} = 72s$. The analytical solution is truncated at $k=3$. In the right panel, the Normal approximation computed from the truncated analytical solution is shown in blue, while the Normal approximation computed with the empirical mean and variance of $\Delta T_i$ from the simulation is shown in dashed black.