Recovering the O VII Absorption Distributions from X-Ray Data

TL;DR

This paper develops methods to recover true cosmic gas absorption line distributions from noisy X-ray data, demonstrating that both parametric and non-parametric approaches effectively estimate the distribution near detection thresholds.

Contribution

It introduces and compares two novel methods for recovering absorption line distributions from noisy data, applicable to cosmic gas studies.

Findings

Both methods perform equally well in recovering distributions.

Approximately 14% accuracy in recovering the number of absorption systems.

Approach is broadly applicable to well-behaved statistical data.

Abstract

The absorption by gas toward background continuum sources informs us about the cosmic density of gas components as well as the hosts responsible for the absorption (galaxies, clusters, cosmic filaments). Cosmic absorption line distributions are distorted near the detection threshold (S/N $\approx 3$) due to true lines being scattered to lower S/N and false detections occurring at the same S/N. We simulate absorption line distributions in the presence of noise and consider two models for recovery: a parametric fitting of the noise plus a cut-off power law absorption line distribution; a non-parametric fit where the negative absorption line distribution (emission lines) is subtracted from the positive S/N absorption line distribution (flip and subtract). We show that both approaches work equally well and can use data where S/N$\gtrsim$3 to constrain the fit. For an input of about 100 absorption line systems, the number of systems is recovered to $\approx$14%. This investigation examined the O VII X-ray absorption line distribution, but the approach should be broadly applicable for statistically well-behaved data.