COSMOGRAIL: the COSmological MOnitoring of GRAvItational Lenses XV. Assessing the achievability and precision of time-delay measurements

TL;DR

This paper evaluates the accuracy and reliability of time-delay measurement techniques for gravitational lensing studies using simulated data, with implications for future cosmological surveys like LSST.

Contribution

It assesses the performance of current time-delay measurement methods and provides estimates for their precision in upcoming monitoring campaigns.

Findings

Techniques show no significant biases and have reliable uncertainty estimates.

Automated methods yield chi2 values between 0.5 and 1.0.

Predictions made for future campaigns' measurement precision based on signal-to-noise ratios.

Abstract

COSMOGRAIL is a long-term photometric monitoring of gravitationally lensed QSOs aimed at implementing Refsdal's time-delay method to measure cosmological parameters, in particular H0. Given long and well sampled light curves of strongly lensed QSOs, time-delay measurements require numerical techniques whose quality must be assessed. To this end, and also in view of future monitoring programs or surveys such as the LSST, a blind signal processing competition named Time Delay Challenge 1 (TDC1) was held in 2014. The aim of the present paper, which is based on the simulated light curves from the TDC1, is double. First, we test the performance of the time-delay measurement techniques currently used in COSMOGRAIL. Second, we analyse the quantity and quality of the harvest of time delays obtained from the TDC1 simulations. To achieve these goals, we first discover time delays through a careful inspection of the light curves via a dedicated visual interface. Our measurement algorithms can then be applied to the data in an automated way. We show that our techniques have no significant biases, and yield adequate uncertainty estimates resulting in reduced chi2 values between 0.5 and 1.0. We provide estimates for the number and precision of time-delay measurements that can be expected from future time-delay monitoring campaigns as a function of the photometric signal-to-noise ratio and of the true time delay. We make our blind measurements on the TDC1 data publicly available