[Paper Review] Modelling systematics of ground-based transit photometry I. Implications on transit timing variations

C. von Essen, S. A. Cellone+3 more|arXiv (Cornell University)|Jul 13, 2016

Stellar, planetary, and galactic studies3 citations

TL;DR

This study investigates the reliability of ground-based transit photometry for detecting transit timing variations (TTVs) caused by Earth-sized planets in 3:2 resonance with hot Jupiters. Using synthetic light curves of Qatar-1b with realistic atmospheric and instrumental systematics, it shows that standard detrending methods fail to recover true TTV signals unless light curves meet strict quality criteria—specifically, high signal-to-noise ratio, near-complete transit coverage, and sufficient in-transit data points—highlighting that poor-quality data can produce false positives or inflated mass estimates by up to a factor of two.

ABSTRACT

The transit timing variation technique (TTV) has been widely used to detect and characterize multiple planetary systems. Due to the observational biases imposed mainly by the photometric conditions and instrumentation and the high signal-to-noise required to produce primary transit observations, ground-based data acquired using small telescopes limit the technique to the follow-up of hot Jupiters. However, space-based missions such as Kepler and CoRoT have already revealed that hot Jupiters are mainly found in single systems. Thus, it is natural to question ourselves if we are properly using the observing time at hand carrying out such follow-ups, or if the use of medium-to-low quality transit light curves, combined with current standard techniques of data analysis, could be playing a main role against exoplanetary search via TTVs. The purpose of this work is to investigate to what extent ground-based observations treated with current modelling techniques are reliable to detect and characterize additional planets in already known planetary systems. To meet this goal, we simulated typical primary transit observations of a hot Jupiter mimicing an existing system, Qatar-1. To resemble ground-based observations we attempt to reproduce, by means of physically and empirically motivated relationships, the effects caused by the Earth's atmosphere and the instrumental setup on the synthetic light curves. Therefore, the synthetic data present different photometric quality and transit coverage. In addition, we introduced a perturbation in the mid-transit times of the hot Jupiter, caused by an Earth-sized planet in a 3:2 mean motion resonance. Analyzing the synthetic light curves produced after certain epochs, we attempt to recover the synthetically added TTV signal by means of usual primary transit fitting techniques, and show how these can recover (or not) the TTV signal.

Motivation & Objective

• To assess whether current ground-based photometric techniques reliably detect TTVs from Earth-sized planets in hot Jupiter systems.
• To investigate how atmospheric and instrumental systematics in ground-based observations distort transit timing measurements.
• To determine the minimum photometric quality required for accurate TTV signal recovery using standard data analysis methods.
• To compare the performance of time-dependent detrending versus polynomial detrending in preserving TTV signal fidelity.
• To quantify the impact of underestimated errors and incomplete light curves on false positive TTV detections.

Proposed method

• Simulated primary transit light curves of Qatar-1b using physical and empirical models of atmospheric extinction, airmass, seeing, and instrumental noise.
• Injected a synthetic TTV signal caused by an Earth-sized planet in 3:2 mean motion resonance with the hot Jupiter.
• Applied standard photometric detrending techniques, including time-dependent low-order polynomial fitting and M2 normalization.
• Compared noise characteristics of synthetic light curves with real observational data to validate realism.
• Defined a light curve quality factor based on signal-to-noise ratio, in-transit data point fraction (NDIT/NDTot), and transit coverage (TC).
• Used false alarm probability (FAP) thresholds (≤0.1%) to assess statistical significance of recovered TTV signals.

Experimental results

Research questions

• RQ1 To what extent do ground-based systematics in photometry obscure or mimic true TTV signals from low-mass planets?
• RQ2 How do different detrending methods affect the accuracy of mid-transit time measurements in synthetic light curves?
• RQ3 What minimum photometric quality is required to reliably recover a TTV signal from a 3:2 resonant Earth-sized planet?
• RQ4 How do under-estimated uncertainties in mid-transit times contribute to false positive TTV detections?
• RQ5 Can light curves with incomplete transit coverage and low signal-to-noise ratio still yield consistent O–C diagrams when standard techniques are applied?

Key findings

• Light curves with signal-to-noise ratios below 7, incomplete transit coverage, or low in-transit data point fractions (NDIT/NDTot < 0.7) fail to recover the injected TTV signal reliably.
• Detrending with time-dependent polynomials produced less accurate and inconsistent orbital parameters compared to physically motivated detrending functions that include airmass, seeing, and centroid variations.
• Systematic errors from unaccounted-for systematics can lead to mass estimates of perturbing planets that are up to a factor of two higher than the true value.
• When only light curves with high signal-to-noise (SNR ≥ 7), full transit coverage (TC = 100%), and high NDIT/NDTot (> 0.7) are used, O–C diagrams show consistency with expected variability.
• Standard statistical techniques likely underestimate mid-transit time errors by a factor of up to three, increasing false positive risk in TTV studies.
• A quality factor combining SNR, NDIT/NDTot, and TC is proposed as a practical filter to identify light curves suitable for reliable TTV analysis.