The Lupus Transit Survey For Hot Jupiters: Results and Lessons

We present the results of a deep, wide-field transit survey targeting Hot Jupiter planets in the Lupus region of the Galactic plane conducted over 53 nights concentrated in two epochs separated by a year. Using the Australian National University 40-inch telescope at Siding Spring Observatory (SSO), the survey covered a 0.66 sq. deg. region close to the Galactic Plane (b=11 deg.) and monitored a total of 110,372 stars (15.0<V<22.0). Using difference imaging photometry, 16,134 light curves with a photometric precision of sigma<0.025 mag were obtained. These light curves were searched for transits, and four candidates were detected that displayed low-amplitude variability consistent with a transiting giant planet. Further investigations, including spectral typing and radial velocity measurements for some candidates, revealed that of the four, one is a true planetary companion (Lupus-TR-3), two are blended systems (Lupus-TR-1 and 4), and one is a binary (Lupus-TR-2). The results of this successful survey are instructive for optimizing the observational strategy and follow-up procedure for deep searches for transiting planets, including an upcoming survey using the SkyMapper telescope at SSO.

💡 Research Summary

The Lupus Transit Survey was a deep, wide‑field photometric campaign aimed at detecting Hot Jupiter planets in a low‑latitude Galactic‑plane field (b ≈ 11°) centered on the Lupus region. Using the Australian National University 40‑inch (1‑m) telescope at Siding Spring Observatory, the team observed a 0.66 square‑degree area over 53 nights split into two epochs separated by roughly one year (2013 and 2014). The target sample comprised 110,372 stars with V‑band magnitudes between 15.0 and 22.0. Images were taken roughly every five minutes, yielding more than 3,000 frames per epoch.

Data reduction relied on difference‑image analysis (DIA), which subtracts a high‑quality reference frame from each science image to isolate variable flux while suppressing crowding‑induced systematics. DIA delivered light curves for 16,134 stars with a per‑point photometric precision better than 0.025 mag (2.5 %). These high‑quality time series were then searched for periodic transit‑like dips using the Box‑Least‑Squares (BLS) algorithm. Candidate selection required a depth of 0.5–3 %, a duration of 1.5–5 hours, and at least three observed transit events. Four objects satisfied these criteria and were labeled Lupus‑TR‑1 through Lupus‑TR‑4.

Follow‑up characterization proceeded in three stages. First, multi‑band (UBVRI) photometry was examined for color‑magnitude consistency and signs of blended flux. Second, low‑resolution spectroscopy provided spectral types and allowed the detection of composite spectra indicative of blends or binaries. Finally, high‑resolution radial‑velocity (RV) measurements were obtained for the brightest candidates using echelle spectrographs on 2‑m class telescopes.

The RV campaign revealed that Lupus‑TR‑3 is a bona‑fide Hot Jupiter: it orbits a G‑type dwarf (V ≈ 18.9) with a period of 3.2 days, exhibits a transit depth of ~1.5 % corresponding to a planetary radius of ~1.2 R_J, and shows an RV semi‑amplitude of ~90 m s⁻¹, implying a mass of ~0.8 M_J. The other three candidates turned out to be false positives. Lupus‑TR‑1 and Lupus‑TR‑4 are blended systems where a background eclipsing binary contaminates the aperture, leading to diluted transit depths; their spectra display double‑lined features and their colors are inconsistent with a single star. Lupus‑TR‑2 is an eclipsing binary with a deep (~5 %) eclipse and an RV variation of ~30 km s⁻¹, confirming a stellar companion rather than a planet.

Beyond the scientific result—a single new Hot Jupiter—the paper emphasizes methodological lessons for future deep transit surveys. Difference‑image photometry proved essential for achieving sub‑2.5 % precision in a crowded Galactic‑plane field. Early multi‑color photometry and low‑resolution spectroscopy can efficiently weed out blends, saving valuable high‑resolution RV time. A two‑epoch observing strategy spaced by at least a year helps to diagnose long‑term systematics and improves period determination. Finally, targeting a magnitude range of 15 < V < 22 balances signal‑to‑noise with a sufficiently large stellar sample for statistical relevance.

These insights directly inform the design of an upcoming survey with the SkyMapper telescope, which offers a much larger field of view, built‑in multi‑band capability, and rapid cadence. By integrating automated DIA pipelines, real‑time color‑diagnostics, and a coordinated RV follow‑up network, the SkyMapper program aims to increase the yield of genuine transiting planets while minimizing false‑positive contamination. The Lupus survey thus serves as both a proof‑of‑concept for deep, ground‑based transit searches in dense stellar fields and a practical guide for optimizing observational and follow‑up strategies in the era of large‑scale time‑domain astronomy.