🔭 astrophysics

Cold giant discoveries from a joint radial-velocity and astrometry framework

By upgrading the EMPEROR framework to jointly analyze long-baseline radial velocities and absolute astrometry for metal-rich FGK stars, this study characterizes two known planets and discovers five new cold giants, demonstrating that combining these methods significantly reduces orbital uncertainties and converts minimum mass estimates into true masses.

Original authors: Pablo A. Peña, James S. Jenkins, Fabo Feng, Douglas R. Alves, Florence de Almeida, Frédéric Dux, Guang-Yao Xiao, Joanne M. Rojas M., Jose I. Vines, Rafael I. Rubenstein, R. Ramírez Reyes, Suman Saha

Published 2026-03-13

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Pablo A. Peña, James S. Jenkins, Fabo Feng, Douglas R. Alves, Florence de Almeida, Frédéric Dux, Guang-Yao Xiao, Joanne M. Rojas M., Jose I. Vines, Rafael I. Rubenstein, R. Ramírez Reyes, Suman Saha, Connor J. Cheverall, Matías R. Díaz

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are trying to figure out the weight and path of a giant, invisible elephant walking around a lighthouse in the fog. You can't see the elephant, but you can see the lighthouse swaying slightly as the elephant pulls on it with a rope.

This is essentially what astronomers do when they hunt for exoplanets (planets outside our solar system). For decades, they've used two main tools:

The "Wobble" (Radial Velocity): Measuring how much the star sways back and forth. This tells us the planet exists and gives a minimum weight, but it's like guessing the elephant's weight without knowing how heavy the rope is or how hard the elephant is pulling.
The "Shadow" (Transits): Waiting for the planet to pass in front of the star and block some light. This is great for close planets, but for distant, cold giants (like Jupiter), they rarely pass in front of their stars from our viewpoint.

The Problem:
For a long time, finding "Jupiter analogues"—planets similar to our own Jupiter, orbiting far away from their stars—was like trying to solve a puzzle with half the pieces missing. We knew they were there, but we didn't know their true mass or their exact orbit.

The Solution: A New Detective Team
This paper introduces a new detective team: CHEPS (Chile-Hertfordshire ExoPlanet Survey). They have been watching 240 metal-rich stars for 16 years, collecting data on how those stars wobble. But recently, they realized they needed a second pair of eyes.

They teamed up with Gaia and Hipparcos, two massive space telescopes that act like a giant, high-precision camera taking pictures of the sky over 20+ years. While the wobble data tells the star is moving back and forth, the camera data shows the star moving side to side across the sky.

The "Aha!" Moment: Breaking the Mirror
Think of the wobble data as looking at a reflection in a mirror. You can see the movement, but you don't know if the object is moving left or right, or how deep it is in the room. This is called the "inclination degeneracy."

By combining the wobble (Radial Velocity) with the side-to-side movement (Astrometry), the astronomers finally broke the mirror. They could now see the 3D shape of the orbit. Suddenly, they could stop guessing the planet's weight and start knowing its true mass.

The Discoveries
Using this new "super-scope" method, the team looked at five specific stars and found:

5 New Planets: They confirmed the existence of five giant planets that had been suspected but not fully understood.
True Masses: They turned "minimum weights" into "real weights." For example, one planet was thought to be a heavy Saturn, but with the new data, they realized it's actually a full-sized Jupiter.
Solar System Twins: They found planets that orbit at distances very similar to Jupiter (about 5 times the distance from the Sun to Earth). This is huge because it helps us understand how common our own Solar System's layout really is.

Why This Matters
Imagine you are trying to build a map of all the planets in the universe. Until now, the map had a huge blank spot in the middle: the "Cold Giant" zone. We had too many close-in planets and too many far-out planets, but nothing in between with reliable data.

This paper fills that blank spot. It proves that by combining old-school telescope wobbles with modern space-camera movements, we can find and characterize these distant giants.

The Takeaway
The authors are essentially saying: "We used to be blindfolded, guessing the size of the elephant. Now, with this new combined method, we can see the elephant clearly, weigh it accurately, and know exactly where it's walking."

This is a major step forward in understanding how planetary systems form and whether our own Solar System is a rare gem or just one of many in the cosmic neighborhood. And the best part? This is just the beginning. As new data from the Gaia telescope arrives in the coming years, we will be able to find even more of these "cold giants" and finally complete the map of our galactic neighborhood.

1. Problem Statement

The population of long-period giant planets ("cold Jupiters," defined here as semi-major axes $a \approx 1\text{--}10$ AU and masses $0.3\text{--}13 M_J$ ) remains poorly characterized.

Radial Velocity (RV) Limitations: While RV surveys are effective at detecting planets, they suffer from sensitivity loss at multi-AU separations due to long orbital periods. Crucially, RVs only provide the minimum mass ( $M \sin i$ ) because the orbital inclination ( $i$ ) is unknown.
Transit Limitations: Transit surveys have extremely low geometric probability for detecting long-period planets.
The Gap: There is a demographic gap between the sensitivity limits of RV surveys and direct imaging. Furthermore, without true masses, it is difficult to distinguish between massive planets and brown dwarfs or to accurately model system architectures.
The Opportunity: The synergy between long-baseline RV data and absolute astrometry (from Hipparcos and Gaia) offers a pathway to break the inclination degeneracy, determine true masses, and validate long-period candidates.

2. Methodology

The authors utilized the EMPEROR framework (Exoplanet Mcmc Parallel tEmpering for Rv Orbit Retrieval), a modular Python-based Bayesian pipeline, to perform a joint analysis of RV and astrometric data.

Target Selection: The study focused on five metal-rich, inactive FGK stars from the Chile-Hertfordshire ExoPlanet Survey (CHEPS) that exhibited long-term RV curvature. The sample includes targets with RV baselines extending up to 16 years (2009–2025).
Data Sources:
- RVs: High-precision data from HARPS, CORALIE, UCLES, and HIRES. The dataset was segmented to account for instrumental offsets (e.g., HARPS fiber upgrades, CORALIE 2014 upgrade).
- Astrometry: The study employed a Hipparcos-Gaia differencing approach. Instead of relying solely on catalog-level proper motion anomalies, the framework:
  1. Fits the Hipparcos Intermediate Astrometric Data (IAD) directly.
  2. Projects the orbital model to the Gaia reference epoch using the Gaia Observation Forecast Tool (GOST) to match scan angles.
  3. Jointly fits the IAD and Gaia (DR2/DR3) position/proper motion differences.
Statistical Framework:
- Bayesian Inference: Used the redemcee sampler (adaptive parallel tempering MCMC) to explore multi-modal posteriors.
- Model Comparison: Employed Bayesian evidence ( $\ln \mathcal{Z}$ ) and Bayes' factors ( $\Delta \ln \mathcal{Z}$ ) to select the optimal number of planets (0K to 2K models).
- Priors: Used truncated normal priors for eccentricity to avoid spurious high- $e$ solutions and isotropic priors for inclination. For multi-planet systems, a quasi-coplanarity prior was applied.
Metrics: Introduced phase-coverage ( $\Phi_c$ ) and baseline-coverage ( $B_c$ ) metrics to quantify how astrometry improves the temporal and phase sampling of the orbit.

3. Key Contributions

Framework Validation: Validated the EMPEROR joint RV+astrometry framework by re-analyzing the well-known system HIP 21850, confirming consistency with previous studies while refining parameters.
New Discoveries: Identified and characterized five new planetary systems (six planets total) around metal-rich stars.
True Mass Determination: Successfully converted minimum masses ( $M \sin i$ ) into true dynamical masses ( $M$ ) for all detected planets by breaking the inclination degeneracy.
Quantitative Impact Analysis: Provided a rigorous quantification of how astrometry improves model evidence and reduces parameter uncertainties, demonstrating that even sparse astrometric data significantly boosts detection confidence.

4. Key Results

The study confirmed two known planets in HIP 21850 and discovered five new planets:

Host Star	Planet	Period ( $P$ )	True Mass ( $M$ )	Semi-major Axis ( $a$ )	Classification
HIP 21850	b	$2539 \pm 1$ d	$8.25^{+1.05}_{-0.85} M_J$	$3.49$ AU	Jupiter Analogue
	c	$11320^{+490}_{-940}$ d	$4.67^{+0.44}_{-0.43} M_J$	$9.90$ AU	Cold Giant
HIP 8923	b	$5160^{+150}_{-240}$ d	$9.98^{+0.78}_{-0.16} M_J$	$5.90$ AU	Jupiter Analogue
HIP 10090	b	$2960^{+120}_{-100}$ d	$3.87^{+0.65}_{-0.60} M_J$	$3.87$ AU	Jupiter Analogue
	c	$321.8^{+0.3}_{-0.6}$ d	$0.85^{+0.03}_{-0.12} M_J$	$0.90$ AU	Warm Jupiter
HIP 39330	b	$4650^{+210}_{-250}$ d	$1.67^{+0.19}_{-0.08} M_J$	$5.05$ AU	Jupiter Analogue
	c*	$257^{+160}_{-75}$ d	Unconstrained	$\sim 1$ AU	Warm Saturn Candidate
HIP 98599	b	$2656^{+40}_{-16}$ d	$6.85^{+0.10}_{-0.22} M_J$	$3.96$ AU	Jupiter Analogue

Note: HIP 39330c presents a multimodal solution where inclination could not be uniquely resolved with current data; the authors conservatively report the 1-planet model as the primary solution.

Performance Improvements:

Bayes' Factors: The inclusion of astrometry increased the Bayes' factor ( $\Delta \ln \mathcal{Z}$ ) by factors ranging from 1.7 to 58.6 compared to RV-only models.
Uncertainty Reduction: Period and mass uncertainties were reduced by factors of 3 to 10. For example, the period uncertainty for HIP 21850c dropped from $\sim 2500$ days to $\sim 500$ days.
Mass Conversion: The most significant result is the conversion of minimum masses to true masses. For instance, HIP 10090c was confirmed as a bona-fide Jupiter-mass planet ( $0.85 M_J$ ) rather than a Saturn-mass candidate ( $0.39 M_J \sin i$ ).

5. Significance and Future Outlook

Filling the Demographic Gap: The discovered planets occupy the critical transition zone between RV and direct imaging sensitivities (3–7 AU). This provides the first robust sample of "Jupiter analogues" with dynamically measured masses, essential for testing planet formation theories (e.g., core accretion vs. migration).
Solar System Analogues: The metal-rich nature of the host stars and the orbital architectures (multiple giants, some near-coplanar) suggest these systems are strong candidates for Solar System analogues.
Preparation for Gaia DR4/DR5: This work serves as a proof-of-concept for the upcoming era of routine astrometric vetting. The methodology demonstrated here (joint RV+astrometry) is directly applicable to future data releases, which will enable the measurement of true masses for thousands of long-period planets.
Follow-up Targets: These systems are prime targets for future direct imaging campaigns (e.g., with the Habitable Worlds Observatory or ELTs) and atmospheric characterization, as their true masses and orbital geometries are now known.

In conclusion, the paper demonstrates that the synergy of long-baseline RV monitoring and absolute astrometry is a robust, necessary pathway to fully characterize cold giant planets, transforming statistical detections into precise dynamical systems.

1. Problem Statement

2. Methodology

3. Key Contributions

4. Key Results

5. Significance and Future Outlook

More like this