Earth 2.0 could reveal shrinking orbits of three more hot Jupiters

A planet’s repeated crossings in front of its star can act like a celestial clock. If those crossings arrive progressively earlier, the planet’s orbit may be shrinking. A new simulation study explores how revisiting Kepler’s original observing field with the planned Earth 2.0 mission could make that slow drift easier to measure.

Looking for a clock that speeds up

Hot Jupiters are giant planets orbiting close to their stars. Their strong tidal interactions can dissipate energy inside the host star, potentially shrinking the planet’s orbit and shortening its orbital period. Measuring that change would help astronomers understand both stellar interiors and the long-term evolution of close-in planets.

The observational challenge is substantial. According to the preprint by Wenqin Wang and colleagues, WASP-12 b remains the only hot Jupiter with a secure orbital-decay detection, its period shortening by roughly 29 milliseconds per year. Whether the tidal dissipation inferred for that system is typical remains uncertain.

The researchers investigate a future opportunity rather than announce new detections: using the planned Earth 2.0 mission, or ET, to revisit planets first observed by Kepler.

A longer view of familiar planets

As described in the paper, ET’s transit instrument would use six 28-centimetre telescopes at the Sun–Earth L2 point. Its planned four-year primary mission, spanning 2029–2033, would include the original Kepler field. Combined with observations beginning in 2009, that would create a timing baseline of approximately 24 years.

The team selected 28 Kepler hot Jupiters and measured individual transit times from archival Kepler and TESS light curves. They excluded incomplete transits and signals below their adopted quality threshold. Kepler-91 b was left out because its red-giant host produces stellar variability and residual noise that make reliable timing difficult.

They then generated four years of simulated ET light curves, normally sampled every two minutes, using the mission’s expected photometric performance. These synthetic observations included assumed orbital decay and were analysed with the same transit-fitting procedure as the archival data.

The distinction matters: Kepler and TESS supplied real measurements; the ET measurements and their decay signals were simulated.

What the simulations predict

The researchers compared a constant-period model with one allowing a steady change in orbital period. Their candidate criteria required a period change at least three times its uncertainty, together with a strong preference for the changing-period model under the Bayesian information criterion, which penalizes extra model complexity.

Among the 23 systems with converged Kepler-plus-TESS solutions used for the precision comparison, adding simulated ET observations reduced the median uncertainty in the period-change rate from about 7.2 to 1.0 milliseconds per year. That is the reported factor-of-7.2 improvement—not a uniform gain for every planet.

Three systems stood out as potentially detectable: Kepler-13 Ab, Kepler-17 b and Kepler-76 b. Their input decay rates were approximately 7.44, 1.09 and 2.41 milliseconds per year, respectively, all representing shortening periods.

For Kepler-17 b, the combined real-and-simulated dataset recovered a shortening of 1.01 ± 0.13 milliseconds per year and strongly favoured the changing-period model. This demonstrates the forecast sensitivity under the simulation assumptions; it is not a measurement from future ET observations.

The assumptions behind three targets

The forecast depends critically on the stellar tidal quality factor, written Q′. This effective parameter describes how efficiently a star dissipates tidal energy: smaller values imply stronger dissipation and faster predicted decay in the model used here.

The simulations adopted a central value corresponding to log₁₀ Q′ = 5.7, with a literature-estimated uncertainty of +0.4 and −0.3. That estimate comes from previous work on changes in hot-Jupiter occurrence rates. It does not establish the dissipation efficiency of each individual host star. The paper notes that tidal behaviour depends on stellar structure, rotation and forcing frequency, so the number of detectable planets remains conditional.

There is also an interpretation problem. An additional companion can change the distance light travels from a system, producing apparent transit-timing drifts without tidal decay. The paper highlights this ambiguity, but its reported model comparison tests constant periods against steadily changing periods—not explicit companion-induced timing models. A significant drift would therefore need further assessment before being called secure tidal decay.

More years matter more than faster sampling

For four-year ET simulations, changing the sampling interval from one to ten minutes left the median period-change uncertainty near 1.0 millisecond per year. Extending ET observations to eight years at two-minute sampling reduced it to about 0.4 milliseconds per year, an improvement of roughly 2.4 relative to the four-year case.

The explanation is that a steady period change produces a timing offset that grows approximately with elapsed time squared. For these targets and assumed instrument performance, additional years offer more leverage than faster sampling. The study therefore supports longer monitoring or a later revisit as promising strategies, while leaving actual detections dependent on future observations and the stars’ true tidal behaviour.