CS20 Splinter session

From light curves to rotation periods, and then ages via gyrochronology

- where do we stand, and where do we go from here?

July 30, 2018



This splinter will likely have 9 – 10 speakers, selected to represent the full diversity of the topic (more below), with each speaker permitted a 11 – 12 min presentation, including immediate questions. There will also be a 45 min segment after the talks, reserved for additional audience comments and questions, enabling more in-depth discussions.


Themes and topics:

A detailed set of themes and topics is given below, to enable speakers to submit responsive abstracts. (Prospective speakers should feel free to propose relevant topics beyond these.)

A. What are the prospects for establishing age-rotation relations for M dwarfs?

This topic is largely driven by efforts to detect earth mass planets around M dwarfs, with the consequent need to determine their ages. A relevant wrinkle is that their intrinsic faintness makes them improbable seismic age targets, and chromospheric ages are also not hugely promising. Fortunately, there are emerging rotation period measurements that could point a way towards a gyrochronology for M dwarfs.

B. Can age-rotation relations be extended to the age of the Sun and beyond?

How can we obtain stable observations (both from space and from the ground) over long-enough baselines to enable derivation of long periods, or periods for stars with low activity? Is chromospheric monitoring possible for large-enough samples? What are the prospects for long-term (over years) photometric monitoring? A related topic is that of the Sun, our archetype for the stars, which provides insights into successful means to query older stars for their rotation periods.

C. Is binarity/exoplanet architecture a problem, and if so, under what circumstances?

Tidally-locked systems are unsuitable for gyrochronology. Conversely, wide binaries are effectively single, and therefore usable. But the applicability of gyrochronology to the region in-between (which contains many systems of interest) remains in question. Similar questions pertain to certain exoplanet architectures with giant planets. We also envision a discussion about relevant efforts on wide binary systems.

D. Does metallicity affect the rotational evolution of a star?

The applicability of gyrochronology to systems with metallicity far from solar has been questioned. Because open clusters with measured rotation periods all have near-solar metallicities, there is little empirical leverage to address this issue. Yet, periods are indeed routinely measured for non-solar metallicity field dwarfs. Both observational and theoretical studies could be relevant.

E. Additional topics (Comparision with other age indicators, supporting data, etc.)

A wide range of subjects could potentially fit in this category. Examples include comparisons against other age indicators and supporting data.


Speakers (alphabetical):

Eliana Amazo-Gomez 5+2

From the Sun to other CoolStars: Accurate rotational periods from GPS method

Reach a clear idea of the physics involved on the activity behaviour of late- type stars is not an easy task. Even when such stars are analogs to the most analysed star, our Sun. There are many variables and degeneracies working simultaneously. Nevertheless, if there is a keystone parameter that help us to maximise the understanding of stellar dynamo action, magnetic fields, chem- ical abundances, age, internal structure, inclination, etc., is the knowledge of the stellar rotational period. High-quality photometric data acquired from Ke- pler mission, high-stability and high-accuracy measurements of the solar total irradiance, TSI, from SORCE/TIM and SOHO/VIRGO missions, and detailed models of solar brightness variations allow better insights into the variability and activity of Sun-like stars. Based on the Gradient of the Power Spectra, GPS, of time series photometric data, we developed a new method to retrieve accurate values of stellar rotational periods. We apply GPS method to the Sun and to 471 Sun-like stars observed by Kepler, calculate their rotation periods and compare our results to earlier works based on autocorrelation functions, Lomb-Scargle periodograms, and wavelet maximum peaks. GPS gives results consistent the methods in comparison for active stars and allows, for first time, the rotation periods of 705 less variable Sun-like Kepler stars to be determined. We present our results and a detailed characterisation of solar analogs in terms of its variability and rotation period.

Paul Beck 10+2

Is there ‘one’ age for a star?

Stellar ages are a difficult topic of stellar astrophysics. It cannot be measured but only inferred from observational quantities that relate to age. These are for example the fundamental spectroscopic parameters, its rotation rate, stellar activity and specific tracer elements as chemical clocks. The unparalleled photometric from the NASA space telescope Kepler opened asteroeismology as a tool to investigate ages of stars. In this talk, we discuss these different classes of age indicators on the example of 18 solar-like analogues, well characterized through asteroseismology and high-resolution ground-based spectroscopy. As an outlook we will discuss the prospects of the ongoing missions GAIA and TESS on stellar ages and solar analogues.

Stephanie Douglas 10+2

The impact of companions on stellar rotation

Observations of open clusters show that among stars of a given age, the rapid rotators tend to be binaries. This could be due to tidal effects acting over 0.01 – 10 billion years or interactions between the binary companion and the protoplanetary disk during the first 10 million years. Both effects would cause binary stars to rotate faster than single stars, thus altering angular momentum evolution in binary systems. I will discuss recent results that indicate a dependence of rotation period and planet occurrence rate on binary orbital separation.

Rebecca Esselstein 5+2

Understanding the Limits of K2 Data in Acquiring Rotation Periods for M67

Determining the rotation periods of solar-age cool stars is a difficult task. Due to intrinsically low amplitudes of variability and expected periods of around 25 days and longer, one must acquire high-precision photometry of each target coupled with an observation window of sufficient duration. The latter is particularly important for discerning between the true period and one of its harmonics, or even systematic-induced variation that may appear astrophysical at a glance.Simulating realistic star-spot signals across the equivalent of two K2 campaigns, we discuss a couple methods that utilize Gaussian processes (GPs) to back out rotation periods, uncertainties on the period, and a physically motivated quality factor that expresses whether the modeled signal is actually periodic. The first of these methods integrates the autocorrelation function and normalized Lomb-Scargle periodogram of Esselstein et al. 2018 with a GP, while the second involves the construction of a GP ‘periodogram,’ though this approach is still very much in its infancy. We show the preliminary results of applying the former method to a sample of K2 Campaign 5 and 16 light curves.

Kenneth Janes 10+2

Gyrochronology of wide binaries in the Kepler field

In a recent paper (Janes 2017), I identified almost 100 pairs of likely common proper motion (CPM) stars in the Kepler spacecraft field of view. My goal was to extend the gyrochronology relation known for solar-type stars to stars of  later spectral type by matching the earlier-type member of each pair with its later-type companion. My identification of two stars as physically-related, coeval objects was based on data from existing proper motion and photometry  catalogs, but a major limitation of my analysis was the uncertainty in the physical connection between the two presumed CPM components. Now, the precise parallaxes and proper motions in Gaia DR2 have made it possible to verify the link between components as well as to add additional CPM systems. About 2/3 of the original CPM candidates are confirmed as likely CPM pairs by the Gaia parallaxes and proper motions, and in a new search using the Gaia data, I found over 200 wide binary systems.

Alessandro Lanzafame 10+2

The Gaia DR2 rotational modulation amplitude—period diagram: a new paradigm for the rotational and magnetic evolution of cool stars

The Gaia second data release includes rotational modulation parameters for some 140,000 late-type stars close to the main sequence. The richness of the sample, the range of periods, and the photometric precision make it possible to unveil, for the first time, the existence of three distinct populations in the amplitude-period diagram. The data show an amplitude bimodality with an evident gap at periods less than 2 days. The low-amplitude branch, in turn, shows a period bimodality, with a minimum density at periods of approximately 1 day. The segregation in the amplitude-period density diagram of these high-amplitude, low-amplitude fast-rotator, and low-amplitude slow-rotator populations hints at rather rapid transitions from one population to another in the course of the early stellar rotational evolution. The transition from high-amplitude to ultra-fast-rotator (which are found to have low amplitude) is very rapid. The high-amplitude population is more prominent at lower mass, and it is gradually depleted moving towards higher masses, until it almost disappears in the 1.15 – 1.25 M⊙ range. Conversely, the low-amplitude fast-rotator population is very poorly populated for M<0.85M⊙, and increasingly populated at higher masses. The existence of a low-amplitude fast-rotator population indicates that the detectability of rotational modulation in a significant fraction of young stars close to the ZAMS requires high photometric precision. We argue that this new observational evidence changes dramatically our view on the evolution of magnetic surface inhomogeneities of cool stars in the first 600 Myr, with magnetic transitions playing a bigger role than previously thought. They also provide an empirical criterion for identifying young field stars to which gyrochronological dating can be applied.

Elizabeth Newton 10+2

An abundance of slowly spinning M dwarfs: insights from ground-based measurements of their rotation periods

Development of an age-rotation (“gyrochronoloy”) relationship for M dwarfs has been inhibited by two factors. First, there has been a paucity of rotation period measurements for the lowest-mass stars. Second, much of the rotational evolution of very low-mass stars occurs at field ages, when precise ages for M dwarfs are challenging to determine. We use  ground-based photometric data to measure new rotation periods for field-aged M dwarfs. Our high-cadence, long-baseline photometry derives from the MEarth Transiting Planet Survey. We present 241 rotation periods for M dwarfs in the Southern hemisphere, which join the 347 periods that we previously measured in the Northern hemisphere. The overall period distribution shows an increase in the longest periods at lower stellar masses and a persistent gap in the distribution at intermediate rotation periods.  We will discuss the ages at which gyrochronology may be applicable to M dwarfs, and prospects for obtaining new measurements of rotation and age in M dwarfs.

Stefanie Raetz 5+2

Exploring the Rotation-Activity-Age Relation of M dwarfs

Measuring stellar rotation periods at different ages is essential to enhance our understanding of angular momentum evolution especially for M dwarfs where spin-down is slow and poorly understood. However, ages of M dwarfs with measurable rotation period are not known. We use magnetic activity as link. By calibrating the (1) activity-rotation relation and the (2) activity-age relation separately, the rotation-age relation will be known. Here I present results from both studies:

(1) We study the rotation-activity relation with the Kepler Two-Wheel (K2) mission for a sample of bright and nearby M dwarfs from the Superblink proper motion catalog by Lepine and Gaidos (2011). Photometric observations with space telescopes provide rotation periods even with low amplitudes as well as a wealth of activity diagnostics. We find a sharp drop of activity at a critical rotation period of ~10d. This phenomenon represents an open problem within the framework of dynamo theory.

(2) The age-activity relation is investigated through deep X-ray observations of a sample of M dwarfs in wide binaries with white dwarfs. Here the white dwarf serves as a chronometer for the age of the M dwarf allowing us to examine the evolution of activity for ages between 1-10 Gyrs.

Jürgen Schmitt 10+2

Rotation measurements of late-type stars in Ca II

Chromospheric plage regions produce emissions in the cores of the Ca II H&K lines and in the Ca II IRT lines. In the case of the Sun, a rather inactive star in a stellar context, such plage regions attain filling factors much larger than those of sun spots. Given an inhomogeneous surface distribution of plage and a sufficient life time of “active regions” on the stellar surface, rotation periods of inactive solar-like stars can actually be measured. Since one expects rotation periods on the order of 30-40 days or more, such measurements require quasi-regular monitoring over extended periods of time, which are best carried out with robotic telescopes. Measurements of rotation periods of solar-like stars is one of the key projects we are pursuing with our robotic spectroscopic TIGRE telescope. I will discuss the prospects and challenges of this project and present first results.

Jörg Weingrill 5+2

A primer on reliable rotation period determination

I describe various mathematical and statistical methods to derive the correct rotation period of a star based on a light curve acquired by the Kepler K2 mission.

The Fourier transform, Lomb-Scargle and power spectral density produce similar results belonging to the same family of trigonometrical analysis tools. Other statistical methods like Gregory-Loredo, string-length minimization and phase dispersion minimization are more suitable for the period determination of erratic light curves that are dominated by one or more spot groups. As a detailed example given in this talk, I compare the EPIC 211394185 from Esselstein et al. (2018) with the results obtained in Barnes et al. (2016). By just visual examination of the light curve it becomes evident, that a single method e.g. Lomb-Scargle is not sufficient to derive the correct period.



Starts at 2pm, sharp

L stands for long, S for short

00 Introduction


01 Amazo-Gomez (S)

02 Esselstein (S)

03 Weingrill (S)

04 Douglas (L)

05 Beck (L)

06 Raetz (S)

07 Newton (L)


Coffee (30 min)


08 Schmitt (L)

09 Lanzafame (L)

10 Janes (L)


Discussion (45 minutes)

Invited speaker Elisabeth Newton

Splinter session speakers

Invited speaker Kenneth Janes (left) with splinter session convenor Sydney Barnes (right).