arXiv subjects

Daniel Foreman-Mackey

Publications and source records attributed to Daniel Foreman-Mackey.

At least 55 records · Page 3Linked to original sources

arXiv2021

PyTorchDIA: A flexible, GPU-accelerated numerical approach to Difference Image Analysis

We present a GPU-accelerated numerical approach for fast kernel and differential background solutions. The model image proposed in the Bramich (2008) difference image analysis algorithm is analogous to a very simple Convolutional Neural Network (CNN), with a single convolutional filter (i.e. the kernel) and an added scalar bias (i.e. the differential background). Here, we do not solve for the discrete pixel array in the classical, analytical linear least-squares sense. Instead, by making use of PyTorch tensors (GPU compatible multi-dimensional matrices) and associated deep learning tools, we solve for the kernel via an inherently massively parallel optimisation. By casting the Difference Image Analysis (DIA) problem as a GPU-accelerated optimisation which utilises automatic differentiation tools, our algorithm is both flexible to the choice of scalar objective function, and can perform DIA on astronomical data sets at least an order of magnitude faster than its classical analogue. More generally, we demonstrate that tools developed for machine learning can be used to address generic data analysis and modelling problems.

astro-ph.IM↗

arXiv2021

Warm Jupiters in TESS Full-Frame Images: A Catalog and Observed Eccentricity Distribution for Year 1

Warm Jupiters -- defined here as planets larger than 6 Earth radii with orbital periods of 8--200 days -- are a key missing piece in our understanding of how planetary systems form and evolve. It is currently debated whether Warm Jupiters form in situ, undergo disk or high eccentricity tidal migration, or have a mixture of origin channels. These different classes of origin channels lead to different expectations for Warm Jupiters' properties, which are currently difficult to evaluate due to the small sample size. We take advantage of the \TESS survey and systematically search for Warm Jupiter candidates around main-sequence host stars brighter than the \TESS-band magnitude of 12 in the Full-Frame Images in Year 1 of the \TESS Prime Mission data. We introduce a catalog of 55 Warm Jupiter candidates, including 19 candidates that were not originally released as \TESS Objects of Interest (TOIs) by the \TESS team. We fit their \TESS light curves, characterize their eccentricities and transit-timing variations (TTVs), and prioritize a list for ground-based follow-up and \TESS Extended Mission observations. Using hierarchical Bayesian modeling, we find the preliminary eccentricity distributions of our Warm-Jupiter-candidate catalog using a Beta distribution, a Rayleigh distribution, and a two-component Gaussian distribution as the functional forms of the eccentricity distribution. Additional follow-up observations will be required to clean the sample of false positives for a full statistical study, derive the orbital solutions to break the eccentricity degeneracy, and provide mass measurements.

astro-ph.EP↗

arXiv2021

Stellar Rotation in the K2 Sample: Evidence for Modified Spindown

We analyze light curves of 284,834 unique K2 targets using a Gaussian process model with a quasi-periodic kernel function. By crossmatching K2 stars to observations from Gaia Data Release 2, we have identified 69,627 likely main-sequence stars. From these we select a subsample of 8,977 stars on the main-sequence with highly precise rotation period measurements. With this sample we recover the gap in the rotation period-color diagram first reported by \cite{McQuillan2013}. While the gap was tentatively detected in \cite{Reinhold2020}, this work represents the first robust detection of the gap in K2 data for field stars. This is significant because K2 observed along many lines of sight at wide angular separation, in contrast to Kepler's single line of sight. Together with recent results for rotation in open clusters, we interpret this gap as evidence for a departure from the $t^{-1/2}$ Skumanich spin down law, rather than an indication of a bimodal star formation history. We provide maximum likelihood estimates and uncertainties for all parameters of the quasi-periodic light curve model for each of the 284,834 stars in our sample.

astro-ph.SR↗

arXiv2021

Evolution of the Exoplanet Size Distribution: Forming Large Super-Earths Over Billions of Years

The radius valley, a bifurcation in the size distribution of small, close-in exoplanets, is hypothesized to be a signature of planetary atmospheric loss. Such an evolutionary phenomenon should depend on the age of the star-planet system. In this work, we study the temporal evolution of the radius valley using two independent determinations of host star ages among the California-Kepler Survey (CKS) sample. We find evidence for a wide and nearly empty void of planets in the period-radius diagram at the youngest system ages ($\lesssim$2-3 Gyr) represented in the CKS sample. We show that the orbital period dependence of the radius valley among the younger CKS planets is consistent with that found among those planets with asteroseismically determined host star radii. Relative to previous studies of preferentially older planets, the radius valley determined among the younger planetary sample is shifted to smaller radii. This result is compatible with an atmospheric loss timescale on the order of gigayears for progenitors of the largest observed super-Earths. In support of this interpretation, we show that the planet sizes which appear to be unrepresented at ages $\lesssim$2-3 Gyr are likely to correspond to planets with rocky compositions. Our results suggest the size distribution of close-in exoplanets, and the precise location of the radius valley, evolves over gigayears.

astro-ph.EP↗

arXiv2021

Analytic Light Curves in Reflected Light: Phase Curves, Occultations, and Non-Lambertian Scattering for Spherical Planets and Moons

We derive efficient, closed form, differentiable, and numerically stable solutions for the flux measured from a spherical planet or moon seen in reflected light, either in or out of occultation. Our expressions apply to the computation of scattered light phase curves of exoplanets, secondary eclipse light curves in the optical, or future measurements of planet-moon and planet-planet occultations, as well as to photometry of solar system bodies. We derive our solutions for Lambertian bodies illuminated by a point source, but extend them to model illumination sources of finite angular size and rough surfaces with phase-dependent scattering. Our algorithm is implemented in Python within the open-source starry mapping framework and is designed with efficient gradient-based inference in mind. The algorithm is 4-5 orders of magnitude faster than direct numerical evaluation methods and about 10 orders of magnitude more precise. We show how the techniques developed here may one day lead to the construction of two-dimensional maps of terrestrial planet surfaces, potentially enabling the detection of continents and oceans on exoplanets in the habitable zone.

astro-ph.EP↗

arXiv2021

Occultation mapping of Io's surface in the near-infrared I: Inferring static maps

Jupiter's moon Io is the most volcanically active body in the Solar System with hundreds of active volcanoes varying in intensity on different timescales. Io has been observed during occultations by other Galilean moons and Jupiter since the 1980s, using high-cadence near infrared photometry. These observations encode a wealth of information about the volcanic features on its surface. We built a generative model for the observed occultations using the code starry which enables fast, analytic, and differentiable computation of occultation light curves in emitted and reflected light. Our probabilistic Bayesian model is able to recover known hotspots on the surface of Io using only two light curves and without any assumptions on the locations, shapes or the number of spots. The methods we have developed are also directly applicable to the problem of mapping the surfaces of stars and exoplanets.

astro-ph.EP↗

arXiv2021

starry_process: Interpretable Gaussian processes for stellar light curves

In this note we present the starry_process code, which implements an interpretable Gaussian process (GP) for modeling variability in stellar light curves. As dark starspots rotate in and out of view, the total flux received from a distant star will change over time. Unresolved flux time series therefore encode information about the spatial structure of features on the stellar surface. The starry_process software package allows one to easily model the flux variability due to starspots, whether one is interested in understanding the properties of these spots or marginalizing over the stellar variability when it is treated as a nuisance signal. The main difference between the GP implemented here and typical GPs used to model stellar variability is the explicit dependence of our GP on physical properties of the star, such as its period, inclination, and limb darkening coefficients, and on properties of the spots, such as their radius and latitude distributions. This code is the Python implementation of the interpretable GP algorithm developed in Luger, Foreman-Mackey, and Hedges (2021).

astro-ph.SR↗

arXiv2021

Mapping stellar surfaces I: Degeneracies in the rotational light curve problem

Thanks to missions like Kepler and TESS, we now have access to tens of thousands of high precision, fast cadence, and long baseline stellar photometric observations. In principle, these light curves encode a vast amount of information about stellar variability and, in particular, about the distribution of starspots and other features on their surfaces. Unfortunately, the problem of inferring stellar surface properties from a rotational light curve is famously ill-posed, as it often does not admit a unique solution. Inference about the number, size, contrast, and location of spots can therefore depend very strongly on the assumptions of the model, the regularization scheme, or the prior. The goal of this paper is twofold: (1) to explore the various degeneracies affecting the stellar light curve "inversion" problem and their effect on what can and cannot be learned from a stellar surface given unresolved photometric measurements; and (2) to motivate ensemble analyses of the light curves of many stars at once as a powerful data-driven alternative to common priors adopted in the literature. We further derive novel results on the dependence of the null space on stellar inclination and limb darkening and show that single-band photometric measurements cannot uniquely constrain quantities like the total spot coverage without the use of strong priors. This is the first in a series of papers devoted to the development of novel algorithms and tools for the analysis of stellar light curves and spectral time series, with the explicit goal of enabling statistically robust inference about their surface properties.

astro-ph.SR↗

arXiv2021

Mapping stellar surfaces II: An interpretable Gaussian process model for light curves

The use of Gaussian processes (GPs) as models for astronomical time series datasets has recently become almost ubiquitous, given their ease of use and flexibility. GPs excel in particular at marginalization over the stellar signal in cases where the variability due to starspots rotating in and out of view is treated as a nuisance, such as in exoplanet transit modeling. However, these effective models are less useful in cases where the starspot signal is of primary interest since it is not obvious how the parameters of the GP model are related to the physical properties of interest, such as the size, contrast, and latitudinal distribution of the spots. Instead, it is common practice to explicitly model the effect of individual starspots on the light curve and attempt to infer their properties via optimization or posterior inference. Unfortunately, this process is degenerate, ill-posed, and often computationally intractable when applied to stars with more than a few spots and/or to ensembles of many light curves. In this paper, we derive a closed-form expression for the mean and covariance of a Gaussian process model that describes the light curve of a rotating, evolving stellar surface conditioned on a given distribution of starspot sizes, contrasts, and latitudes. We demonstrate that this model is correctly calibrated, allowing one to robustly infer physical parameters of interest from one or more stellar light curves, including the typical radii and the mean and variance of the latitude distribution of starspots. Our GP has far-ranging implications for understanding the variability and magnetic activity of stars from both light curves and radial velocity (RV) measurements, as well as for robustly modeling correlated noise in both transiting and RV exoplanet searches. Our implementation is efficient, user-friendly, and open source, available as the Python package starry-process.

astro-ph.SR↗

arXiv2021

Multi-Wavelength Photometry Derived from Monochromatic Kepler Data

The Kepler mission has provided a wealth of data, revealing new insights in time-domain astronomy. However, Kepler's single band-pass has limited studies to a single wavelength. In this work we build a data-driven, pixel-level model for the Pixel Response Function (PRF) of Kepler targets, modeling the image data from the spacecraft. Our model is sufficiently flexible to capture known detector effects, such as non-linearity, intra-pixel sensitivity variations, and focus change. In theory, the shape of the Kepler PRF should also be weakly wavelength dependent, due to optical chromatic aberration and wavelength dependent detector response functions. We are able to identify these predicted shape changes to the PRF using the residuals between Kepler data and our model. In this work, we show that these PRF changes correspond to wavelength variability in Kepler targets using a small sample of eclipsing binaries. Using our model, we demonstrate that pixel-level light curves of eclipsing binaries show variable eclipse depths, ellipsoidal modulation and limb darkening. These changes at the pixel level are consistent with multi-wavelength photometry. Our work suggests each pixel in the Kepler data of a single target has a different effective wavelength, ranging from $\approx$ 550-750 $nm$. In this proof of concept, we demonstrate our model, and discuss possible use cases for the wavelength dependent Pixel Response Function of Kepler. These use cases include characterizing variable systems, and vetting exoplanet discoveries at the pixel level. The chromatic PRF of Kepler is due to weak wavelength dependence in the optical systems and detector of the telescope, and similar chromatic PRFs are expected in other similar telescopes, notably the NASA TESS telescope.

astro-ph.IM↗

arXiv2021

Refining the transit timing and photometric analysis of TRAPPIST-1: Masses, radii, densities, dynamics, and ephemerides

We have collected transit times for the TRAPPIST-1 system with the Spitzer Space Telescope over four years. We add to these ground-based, HST and K2 transit time measurements, and revisit an N-body dynamical analysis of the seven-planet system using our complete set of times from which we refine the mass ratios of the planets to the star. We next carry out a photodynamical analysis of the Spitzer light curves to derive the density of the host star and the planet densities. We find that all seven planets' densities may be described with a single rocky mass-radius relation which is depleted in iron relative to Earth, with Fe 21 wt% versus 32 wt% for Earth, and otherwise Earth-like in composition. Alternatively, the planets may have an Earth-like composition, but enhanced in light elements, such as a surface water layer or a core-free structure with oxidized iron in the mantle. We measure planet masses to a precision of 3-5%, equivalent to a radial-velocity (RV) precision of 2.5 cm/sec, or two orders of magnitude more precise than current RV capabilities. We find the eccentricities of the planets are very small; the orbits are extremely coplanar; and the system is stable on 10 Myr timescales. We find evidence of infrequent timing outliers which we cannot explain with an eighth planet; we instead account for the outliers using a robust likelihood function. We forecast JWST timing observations, and speculate on possible implications of the planet densities for the formation, migration and evolution of the planet system.

astro-ph.EP↗

arXiv2020

Stability Constrained Characterization of Multiplanet Systems

Many discovered multiplanet systems are tightly packed. This implies that wide parameter ranges in masses and orbital elements can be dynamically unstable and ruled out. We present a case study of Kepler-23, a compact three-planet system where constraints from stability, transit timing variations (TTVs), and transit durations can be directly compared. We find that in this tightly packed system, stability can place upper limits on the masses and orbital eccentricities of the bodies that are comparable to or tighter than current state of the art methods. Specifically, stability places 68% upper limits on the orbital eccentricities of 0.09, 0.04, and 0.05 for planets $b$, $c$ and $d$, respectively. These constraints correspond to radial velocity signals $\lesssim 20$ cm/s, are significantly tighter to those from transit durations, and comparable to those from TTVs. Stability also yields 68% upper limits on the masses of planets $b$, $c$ and $d$ of 2.2, 16.1, and 5.8 $M_\oplus$, respectively, which were competitive with TTV constraints for the inner and outer planets. Performing this stability constrained characterization is computationally expensive with N-body integrations. We show that SPOCK, the Stability of Planetary Orbital Configurations Klassifier, is able to faithfully approximate the N-body results over 4000 times faster. We argue that such stability constrained characterization of compact systems is a challenging "needle-in-a-haystack" problem (requiring removal of 2500 unstable configurations for every stable one for our adopted priors) and we offer several practical recommendations for such stability analyses.

astro-ph.EP↗

arXiv2020

A Fast, 2D Gaussian Process Method Based on Celerite: Applications to Transiting Exoplanet Discovery and Characterization

Gaussian processes (GPs) are commonly used as a model of stochastic variability in astrophysical time series. In particular, GPs are frequently employed to account for correlated stellar variability in planetary transit light curves. The efficient application of GPs to light curves containing thousands to tens of thousands of datapoints has been made possible by recent advances in GP methods, including the celerite method. Here we present an extension of the celerite method to two input dimensions, where, typically, the second dimension is small. This method scales linearly with the total number of datapoints when the noise in each large dimension is proportional to the same celerite kernel and only the amplitude of the correlated noise varies in the second dimension. We demonstrate the application of this method to the problem of measuring precise transit parameters from multiwavelength light curves and show that it has the potential to improve transit parameters measurements by orders of magnitude. Applications of this method include transit spectroscopy and exomoon detection, as well a broader set of astronomical problems.

astro-ph.IM↗

arXiv2020

The Assembly History of M87 Through Radial Variations in Chemical Abundances of its Field Star And Globular Cluster Populations

We present an extensive study of spectroscopically-derived chemical abundances for M87 and its globular cluster (GC) system. Using observations from the Mitchell spectrograph at McDonald, LRIS at Keck, and Hectospec on the MMT, we derive new metallicity gradients from $\sim 2$ to $140$ kpc. We use a novel hierarchical statistical framework to simultaneously separate the GC system into subpopulations while measuring the metallicity gradients of those subpopulations. We create physically-motivated spectral stacks of the GC subpopulations by leveraging the output of this statistical framework to perform the first application of abundance tagging in a massive ETG to better constrain the origins of the GC subpopulations and, thus, the assembly history of M87. We find a metal-poor, $α$-enhanced population of GCs in both in the inner and outer halo unanticipated by current cosmological simulations of galaxy evolution. We use the remarkably flat metallicity gradients we find for both the metal-rich and metal-poor GC subpopulations in the inner halo as tentative evidence that some amount of the metal-poor GCs formed directly in the halo of M87 at high redshift.

astro-ph.GA↗

arXiv2020

A planet within the debris disk around the pre-main-sequence star AU Microscopii

AU Microscopii (AU Mic) is the second closest pre main sequence star, at a distance of 9.79 parsecs and with an age of 22 million years. AU Mic possesses a relatively rare and spatially resolved3 edge-on debris disk extending from about 35 to 210 astronomical units from the star, and with clumps exhibiting non-Keplerian motion. Detection of newly formed planets around such a star is challenged by the presence of spots, plage, flares and other manifestations of magnetic activity on the star. Here we report observations of a planet transiting AU Mic. The transiting planet, AU Mic b, has an orbital period of 8.46 days, an orbital distance of 0.07 astronomical units, a radius of 0.4 Jupiter radii, and a mass of less than 0.18 Jupiter masses at 3 sigma confidence. Our observations of a planet co-existing with a debris disk offer the opportunity to test the predictions of current models of planet formation and evolution.

astro-ph.EP↗

arXiv2020

Exploring the evolution of stellar rotation using Galactic kinematics

The rotational evolution of cool dwarfs is poorly constrained after around 1-2 Gyr due to a lack of precise ages and rotation periods for old main-sequence stars. In this work we use velocity dispersion as an age proxy to reveal the temperature-dependent rotational evolution of low-mass Kepler dwarfs, and demonstrate that kinematic ages could be a useful tool for calibrating gyrochronology in the future. We find that a linear gyrochronology model, calibrated to fit the period-Teff relationship of the Praesepe cluster, does not apply to stars older than around 1 Gyr. Although late-K dwarfs spin more slowly than early-K dwarfs when they are young, at old ages we find that late-K dwarfs rotate at the same rate or faster than early-K dwarfs of the same age. This result agrees qualitatively with semi-empirical models that vary the rate of surface-to-core angular momentum transport as a function of time and mass. It also aligns with recent observations of stars in the NGC 6811 cluster, which indicate that the surface rotation rates of K dwarfs go through an epoch of inhibited evolution. We find that the oldest Kepler stars with measured rotation periods are late-K and early-M dwarfs, indicating that these stars maintain spotted surfaces and stay magnetically active longer than more massive stars. Finally, based on their kinematics, we confirm that many rapidly rotating GKM dwarfs are likely to be synchronized binaries.

astro-ph.SR↗

arXiv2020

Forward modeling the orbits of companions to pulsating stars from their light travel time variations

Mutual gravitation between a pulsating star and an orbital companion leads to a time-dependent variation in path length for starlight traveling to Earth. These variations can be used for coherently pulsating stars, such as the δ Scuti variables, to constrain the masses and orbits of their companions. Observing these variations for δ Scuti stars has previously relied on subdividing the light curve and measuring the average pulsation phase in equally sized subdivisions, which leads to under-sampling near periapsis. We introduce a new approach that simultaneously forward-models each sample in the light curve and show that this method improves upon current sensitivity limits - especially in the case of highly eccentric and short-period binaries. We find that this approach is sensitive enough to observe Jupiter mass planets around δ Scuti stars under ideal conditions, and use gravity-mode pulsations in the subdwarf B star KIC 7668647 to detect its companion without radial velocity data. We further provide robust detection limits as a function of the SNR of the pulsation mode and determine that the minimum detectable light travel time amplitude for a typical Kepler δ Scuti is around 2 s. This new method significantly enhances the application of light travel time variations to detecting short period binaries with pulsating components, and pulsating A-type exoplanet host stars, especially as a tool for eliminating false positives.

astro-ph.SR↗

arXiv2019

NGTS clusters survey. I. Rotation in the young benchmark open cluster Blanco 1

We determine rotation periods for 127 stars in the ~115 Myr old Blanco 1 open cluster using ~200 days of photometric monitoring with the Next Generation Transit Survey (NGTS). These stars span F5-M3 spectral types (1.2 $\gtrsim M \gtrsim$ 0.3 M$_{\odot}$) and increase the number of known rotation periods in Blanco 1 by a factor of four. We determine rotation periods using three methods: Gaussian process (GP) regression, generalised autocorrelation (G-ACF) and Lomb-Scargle (LS) periodograms, and find that GPs and G-ACF are more applicable to evolving spot modulation patterns. Between mid-F and mid-K spectral types, single stars follow a well-defined rotation sequence from ~2 to 10 days, whereas stars in photometric multiple systems typically rotate faster. This may suggest that the presence of a moderate-to-high mass ratio companion inhibits angular momentum loss mechanisms during the early pre-main sequence, and this signature has not been erased at ~100 Myr. The majority of mid-F to mid-K stars display evolving modulation patterns, whereas most M stars show stable modulation signals. This morphological change coincides with the shift from a well-defined rotation sequence (mid-F to mid-K stars) to a broad rotation period distribution (late-K and M stars). Finally, we compare our rotation results for Blanco 1 to the similarly-aged Pleiades: the single star populations in both clusters possess consistent rotation period distributions, which suggests that the angular momentum evolution of stars follows a well-defined pathway that is, at least for mid-F to mid-K stars, strongly imprinted by ~100 Myr.

astro-ph.SR↗