BayesComp talk

Describing my research exoplanet populations across time and space to a statistics audience. Warning: this talk contained a lot of videos so this is an incomplete representation!

Transcript

1. Planetary systems across space and time Ruth Angus Simons Fellow, Columbia University @ruthangus
2. Image credit: NASA Kepler
3. Image credit: ESO @ruthangus
4. Image credit: NASA \ @ruthangus
5. \ @ruthangus
6. \ @ruthangus
7. Time ΔBrightness @ruthangus
8. 2341 confirmed transiting exoplanets (2/23/18) Data from the NASA exoplanet archive @ruthangus
9. Planet radius Γ @ruthangus (rate) The radius occurrence rate distribution of exoplanets
10. Planetary systems across space and time Ruth Angus Simons Fellow, Columbia University @ruthangus
11. Planet radius @ruthangus The radius occurrence rate distribution of exoplanets Γ (rate)
12. Time/Age The evolving occurrence rate of exoplanets ΓAge @ruthangus
13. Time/Age The evolving occurrence rate of exoplanets ΓAge @ruthangus
14. @ruthangus Time/Age
15. Image: Rudolf Dobesberger @ruthangus
16. Video credit: The Science Channel @ruthangus
17. \ @ruthangus
18. \ @ruthangus
19. Time ΔBrightness @ruthangus
20. The time signature of rotating stars @ruthangus
21. The time signature of rotating stars @ruthangus Prot
22. The time signature of rotating stars @ruthangus
23. NSPOTS SPOT Ω i SIZE1…N TEMPERATURE F LIMB DARKENING λ1…N, t… PROT ϕ1…N, t… models Spot @ruthangus
24. Gaussian process rotation model l A Γ Prot @ruthangus
25. A Gaussian process model Prot @ruthangus
26. Célérité: (N) Foreman-Mackey et al. +RA (2017) ArXiv: 1703.09710 george: (N log(N)2) george.readthedocs.io celerite.readthedocs.io emcee Foreman-Mackey et al. (2013) ArXiv: 1202.3665 @ruthangus
27. 4 x more accurate than sine-periodograms 18 x more accurate than autocorrelation functions The Gaussian process rotation period method: Angus et al. (2018) ArXiv: 1706.05459 @ruthangus
28. Γtrue Dn n = 1, …, N
29. m = 1, …, M Γtrue Noise tn,m fn,m n = 1, …, N
30. m = 1, …, M Γobs Γtrue Noise tn,m fn,m det eff n = 1, …, N
31. m = 1, …, M ΓA, obs ΓA, true Noise tn,m fn,m det eff n = 1, …, N
32. m = 1, …, M ΓA, obs ΓA, true Noise tn,m fn,m Prot gyro det eff n = 1, …, N
33. Improving stellar ages… @ruthangus
34. Rotation & magnetism Core H fraction Chemistry Age Galactic orbit @ruthangus
35. ’Chronometer’ example: core H fraction & rotation Angus et al. (in prep) 5x precision improvement @ruthangus @ruthangus
36. m = 1, …, M ΓA, obs ΓA, true Noise tn,m fn,m Prot gyro det eff n = 1, …, N
37. m = 1, …, M ΓA, obs ΓA, true n = 1, …, N Noise tn,m fn,m Prot gyro det eff H M C G
38. Kepler PLATO Gaia LSST TESS WFIRST Time #Stars ~mid 2020s 2009 Billions 100-thousands April 2018 @ruthangus
39. @ruthangus
40. Planet occurrence rates Orbital periods Radii Temperature/habitability Climate Biosignatures The evolving properties of planets Stability @ruthangus
41. • I am using improved stellar ages to characterize the exoplanet population across time and space. Summary • I use GP regression & auxiliary information to infer stellar ages. @ruthangus
42. Planetary systems across space and time Ruth Angus Simons Fellow, Columbia University @ruthangus
43. @ruthangus
44. Model from Kawaler (1988) Rotation periods converge after ~500 million years Solar mass stars Age [Myr] @ruthangus
45. A Quasi-periodic GP kernel function χ χ @ruthangus
46. 6.4 6.5 6.6 6.7 ln(l) 0.75 0.60 0.45 0.30 0.15 ln( ) 19.14 19.08 19.02 18.96 18.90 ln( ) 14.8 14.6 14.4 14.2 ln(A) 3.015 3.030 3.045 3.060 ln(P) 6.4 6.5 6.6 6.7 ln(l) 0.75 0.60 0.45 0.30 0.15 ln( ) 19.14 19.08 19.02 18.96 18.90 ln( ) 3.015 3.030 3.045 3.060 ln(P) Plot made using corner.py (Foreman-Mackey, 2016) Affine invariant MCMC sampler: emcee (Foreman-Mackey, 2013; Goodman & Weare, 2013)