Filtering¶

Let's start from all the transiting exoplanets in the NASA Exoplanet Archive, and then filter down into a Population of planets we might want to consider observing.

In [1]:

import exoatlas as ea
import exoatlas.visualizations as vi

ea.version()

import exoatlas as ea import exoatlas.visualizations as vi ea.version()

Out[1]:

'0.6.6'

Download Exoplanet Data¶

Let's download a recent list of transiting exoplanets from the NASA Exoplanet Archive. We should, as always, be careful not to completely trust every detail of this database. For any particular planet, there's a possibility that details might be wrong or missing so for anything super crucial, it would be very wise to double check.

In [2]:

e = ea.TransitingExoplanets()
s = ea.SolarSystem()

e = ea.TransitingExoplanets() s = ea.SolarSystem()

Throughout, let's use PlanetGallery to visualize how the various filtering cuts remove planets from our observing sample.

In [3]:

vi.PlanetGallery().build(e);

vi.PlanetGallery().build(e);

Define Filtering Criteria¶

Let's make some cuts to this sample in order to select systems that might be worth trying to observe at the telescope. We'll start by defining an unfiltered populatioin, for comparison.

In [4]:

unfiltered = e[:]
unfiltered.color = "orchid"
unfiltered.label = "Unfiltered"

unfiltered = e[:] unfiltered.color = "orchid" unfiltered.label = "Unfiltered"

Is it big enough?¶

To ensure that a planet has a hydrogen-dominated atmosphere (which makes interpreting a transmission spectrum a lot easier), we could place a cut on the planet's radius. "Most 1.6 Earth-radius planets are not rocky", so let's try cutting there.

In [5]:

import astropy.units as u

is_big = e.radius() > 1.6 * u.Rearth
filtered = e[is_big]
filtered.label = "Big"
filtered

import astropy.units as u is_big = e.radius() > 1.6 * u.Rearth filtered = e[is_big] filtered.label = "Big" filtered

Out[5]:

✨ Big | 3060 elements ✨

In [6]:

vi.PlanetGallery().build([unfiltered, filtered]);

vi.PlanetGallery().build([unfiltered, filtered]);

Is the duration short enough?¶

An ideal transit observation would include at least some baseline before the start and after the end of the transit itself. When observing from a ground-based telescope, practically that means it's really hard to observe (complete) transits with durations longer than about a few hours.

In [7]:

is_short = e.transit_duration() < 3 * u.hour
filtered = e[is_short]
filtered.label = "Short"
filtered

is_short = e.transit_duration() < 3 * u.hour filtered = e[is_short] filtered.label = "Short" filtered

Out[7]:

✨ Short | 2002 elements ✨

In [8]:

vi.PlanetGallery().build([unfiltered, filtered]);

vi.PlanetGallery().build([unfiltered, filtered]);

Is transmission spectroscopy possible?¶

The best precision we could possibly acheive on measuring a transit depth is set by the number of photons that we can gather with our telescope. If the signal we're hoping to detect for transmission spectroscopy (see Observing) is smaller than that expected precision, it's basically hopeless that we'll see anything. So, let's at least limit our sample to targets with a reasonably high predicted signal-to-noise ratio.

In [9]:

import matplotlib.pyplot as plt
import numpy as np

signal = e.transmission_signal()
wavelength = 0.7 * u.micron
dt = 1 * u.hour
R = 10
noise = e.depth_uncertainty(telescope="APO", wavelength=wavelength, dt=dt, R=R)
snr = signal / noise

plt.figure(figsize=(8, 5))
plt.scatter(noise, signal, c=snr, vmin=1, vmax=10)
plt.xscale("log")
plt.yscale("log")
plt.axis("scaled")
plt.xlabel("")
plt.ylim(1e-6, 1e-2)
plt.xlim(1e-6, 1e-2)
plt.plot(
    [1e-6, 1e-2], [1e-6, 1e-2], color="black", linewidth=3, linestyle="--", alpha=0.3
)
plt.colorbar(label="S/N")
plt.xlabel(rf"Predicted Photon Noise\n($\lambda=${wavelength}, dt={dt}, R={R})")
plt.ylabel(r"Possible Transmission\nSignal Size ($2HR_p/R_{\star}^2$)");

import matplotlib.pyplot as plt import numpy as np signal = e.transmission_signal() wavelength = 0.7 * u.micron dt = 1 * u.hour R = 10 noise = e.depth_uncertainty(telescope="APO", wavelength=wavelength, dt=dt, R=R) snr = signal / noise plt.figure(figsize=(8, 5)) plt.scatter(noise, signal, c=snr, vmin=1, vmax=10) plt.xscale("log") plt.yscale("log") plt.axis("scaled") plt.xlabel("") plt.ylim(1e-6, 1e-2) plt.xlim(1e-6, 1e-2) plt.plot( [1e-6, 1e-2], [1e-6, 1e-2], color="black", linewidth=3, linestyle="--", alpha=0.3 ) plt.colorbar(label="S/N") plt.xlabel(rf"Predicted Photon Noise\n($\lambda=${wavelength}, dt={dt}, R={R})") plt.ylabel(r"Possible Transmission\nSignal Size ($2HR_p/R_{\star}^2$)");

In [10]:

is_snr = snr > 5
filtered = e[is_snr]
filtered.label = "High S/N"
filtered

is_snr = snr > 5 filtered = e[is_snr] filtered.label = "High S/N" filtered

Out[10]:

✨ High S/N | 253 elements ✨

In [11]:

vi.PlanetGallery().build([unfiltered, filtered]);

vi.PlanetGallery().build([unfiltered, filtered]);

Is it above the horizon at night?¶

If we're observing at a particular time of year from a particular observatory, only some targets will be high enough in the sky to be observable at night. Let's place a cut on the airmass) at midnight on some particular night, so we don't waste time considering planets that are beneath the Earth when we might be trying to observe.

In [12]:

from astroplan import Observer
from astropy.time import Time

date = Time.now()
observatory = Observer.at_site("APO", timezone="US/Mountain")

from astroplan import Observer from astropy.time import Time date = Time.now() observatory = Observer.at_site("APO", timezone="US/Mountain")

In [13]:

local_midnight = observatory.midnight(date)
sidereal_time_at_midnight = observatory.local_sidereal_time(local_midnight)
ra_at_midnight = sidereal_time_at_midnight
ra_at_midnight

local_midnight = observatory.midnight(date) sidereal_time_at_midnight = observatory.local_sidereal_time(local_midnight) ra_at_midnight = sidereal_time_at_midnight ra_at_midnight

Out[13]:

$17^{\mathrm{h}}48^{\mathrm{m}}15.61675714^{\mathrm{s}}$

In [14]:

declination_at_zenith = observatory.latitude
declination_at_zenith

declination_at_zenith = observatory.latitude declination_at_zenith

Out[14]:

$32^\circ46{}^\prime48{}^{\prime\prime}$

In [15]:

airmass = e.airmass(where=observatory, when=local_midnight)
is_up = airmass < 2

filtered = e[is_up]
filtered.name = "Up at Midnight"

plt.figure()
plt.scatter(e.ra(), e.dec(), color="black", marker=".", s=2, alpha=0.2)
plt.scatter(filtered.ra(), filtered.dec(), color="black", marker="o", facecolor="none")
plt.scatter(e.ra(), e.dec(), c=airmass, cmap="viridis_r", vmin=1, vmax=10, zorder=-1)
plt.axvline(ra_at_midnight, color="gray")
plt.axhline(declination_at_zenith, color="gray")
plt.xlabel("Right Ascension (degree)")
plt.ylabel("Declination (degree)")
plt.axis("scaled")
plt.ylim(-90, 90)
plt.xlim(360, 0)
plt.colorbar(label="airmass")

plt.title(f"{observatory.name} | {local_midnight.iso} UTC");

airmass = e.airmass(where=observatory, when=local_midnight) is_up = airmass < 2 filtered = e[is_up] filtered.name = "Up at Midnight" plt.figure() plt.scatter(e.ra(), e.dec(), color="black", marker=".", s=2, alpha=0.2) plt.scatter(filtered.ra(), filtered.dec(), color="black", marker="o", facecolor="none") plt.scatter(e.ra(), e.dec(), c=airmass, cmap="viridis_r", vmin=1, vmax=10, zorder=-1) plt.axvline(ra_at_midnight, color="gray") plt.axhline(declination_at_zenith, color="gray") plt.xlabel("Right Ascension (degree)") plt.ylabel("Declination (degree)") plt.axis("scaled") plt.ylim(-90, 90) plt.xlim(360, 0) plt.colorbar(label="airmass") plt.title(f"{observatory.name} | {local_midnight.iso} UTC");

In [16]:

vi.PlanetGallery().build([unfiltered, filtered]);

vi.PlanetGallery().build([unfiltered, filtered]);

Make Target Sample¶

After looking at each of them one by one, let's put all of those filtering criteria together.

In [17]:

targets = e[is_big * is_short * is_snr * is_up]
targets.label = "Observable?!"
targets

targets = e[is_big * is_short * is_snr * is_up] targets.label = "Observable?!" targets

Out[17]:

✨ Observable?! | 31 elements ✨

In [18]:

vi.PlanetGallery().build([unfiltered, targets]);

vi.PlanetGallery().build([unfiltered, targets]);

Let's print out the names of the planets that made it through all our cuts.

In [19]:

print(targets.name())

print(targets.name())

['GJ 1214 b' 'HAT-P-11 b' 'HAT-P-12 b' 'HAT-P-18 b' 'HAT-P-26 b'
 'HD 189733 b' 'HD 191939 b' 'HD 219134 b' 'K2-31 b' 'KELT-23 A b'
 'Kepler-51 c' 'NGTS-5 b' 'Qatar-10 b' 'Qatar-6 b' 'TOI-1173 b'
 'TOI-1248 b' 'TOI-1259 A b' 'TOI-1273 b' 'TOI-1408 b' 'TOI-1431 b'
 'TOI-1855 b' 'TOI-2134 b' 'TOI-4463 A b' 'TrES-1 b' 'WASP-39 b'
 'WASP-67 b' 'WASP-69 b' 'WASP-74 b' 'WASP-80 b' 'XO-1 b' 'XO-7 b']

Let's also make a table that includes more useful information, like positions and transit ephemerides, which we could use as an input to plan an observing run.

In [20]:

table = targets.create_planning_table()
table

table = targets.create_planning_table() table

Out[20]:

QTable length=31

name	ra	dec	period	transit_midpoint	transit_duration	radius	relative_insolation	stellar_radius	stellar_teff	distance
	deg	deg	d	d	h	earthRad		solRad	K	pc
str29	float64	float64	float64	float64	float64	float64	float64	float64	float64	float64
GJ 1214 b	258.8313991	4.9606795	1.580404531	2459639.7812619	0.86966	2.733	17.21817034577689	0.2162	3101.0	14.6427
HAT-P-11 b	297.7101763	48.0818635	4.888	2454957.8132067	2.35562	4.36	98.02544814710485	0.683	4653.0	37.7647
HAT-P-12 b	209.3886452	43.49331	3.2130598	2454419.19556	2.3376	10.749	142.34347084035403	0.701	4650.0	142.751
HAT-P-18 b	256.346376	33.0123251	5.508023	2454715.02174	2.7144	11.153	86.33291087921891	0.749	4803.0	161.4
HAT-P-26 b	213.1565508	4.0594175	4.23452	2456901.059458	2.4552	7.06167	190.5675599152924	0.87	5079.0	141.837
HD 189733 b	300.1821223	22.7097759	2.21857567	2453955.5255511	1.8233621	12.66617	353.86047213972984	0.75	5052.0	19.7638
HD 191939 b	302.0256247	66.8503014	8.8803256	2459443.54236	2.939	3.41	100.5536645439072	0.94	5348.0	53.6089
HD 219134 b	348.3372026	57.1696255	3.092926	2457126.69913	0.945	1.602	176.12559092759645	0.778	4699.0	6.53127
K2-31 b	245.4405531	-23.548273	1.25785	2457191.70889	0.9816	11.88154	1602.7826713097115	0.78	5280.0	110.527
...	...	...	...	...	...	...	...	...	...	...
TOI-4463 A b	279.3581286	18.7299522	2.8807198	2459291.64004	1.85	13.26022417	629.8653023803159	1.062	5640.0	173.117
TrES-1 b	286.0408755	36.632536	3.03007	2453186.806341	2.508	12.66617	324.79600430434164	0.85	5230.0	159.658
WASP-39 b	217.3266477	-3.4444994	4.0552842	2459783.5015	2.8032	14.930388	315.36062051908334	0.918	5485.0	213.982
WASP-67 b	295.7438485	-19.9497314	4.61442	2456618.0537	1.896	12.89035	189.27714560988787	0.88	5200.0	189.469
WASP-69 b	315.0259661	-5.094857	3.86814	2459798.775459	2.1610195	12.44199	166.35384100958626	0.86	4700.0	49.9605
WASP-74 b	304.538843	-1.0760033	2.13775	2457103.325971	2.292	15.24424	2117.5611209480976	1.42	5990.0	149.216
WASP-80 b	303.1667996	-2.1444369	3.06785234	2456487.425006	2.131	11.197791	77.07315102540811	0.586	4143.0	49.7876
XO-1 b	240.5492745	28.1696248	3.94153	2455385.51978	2.9712	12.77826	344.2550786936063	0.88	5750.0	163.553
XO-7 b	277.4780753	85.2333208	2.8641424	2457917.47503	2.772	15.389957	1381.2712428658124	1.48	6250.0	234.149

Let's save that table out to a text file, and also make sure we can load it back in again.

In [21]:

observatory_string = observatory.name.replace(" ", "-")
date_string = local_midnight.iso.split()[0]
table.write(
    f"planets-to-observe-{observatory_string}-{date_string}.ecsv", overwrite=True
)

observatory_string = observatory.name.replace(" ", "-") date_string = local_midnight.iso.split()[0] table.write( f"planets-to-observe-{observatory_string}-{date_string}.ecsv", overwrite=True )

In [22]:

from astropy.io.ascii import read

read(f"planets-to-observe-{observatory_string}-{date_string}.ecsv")

from astropy.io.ascii import read read(f"planets-to-observe-{observatory_string}-{date_string}.ecsv")

Out[22]:

QTable length=31

name	ra	dec	period	transit_midpoint	transit_duration	radius	relative_insolation	stellar_radius	stellar_teff	distance
	deg	deg	d	d	h	earthRad		solRad	K	pc
str12	float64	float64	float64	float64	float64	float64	float64	float64	float64	float64
GJ 1214 b	258.8313991	4.9606795	1.580404531	2459639.78126	0.86966	2.733	17.2181703458	0.2162	3101.0	14.6427
HAT-P-11 b	297.7101763	48.0818635	4.888	2454957.81321	2.35562	4.36	98.0254481471	0.683	4653.0	37.7647
HAT-P-12 b	209.3886452	43.49331	3.2130598	2454419.19556	2.3376	10.749	142.34347084	0.701	4650.0	142.751
HAT-P-18 b	256.346376	33.0123251	5.508023	2454715.02174	2.7144	11.153	86.3329108792	0.749	4803.0	161.4
HAT-P-26 b	213.1565508	4.0594175	4.23452	2456901.05946	2.4552	7.06167	190.567559915	0.87	5079.0	141.837
HD 189733 b	300.1821223	22.7097759	2.21857567	2453955.52555	1.8233621	12.66617	353.86047214	0.75	5052.0	19.7638
HD 191939 b	302.0256247	66.8503014	8.8803256	2459443.54236	2.939	3.41	100.553664544	0.94	5348.0	53.6089
HD 219134 b	348.3372026	57.1696255	3.092926	2457126.69913	0.945	1.602	176.125590928	0.778	4699.0	6.53127
K2-31 b	245.4405531	-23.548273	1.25785	2457191.70889	0.9816	11.88154	1602.78267131	0.78	5280.0	110.527
...	...	...	...	...	...	...	...	...	...	...
TOI-4463 A b	279.3581286	18.7299522	2.8807198	2459291.64004	1.85	13.26022417	629.86530238	1.062	5640.0	173.117
TrES-1 b	286.0408755	36.632536	3.03007	2453186.80634	2.508	12.66617	324.796004304	0.85	5230.0	159.658
WASP-39 b	217.3266477	-3.4444994	4.0552842	2459783.5015	2.8032	14.930388	315.360620519	0.918	5485.0	213.982
WASP-67 b	295.7438485	-19.9497314	4.61442	2456618.0537	1.896	12.89035	189.27714561	0.88	5200.0	189.469
WASP-69 b	315.0259661	-5.094857	3.86814	2459798.77546	2.1610195	12.44199	166.35384101	0.86	4700.0	49.9605
WASP-74 b	304.538843	-1.0760033	2.13775	2457103.32597	2.292	15.24424	2117.56112095	1.42	5990.0	149.216
WASP-80 b	303.1667996	-2.1444369	3.06785234	2456487.42501	2.131	11.197791	77.0731510254	0.586	4143.0	49.7876
XO-1 b	240.5492745	28.1696248	3.94153	2455385.51978	2.9712	12.77826	344.255078694	0.88	5750.0	163.553
XO-7 b	277.4780753	85.2333208	2.8641424	2457917.47503	2.772	15.389957	1381.27124287	1.48	6250.0	234.149

You can try filtering populations for your own devious and wonderful purposes, using whatever criteria you like!