1
00:00:05,460 --> 00:00:06,293
- [Scott] This update on

2
00:00:06,293 --> 00:00:08,700
the extreme precision
radical velocity initiative.

3
00:00:09,730 --> 00:00:12,570
Most of this is gonna be given by Jenn

4
00:00:12,570 --> 00:00:15,270
but I'm just gonna kick it off

5
00:00:15,270 --> 00:00:20,180
by giving the motivation for
why we did this work in group.

6
00:00:20,180 --> 00:00:21,943
So next slide, Jenn.

7
00:00:22,990 --> 00:00:23,823
So like I said,

8
00:00:23,823 --> 00:00:25,780
I'll just start off with the motivation.

9
00:00:25,780 --> 00:00:28,170
Why did we need extreme
precision radial velocity?

10
00:00:28,170 --> 00:00:30,000
I'll define what I mean by that.

11
00:00:30,000 --> 00:00:32,720
And then Jenn is basically
gonna finish this off

12
00:00:32,720 --> 00:00:35,750
by talking about the state
of the art methodology,

13
00:00:35,750 --> 00:00:38,553
architecture simulations and finally,

14
00:00:39,390 --> 00:00:42,930
the basic outcome, which is
the proposed research program.

15
00:00:42,930 --> 00:00:43,763
Next slide.

16
00:00:44,900 --> 00:00:47,330
Okay, so the motivation for EPRV,

17
00:00:47,330 --> 00:00:49,040
or why do we need to measure masses

18
00:00:49,040 --> 00:00:51,973
of Earth-like planets orbiting
nearby sun-like stars?

19
00:00:53,240 --> 00:00:55,243
A lot of this is basically old news,

20
00:00:56,124 --> 00:00:57,540
so I'm gonna go through it fairly quickly.

21
00:00:57,540 --> 00:01:00,653
We all know this mass, of course,

22
00:01:03,117 --> 00:01:04,250
is one of the most fundamental,

23
00:01:04,250 --> 00:01:07,210
if not the most fundamental
property of a planet.

24
00:01:07,210 --> 00:01:09,050
And knowing that planet's mass,

25
00:01:09,050 --> 00:01:11,490
along with the knowledge of its radius,

26
00:01:11,490 --> 00:01:13,150
you can tell what kind of a planet it is,

27
00:01:13,150 --> 00:01:15,520
a rocky planet, a super Earth,

28
00:01:15,520 --> 00:01:18,210
mini Neptune or gas giant.

29
00:01:18,210 --> 00:01:20,640
And having a mass
measurement also allows you

30
00:01:20,640 --> 00:01:23,750
to better interpret any
spectroscopic features

31
00:01:23,750 --> 00:01:25,090
that you get when you measure it

32
00:01:25,090 --> 00:01:29,253
in a reflected light or thermal emission.

33
00:01:30,160 --> 00:01:31,320
But the problem is that,

34
00:01:31,320 --> 00:01:35,010
we need to push the mass
measurement precision down

35
00:01:35,010 --> 00:01:37,130
to the level where we can
check Earth-mass planets

36
00:01:37,130 --> 00:01:38,760
around sun-like stars.

37
00:01:38,760 --> 00:01:39,593
Next slide.

38
00:01:40,770 --> 00:01:45,440
Okay, so, it's a very wordy slide,

39
00:01:45,440 --> 00:01:48,400
but I will go through the logic

40
00:01:48,400 --> 00:01:50,650
why I think this is a
pretty airtight argument

41
00:01:50,650 --> 00:01:53,510
why we should be starting
this project right now.

42
00:01:53,510 --> 00:01:55,190
So I call it, learn it, love it

43
00:01:55,190 --> 00:01:58,393
and use it, the precision
radial velocity, next.

44
00:01:59,690 --> 00:02:02,280
So, as I've already said,

45
00:02:02,280 --> 00:02:03,410
we need to measure the masses

46
00:02:03,410 --> 00:02:05,740
of directly-imaged habitable planets,

47
00:02:05,740 --> 00:02:07,550
as well as the masses

48
00:02:07,550 --> 00:02:10,060
of rocky terrestrial transiting planets.

49
00:02:10,060 --> 00:02:13,500
Next, we have basically
two choices to do this.

50
00:02:13,500 --> 00:02:16,620
Next, the first is astrometry
with the systematic floor

51
00:02:16,620 --> 00:02:20,390
of a few 10s of astrometry, sorry,

52
00:02:20,390 --> 00:02:21,310
with the systematic floor

53
00:02:21,310 --> 00:02:24,720
of a few 10s of nanoarcseconds,

54
00:02:24,720 --> 00:02:28,060
or precision radical velocity
with systematic floor

55
00:02:28,060 --> 00:02:30,400
of a few centimeters per second.

56
00:02:30,400 --> 00:02:32,280
And I'll motivate those numbers, letters.

57
00:02:32,280 --> 00:02:35,300
So, astrometry, we all agree
it must be done in space.

58
00:02:35,300 --> 00:02:37,230
And so if it's a stand-alone mission,

59
00:02:37,230 --> 00:02:40,603
it's likely greater than
a billion dollars, next.

60
00:02:41,810 --> 00:02:43,030
Although you could imagine

61
00:02:43,030 --> 00:02:45,430
a specifically designed instrument

62
00:02:45,430 --> 00:02:47,530
on another large aperture space mission

63
00:02:47,530 --> 00:02:48,667
such as LUVOIR,

64
00:02:49,560 --> 00:02:52,050
and that's plausible, but
would still be expensive

65
00:02:52,050 --> 00:02:53,410
with several hundred million dollars

66
00:02:53,410 --> 00:02:56,430
that would require significant
technology development.

67
00:02:56,430 --> 00:02:58,730
And of course, you would need the mission.

68
00:02:58,730 --> 00:03:01,420
On the other hand EPRV at a
few centimeters per second,

69
00:03:01,420 --> 00:03:03,160
may be doable from the ground

70
00:03:03,160 --> 00:03:06,210
and so it will likely be
cheaper than any other options.

71
00:03:06,210 --> 00:03:09,830
Now, I recognize that this
is a controversial statement.

72
00:03:09,830 --> 00:03:11,550
Some people will tell you it's impossible.

73
00:03:11,550 --> 00:03:12,860
Some people will tell you,

74
00:03:12,860 --> 00:03:14,680
it's certainly possible.

75
00:03:14,680 --> 00:03:17,400
But, the simple fact is
that, no one really knows.

76
00:03:17,400 --> 00:03:19,070
And so, we actually
have to do the research

77
00:03:19,070 --> 00:03:20,933
to find out, next.

78
00:03:22,330 --> 00:03:23,540
And I will mention that there have been

79
00:03:23,540 --> 00:03:26,660
some promising progress along the way

80
00:03:26,660 --> 00:03:29,453
and I'm sure we heard
about a lot of those today.

81
00:03:30,690 --> 00:03:32,930
Okay, so given that, we should try, first,

82
00:03:32,930 --> 00:03:35,690
what is likely to be the cheapest option,

83
00:03:35,690 --> 00:03:37,600
we should form the RNA needed

84
00:03:37,600 --> 00:03:40,150
to determine if we can
actually achieve this

85
00:03:40,150 --> 00:03:44,620
few centimeter per second
systematic floor, next.

86
00:03:44,620 --> 00:03:46,740
And so, if we can do this from the ground,

87
00:03:46,740 --> 00:03:48,730
then we can also improve the efficiency

88
00:03:48,730 --> 00:03:49,980
of direct-imaging missions,

89
00:03:49,980 --> 00:03:52,200
as well as increase the yield,

90
00:03:52,200 --> 00:03:54,500
perhaps by an amount

91
00:03:54,500 --> 00:03:57,610
that depends on the specific architecture

92
00:03:57,610 --> 00:03:59,780
of the direct-imaging mission.

93
00:03:59,780 --> 00:04:02,830
Next, so that's the basic argument

94
00:04:02,830 --> 00:04:06,570
for why we should be
starting this initiative now.

95
00:04:06,570 --> 00:04:09,060
In terms of the value of
precursor observations,

96
00:04:09,060 --> 00:04:12,630
versus concurrent, or post observations

97
00:04:14,160 --> 00:04:17,840
for direct-imaging mission,

98
00:04:17,840 --> 00:04:19,940
generally precursor observations help

99
00:04:19,940 --> 00:04:22,760
if the time to detect a planet,

100
00:04:22,760 --> 00:04:24,770
is greater than time to characterize it.

101
00:04:24,770 --> 00:04:28,000
For example, if you have
low completeness per visit,

102
00:04:28,000 --> 00:04:31,150
due to the fact that you
have a small dark hole,

103
00:04:31,150 --> 00:04:32,640
a large inner working angle,

104
00:04:32,640 --> 00:04:34,423
or if ETA Earth is small,

105
00:04:35,530 --> 00:04:36,363
in that case,

106
00:04:36,363 --> 00:04:39,240
if the yield is resource-limited,

107
00:04:39,240 --> 00:04:41,550
in other words, you have
a limited number of slews

108
00:04:41,550 --> 00:04:42,440
for a Starshade

109
00:04:42,440 --> 00:04:45,440
or long integration time
for characterization,

110
00:04:45,440 --> 00:04:48,410
then precursor observations
can dramatically improve

111
00:04:48,410 --> 00:04:50,410
the efficiency of direct imaging missions,

112
00:04:50,410 --> 00:04:53,200
'cause basically you know where to look

113
00:04:53,200 --> 00:04:55,630
and you don't have to spend
a lot of time searching

114
00:04:55,630 --> 00:04:58,560
for the Earth-like planets
around sun-like stars.

115
00:04:58,560 --> 00:05:00,330
That allows time for other science,

116
00:05:00,330 --> 00:05:02,430
be it exoplanet science,

117
00:05:02,430 --> 00:05:05,450
comparative exoplanetology
with direct energy,

118
00:05:05,450 --> 00:05:10,450
or non-exoplanet science as is advocated

119
00:05:10,540 --> 00:05:13,150
by both LUVOIR and HabEX.

120
00:05:13,150 --> 00:05:14,570
And in certain circumstances,

121
00:05:14,570 --> 00:05:17,580
you can also improve the yield
of characterized planets.

122
00:05:17,580 --> 00:05:18,523
Next slide.

123
00:05:20,470 --> 00:05:22,120
So this just kind of demonstrates

124
00:05:22,120 --> 00:05:23,370
the efficiency improvement.

125
00:05:23,370 --> 00:05:26,500
This is from Rhonda
Morgan's using exoSIMs.

126
00:05:26,500 --> 00:05:29,250
The blue curve is with
precursor observations,

127
00:05:29,250 --> 00:05:30,820
it's the number of Earth candidates

128
00:05:30,820 --> 00:05:35,050
you detect as a function of
the mission duration and days.

129
00:05:35,050 --> 00:05:37,970
And so, you can see that you get your 50%

130
00:05:37,970 --> 00:05:40,320
of your yield within 200 days.

131
00:05:40,320 --> 00:05:42,390
If you don't have precursor observations,

132
00:05:42,390 --> 00:05:47,140
then you get 50% of your
yield within roughly 700 days.

133
00:05:47,140 --> 00:05:49,790
And so that means that you
can improve your efficiency

134
00:05:49,790 --> 00:05:52,700
by a factor of three.

135
00:05:52,700 --> 00:05:53,700
And that's important

136
00:05:53,700 --> 00:05:56,750
because high-impact science
occurs earlier in the mission,

137
00:05:56,750 --> 00:05:59,650
allowing time for characterization.

138
00:05:59,650 --> 00:06:01,830
You get immediate science results

139
00:06:01,830 --> 00:06:04,320
that excite the public
in the science community,

140
00:06:04,320 --> 00:06:07,713
and it mitigates any risk
of early mission failure.

141
00:06:08,650 --> 00:06:13,650
So, EPRV generally makes missions
more nimble and powerful.

142
00:06:14,240 --> 00:06:17,623
And we'll skip to next slide.

143
00:06:19,810 --> 00:06:21,950
Okay, so we all know
that the basic problem

144
00:06:21,950 --> 00:06:24,020
is that we've been stuck
at one meter per second

145
00:06:24,020 --> 00:06:26,120
for roughly the last decade.

146
00:06:26,120 --> 00:06:30,183
This has been documented
in several, many papers,

147
00:06:31,770 --> 00:06:36,120
I think of the most prescient
are Fisher et al review

148
00:06:36,120 --> 00:06:38,900
of the first EPRV meeting.

149
00:06:38,900 --> 00:06:42,630
And Dumusque et al
describes a data challenge

150
00:06:42,630 --> 00:06:47,630
where he sent fake data with planets

151
00:06:47,790 --> 00:06:50,260
and without planets of various amplitudes

152
00:06:50,260 --> 00:06:53,570
and asked people to try to fit

153
00:06:53,570 --> 00:06:54,980
and see if they could detect planets.

154
00:06:54,980 --> 00:06:57,010
And in general, the result was that,

155
00:06:57,010 --> 00:06:59,030
if the amplitude of the planets

156
00:06:59,030 --> 00:07:01,350
was below one meter per second,

157
00:07:01,350 --> 00:07:03,250
it was generally not
successfully detected.

158
00:07:03,250 --> 00:07:04,760
And in a large fraction of cases

159
00:07:04,760 --> 00:07:06,840
where it was above one meter per second,

160
00:07:06,840 --> 00:07:09,450
it was successfully detected.

161
00:07:09,450 --> 00:07:11,560
So their primary conclusion was,

162
00:07:11,560 --> 00:07:13,340
even with the best models
of stellar signals,

163
00:07:13,340 --> 00:07:14,690
planetary signals with amplitude

164
00:07:14,690 --> 00:07:16,520
with less than a meter per second,

165
00:07:16,520 --> 00:07:19,450
are rarely extracted correctly
with current precision

166
00:07:19,450 --> 00:07:21,194
and current techniques.

167
00:07:21,194 --> 00:07:22,027
In other words,

168
00:07:22,027 --> 00:07:23,870
we have to do something
fundamentally different

169
00:07:23,870 --> 00:07:26,400
than what we've been doing
for the last 10 years

170
00:07:26,400 --> 00:07:30,040
if we wanna get to 10
centimeters per second precision

171
00:07:30,040 --> 00:07:33,730
and at least a few meters
per second accuracy.

172
00:07:33,730 --> 00:07:34,563
Next slide.

173
00:07:36,390 --> 00:07:38,160
So this led to the National Academy

174
00:07:38,160 --> 00:07:42,070
of Sciences Exoplanet Science
Strategy report finding,

175
00:07:42,070 --> 00:07:43,480
that improving the precision

176
00:07:43,480 --> 00:07:45,230
of radical velocity measurements,

177
00:07:45,230 --> 00:07:47,343
will support exoplanet missions.

178
00:07:48,280 --> 00:07:50,390
And the recommendation that NASA

179
00:07:50,390 --> 00:07:55,390
and NSF should establish a
strategic initiative in EPRVs

180
00:07:55,620 --> 00:07:58,760
to develop a method of facilities
for measuring the masses

181
00:07:58,760 --> 00:08:01,900
of temperature terrestrial
planets orbiting sun-like stars.

182
00:08:01,900 --> 00:08:04,160
And you might wanna ask why NASA and NSF?

183
00:08:04,160 --> 00:08:06,850
Well NSF is very good,

184
00:08:06,850 --> 00:08:08,030
it has lots of experience

185
00:08:08,030 --> 00:08:10,160
at running ground-based telescopes.

186
00:08:10,160 --> 00:08:14,970
NASA is very good at
project-level organization.

187
00:08:14,970 --> 00:08:16,050
And so working together,

188
00:08:16,050 --> 00:08:18,810
they can play off of
each other's strengths

189
00:08:18,810 --> 00:08:20,830
to achieve this initiative,

190
00:08:20,830 --> 00:08:23,180
which is definitely challenging

191
00:08:23,180 --> 00:08:28,180
and requires a very broad range
of resources and expertise.

192
00:08:30,570 --> 00:08:31,403
Next slide.

193
00:08:33,160 --> 00:08:35,390
So, I just wanna motivate this

194
00:08:35,390 --> 00:08:38,393
with a few centimeters
per second accuracy.

195
00:08:39,400 --> 00:08:42,140
So you can see a simulation here

196
00:08:42,140 --> 00:08:46,230
from Patrick Newman at
George Mason University

197
00:08:47,640 --> 00:08:49,070
and Peter Plavchan,

198
00:08:49,070 --> 00:08:50,750
you can see individual measurements

199
00:08:50,750 --> 00:08:52,830
are basically never gonna be at the level

200
00:08:52,830 --> 00:08:55,160
of a few centimeters per second,

201
00:08:55,160 --> 00:08:57,290
they're gonna be at the level of, maybe,

202
00:08:57,290 --> 00:08:58,630
20 centimeters per second.

203
00:08:58,630 --> 00:09:01,740
So that means, you have
to actually route down

204
00:09:01,740 --> 00:09:05,680
to get down to the
precision probe endpoint

205
00:09:05,680 --> 00:09:06,953
that allows you to detect the planet

206
00:09:06,953 --> 00:09:09,123
that signals noise of 10 or 20.

207
00:09:10,200 --> 00:09:12,720
And so, that means your
systematic accuracy

208
00:09:12,720 --> 00:09:13,840
or your systematic floor,

209
00:09:13,840 --> 00:09:17,083
must be at the level of a
few centimeters per second.

210
00:09:18,080 --> 00:09:18,913
Next slide.

211
00:09:20,450 --> 00:09:23,170
So, this is a slide from Deborah Fisher,

212
00:09:23,170 --> 00:09:24,390
which I really like

213
00:09:24,390 --> 00:09:28,200
and it kind of illustrates
the problem that we have.

214
00:09:28,200 --> 00:09:30,250
There are many different sources

215
00:09:30,250 --> 00:09:31,880
of the stellar variability,

216
00:09:31,880 --> 00:09:36,277
to instrumental stability
that we have to deal with.

217
00:09:38,620 --> 00:09:42,620
Then they contribute
various different amounts

218
00:09:42,620 --> 00:09:44,400
to the error budget

219
00:09:45,450 --> 00:09:48,200
and the more nails that we nail down,

220
00:09:48,200 --> 00:09:50,853
the more nails we find at lower precision.

221
00:09:51,720 --> 00:09:54,393
And sometimes we even
have the wrong tools,

222
00:09:55,580 --> 00:09:57,170
as shown here by the screw,

223
00:09:57,170 --> 00:09:59,730
where we need a screwdriver
and not a hammer.

224
00:09:59,730 --> 00:10:02,740
And so, it's gonna take
coordinated initiative

225
00:10:02,740 --> 00:10:05,700
to actually allow us to figure out

226
00:10:05,700 --> 00:10:07,590
what the sources of noise are

227
00:10:08,689 --> 00:10:09,522
and beat them down,

228
00:10:09,522 --> 00:10:10,430
so that we can get down

229
00:10:10,430 --> 00:10:14,071
to the few centimeters per
second systematic limit.

230
00:10:14,071 --> 00:10:14,904
And I think with that,

231
00:10:14,904 --> 00:10:16,310
I'm gonna turn it over to Jenn.

232
00:10:16,310 --> 00:10:17,470
- [Coordinator] Okay, I was just gonna say

233
00:10:17,470 --> 00:10:19,328
we're halfway done, Scott.

234
00:10:19,328 --> 00:10:20,161
- [Scott] Perfect.

235
00:10:20,161 --> 00:10:22,326
- [Jenn] Okay, so I'm
gonna take over from here.

236
00:10:22,326 --> 00:10:24,590
I'm gonna talk about the
current state of the art.

237
00:10:24,590 --> 00:10:26,147
So with the latest generation of precision

238
00:10:26,147 --> 00:10:28,180
of the spectrograph, or timeline,

239
00:10:28,180 --> 00:10:31,476
we're starting to push below
that meter per second level.

240
00:10:31,476 --> 00:10:33,440
Instruments like EXPRESSO
and WIYN/NEID and EXPRES,

241
00:10:33,440 --> 00:10:35,750
are showing that breaching
instrumental precision

242
00:10:35,750 --> 00:10:39,180
of 30 to 50 centimeters per
second, does seem doable.

243
00:10:39,180 --> 00:10:40,590
And actually they've been in operation

244
00:10:40,590 --> 00:10:41,530
for at least a few years.

245
00:10:41,530 --> 00:10:42,820
There's hope that we can use

246
00:10:42,820 --> 00:10:45,000
their individual design decisions

247
00:10:45,000 --> 00:10:46,440
and on-sky performance

248
00:10:46,440 --> 00:10:48,850
to help decide what
areas of instrumentation

249
00:10:48,850 --> 00:10:50,750
to focus on next.

250
00:10:50,750 --> 00:10:53,100
But addressing the
instruments stability alone

251
00:10:53,100 --> 00:10:54,190
is not enough.

252
00:10:54,190 --> 00:10:56,890
As you can see here, it
has been mentioned earlier,

253
00:10:56,890 --> 00:11:00,080
there's a veritable pileup of
stellar phenomena occurring

254
00:11:00,080 --> 00:11:01,900
on a variety of timescales

255
00:11:01,900 --> 00:11:04,930
that produce signals at the
one meter per second level.

256
00:11:04,930 --> 00:11:07,340
And so one of our largest
challenges going forward,

257
00:11:07,340 --> 00:11:08,480
we'll be transitioning away

258
00:11:08,480 --> 00:11:10,170
from treating these different signals

259
00:11:10,170 --> 00:11:12,250
and sources as noise,

260
00:11:12,250 --> 00:11:14,820
and instead obtaining data
with a signal to noise

261
00:11:14,820 --> 00:11:17,330
and cadence and spectral
resolution necessary

262
00:11:17,330 --> 00:11:19,000
to treat them as individual signals

263
00:11:19,000 --> 00:11:22,570
that can be modeled and
removed from the RV data set.

264
00:11:22,570 --> 00:11:25,530
So with those, somewhat
intimidating challenges in mind,

265
00:11:25,530 --> 00:11:28,490
we convened the EPRV working group.

266
00:11:28,490 --> 00:11:29,780
The working group was comprised

267
00:11:29,780 --> 00:11:31,660
of roughly 40 precision RV experts

268
00:11:31,660 --> 00:11:33,610
from around the country
and around the world,

269
00:11:33,610 --> 00:11:35,630
with people specializing in everything,

270
00:11:35,630 --> 00:11:37,700
from instrumentation,
to observing strategies

271
00:11:37,700 --> 00:11:38,943
and stellar variability.

272
00:11:39,940 --> 00:11:40,880
We split participants

273
00:11:40,880 --> 00:11:42,440
into eight different subgroups,

274
00:11:42,440 --> 00:11:43,920
targeting many of the areas

275
00:11:43,920 --> 00:11:45,840
that we thought were most pressing.

276
00:11:45,840 --> 00:11:47,570
Things from defining a cohesive set

277
00:11:47,570 --> 00:11:49,100
of science mission drivers,

278
00:11:49,100 --> 00:11:50,740
to fitting potential improvements

279
00:11:50,740 --> 00:11:52,530
to instrument performance calibration,

280
00:11:52,530 --> 00:11:53,720
to trying to better understand

281
00:11:53,720 --> 00:11:56,010
how Earth-based effects like caloric lines

282
00:11:56,010 --> 00:11:57,410
will complicate the process.

283
00:11:58,800 --> 00:11:59,910
And our objective was,

284
00:11:59,910 --> 00:12:02,470
to synthesize the results
from those different subgroups

285
00:12:02,470 --> 00:12:05,690
and then recommend to NASA
headquarters and the NSF,

286
00:12:05,690 --> 00:12:08,750
the best ground-based
observing program architecture

287
00:12:08,750 --> 00:12:10,310
and corresponding research

288
00:12:10,310 --> 00:12:12,650
and development program
that we could think of,

289
00:12:12,650 --> 00:12:15,230
to try and achieve the goal
of measuring the masses

290
00:12:15,230 --> 00:12:18,180
of Earth analog planets
around sun-like stars

291
00:12:18,180 --> 00:12:21,003
in time for mission concepts
such as HabEX or LUVOIR.

292
00:12:22,370 --> 00:12:25,710
So, we identified a
number of success criteria

293
00:12:25,710 --> 00:12:27,563
that the must that are listed here,

294
00:12:28,830 --> 00:12:31,300
and that are things like making sure

295
00:12:31,300 --> 00:12:32,880
that we learn from the current instruments

296
00:12:32,880 --> 00:12:34,620
that are in operation right now,

297
00:12:34,620 --> 00:12:36,890
and setting up surveys in the current era

298
00:12:36,890 --> 00:12:40,110
to start better characterizing
stellar variability.

299
00:12:40,110 --> 00:12:41,470
And then also, some additional things

300
00:12:41,470 --> 00:12:42,870
that would be nice, if possible,

301
00:12:42,870 --> 00:12:45,327
the wants that are on the
bottom part of the slide.

302
00:12:45,327 --> 00:12:46,170
And these are things like,

303
00:12:46,170 --> 00:12:48,060
having a relatively low estimated cost,

304
00:12:48,060 --> 00:12:50,300
'cause it turns out
you can't just ask NASA

305
00:12:50,300 --> 00:12:51,480
to build six new techs

306
00:12:51,480 --> 00:12:53,803
to achieve your science
goals, they don't like that.

307
00:12:55,040 --> 00:12:57,480
So with all those musts and wants in mind,

308
00:12:57,480 --> 00:12:59,300
our working groups started contemplating

309
00:12:59,300 --> 00:13:00,950
what's some notional architectures

310
00:13:00,950 --> 00:13:02,420
for detecting Earth analogues

311
00:13:02,420 --> 00:13:04,870
in the late 2030s might look like.

312
00:13:04,870 --> 00:13:06,390
The ultimate goal was to see

313
00:13:06,390 --> 00:13:07,840
whether we could come up with one

314
00:13:07,840 --> 00:13:09,330
or more observing strategies

315
00:13:09,330 --> 00:13:10,950
that serve as an existent proof

316
00:13:10,950 --> 00:13:13,760
for the first two columns
of the RV uncertainty chart

317
00:13:15,360 --> 00:13:16,790
that John showed you yesterday,

318
00:13:16,790 --> 00:13:19,190
the effects of the
observing facility itself

319
00:13:19,190 --> 00:13:21,550
and the RV spectrograph.

320
00:13:21,550 --> 00:13:23,230
Though, for right now,
we're going to assume

321
00:13:23,230 --> 00:13:25,190
that the third column on that chart,

322
00:13:25,190 --> 00:13:27,663
the stellar variability
part, is mitigated.

323
00:13:28,520 --> 00:13:31,130
But even when I'm putting
stellar variability,

324
00:13:31,130 --> 00:13:33,400
the decisions of how to
observe your targets,

325
00:13:33,400 --> 00:13:34,700
of course, depends on what types

326
00:13:34,700 --> 00:13:36,730
of stars you're interested in.

327
00:13:36,730 --> 00:13:38,440
And so, with much help
from Eric Mammon Jack

328
00:13:38,440 --> 00:13:39,440
and Chris Stark,

329
00:13:39,440 --> 00:13:42,500
we started by assembling
a list of all of the stars

330
00:13:42,500 --> 00:13:44,920
that LUVOIR or HabEX or Starshade,

331
00:13:44,920 --> 00:13:47,980
might consider as targets
in the 2030s and 40s.

332
00:13:47,980 --> 00:13:49,723
And then we called down
to select the stars

333
00:13:49,723 --> 00:13:53,510
that were also amenable to
precision RV observations.

334
00:13:53,510 --> 00:13:56,680
So that meant selecting stars
cooler than 6200 Kelvin,

335
00:13:56,680 --> 00:13:58,990
that they would have a sufficient number

336
00:13:58,990 --> 00:14:02,630
of absorptions line to use
for measuring miniscule,

337
00:14:02,630 --> 00:14:05,010
or nine centimeter per second RV shifts

338
00:14:05,010 --> 00:14:06,350
that Scott talked about.

339
00:14:06,350 --> 00:14:08,250
And then also start a rotational velocity

340
00:14:08,250 --> 00:14:10,460
is less than about five
kilometers per second,

341
00:14:10,460 --> 00:14:11,810
so that those absorption lines

342
00:14:11,810 --> 00:14:13,300
are not rotationally brought in,

343
00:14:13,300 --> 00:14:17,090
which decreases the RV information
content of the spectra.

344
00:14:17,090 --> 00:14:18,840
The resulting list is plotted here,

345
00:14:18,840 --> 00:14:21,830
we call this our primary
list or green star list,

346
00:14:21,830 --> 00:14:24,110
and it's comprised about 101 stars.

347
00:14:24,110 --> 00:14:25,200
And as you can see,

348
00:14:25,200 --> 00:14:26,960
we're not really target-limited

349
00:14:26,960 --> 00:14:28,290
and the targets that we end up with,

350
00:14:28,290 --> 00:14:30,460
are spread pretty evenly across the sky,

351
00:14:30,460 --> 00:14:32,493
both in terms of array and declination.

352
00:14:33,980 --> 00:14:36,360
We then designed eight
notional architectures

353
00:14:36,360 --> 00:14:39,050
that form a basis set of
different telescope apertures

354
00:14:39,050 --> 00:14:41,340
and instrument types that
we envision being available

355
00:14:41,340 --> 00:14:43,470
to the community in the 2030s.

356
00:14:43,470 --> 00:14:45,630
Each architecture contains
a hand-crafted set

357
00:14:45,630 --> 00:14:47,720
of system properties,
ranging from the size

358
00:14:47,720 --> 00:14:48,830
of the telescopes,

359
00:14:48,830 --> 00:14:50,110
to the light coverage,

360
00:14:50,110 --> 00:14:52,793
observing cadence and analysis process,

361
00:14:54,530 --> 00:14:55,363
along with things like

362
00:14:55,363 --> 00:14:57,930
the RV single measurement
precision among others.

363
00:14:57,930 --> 00:15:00,070
Now while all of the architectures
were simulated in detail,

364
00:15:00,070 --> 00:15:02,110
I'm gonna focus on just one for right now

365
00:15:02,110 --> 00:15:03,620
to give you a better sense of our design

366
00:15:03,620 --> 00:15:05,270
and analysis process that we use.

367
00:15:05,270 --> 00:15:07,370
Architecture one, which is comprised

368
00:15:07,370 --> 00:15:09,503
of six, 2.4 meter telescopes.

369
00:15:11,360 --> 00:15:13,270
Those telescopes are
spread across the globe

370
00:15:13,270 --> 00:15:15,230
to provide both north-south coverage

371
00:15:15,230 --> 00:15:17,230
and even longitudinal coverage,

372
00:15:17,230 --> 00:15:18,540
at least to the best of our abilities,

373
00:15:18,540 --> 00:15:20,850
based on current observing life.

374
00:15:20,850 --> 00:15:22,460
Because of the inputs that we receive

375
00:15:22,460 --> 00:15:24,220
from the stellar variability community

376
00:15:24,220 --> 00:15:26,120
emphasize the need for
high-cadence observing

377
00:15:26,120 --> 00:15:27,920
in order to disentangle the variety

378
00:15:27,920 --> 00:15:29,910
of signals caused by the star itself,

379
00:15:29,910 --> 00:15:32,020
we spread the telescope
as evenly as possible

380
00:15:32,020 --> 00:15:33,140
across the globe to ensure

381
00:15:33,140 --> 00:15:35,900
that even if a weather
system shuts down one site

382
00:15:35,900 --> 00:15:37,290
for a few days in a row,

383
00:15:37,290 --> 00:15:39,210
the other sites will be
able to pick up the slack

384
00:15:39,210 --> 00:15:42,280
and try to get nightly
observations of our stars.

385
00:15:42,280 --> 00:15:43,800
- [Coordinator] Jenn, we're at 15 minutes.

386
00:15:43,800 --> 00:15:45,400
- [Jenn] Okay, spread also ensures

387
00:15:45,400 --> 00:15:47,430
that if we do start to see
signs of an Earth analog

388
00:15:47,430 --> 00:15:49,600
in the data that we've taken
with one of the facilities,

389
00:15:49,600 --> 00:15:51,580
we can ensure that that
same signal shows up

390
00:15:51,580 --> 00:15:52,490
in other facilities

391
00:15:52,490 --> 00:15:54,290
and it's not something
based on the instrument

392
00:15:54,290 --> 00:15:55,523
or the telescope itself.

393
00:15:57,070 --> 00:15:58,630
Each of the telescope is linked

394
00:15:58,630 --> 00:16:03,340
with a kind of notional
next-generation EPRV instrument

395
00:16:03,340 --> 00:16:05,240
of high resolution, broad wavelength,

396
00:16:05,240 --> 00:16:06,900
grasping an instrumental noise floor

397
00:16:06,900 --> 00:16:08,450
of 10 centimeters per second,

398
00:16:08,450 --> 00:16:10,090
as well as a small solar telescope

399
00:16:10,090 --> 00:16:12,030
so it can look at the sun during the day

400
00:16:12,030 --> 00:16:14,003
and produce useful science 24/7.

401
00:16:15,070 --> 00:16:17,510
We created a simulated dispatch scheduler

402
00:16:17,510 --> 00:16:20,353
that ties information about the facility

403
00:16:20,353 --> 00:16:22,640
and the instrument, as
well as the target lists

404
00:16:22,640 --> 00:16:24,690
and things like historical
weather patterns,

405
00:16:24,690 --> 00:16:28,490
and then produces a
decade-long simulated survey,

406
00:16:28,490 --> 00:16:31,290
observing loss of
information on the targets

407
00:16:31,290 --> 00:16:34,290
when they were observed and
what RV precision they achieved.

408
00:16:35,810 --> 00:16:37,100
And after the simulations are run,

409
00:16:37,100 --> 00:16:38,350
we quantify our ability

410
00:16:38,350 --> 00:16:41,460
to recover an Earth analog
around each target star.

411
00:16:41,460 --> 00:16:43,660
Now to do so, we calculate
the detection significance

412
00:16:43,660 --> 00:16:46,770
of an Earth analog and
each star simulated RV data

413
00:16:46,770 --> 00:16:48,960
in the absence of any stellar activity.

414
00:16:48,960 --> 00:16:50,990
That is, if there were Earth analog

415
00:16:50,990 --> 00:16:52,790
around each star that we observe,

416
00:16:52,790 --> 00:16:55,120
and we managed to mitigate
all of the variability

417
00:16:55,120 --> 00:16:56,900
from the stars observed in our data,

418
00:16:56,900 --> 00:17:00,270
how significant would our
detection of that planet be?

419
00:17:00,270 --> 00:17:01,790
The results for architecture one,

420
00:17:01,790 --> 00:17:03,520
are shown here with a red line marking

421
00:17:03,520 --> 00:17:04,580
the 10 sigma level

422
00:17:04,580 --> 00:17:06,010
that we believe to be necessary

423
00:17:06,010 --> 00:17:08,710
for correctly interpreting
atmospheric spectra.

424
00:17:08,710 --> 00:17:10,380
And as you can see, and
it's ideal in a case

425
00:17:10,380 --> 00:17:12,300
of there being no stellar activities,

426
00:17:12,300 --> 00:17:15,940
all of the stars do pass
that 10 sigma threshold.

427
00:17:15,940 --> 00:17:17,790
We repeated this for each
of the other architectures,

428
00:17:17,790 --> 00:17:19,280
like showed so.

429
00:17:19,280 --> 00:17:21,580
And you can see here that in most cases,

430
00:17:21,580 --> 00:17:22,850
the architecture is achieved

431
00:17:22,850 --> 00:17:24,210
at 10 sigma detection level,

432
00:17:24,210 --> 00:17:26,910
on almost all of the stars they surveyed.

433
00:17:26,910 --> 00:17:28,010
Now, this is not the moment

434
00:17:28,010 --> 00:17:30,230
that any of these architectures
are guaranteed to work,

435
00:17:30,230 --> 00:17:31,780
there is still a variety of things

436
00:17:31,780 --> 00:17:33,660
that should be simulated in more detail,

437
00:17:33,660 --> 00:17:36,160
and the assumption that
we'll be able to 100% correct

438
00:17:36,160 --> 00:17:39,190
for the stellar variability, is a big one.

439
00:17:39,190 --> 00:17:40,890
But what this does say is that,

440
00:17:40,890 --> 00:17:43,340
we think there are paths forward.

441
00:17:43,340 --> 00:17:44,173
And so, with knowledge

442
00:17:44,173 --> 00:17:46,530
that there are multiple potential
ways to address our must

443
00:17:46,530 --> 00:17:48,930
and to close the Kepner-Tregoe matrix

444
00:17:48,930 --> 00:17:51,470
that we use during these
first class architectures,

445
00:17:51,470 --> 00:17:53,840
we did get an excellent
reason to push forward.

446
00:17:53,840 --> 00:17:55,150
But, as mentioned earlier,

447
00:17:55,150 --> 00:17:56,550
these simulations assumed first

448
00:17:56,550 --> 00:17:57,940
that we could reach a systematic noise

449
00:17:57,940 --> 00:17:59,600
for our 10 centimeters per second

450
00:17:59,600 --> 00:18:01,030
and that we could mitigate all

451
00:18:01,030 --> 00:18:03,730
of the stellar variability
inherent in our data,

452
00:18:03,730 --> 00:18:05,150
and neither of those are things

453
00:18:05,150 --> 00:18:06,910
we can actually do right now.

454
00:18:06,910 --> 00:18:08,330
So our crucial next step,

455
00:18:08,330 --> 00:18:11,090
is both to increase the
realism of the simulations,

456
00:18:11,090 --> 00:18:13,570
and to work on achieving those precision

457
00:18:13,570 --> 00:18:17,000
and variability steps
forward in the real world,

458
00:18:17,000 --> 00:18:18,400
both of which we think can be addressed,

459
00:18:18,400 --> 00:18:20,993
to be dedicated and sustained R&A program.

460
00:18:21,890 --> 00:18:23,530
Now when thinking about
stellar variability,

461
00:18:23,530 --> 00:18:25,740
we need to develop a much
more nuanced understanding

462
00:18:25,740 --> 00:18:28,890
of the many phenomena that
produce signals within RV data.

463
00:18:28,890 --> 00:18:30,680
This includes delving into topics like,

464
00:18:30,680 --> 00:18:33,160
how does convection interact
with magnetic fields,

465
00:18:33,160 --> 00:18:36,480
and how to stellar surface
phenomena drive to sun as a star,

466
00:18:36,480 --> 00:18:38,080
are the variations?

467
00:18:38,080 --> 00:18:40,670
We're also working to
understand line formation

468
00:18:40,670 --> 00:18:42,670
and behavior to a level of detail,

469
00:18:42,670 --> 00:18:43,970
necessary to create an exhibition

470
00:18:43,970 --> 00:18:46,740
of physically, motivated stellar models.

471
00:18:46,740 --> 00:18:48,660
And we're also highlighting the importance

472
00:18:48,660 --> 00:18:50,830
of putting more time and
energy into reaching out

473
00:18:50,830 --> 00:18:53,280
to the Heliophysics side of astronomy,

474
00:18:53,280 --> 00:18:54,910
because these are folks
who've been thinking

475
00:18:54,910 --> 00:18:58,180
about how variability shows
up on stars like the sun,

476
00:18:58,180 --> 00:18:59,047
for many decades now

477
00:18:59,047 --> 00:19:01,520
and we should take advantage that.

478
00:19:01,520 --> 00:19:04,740
But, also we need to improve
our approach-to-data analysis.

479
00:19:04,740 --> 00:19:05,573
And this includes,

480
00:19:05,573 --> 00:19:08,960
both how we go from office
files, to measured RVs,

481
00:19:08,960 --> 00:19:10,500
and then how we go from measured RVs,

482
00:19:10,500 --> 00:19:12,690
to fitted Kepler in signals.

483
00:19:12,690 --> 00:19:13,650
And so this is a screenshot

484
00:19:13,650 --> 00:19:16,960
of only part of the immensely
detailed rubric constructed

485
00:19:16,960 --> 00:19:19,520
by our pixels to planets subgroup

486
00:19:19,520 --> 00:19:21,060
that details some best practices

487
00:19:21,060 --> 00:19:22,550
and necessary improvements

488
00:19:22,550 --> 00:19:25,573
for new RV data reduction
pipelines in the coming years.

489
00:19:26,530 --> 00:19:27,940
And, of course, the
instruments themselves,

490
00:19:27,940 --> 00:19:29,500
will also need to be pushed even further

491
00:19:29,500 --> 00:19:31,200
than what's been achieved so far.

492
00:19:31,200 --> 00:19:33,220
So we'll need to work on
advancing things like,

493
00:19:33,220 --> 00:19:35,820
wavelength calibration and
improve fiber coupling.

494
00:19:35,820 --> 00:19:37,180
We're also thinking about ways

495
00:19:37,180 --> 00:19:41,103
to further auxiliary technologies
like visible light AO.

496
00:19:42,390 --> 00:19:44,180
Now, those are just a
few of the many topics

497
00:19:44,180 --> 00:19:45,450
that were discussed in the working group,

498
00:19:45,450 --> 00:19:46,740
over our nine-month sprint

499
00:19:46,740 --> 00:19:49,960
to try and gather the
field's existing knowledge

500
00:19:49,960 --> 00:19:51,820
and best practices for how to move forward

501
00:19:51,820 --> 00:19:53,880
towards detecting Earth analog.

502
00:19:53,880 --> 00:19:54,770
We're in the final stages

503
00:19:54,770 --> 00:19:56,780
of preparing our written report right now,

504
00:19:56,780 --> 00:19:57,613
and once that's done,

505
00:19:57,613 --> 00:20:00,210
it will appear both on the
website that's listed here

506
00:20:00,210 --> 00:20:03,720
and come out via Eric's
exoplanet email services,

507
00:20:03,720 --> 00:20:05,600
you should keep your eyes out for that.

508
00:20:05,600 --> 00:20:06,610
And then of course,

509
00:20:06,610 --> 00:20:09,100
you did look at this
new ROSES solicitation

510
00:20:09,100 --> 00:20:10,890
that was announced by Doug yesterday.

511
00:20:10,890 --> 00:20:12,290
And if you wanna get involved

512
00:20:12,290 --> 00:20:13,370
in helping us with this effort,

513
00:20:13,370 --> 00:20:15,550
make sure to apply for that.

514
00:20:15,550 --> 00:20:16,383
So with that, I will say,

515
00:20:16,383 --> 00:20:17,540
thank you so much for your attention.

516
00:20:17,540 --> 00:20:18,890
And is there any questions

517
00:20:18,890 --> 00:20:21,380
that Scott and I can try to answer?

518
00:20:21,380 --> 00:20:22,830
- [Coordinator] Thanks Jenn, thanks Scott.

519
00:20:22,830 --> 00:20:24,509
We are out of time.

520
00:20:24,509 --> 00:20:26,030
But because this is so important,

521
00:20:26,030 --> 00:20:27,210
I will ask at least one question

522
00:20:27,210 --> 00:20:28,733
that has come up.

523
00:20:30,220 --> 00:20:33,010
These EPRV measurements
you're talking about,

524
00:20:33,010 --> 00:20:35,840
do they require a brand-new
purpose built telescope?

525
00:20:35,840 --> 00:20:37,330
Or could an instrument be made

526
00:20:37,330 --> 00:20:39,040
and installed on current platforms?

527
00:20:39,040 --> 00:20:41,400
And I think that was embedded
in your presentation,

528
00:20:41,400 --> 00:20:42,950
but maybe you just clarify the difference

529
00:20:42,950 --> 00:20:46,630
between brand-new versus
adapting existing?

530
00:20:46,630 --> 00:20:48,100
- [Jenn] Absolutely,
and so one of the things

531
00:20:48,100 --> 00:20:50,400
that was discussed, are
both of those approaches.

532
00:20:50,400 --> 00:20:52,230
We think both of them have merits.

533
00:20:52,230 --> 00:20:54,030
I think that what need to happen is,

534
00:20:54,030 --> 00:20:55,900
site-visits to existing facilities

535
00:20:55,900 --> 00:20:58,820
to make sure that they are
compatible with the abilities.

536
00:20:58,820 --> 00:21:00,390
We would want like having an instruments

537
00:21:00,390 --> 00:21:01,400
that talk to each other

538
00:21:01,400 --> 00:21:03,410
and try to network their
observing strategies

539
00:21:03,410 --> 00:21:04,350
throughout the night.

540
00:21:04,350 --> 00:21:06,380
And so, I think there is
certainly a possibility,

541
00:21:06,380 --> 00:21:09,350
but saying an official
yes no for each facility

542
00:21:09,350 --> 00:21:10,270
that exists right now,

543
00:21:10,270 --> 00:21:11,515
would require a more in-depth..

544
00:21:11,515 --> 00:21:13,369
(phone ringing)