Decadal Survey: 2010

Decadal eTownhall meeting is about to start, and apparently some astronomy departments “forgot” to sign up for a webcast slot, so, like modern finance, those of us with the millisecond time advantage will leverage the advantage.
For the rest, here is the liveblog of the webcast, or find a tweet with no delay
(Derek is tweeting).

First question, no doubt, will be: “Roger, WHO chose the Muzak for the people waiting on hold…?”

and we’re off…

apparently proceeds of the popcorn sales at the webcast sites will fund the new projects… hah!
Oh, that was not a joke.

We start off by emphasizing, again, that consensus and everyone being onboard is critical.
No whinging, please.

Hm: “New World, New Horizon” – is the title – wonder if that is a hint.

Roger is on.

Here is the NAP draft of the report
PDF is free, 81 Mb – will the rest of you please pause your download for 3 more minutes…

Ok, Exec Summary:

“The priority science objectives chosen by the survey committee for the decade 2012-2021 are searching for the first stars, galaxies, and black holes; seeking nearby habitable planets; and advancing understanding of the fundamental physics of the universe.”

Ok, here are the priorities:


  1. Wide Field Infrared Survey Telescope – WFIRST
    Hmm. ExoplanetsDark Energy as priority, then dark energy exoplanets – from same instrument.
    Sounds like a near-IR SNAP with a coronographa near-IR wide field “New Worlds Observer”? – nope, exoplanet microlensing…

  2. Explorer Progam (small/medium sized rapid(ish) missions)
  3. Laser Interferometer Space Antenna!
  4. IXO – next generation large x-ray observatory

    EDIT: the decadal panel asked me to clarify that the ranking of large and medium categories is independent – that is to say, the medium sized development programs below are ranked relative to each other, not to the large programs

    • New Worlds Technology Development Program–a competed program to lay the technical and scientific foundation for a future mission to study nearby Earth-like planets
    • Inflation Probe Technology Development Program–a competed program designed to prepare for a potential next-decade cosmic microwave-background mission to study the epoch of inflation.


    and the winner is…

  1. LSST
  2. Mid-Scale Innovations – ie instrumentation for existing telescopes
  3. Giant Segmented Mirror Telescope – ie TMT or GMT – Congress has tipped the scales to TMT – recommend 25% share of $1+G telescope. Committee won’t choose, NSF please do so, with subtle input from Congress, no doubt.
  4. ACTA – Atmospheric Cherenkov Array

EDIT: as with space – the mid scale ground project is ranked independent of the large

  1. CCAT – Atacama 25m submillimeter telescope

Strong recommendation buried to go in on JAXA’s SPICA mission
Hmm, is SPICA really W-FIRST?
Ah, Caltech wants SPICA partnership.

No, WFIRST is “JDEM-Omega” – surprise.
With exoplanet microlensing (congratulation, you know who) as the secondary science, and W, W’ as primary.
Feed crashed as Roger said this…
Figure 7.12 – page 199 – WFIRST eats everyone’s lunch in mid-decade.
[Pages 201 and 202 have the analogous "sandcharts" for NSF and DoE.
I don't see DoE managing the funding "hump" at the end of the decade.
Looks like something will have to stretch or give.]

$1.6 billion estimated cost for WFIRST. Seems optimistic.
This is the flagship mission.
Wonder what happened to optical JDEM which was the last priority…

PS: EUCLID – something to contemplate.
More later.
More: Wide Field Imager in Space for Dark Energy and Planets

Big expansion of Explorer mission proposed.

LISA ranked third – contingent on LISA Pathfinder in 2012, and ESA keeping LISA as a flagship with 50% cost share.
That’d be post-2020.

IXO is 4th. Sorry.
Committee estimates cost a $5G and medium-high risk.
Recommend $180M for tech for the decade.
IXO is dead.

Tech development for Exoplanet finder – for 2020 – $1-200M for the decade;
and tech development for Inflation Probe.

So no exoplanet imager till next decade.

Nature News take on the report

Good set of discussion by Julianne at CV
overview, space and ground

Ok: x-ray astronomy is in deep doo-doo (that’s a technical term).

Exoplanets are subtly undermined, in my ever so humble opinion.
Exoplanet microlensing is neat, and a good secondary science for JDEMWFIRST, but it is somewhat orthogonal to where the main effort has been.
Exoplanet imaging/spectroscopy which was going clappers is slammed to a halt, in the US, by this.
Opportunity for ESA to regain the lead in exoplanets, methinks.

My other thought is that DoE really did a good job buying into the system.
Decadal reports used to be NSF/NASA, DoE bought into this one.

Everything is budget constrained.
They recommend what to cut if budgets tighten.

Some good recommendations on smaller programs – new instruments, computation, theory, databases – these are small budget items, but need to be anchored in the report so as to not be eaten by the $ starved big projects.

I didn’t even notice, but SIM was not even mentioned.

SIM is dead.
SIMlite also.

Ah, page 34 – footnote:
“Two space missions recommended in the 2001 decadal survey Astronomy and Astrophysics in the New Millennium–namely ARISE and EXIST–and one recommended by the 1991 The Decade of Discovery in Astronomy and Astrophysics survey, SIM, do not appear in this survey’s priorities. The goals of ARISE have been largely subsumed by JAXA’s VSOP-2 project and the SAMURAI proposal. EXIST and SIM (now SIMLite) are not included in the recommended program for the decade, following the committee’s consideration of the strengths of competing compelling scientific opportunities and the highly constrained budget scenarios described in this report.”

Also page 168 footnote:
“In considering possible exoplanet missions for the next decade, the committee gave serious consideration to SIMLite but decided against recommending it. SIMLite is technically mature and would provide an important new capability (interferometry). Through precision astrometry it could characterize the architectures of 50 or so nearby planetary systems, provide targets for future imaging missions, and carry out other interesting astrophysics measurements. However, the committee considered that its large cost (appraised by the CATE process at $1.9 billion) and long time to launch (estimated at 8.5 years) make it uncompetitive in the rapidly changing field of exoplanet science. The planetary architecture science can be more efficiently carried out by the committee’s exoplanet strategy involving Kepler, WFIRST, and the ground-based program. The role of target-finding for future direct-detection missions, one not universally accepted as essential, can be done at least partially by pushing ground- based radial-velocity capabilities to a challenging but achievable precision below 10 centimeters per second. Finally, the ancillary astrophysics promised by SIMLite was not judged to be competitive.”

Radio is also downplayed and the panel just said US is NOT going in on SKA.
Might play along with incremental SKA developments

They like anything that would look at 21cm H at z > 10

Not to undermine our community consensus, or anything, but would the dark energy folks please stop moving the playing field (moving the goal is ok) – the JDEM/SNAP concept, in particular, is really stretching things out from what they claimed just a few years ago – time to deliver or get out of the way.
In my ever so humble opinion.


  1. #1 ct
    August 13, 2010

    - Glad dark energy was supported

    - “With exoplanet microlensing (congratulation, you know who) …”


    The implicit decision to put microlensing-driven exoplanet science (even if it’s just piggy-backing on WFIRST) on the front burner as opposed to exoplanet imaging/spectroscopy is just baffling to me and a triumph of marketing. Yes, we’ve just invited the Europeans to overtake the US on exoplanet science.

    I guess the only silver lining here is that you can still push the boundaries with imaging/spectroscopy by building better coronographic masks/Lyot stops, better detectors, slightly better AO systems, and by using better image processing techniques. Or in other words, sometime in the next 5–10 years we’ll start getting spectra of Jupiter-like planets out the wazoo or even image RV planets, and all this will be at a fraction of the cost of a space mission.

  2. #2 The AstroDyke
    August 13, 2010

    Am I reading the report correctly, that the committee felt that exoplanet probes and IXO were not technologically ready for prime time? And that in the exoplanet case, the field is moving so fast that they were afraid of approving a “Gravity Probe B” style mission, that would be obsolete before launch?

  3. #3 Steinn Sigurdsson
    August 13, 2010

    Yeah – I’m thinking they should maybe have worried a bit more about JDEM becoming the next SIM.

  4. #4 nick
    August 13, 2010

    what’s your take on theory’s immediate future? Some minimal mention in the report, but nothing too concrete I thought. I’ve read mixed thoughts from others and I’m not very used to reading between the lines on stuff like this…

  5. #5 Steinn Sigurdsson
    August 13, 2010

    I think theory will limp along.
    It is such a small piece of the action that the main danger is that it becomes collateral damage, as it has in the past.

    The report did theory some favour by a) putting down a marker noting its importance and b) recommending some increase in funding

    Don’t know if there’ll be any new money, but theory can fight back against attempts to take funding to cover overruns on large projects by pointing to it being ranked in the decadal as a priority.

    Computation will also be interesting – the shortage there, from my perspective, is people, not hardware, though the hardware of course needs constant new money, not a one time injection.

  6. #6 Dave
    August 13, 2010

    The decision to put microlensing-driven exoplanet on the front burner is hardly a surprise. Space-based microlensing had a strong recommendation from the ExoPlanet Task Force. The ExoPTF and the exoplanet community certainly prefers SIM, but SIM costs 10x as much as an exoplanet program on JDEM. The ExoPTF judged the exoplanet science from very small imaging spectroscopy missions as not very good, and large missions are very expensive. Plus. the technology is not ready. Space-microlensing has a good science case and is the only thing that can fit in the budget once JDEM is ranked above SIM.

    Anyone who thinks that the Europeans will overtake the US on exoplanet science hasn’t been paying attention to what is going on with ESA. They rejected DARWIN, and are considering PLATO, which is an astroseismology mission being sold as an exoplanet mission. If they actually fly the proposed mission, it will find little that Kepler won’t find first. And, if NASA selects a transit mission like TESS in the current Explorer round, PLATO will be completely scooped, unless they total change their design.

  7. #7 ct
    August 13, 2010

    I don’t disagree that microlensing provides a useful probe of exoplanet phase space, or that it enhances the merit of a JDEM-ish mission like WFIRST to the point of making it a clear superior to SIM (which it does). Rather, I just don’t think that it should drive exoplanet science.

    The goal is to directly image an extrasolar Earth-mass and demonstrate (via spectroscopy) that it has biosignatures. But we aren’t going to be imaging an Earth in the galactic center, so lensing just strikes me as a bit tangential to achieving this goal.

    Lensing helps fill out the phase space for exoplanets, but I don’t see how it could do much better than lower-jitter and longer-baseline RV surveys (to detect low-mass planets and massive long-period planets, respectively) plus a bleeding-edge high-contrast imaging project (think GPI on steroids). Since we will need RV to detect an Earth around a nearby star anyway (either on its own or via transit follow up) and imaging/spectroscopy of an Earth is the goal, it makes a lot more sense to focus on the latter.

    To me, it would have been more reasonable trying to maximize what can be done with RV and imaging/spectroscopy from the ground (the “Europeans taking the lead” was more focused on the ground, not space). In the first case, a laser frequency comb like HARPS except better and used on a 8-10m class telescope or dedicated smaller facility just for RV. In the second case, support the instrumentation for TMT and GMT for a GPI-like project; put a well-corrected subaperture/vector vortex coronograph system on Keck, etc. There are a lot of low/medium-cost ideas that could work that when combined would yield an effective demographic probe of planetary systems and specifically target systems we could later revisit when it’s time to do a TPF-like mission, unlike lensing.

    Proposals like these could soak up a lot of the mid-range 10–100m funding, but I would have felt better had the panel specifically recommended projects along these lines. Perhaps this is buried somewhere in the text of the review, but it certaintly wasn’t emphasized in the presentation.

  8. #8 Dave
    August 14, 2010

    Dear tc,
    Perhaps you should read some of the microlensing white papers or the ExoPTF report. RV is not at all competitive with ground-based microlensing for planets beyond the snow line, and space-based microlensing is sensitive to planets down to to 0.1 Earth masses. It will be a big challenge for RV to get down to 100 Earth masses beyond the snow line. If, as core accretion theory suggests, most planetary systems are like our own but with Neptunes in place of Jupiter and Saturn, RV and a GPI-like project will see nothing. In terms of a statistical census of planet properties in orbits beyond what Kepler can see, RV isn’t likely to get within a factor of 100 of the mass limit of a space-based microlensing survey.

    It would be great if RV could detect an Earth around a nearby star, but this seems pretty unlikely. My understanding is that some of the RV folks are wondering if some of the Geneva team’s multi-planet systems are actually RV’s due to spots and stellar rotations. It does seem a bit odd to have so many systems in which several planets contribute a similar RV signal. They can be modeled as multiplanet systems, but the exoplanet interpretation would seem more secure if the signal was dominated by a clear Keplerian orbit.

    In any case, the ground based programs are not in direct competition with the space missions.

  9. #9 Steinn Sigurdsson
    August 14, 2010

    Hey Dave, and congratulations.

    Microlensing is great, and it has a nice discovery space where it will give good sampling – but followthrough on individual objects is near impossible and it doesn’t grab the public attention the way direct detection nearby does.

    My worry is that the committee “threw the dog a bone” – putting exoplanet microlensing as secondary to irJDEM is supposed to take care of the exoplanet community, except it doesn’t, it slams to a halt major lines of investigation where we were making much more rapid progress than anticipated – the radial velocity, transit and direct imaging efforts are all succeeding, and by omission the panel shuts the door on making a push on these efforts.
    Which I think is very damaging.

    My viewpoint may be somewhat distorted by my irritation with the JDEM crowd, they bumped everyone in Beyond Einstein because they claimed to be mission ready, and here we are, years later, with powerpoints and another mission concept – with no flight this decade.

    ESA has an interesting window here – fly Euclid quickly and scoop the primary science of WFIRST, and ramp up Cosmic Visions for a super-earth detector and be going before NASA gets to any sort of TPF concept.
    Depends on politics – like whether the ESA nations go for austerity and stimulus, and whether ESA can react that quickly.

    In the meantime I weep for poor LISA.

  10. #10 Steinn Sigurdsson
    August 14, 2010

    @ct – I don’t think HARPS has a lasercomb yet.
    Codex has been under development but I haven’t heard they did a science run.

  11. #11 ct
    August 14, 2010


    You’re missing the point (and yes I *am* familiar with the microlensing white papers and all too familiar with your talks). We aren’t going to be able to image or do spectroscopy for an Earth in the galactic center. You can’t characterize the orbital properties, luminosity, mass, atmospheric chemistry, etc etc like you can by combining RV, transits, and direct imaging. If the goal is to generally place the solar system in context by identifying broad statistical trends for the frequency of objects with a given mass over a given range in separation in a certain region of the galaxy then microlensing is not just fine but essential. But if you really want to start to do serious comparative exoplanetology of systems we can study later, let alone study systems that may ultimately contain a detectable Earth-like (not just Earth *mass*) planet, then it’s not. Still, congratulations: WFIRST looks like a very good project.

    For what it’s worth, the yield of direct imaging surveys is probably going to increase with NICI and is going to dramatically improve with GPI and SPHERE. The inner working angle for these systems will be far smaller, which not only pushes the sensitivity of exoplanet detection closer to the snow line but puts a lot of more distant, extremely young systems (with wider, luminous planets) into play.

    @Steinn, your point about “throwing the exoplanet community a bone” was exactly the one I was trying to make. Sorry my text regarding HARPS there was garbled. I meant that we need to do better than something like HARPS, and this could be achieved with a laser-frequency comb.

  12. #12 Steinn Sigurdsson
    August 14, 2010

    Ok, I agree that lasercombs are a promising way forward.
    Very very promising and moving very quickly.

  13. #13 Dave
    August 15, 2010

    Dear ct,
    Sorry for misreading your post. The main selling point of the space microlensing mission is its relative cheapness and technological readiness. A standalone MPF mission is perhaps 1/4 the cost of WFIRST. It does give an important part of the exoplanet picture that can’t be obtained by other methods, but we sure aren’t going to find TPF targets.

    While there are serious difficulties with trying to measure RV signals of Earths, that doesn’t mean that it isn’t worth trying. The laser combs will be great, but I suspect that the key is figuring out better ways to characterize the RV signals intrinsic to the star.

  14. #14 Rob
    August 18, 2010

    I see the recommendations as a wonderful public relations success
    for DOE and the DE crowd. But really, how important is it to move
    by one the decimal point on the “determination” of DE. After
    spending 1,600,000,000+ dollars on WFIRST, there will still be
    way too many theories that will match the data.

    Why not fly WFIRST-like missions from long-duration balloon
    flights: this stuff does not need to be done from space.


    Talking about new technology: that they let SOFIA continue,
    that’s the biggest waste of astronomical money ever. The things
    costs as much as a space mission (3.75 G$) but observes only 9%
    of the time (33 days or ~5x less than a ground based
    telescope). Long duration balloon flights can last easily 30 days
    and they fly already 2 meter telescopes and cost just a couple of
    million dollars (check the 22 July 2010 edition of Nature)


    Then there is the curious emphasis on “Cosmic Paleontology:” this
    is a new version of “Galactic Archeology.” The difference is that
    the former is done at large redshift where not much is learned
    about the physics of the first stars (except that they shine and
    explode: revolutionary indeed). The latter is done in the Milky
    Way (with SIMLite/Gaia & ground-based spectroscopy) and would
    allow us to actually learn something about these stars: how they
    explode, what elements they made (via the abundance patterns in
    the observable second generation stars) and when this actually
    happened (via the ages of the oldest stars in the MW). Yes,
    after Gaia & SIMLite, age determination of stars will be
    _accurate_ to +/- 1% or so: much more useful that the model
    dependent _precision_ obtained via the cosmological route. And
    then there are all the other astrophysics that SIM would enable:
    DM in dwarf galaxies, DM in the Local Group; exact masses of
    stellar black holes; equation of state of neutron stars; test of
    GR in the strong gravity regime (10^12 times stronger for stellar
    black holes than for supermassive black holes); 1% distances (H0)
    for nearby galaxies; formation history of the MW (and M31);
    masses & luminosities of all rare/important stars; final
    calibration of stellar evolutionary models ==> accurate mass/age
    inferences for high-z galaxies; a host of AGN science (origin of
    jet; collimation of jet; binary-induced motion of supermassive
    BHs; luminosity-dependent origin of AGN emission, …); etc. Many
    of them >= as worthy as the WFIRST micro-lensing program. All of
    them are deemed unworthy, though.


    With regards to exoplanets, the report states that it is:
    “… seeking nearby habitable planets …” (page ES-1)

    and “…Perhaps most exciting of all, researchers will identify
    which nearby stars are orbited by planets on which life could
    also have developed.” (page ES-2)

    and “… will identify which nearby stars that are orbited by
    planets on which life could also have developed.” (page ES-2)

    and “… discovery of the closest planets beyond our solar
    system.” (page 1-1) and so on.

    Postulating that more emphasis on more accurate RV will lead to
    the detection of Earth-like planets, is not quite enough to make
    it happen. The evidence is currently that by beating down on a
    given star you can eventually “average out” some of the intrinsic
    “RV variations” (line shape/location variations due to stellar
    activity/pulsations etc). But who knows by how much these
    variations can be beaten down, to 6 cm/s? This idea has been
    ridiculed for many years. Laser combs won’t help very much with

    In my opinion, by killing SIM(Lite), the panel has actually
    ensured that Earth-like planets IN the HABITABLE ZONE will NOT be
    found during the next decade. People have been working for
    centuries on astrometry: we KNOW that this must be done from
    space to reach micro-arcsec results.

    Also, SIMLite would space qualify interferometry: the only way to
    gain substantial substantial resolution without the increase in
    cost associated with a filled aperture. To get the so-desirable
    high resolution, interferometry is the only way to go. Now we’re
    stuck yet another decade with filled apertures.


    And then there is the recommendation for a 30m telescope without
    specifying that it _has_ to be in the northern hemisphere. Did
    they notice that ESO is building a 42 meter telescope in the
    south? If you’re talking about scarce resources, the north is a
    no brainer.