Meetings

I just arrived home after a week-long visit to Aspen, CO to attend the Formation and Dynamical Evolution of Exoplanets Conference at the Aspen Center for Physics.

This conference was a cozy affair, with just over 100 attendees, and was narrowly focused on dynamical questions and approaches related to the origins and fates of exoplanet systems.

Researchers from around the world gave presentations on topics ranging from the dynamics of debris disks to observations of planet-hosting binary star systems. Blocks of presentations were punctuated by lengthy coffee breaks, when the real scientific give-and-take takes place. These interludes often give rise to groundbreaking, thesis-motivating, all-nighter-pulling research ideas.

Most of the presentations and conversations were excellent and inspiring, and I can’t do them all justice in a short blog post. So I’ll just talk about one that struck me in particular.

On Tuesday, Hanno Rein at Toronto spoke about a new N-body integrator his team has been developing in recent years, called REBOUND. This new framework may spur a revolution in dynamical modeling of astrophysical systems.

In astronomy, “n-body integration” is jargon for the numerical simulation of interactions among multiple (“n” of them) gravitating bodies. For hundreds of years, astronomers have been able to describe the orbital of two gravitating bodies quite easily, thanks to Johannes Kepler.

But as soon as you add another body to the system, there is no exact way to solve for the orbital motion of the bodies (except in very specific and limited circumstances). Even in the case of two bodies, if you want to include more complicated forces than simple gravity, solving for the orbital motion can be quite difficult.

To surmount these difficulties, scientists have turned to computer simulations to model in an approximate way the evolution of n-body systems. Although scientists have spent decades coming up with better and better models and algorithms, n-body simulations can still take a lot of computing power, and the often complicated codes can be cumbersome to set up and run. More than that, it’s often difficult or impossible for scientists to share results because there’s no good agreed-upon format for simulation output.

Rein’s REBOUND open-source code solves several of these problems at once: it employs latest modeling schemes to track orbital motions and gravitational interactions; it can be run using inside of an iPython Notebook; and it provides a uniform format for simulation output which anyone can use to re-run or re-analyze another scientists work – critical for scientific reproducibility. The iPython Notebook also provides a really neat visualization capability so you can directly watch the evolution of your astronomical system.

Time evolution of the orbits of stars in Leela’s constellation.

The code is so easy to run, in fact, that I installed and began running it immediately after Rein’s presentation. And all of its capabilities allowed me to finally simulate and visualize the evolution of a system I’ve wanted to look at for a long time – see animation at left (see here for how I created it).

I also gave a presentation on our group’s work looking at disruption of gaseous exoplanets.

And so, the combination of beautiful scenery and beautiful science made the Aspen Exoplanets conference one of the best in recent memory.

LPSC 2017 Meeting

I’m sitting in my hotel room on the morning after the 2017 LPSC Meeting, trying process all the science that washed over me in the past week. And it was a lot, as you can see by checking out the twitter hashtag #lpsc2017.

From the migration of Martian dune ripples to the global seas on Enceladus, there’s no way for me to do it all justice in one blog post. So instead, I’ll talk about one of the talks that stood out most for me.

On the last day of the conference, Dr. Colin Dundas of USGS gave a mic-droppingly good talk in which he argued very convincingly that Mars’ recurring slope lineae are NOT the result of flowing water. This is a big deal for folks who study Mars but might sound a little arcane for non-Martians.

The animated image above consists of several pictures taken by the HiRISE camera onboard the Mars Reconnaissance Orbiter over several months. The very pronounced dark streaks extending out from the cliff face – those are the recurring slope lineae or RSLs. In many locations on Mars, they recur from one year to the next; they always show up on sloped surfaces; and they are long lines. Hence recurring slope lineae.

Under the current conditions at Mars’ surface, liquid water is not stable over long times, but these dark streaks sure look like small amounts of water running out over the surface, which got people very excited when they were first discovered.

In fact, just a few years ago, some scientists analyzed the spectra of some RSLs and argued they were highly salty brine flows. The salt is important because, even though pure water can’t exist on the surface of Mars, large amounts of salt can stabilize the water, at least for a little while.

This was a huge result, not just because it meant RSLs were small amounts of running water on Mars but because it raised the possibility of much larger sub-surface reservoirs of water. And where there’s lots of water, there could be life.

Well, Dundas’s talk throws all that into doubt. By analyzing the angles of the slopes on which the RSLs occur, Dundas showed that they almost every one of them ended on a slope of about 30 degrees.

Now, if RSLs are flowing water, they should keep flowing, even on small slopes. But if they are some sort of dry granular flow instead, you would expect them to stop once they reached the angle of repose, which is about 30 degrees.

Dark streaks on the sides of Martian dunes seen by HiRISE. These are NOT due to liquid water but resemble in many ways the RSLs.

To bolster his argument, Dundas showed several examples of dry granular flows on Mars that exhibited many of the same properties as RSLs – recurring dark streaks, running along sloped surfaces.

Continued work will either corroborate Dundas’s striking result or circumvent it, and given the number of folks lined up to talk with him after his talk, I’m sure the fans of liquid-flow RSLs will work hard to counter his arguments. And that’s how science progresses – one falsified hypothesis at a time.

Not quite as splashy, I also gave a presentation on our work trying to de-bias dust devil surveys. I’ve posted the presentation below.

All in all, my first LPSC was a great mix of science, warm weather, and warm friends.

Figure from Gao’s talk, showing the spectrum of a planet before (blue and black lines) hazes and after (red line). The molecular features are almost completely wiped out by the hazes.

On day two of the AAS 229th meeting, I attended the morning session on exoplanet characterization and theory, which focused on atmospheric characterization. Several great talks, but one that made an impression on me was Peter Gao‘s presentation on sulfur hazes in hot Jupiter atmospheres.

Gao discussed work from Kevin Zahnle at NASA Ames showing that UV photolysis can transform even small amounts of gaseous sulfur in a hot Jupiter’s atmosphere into significant amounts of polymer haze. Something that has become a running motif in exoplanet atmosphere studies, these hazes discombobulate the spectrum of light emerging from a planet’s atmosphere, completely masking the signatures of other atmospheric components. This is bad news if, say, you wanted to determine the composition of a planet using light reflected from its atmosphere.

In the afternoon, I was fortunate to attend Sean Carroll‘s plenary talk “What We (Don’t) Know About the Beginning of the Universe”. It was a fascinating tour of all the different ideas about the origin of the universe, including The Big Bounce, baby universes hidden inside black holes, and the idea that the universe may have no beginning and no end. The best part of the talk, for me, was the end, when Carroll showed us a stern letter from a 10 year old skeptic sent to him in a response to an NYT article in which Carroll was quoted:

I Don’t know if you Exist But I Do! I bo not Agree with your Articl and I Do not Beleave that “MOMBO-JOMBO” if you do … Well! it’s Disturbing thought But I know How to Deal with it! I will Not let the Wolb Disiper under My Nose But if you Do I cant say I’m sorry!

Sincerely

a ten year old who knows a little more than some Pepeol!

George Wing

ps. some peopl Have a little to Much time.

Just brilliant!

Unfortunately, I had to depart for home shortly thereafter, so I’m missing the rest of the meeting. So here endeth my blog series on the meeting. Our fall semester at Boise State starts next week, though, and I plan to have weekly (maybe even semi-weekly) updates on the blog, so stay tuned.

A very active and engaging morning session on detecting exoplanets via the transit method on AAS 229 Day 1. Lots of good talks (although all of the talks were by male astronomers) and probing but polite questions (again, mostly by male astronomers – interesting study on these trends here). A few talks stuck out in my mind.

Aaron Rizzuto from UT Austin looked for transiting planets in stellar clusters spanning a range of ages using data from the K2 Mission and found there seem to be fewer hot Jupiters in clusters 10 million years old than there are in older (hundreds of millions of years old) clusters. He suggested that this may mean it takes 100s of millions of years for the migration that makes hot Jupiters to take place. That would probably rule out one standard model for making hot Jupiters, namely gas disk migration, since that probably takes place within a few 10s of millions of years.

Dave Kipping of Columbia University discussed his search for transits of Proxima b, the recently discovered, Earth-sized planet just four light years from Earth. Unfortunately, the host star, Proxima, is a highly variable star due to almost constant flaring. As a result, it would be very difficult to see the planet’s transit – as Kipping said, it requires wading through “a sea of variability”. However, Kipping and his group struggled valiantly to recover the transit using data from the Canadian MOST satellite, and it looks like the planet just does not transit. So we probably won’t know the planet’s radius (at least not for a long time). Bummer.

The last talk of the session was from George Ricker, the PI of the TESS mission, the follow-up mission to Kepler, about TESS’s status and prospects. Apparently, the mission will observe more than 2 million stars! Orbiting many of those stars will be nearby habitable planets, and Ricker showed an amazing simulation of where those stars might be found (courtesy of Zach Berta-Thompson of UC Boulder), a still from which is shown below.

I’m at the beautiful Gaylord Convention Center in Grapevine TX, attending the 229th meeting of the American Astronomical Society.

Today’s the first day, and it will be chockfull of presentations, posters, and press releases.

Given my research focus on exoplanets, I’ll mostly be attending talks on that topic, but I’m going to duck out at one point to hear about the solar eclipse, coming up in August this year.

Fortunately for me this year, my presentation was scheduled on the first day of the meeting, in the first session of contributed talks. That means I’ll be able to focus on the rest of the meeting without being preoccupied by preparing for my own presentation. I’ve posted my slides below.

Stay tuned this week for more.

The last day of the meeting is always to hardest to write about because I’m usually so busy wrapping things up, I don’t have time to write (hence my writing this post from Boise on the Sunday AFTER the conference).

In any case, lots of talks and goodbyes on the last day, but one talk that stands out for me came from Andrew Hesselbeck Hesselbrock, one of David Minton‘s grad students at Purdue’s EAPS. The talk tackled one of the longest-standing mysteries in solar system science: Why hasn’t Phobos crashed into Mars yet?

Phobos (left) and Deimos (right). From http://www.planetary.brown.edu/planetary/geo287/PhobosDeimos/images/Mars%20and%20Moons.jpg.

Phobos (left) and Deimos (right).

Mars has two tiny moons, Phobos and Deimos, which visibly resemble asteroids but are probably not for a long list of reasons.

Phobos is close enough to Mars that Mars’ gravity is dragging the moon inward, similar to but in the opposite direction as the effect of the Earth’s gravity on the Moon. Phobos is so close, in fact, that astronomers expect it will spiral into Mars in just a few million years.

Phobos and Deimos have probably been orbiting Mars for about the age of the solar system, 4.6 billion year. So if this orbital decay were the whole story, it would be mean we just caught Phobos right at the end of its life, about as likely as catching someone driving from Boise to New York City right as they pass through the Holland Tunnel*. Hesselbeck Hesselbrock suggested in this talk that we’re actually seeing a recurring phase in a much more dramatic story for Phobos.

Instead of steadily spiraling in toward Mars for 4.6 billion years, Phobos (or at least a proto-Phobos) already spiraled in toward Mars before, millions of years ago. But when the satellite got close enough to Mars, Mars’ gravity ripped it apart and formed a disk of rubble around the planet. Soon after forming, this disk spread out, some moving toward Mars (and ultimately impacting the surface) and some moving away. Eventually, the bits that moved outward moved far enough away from Mars that they re-coalesced. In fact, Hesselbeck Hesselbrock speculated that Phobos has actually been reincarnated many times in this way, every time a little smaller than before, until we were left with the bitty moon we see today.

As crazy as this hypothesis sounds, it could answer several puzzles of the Martian system, including accounting for cyclic sediment deposits on Mars’ surface — the deposits form every time Phobos falls aparts and bits rain down on Mars’ surface.

Again, the annual DPS meeting astounds and amazes. Looking forward to Provo next year.


The distance from Boise to New York City is about 2,475 miles, and the Holland Tunnel is about 9,000 feet long. Assuming a uniform driving speed, the probability of catching our driver in the tunnel is roughly equal to 9,000 feet/2,475 miles ~ 0.1%. The probability of catching Phobos during a 10 million year window over the age of the solar system is about 0.2%. Of course, you’re a little more likely to catch our driver in the Holland Tunnel, given NYC’s traffic.

Fourth day of the DPS meeting, and I found myself sitting through some great plenary talks.

cvprxvsvuaeikj1First up was Kleomenis Tsiganis‘s Farinella Prize lecture “Flavors of Chaos”, a rapid-fire tour of the intricate and complex web of gravitational interactions among planets and asteroids in our solar system.

Tsiganis’s described how, using a combination of computational and pencil-and-paper techniques, we can pick at the threads in this cosmic network to tease out the early history and evolution of our solar system.

For instance, the orbits of asteroids in the asteroid belt provide subtle clues that, billions of years ago, Jupiter moved inward almost to the orbit of Mars before backing out near to its current orbit, a celestial maneuver referred to as “The Grand Tack“.

cvpbr0kusaaqdlgThis presentation was followed by Leigh Fletcher‘s Urey Prize talk about the menagerie of seasonal changes we observe in the atmospheres for all the outer planets, from Jupiter to Neptune.

The talk was full of beautiful images of the roiling and boiling of planetary atmospheres and concluded with Fletcher’s plea to send another mission to the Uranus or Neptune before he’s too old to participate (some plans from NASA have a mission launching to Uranus or Neptune sometime in the late 2020s/mid-2030s).

Finally, we had a tag-team talk from Ashwin Vasaveda and Sanjeev Gupta about new results from Mars Curiosity rover. In addition to the stupefying images, the thing that impressed me most about the talk was just the level of detail to which we can infer the geological history of Gale Crater, where Curiosity landed.

cvpm7vnuiaarkpeGupta described how the tilt of beds of sedimentary rock could be used to infer the presence of a river delta spilling out into the crater, which suggests the existence of a long-lived (millions of years) lake in the crater, probably billions of years ago when Mars was warmer and wetter.

svmregression

Animation showing how a machine-learning algorithm decides where lies the boundary between two classes of objects.

Third day of the DPS Meeting was full of fascinating talks about the orbital architectures of exoplanet systems.

One that caught my attention was Dan Tamayo‘s talk on using machine-learning to classify the stability of a planetary system.

As astronomers have discovered more potential planetary systems, it’s becoming more time-consuming to decide whether what we see are actually planets or some other thing that has fooled us into thinking they’re planets.

When astronomers find what they think might be a planetary system, one of the first things they check is whether the putative planetary system is actually stable — that is, whether the gravitational tugs among the putative planets would cause the objects to crash into one another or be thrown out of the system.

Since most of the planetary systems we find are probably billions of years, astronomers expect that real planetary systems are stable for billions of years, so if the system we’re looking out turns out to be unstable on short timescales (less than billions of years), we usually decide that it’s not really a planetary system (or that we mis-estimated the planetary parameters).

Unfortunately, doing this check usually requires running big, complicated computer codes, called N-body simulations (“N” for the number of planets or bodies in the system) for hundreds or thousands of computer-hours. That can be a problem if you’ve got planetary candidates flooding in, as with the Kepler or upcoming TESS missions.

Tamayo wanted to try a different approach: what if the same machine-learning techniques that allow Google or Facebook to decide whether someone is likely to buy an iPhone could be used to more quickly decide whether a putative planetary system was stable

So Tamayo created many, many synthetic planetary systems, some stable, some not, and had his machine-learning algorithm sort through them. According to Tamayo, his scheme was able to pick up on subtle features that helped distinguish stable systems from unstable ones with very high accuracy in a fraction of the time it would take to run an N-body simulation.

aaeaaqaaaaaaaamaaaaajdjmowm5yzjjlwjhzgitnge4ys05ogi3lwu4mdjmmmi4zgexyqI also attended an eye-opening talk from Patricia Knezek of NSF about unconscious biases and their effects in astronomy and planetary science. Knezek explained that several studies have shown how these biases cause everyone to draw unconscious conclusions about someone based on very cursory information, such as their first name, race, gender, etc.

For instance, one study showed that the same application for a faculty position did much better if the applicant’s first name was “Brian” instead of “Karen”, even when women were evaluating the application.

Fortunately, these same studies have shown several ways to mitigate the effects of these biases, and being aware of them is a big first step.