Friday, February 26, 2016

Hawaii 2016


It is my pleasure to relate yet another crazy weekend, in which I did Hawaii wrong. I spent three days running around razor sharp lava flows in sandals, and didn't even draw blood. Of course, I did not go to the beach.

Also, in case it's not entirely obvious, I do not recommend following my example in all respects. I take calculated risks and take full responsibility for the silly things I do or do not describe myself doing, possibly in an alternate reality. 

Friday afternoon

I picked up my pack and made it to the airport. My Uber driver was a French/Algerian software engineer, and I spent a lot of the flight reading or writing. The landing was particularly rough in Hilo, after which the stewardess said "Ladies and gentlemen, United is pleased to report we have most definitely arrived in Hawaii," and the captain said "The island was at a lower altitude than it at first seemed. Thank you for flying with us today."

I picked up a bright red Hyundai rental (goes extra fast) and, at about 9:30pm, headed south towards Volcano. On the way I picked up two local kids heading home and got some good information about the proliferation of violent crime on the island - mostly smoke without fire, at least calibrated on a scale of zero to Los Angeles. C was on the phone from Antarctica, and was pleased to report the local temperature (-58C with windchill), to the complete shock and horror of my small town passengers.

At length I made it to my destination, where I met E and C, two volcanologist researchers, with whose evil ways I was to fall in. C has a robot for exploring volcanoes, and it uses an IR depth sensor to map the shapes of various fissures and vents. Unfortunately it doesn't work well in direct sunlight, so E and C went out at night to map the openings of the vents. A late dinner was procured, hardware was fixed, and sleep was had.


Early start and drove into the national park to find a suitable crack. The robot has controllable wheels to help it drive past obstructions that would ordinarily stop a plumb bob, and we quickly set up the gear. Unfortunately, a series of technical issues and one false start led to a complete breakdown, so after some more surveying and baking in the sun, we schlepped all the gear back to the 1998 just-barely-working budget rental car and headed to the art gallery. Moods improved rapidly.

Later, C and I debugged the whatsits out of the robot, isolating and fixing the problem, before reassembling the chassis with all the cables in the right place. Not long after that K showed up, at which point things got slightly off the wall. Later that afternoon we convoyed into town, dropped E at the airport, collected A, went shopping, found J at the airport, bought some icecream, and headed back to Volcano.


On Sunday we headed to the primary research site. We set up a tent, measurement line, potential test locations, and got busy with the robot. Nearly all the electronics systems worked, and we took good data (apparently) for the morning. K, A, and J got nice and sunburnt, while I sweated inside my long pants and sleeves. At this test site, the eruption began in 1969 as a series of big fissures, some of which became eruptive. After about a day, eruptions focused about 4km further north, and subsequent lava flows failed to bury the entire initial rift, which is still open. Earlier plumb bobbing revealed its depth to be >100m at some places, not bad for fissures only just big enough to fall into. I thought it would be possible to climb, though extremely rough. So, we send the robot instead, often to its full tether length of 27m.  

Around lunch, an astrobiologist called P showed up. P was full of all kinds of stories, so C gave her a tour of the various vents, including one that seems big enough to rappel into. After P left we dropped four more holes, then packed up and left. K managed to lock himself out of his car, so we piled into the remaining vehicle and went back for dinner. J and I prepared a dinner of spaghetti squash, beans, and lemon/avocado/basil/secret sauce. 

K and I headed back to the park to meet the AAA guy who deftly broke into the car, and after a quick look from Jaggar museum, K headed back while I lingered in the park. A beautiful full moon, clear night, and almost perfect silence, so I drove around, took a bunch of photos, and then headed back in time to sleep.


Monday morning we went to the weekly eruption update at the volcano observatory, which was cool. One presentation focused on helping out Make-A-Wish kids who want to be/see volcanologists. The pros headed for the test site and I drove south to the Great Crack. The 2.5 mile trail was really overgrown and basically guesswork, but soon the sandaled warrior (that's me) made it onto the 1823 flow and thence along a barely marked road towards the National Park boundary. As I approached the fence, a chasm appeared before me. Approaching the edge, it became apparent that it was way, waaaay deeper than it at first seemed. In this place, perhaps 5m wide and 50m deep. Along its walls nested numerous birds, I later saw nene, an owl, and other more generic varieties. 

I walked up and down the crack for about a mile each way, but the whole thing stretches for dozens of miles in each direction. More amusingly, the system has parallel satellite cracks often obscured by the overlying pahoehoe clinker, with just occasional gaps revealing a yawning nothingness beneath. Soon enough I turned around and began the slog back up the hill, focusing on trail finding/keeping, and taking dozens of spiderwebs to the face.

The car, having not been stolen, took me back (past the 'warning, road cracks' sign) and down to Glenwood, where I caught up with an old couch surfer host (V) and his new pack of dogs, discussed solar power systems, water filtration, and how fast lava from Pu'u O'o was eating the jungle between the vent and his house. While I was there it began to pour with rain, so I headed back up to the test site, arriving just as the laptop battery died, and so we packed up and left.

A nice quiet afternoon, tweaking datasets and discussing science ideas for outer planet cryovolcanism. Two parallel pasta dinners were constructed, I packed, and then left for the airport. The drive through fog and rain was quick enough, so I took a quick turn of the waterfront at Hilo before dropping the car, taking a call from the extreme south, and boarding the plane for home.


I slept on the flight back, waking when we began to descend suddenly barely 3 hours into the flight. At times like these, one wonders how well one would float during the post-ditching pre-sharkbait phase. But it turns out we had a good tailwind and arrived ahead of schedule. I took an Uber (Armenian/Georgian/Russian mechanical engineer) to work, cleaned up a bit, and waited for the exhaustion to hit.

It's not every day you get to try a new thing, look in a new direction, or spend three days running around on an active volcano in wildly questionable footwear. Thank you to E, C, K, A, and J for letting me hang out and answering my incessant questions!

Sunday, February 14, 2016

We have found gravitational waves!

If you were watching the news recently, you may have seen droves of excited scientists talking about a recent discovery of gravitational waves. This blog, written primarily for my readership/family, will talk a bit about it, as I was lucky enough to be involved, in a peripheral way, in this research.

First, what are gravitational waves? If you were reading certain German language science journals in 1915 and 1916, you may have come across a series of papers by Albert Einstein laying out the foundations of what we now call General Relativity. Newton's observation that gravitational force is proportional to mass and inversely proportional to distance is 99% correct, but fixing those 1%s is where the fun happens in science. Einstein's insight was that the laws of physics are essentially the same everywhere. Wind the crank on this postulate and out falls General Relativity, an incredibly powerful theory that explains gravitation in terms of geometry.

Geometry, which means literally "Earth measuring", is more than understanding circles and triangles on a page, though that will get you pretty far. But we know that when your surface is curved, like the Earth, you can run into trouble. For instance, my friend and I leave the cinema after a midnight screening of Pitch Perfect 2. If I go a km north and then a km west, and my friend goes a km west then a km north, chances are we'll arrive pretty close to each other, and can continue to sing the numbers in approximate harmony.

But if we go much further, or do so near the poles, we're liable to get entirely lost and never see each other again. It turns out that measuring the distance between us after our respective treks, as a function of the directions we went and how far we went, is a good way to measure the curvature of a surface.

Out in space, the universe is pretty 'flat', which is to say, Pythagoras' theorem works. Closer to Earth or some other massive body that warps space, you need to tweak the theorem a bit to make it consistent. The little factors you add, for instance to the vertical direction, can be converted into curvatures, and vice versa. But it turns out that if you wobble these numbers in just the right way, they keep wobbling, forming a wave that propagates through the universe at the speed of light. This is a gravitational wave.

Gravitational waves are not the same as gravity waves. Ocean waves are an example of gravity waves, where the restoring force is buoyancy and gravity. They are not difficult to detect. Gravitational waves move through space and time (which us technical types call spacetime) and subtly shift the distances between things back and forth as they pass. Which is how we try to measure them.

Since the late 1980s, a herculean effort by thousands of mostly NSF-funded scientists has resulting in the building of the Laser Interferometer Gravitational Wave Observatory, or LIGO. This antenna, and a few similar versions around the world, is designed to measure gravitational waves as they pass through the Earth. 

When I first read about LIGO, in Scientific American way back in the 90s, I thought it rather a lot of effort. A powerful and very fancy laser shoots a beam into a beam splitter which sits at the junction of two 4km-long arms, each with mirrors on each end. Lots of other fancy stuff, other mirrors, etc, are involved. When the mirrors are correctly positioned, the reflected laser light interferes destructively and a nearby light sensor sees nothing. If the mirror is moved by a passing wave, the light doesn't cancel out as well and we see a fringe shift and can detect the passing wave.

Of course, lots of other things move the mirrors too, including cars, planes, lightning strikes, waves crashing on the beach, earthquakes, etc, so the entire set up is suspended by seven levels of active and passive seismic isolation, and kept in a vacuum pumped down to <1 micro pascal, which is, I think, a better vacuum than interplanetary space. I was lucky enough to visit LIGO in Livingstone, Louisiana, in 2013. Some photos from the trip can be seen here: . There is another LIGO in Hanford, Washington, a pathfinder instrument at Caltech and in Western Australia, and similar observatories in Germany, Italy, and Japan, with one planned in India. When these are all up and running, we'll be much more capable of localizing the position of sources much more accurately.

The reason you need such a staggeringly complicated instrument is that spacetime is stiff. Stiffness means how much it warps in response to a force. It's much easier to shake a piece of string than a heavy rope, and spacetime is the ultimate heavy rope. Whirling masses around like poi might generate a gravitational wave, but it's much too weak for the mirrors of LIGO to see. Even though LIGO can sense a strain of one part in 10^21, you need very dense objects moving very close to each other and very, very fast. The best candidates for signals are binary neutron stars and binary black holes, about which more later.

The first LIGO runs were completed in the 2000s, when the instrument was in its initial and enhanced phases, and they detected ... nothing. Except a deliberate (fake) injection in 2010. This was expected - the first phase was designed to inform the design of the second phase, which was built between 2010 and 2015.

Then, on September 14 2015, during an engineering shakedown commissioning run, both detectors saw this astonishingly powerful signal, visible in the data to the naked eye, almost as soon as the instruments were turned on. The intervening 5 months were spent looking for mistakes, other noise sources, running veto checks, waiting for something to go wrong. The signal was verified, and the papers published. 

In the meantime, of course, about 1000 scientists had to keep it a secret, which they mostly did. The field has a not-entirely-undeserved reputation for publishing first, peer reviewing second, and retracting third. Just in 2014, another group had claimed detection of gravitational waves from the big bang, imprinted in the cosmic microwave background, but had to later retract the discovery. The existence of gravitational waves had been previously indirectly inferred by the orbital decay of binary pulsars, but September 14 was the first directly detected gravitational wave, and it was a humdinger, as you can see in the image below.

So, given this wave, where did it come from? We were able to fit a waveform to the observed signal, matching it with a computer simulation. From this, we can determine that the source is a pair of black holes each about 30 solar masses, roughly 150km across, that orbit each other at a steadily increasing rate. They emit gravitational waves, decaying their orbits, until at last their event horizons merge and only one, larger, black hole emerges. The whole thing - 10 orbits, over a distance of less than 1000km, occurs in less than a second - the black holes approach the speed of light. These black holes merged over a billion years ago, back when most of the rocks in the grand canyon hadn't yet been laid down and life was still unicellular. The signal, and LIGO's sensitivity bucket, lies in the human audible range, and can be perceived as a low frequency rumble and thump, if your sound system is good enough. 

In the above image, taken from the primary paper, you can see the signals reach the blindingly powerful strain of 10^-21. Which means that the mirrors moved by about a thousandth of the width of a proton, which is a million times smaller than an atom. Not bad for a system made of atoms and measured with discrete photons of light with a wavelength of a mere 10^-6m. During the last tenth of a second of the black holes' separate existence, they emitted a lot of energy as gravitational waves. The equivalent of 3 solar masses worth, to be precise. This is also an incredible amount of energy. For that tenth of a second, the black hole binary dissipated more energy than every star in every galaxy in the entire visible universe. 50 times more. If that energy were in the visible electromagnetic spectrum, instead of gravitational waves, it would (briefly) be brighter than the full moon, despite being more than 50 times as distant than the Andromeda galaxy, the most distant object normally visible to the unaided eye. 

Already, we know far more than we ever did before. Black hole binaries exist - and might even be common. Intermediate black holes, in the 10s of solar masses, exist! Perhaps they originated from earlier mergers, perhaps from single older massive stars with low metalicities. Metals, generated in the cores of dying stars, make second generation stars significantly less transparent and limit how much mass is left over to form a black hole. There are almost certainly more signals in this, the first science run of Advanced LIGO. Up until now, everything we know about our universe (except for a few stray neutrinos) we have learned using electromagnetic waves - light, radio, x-ray, gamma rays, microwaves. As of 2016, humans can now hear the universe as well as see it, showing us all kinds of things that are there, even if they don't glow. We are sure to be in for some surprises. 

No blog of this length would be complete without a mention of the founders of LIGO. Three venerable academics can claim that title, all still alive and, to an extent, kicking. Kip Thorne, the executive producer of Interstellar and professor emeritus at Caltech, Rai Weiss, of MIT, and Ron Drever, who built the 40m prototype, are names you may hear later this year in connection with a royal trip to Sweden!

Some of my readers may be familiar with my own doctoral work at Caltech, where I was part of the numerical relativity group. My own infinitesimal contribution concerned one aspect of producing simulations of binary black holes, which we use to help detect, and then back out the physical properties of the signals we detect. I don't think my research was used in this particular detection, but if I discover otherwise I'll certainly update this blog!

Perhaps you want to know more? One positive trend in science is the open access movement, whereby publications pertaining to publicly funded research are made publicly available, instead of by subscription only. Physics is generally ahead of the curve in this respect, and all of the (15ish) papers, data, and data processing software related to this discovery is freely available online.

Papers 100-113+ on, click the ArXiV link in the 5th column, download the pdf.

I'm not the first, only, best, or most handsome scientist to also blog on the topic. Here are a few others, often focusing on different aspects of this incredible discovery.

Blog by John Preskill, another Caltech professor. 
Blog by Roger Blandford, a professor at Stanford and occasional coauthor of Kip Thorne.