Last time I gave the background for our third paper, which is now up on the arXiv here.

So what did we find?

We were looking for so-called Type III Kardashev civilizations, galaxy-spanning civilizations that command power equal to all of the stars in their galaxy.  A true Type III civilization would be blocking a significant fraction of all of a galaxy’s starlight from leaving the galaxy, and harnessing it for its own purposes.  Such a galaxy would be very faint in optical light — from all of the solar panels collecting the starlight — but very bright in the mid-infrared, where all of that energy would have to come out when the civilization was done using it.

Looking through the WISE catalog, Roger found that zero galaxies were consistent with more than 85% starlight reprocessing by alien industry.  This is our first-order upper limit: there are no galaxy-spanning civilizations in galaxies resolved by WISE with aliens using more than 85% of their starlight.  This mostly rules out Type III civilizations as Kardashev originally defined them, because they were a sort of maximal case.

We found 50 galaxies (out of 100,000 galaxies surveyed) were consistent with more than 50% starlight reprocessing.  Jessica found that most of these have a heavy literature presence, which is not surprising, since they have such extreme MIR emission.  They are mostly starbursts; indeed one of the reddest objects we list in our catalog is Arp 220, the quintessential local starburst.  We know why they have a lot of MIR emission — the morphology of these galaxies makes it obvious that the emission is from dust illuminated by star formation regions.  Once we have gone through all 50 of these galaxies and confirmed the natural origin of their excessive amounts of MIR emission, then our upper limit in Type III civilizations will be around 50%.

We also identified around 90 galaxies consistent with more than 25% reprocessing that have very little literature presence.  These are more interesting from a SETI perspective, because it’s possible that they are NOT starbursts (that is, without a lot of study, no one would have noticed anything anomalous about them).

Now, you might be thinking, 85% is a pretty weak upper limit.  All those galaxies might be filled with alien civilizations using enormous amounts of power, but just be lost in the glare of all of those stars and dust in the galaxy.  But weak upper limits are how all sensitive experiments begin.  Raymond Davis (who won the Nobel Prize for detecting cosmic neutrinos) quoted the referee of his first paper on neutrino detection, a very weak upper limit in 1955:

Any experiment such as this, which does not have the requisite sensitivity, really has no bearing on the question of the existence of neutrinos. To illustrate my point, one would not write a scientific paper describing an experiment in which an experimenter stood on a mountain and reached for the moon, and concluded that the moon was more than eight feet from the top of the mountain.

Well, one might if one had no idea how high the Moon was! Indeed, I’m sure our ancient forebears asked as much of the first explorers to return from the highest mountaintops.  The idea only seems absurd to us today because we know the answer.  In 1955 people thought they knew what sorts of cosmic neutrino fluxes to expect — but one must always be prepared to be surprised!

So, OUR fist upper limit is 85% of a galaxy’s stellar luminosity, and with some work we can get down to 50%.  What’s next?

We can still look for Type “2.9” (on Sagan’s scale).

We can actually model the SEDs of these galaxies.  Star formation and dust emission have many tracers, and there are relationships among luminosities at different wavelengths for the many components that make up a galaxy.  By combining optical, NIR, MIR, and radio observations of galaxies, we can ask: which galaxies have MIR excesses given the rest of their SEDs.  That will get us even lower.

But we can do much better.  Elliptical galaxies have very little MIR luminosity.  This means they have a low “natural background” for us to search in.  We can get down to a few percent by looking at dust-free elliptical galaxies.

But we can do even better.  Because WISE resolves galaxies, we can see if the midinfrared emission traces the stars!  The stars in an elliptical should be well mixed, so the civilizations harvesting their energy should be well mixed, and the MIR emission should trace the starlight.  MIR emission concentrated in the center is a giveaway that it’s from dust.

Next time: Strange new (non-alien) objects in WISE!

It’s been a while since I’ve blogged about our Ĝ project, and since we have a press release coming out soon, it seems like a good time to start again.

We have our first results coming out in a paper published in ApJS soon; until then here’s the background. This paper is the result of a lot of hard work by our post-baccalaureate Roger Griffith.  Roger comes to us from the WISE team that discovered the first spectrally-classified Y dwarfs; in fact it was a talk by Michael Cushing on that team at Penn State that first gave me the idea to use WISE to do this (since Y dwarfs are at about room temperature!).

Roger came with ready knowledge about how to interrogate not just the WISE All-sky catalog, but also the lower level data that we used to reject artifacts and comets and other false positives.  He also had code all set up to make “at-a-glance” charts of all of our most interesting sources, and in the end he personally inspected something like 100,000 images!  The paper is long, a testament to his hard work.

We decided to start with the red extended sources in WISE that appear outside the Galactic Plane.  Our rationale was that there are lots of reasons an object might appear “red” (i.e., to have a lot of 12 or 22 micron emission).  If all you have to go on is a point of light in the sky and broad-band photometry, there’s not much one can do to diagnose the origins.  But if the source is resolved, then you have some information, and the number of false positives goes way down.

We discovered that when you go looking for extreme objects, all of the rare special cases in the data reduction come flooding into your lists.  The WISE team did a great job of removing artifacts, solar system objects, and other things we weren’t interested in from the list, but when you have 100,000,000 sources, there are always going to be a few that sneak through.  We discovered lots of things that looked a lot like what we were looking for at first glance, but turned out to be reflections from bright stars, latency in the infrared detector array, nebulosity in the Milky Way, and similar artifacts.  Roger had his work cut out for him sorting through all of this stuff, and his work on that constitutes much of his paper.

Another major complication is that the WISE photometry pipeline was not designed to do great photometry on extended sources.  It’s an automated routine, and extended objects tend to need human intervention to get the details right.  We found that there were systematics in the All-sky photometry for most extended sources.  Fortunately, Tom Jarrett has been working on a WISE atlas of extended sources with careful photometry.  He graciously lent us a preliminary version of his catalog for the South Galactic Cap.  We found a good coefficients to a nonlinear combination of All-sky photometry from different apertures, and this allowed us to calibrate the existing WISE photometry to these sources.

After we had made our list, we needed to know what these things were.  Roger cross-matched to NED and SIMBAD, but we still needed someone to hunt down each and every object in the literature and see what it was.  That’s where Jessica Maldonado came in.  Jessica is an undergraduate at Cal Poly Pomona working with Matthew Povich, and she did the yeoman’s work of lots and lots and lots of literature searches for hundreds of our best candidates.

So why would we conduct a search for alien waste heat using resolved sources? Because a galaxy containing a Kardashev Type III civilization is pandemic with intelligent, spacefaring life.  The total energy use of this supercivilization would, unless they were using most of it to send out lasers or create matter or something, have to come out at roughly mid-infrared wavelengths.  The entire galaxy would have a bit too much 12 or 22 micron emission.  If the civilization’s energy output were greater than a few % of the starlight in the galaxy, then this would be easily detectable with WISE as a diffuse MIR excess across the whole galaxy.  We discuss all of this in our first two papers (here and here).

Of course, galaxies have lots of mid-infrared light for many other reasons, too, mostly having to do with dust.  But a typical elliptical galaxy, for instance, has so little dust that even a  few % excess would be obvious.  We found that any civilization preprocessing more than about 10% of the starlight in the galaxy would be anomalously bright — even if the galaxy were a spiral.  So we set out to find what the upper limit was.  After all, no one has done this before, so it was in principle possible, though unlikely, that there was a galaxy with a civilization using most of its starlight right in our own backyard!

Also, we don’t know what energy sources an alien civilization might use.  We tried to think hard about this in our first paper, but for all we know it’s possible to tap zero point energy or generate Romulan warp core singularities or who-knows-what, so in principle an alien civilization could have an energy supply that is 110%, 200% or 1,000% of the starlight in its galaxy.  Has it ever happened?  If it has, WISE would see it.

We used the shorter, W1 and W2 (3-5 micron) bands to get a handle on the stellar photospheres in the galaxies we searched, and the W3 and W4 (8-26 micron) bands to measure the waste heat.  We started by assuming that all of the heat was coming from alien civilizations (that is, that the galaxy had no dust) and calculated the fraction of starlight that would have to be reprocessed to explain it.  We also allowed the waste heat temperature to float as a free parameter.  We then sorted all of the WISE galaxies by this starlight-covering fraction. This gave us upper limits on the amount of alien waste heat that could be in each of our 100,000 galaxies.

Next time:  The results of our survey.