Kepler stuff

About the Kepler mission’s announcement last week of tons of extrasolar planets:

1. A local Richmond TV station had me on the news to talk about the announcement. (The fact that they asked me primarily goes to show that astrophysicists are not exactly thick on the ground in Richmond.) I can’t stand to look at myself on video, but if you want to see it, go ahead.

2. Via Sean Carroll, some cool visualizations of the Kepler data, showing number of planets by size,  distance from star, temperature.

3. Sean says

A back-of-the-envelope calculation implies that there might be a million or so "Earth-like" planets in our Milky Way galaxy.

I’d go much higher than that. Kepler looked at about 150,000 stars and found five Earth-like planets (meaning roughly Earth-sized and in the habitable zone where liquid water could exist).  If you imagine that they had 100% efficiency — that is, that they found all the Earth-like planets there are in the sample — then one in 30,000 stars would have an Earth-like planet. Multiply by 100 billion stars in the galaxy, and you get about 3 million Earths.

But here’s the thing: Kepler’s efficiency can’t be more than about 1% or so. The mission works by looking for eclipses, which occur when the planet passes directly in front of the star as seen from Earth. That means that it only has a chance of detecting a planet if the geometry is fortuitously aligned. The probability of such an alignment occurring depends on the size of the star and of the planet’s orbit (in fact, it’s just the ratio of the two). For the actual Earth and Sun, the probability works out to 1%. Many of Kepler’s planets are closer in and have higher probabilities, but at best the geometrical alignment can only occur a few percent of the time on average.

Even with the right geometry, they don’t have a 100% chance of finding a planet, of course. Once you fold in all sources of inefficiency, I’d be very surprised if they have a better than 1% chance of finding any given Earth-like planet. I wouldn’t be surprised if it’s more like 0.1%

Just to be clear, that’s not a criticism of Kepler. It’s just an acknowledgment that this is a hard task they’ve set themselves!

So my back-of-the-envelope estimate is hundreds of millions, if not billions, of Earth-like planets in the Galaxy.

Paper submitted

I love the feeling right after submitting a paper for publication.  One just went off yesterday. Since my sabbatical is winding down, and I’m now acting department chair, I’m glad to have gotten this finished before the semester starts.

This one’s pretty specialized, likely to be of interest only to people who analyze microwave background polarization maps.  Here’s the abstract, in case you care:

 Separation of the B component of a cosmic microwave background (CMB) polarization map from the much larger E component is an essential step in CMB polarimetry. For a map with incomplete sky coverage, this separation is necessarily hampered by the presence of “ambiguous” modes which could be either E or B modes. I present an efficient pixel-space algorithm for removing the ambiguous modes and separating the map into “pure” E and B components. The method, which works for arbitrary geometries, does not involve generating a complete basis of such modes and scales the cube of the number of pixels on the boundary of the map.

En français

I’m heading back to the US on Sunday (weather and transit strikes permitting), after almost three months in Paris.  I’ve been interested to see what would happen to my French during my time here.  I think language is a fascinating phenomenon: there’s a long list of subjects I wish I knew more about almost, but not quite, enough to do anything about it, and linguistics is pretty much right at the top.

When I arrived, my weak point, by far, was spoken comprehension: I could read and write pretty well, and even my speaking wasn’t too bad, but it was really hard to understand people when they spoke.  My self-evaluation is that I’m now much less terrible at that than I used to be, but still pretty terrible.  It’s  striking how hard it is to improve in this area.  One piece of evidence that I’ve gotten better: The fraction of times that I can get through a commercial transaction without the other person getting exasperated and switching to English is pretty high these days.

I have noticed that I’m much less likely than before to consciously translate what I’m hearing word-by-word into English as I hear it.  This is good for two reasons.  First, of course, because it’s impossible to understand rapid speech in that way.  More importantly,  though, it means that my colleagues no longer sound like Hercule Poirot in my head.  (“You mock yourself at me, my friend!”)

It’s interesting to think about why oral comprehension is so hard.  Comprehending speech is a many-step process: you have to process a continuous stream of sound into phonemes, assemble those into words, and syntactically analyze the result.  You can imagine breakdowns at any stage, but for me the first step is the big problem.  When I’m not understanding someone’s speech, it’s generally because I can’t hear the phonemes: I hear a continuous, undifferentiated stream of sound, rather than discrete consonants and vowels. The problem gets much worse with even low levels of background noise, and if two people are talking at once, I have zero chance of picking up anything.

Grammar’s never a problem: I never fail to understand a sentence because the speaker used the pluperfect subjunctive or something.  Vocabulary’s not much of a problem either.  Sure, sometimes people use words I don’t know, but that rarely stops me from getting the gist of what they’re saying.  (In the context of a restaurant menu, there’s a virtually 100% chance that the unknown word is the name of a fish, which makes things easier.)

One stumbling block for me, ironically, is numbers.  I still have to stop and explicitly translate them into English in my head.  And when the number is a time of day in the afternoon, there’s the additional problem that the French commonly use 24-hour time.  So when someone asks me if I’m free for a meeting at 4:00, there’s a ridiculously long pause while I think, “Seize = 16.  16-12 = 4.”


My colleague Jerry Gilfoyle and I were just awarded an NSF grant to buy a new computing cluster.  In the past, my students and I have mostly worked on problems that could be attacked with ordinary desktop computers.  This grant means that we’ll be able to go after more computationally intensive problems.  It also means I’ll have to learn about supercomputing techniques.  Fortunately, Jerry’s very experienced at this.

This has been a good funding year for me: I submitted three NSF proposals, and all three were funded. That’s at least partly due to the federal stimulus bill: only one of the three is officially stimulus money, but no doubt all the stimulus money washing around freed up more non-stimulus money for other grants.

Correction: Actually, two out of the three, including the computing cluster, are stimulus funds.  I’m nothing if not shovel-ready.


Two papers I submitted a while ago were accepted for publication by the American Journal of Physics this week, within minutes of each other as a matter of fact.  To be precise, they were “conditionally accepted,” meaning that they’ve successfully passed the review by external referees and the science content been deemed acceptable.  There’s a further review by the editors for style, clarity, etc., before they’re  finally accepted.  Because AJP is a journal with a pedagogical slant, they place a heavy emphasis on clarity, which is probably why they have this “conditional acceptance” stage.

Both of these are less technical than the usual research paper: they’re intended for readers who know some physics but are not necessarily specialists in any particular field.  The first one requires a bit of knowledge of relativity (students who took my Physics 479 course should be fine) , and the second one requires just undergraduate-level thermodynamics and statistical mechanics.

The first article is on the correct interpretation to place on the observed redshifts of galaxies in the expanding Universe.  I blogged about it when we originally submitted it.  These redshifts are usually described as being due to the “stretching of space,” but David Hogg and I argue that this conceptual model is misleading.  We claim that, contrary to what you often see in introductory textbooks,  it’s correct to think of the redshift as being due to a plain old Doppler shift.

Here’s the revised version of the paper.  It  doesn’t differ all that much from the one we originally submitted, although some aspects of the argument are expanded and clarified a bit in response to the referees’ comments.

The second article is on the relationship between entropy and the second law of thermodynamics.  It’s a response to a very nice paper by Daniel Styer, which attempts to show quantitatively that the entropy production due to sunlight is more than enough to account for the entropy reduction required for biological evolution (contrary to claims often made by creationists).  The original article had a serious gap in it: it depended on an assumption that was unjustified and, I argue, almost certainly wrong.  My paper presents an argument that doesn’t depend on that assumption.  The new argument shows quite rigorously that there is no conflict between evolution and the second law.

I blogged about the original Styer article and about my response a while back.  Here’s the revised version of the paper.  Thanks to the referees, I think the new version is much clearer than the original.  It’s also much longer.  I was thinking of the original as just a comment on Styer’s earlier paper, but the new version reads more like a stand-alone article.


The paper that I coauthored with Brent Follin (UR undergraduate) and Peter Hyland (Wisconsin grad student turned McGill postdoc) has officially been accepted for publication in Monthly Notices of the Royal Astronomical Society.  I thought it would be, but it’s still nice to make it official.  Congratulations to Brent especially, for becoming a published scientist.

Unlike the last one I posted about, this is a “real” refereed paper.  We decided to submit it to Monthly Notices, not Astronomy and Astrophysics as I wrote in my earlier post, for reasons that aren’t at all interesting.  Monthly Notices is a very good journal, and it has a way cooler name than A&A.