Sunday, May 28, 2006
Now you see 'em, now you don't.... pt.1
The great thing about science, as opposed to religion, is that the Universe is constantly demonstrating to scientists how little they know. And scientists, sometimes gracefully, sometimes pissily, and sometimes with great excitement, just have to put up with their pet theories and ideas being wrong. If you think of the laws of the Universe as set up by some type of god, then scientists are the ones trying to directly understand the mind and intentions of this god by paying attention to what the Universe tells them. (As opposed to many religious leaders who only pay attention to select parts of what some dudes wandering in the desert (or hanging out in Jerusalem) a few thousand years ago said this god wanted. They think of this as "perfect knowledge" despite the abundant evidence against parts of this "perfect knowledge" that this same god placed throughout the Universe for anybody who bothers to look). Anyway, for the vast majority of you who aren't scientists and so don't understand what science is really about, I will tell you. It is about data trashing your preconceived notions and forcing you to modify and change your ideas. That's why many scientists say anything that can't be proved wrong isn't science. This includes some of the sexier theories about gravity, cosmology, and black holes promoted by some of the biggest names in science. I mean, they may be right, but it isn't science until an observational test can be devised. A Good Theory is one that makes a prediction that is afterwards confirmed. A fortune teller that has a perfect record on predicting the previous week's lotto numbers isn't much of a fortune teller. But one that tells you tomorrow's numbers... that's someone to pay attention to. So making predictions and seeing if they come out to be true is what science is all about. Then you go out and get the data and hope it supports your idea. And sometimes, if you make a prediction, you don't even have to get new data. You might just look at data that has been lying around for months or even years in a new way. I'll give a few examples. One from a conference I attended a few weeks ago, and another from some guys at Columbia University in New York.
So first, let us go to the banks of the Rhein to a little town called Bad Honnef. This is a little German resort town that is ridiculously quaint. A tower of a ruined medieval castle watches over the town and river from a steep crag. White houses with those crazy brown wooden beamed criss-crosses and well tended gardens line narrow windy streets, just waiting for some crazed Hansel to come wandering out of the creepy woods on the surrounding hills and start nibbling on the gingerbread. In this town there is a massive stone building built in the
early 1900s but harking back to the middle ages called the physikzentrum. In back, there is a modern lecture hall with a glassed in foyer, where a bunch of posters were hanging up. One, by Isabelle Grenier and collaborators, was talking about a 4th EGRET catalog.
Now, from the last post you will remember that EGRET was a gamma-ray telescope with the majority of the sources it detected being unidentified. These were cataloged in the brilliantly titled 3rd EGRET Catalog. To understand this story, I need to talk a bit about data analysis.
NO, GET AWAY FROM THAT MOUSE, IT IS NOT GOING TO BE BORING!
(Doest I protesteth too much?)
Ahem. Anyway. EGRET data analysis. Two things you need to know. One is that gamma-ray telescopes have terrible resolution, so all the sources get confused with each other and something called liklihood analysis is used to distinguish sources from each other and from the background glow. A key component of this type of analysis is knowing what to include in your model. For example, if there are two sources near each other, but you assume there is only one source in that region of the sky, the liklihood analysis will cheerfully tell you that there is one source located right between the two real sources. A rather painful iterative process is used where you make a guess of a source location, let the liklihood analysis tweak it up for you, then look at where the data says there is some unmodelled excess, try adding another source to your model, let the liklihood analysis tweak again, then see if what you ended up with is any better fit to the data than what you started with.
A key component of the analysis is the diffuse background glow. Look at this map of the sky.
What you are looking at is an image of the entire sky in all directions. It is orientated such that the direction to the center of our Galaxy is in the center of the image, and the thin disk of our Galaxy (the Galactic plane) is horizontal across the middle. This type of image is commonly used when talking about the entire sky. In this particular case, you are looking at what the sky looks like in high-energy gamma-rays. That bright streak across the center is due to very high-energy cosmic-rays interacting with gas and dust in our Galaxy. In order to see discrete, compact sources in this image, you have to have a model of this diffuse emission. We have a pretty good idea of what the Cosmic rays are like, since we are in the plane of the Galaxy and they are hitting the Earth all the time. But what about the distribution of gas and dust?
Well, for that, we have maps of the radio sky. In particular, single, neutral hydrogen atoms (HI) emit light with a wavelength of 21cm, which almost all radio telescopes are designed to be able to detect.
However, much of the time, hydrogen gas is in the form of a molecule with 2 hydrogen atoms (HII) which does not emit at this wavelength. We think, though, that where there is hydrogen molecular gas there is also carbon monoxide (CO), which also emits radio emission. So, using HI and CO radio maps of the sky, we can model that gamma-ray image of the sky as this:
+
And of all of those sources, this is how they were identified:
Now one group of these sources which were particularly mysterious appear to be associated with our Galaxy, but not in the thin disk where most of the stars, gas, and dust hang out. These are the smaller green dots that are mostly between the lines above and below the center line in the figure. Now Isabelle has been staring at those dots for a long time, trying
to figure out what they could be. And what she noticed maybe 8 or 9 years ago was that they are distributed around the sky in a way similar to a bunch of relatively nearby regions of gas and dust. These gas regions are where alot of young, giant stars were fairly recently born and are collectively known as the Gould Belt. Her study led to all sorts of work on how these gamma-ray sources may be a bunch of nearby pulsars, or maybe even black holes, which were born from the most massive of these stars exploding (massive stars, like SUVs, burn up their fuel very quickly, and die very young). Proponents of the two main opposing theories of how gamma-rays are produced by pulsars each came up with ways in which their pet theory could explain these proposed Gould Belt pulsars. Pulsar searches were conducted towards these sources. All sorts of fun.
And then, a year or two ago, Isabelle was watching a presentation of new results from cosmic microwave background instruments. And one of the things that was being discussed was how to get rid of all the foreground radiation from our Galaxy and how alot of extra gas was required in their models. Looking at their models of the gas, Isabelle felt that she had seen them somewhere before. And when she learned how much extra gas was required, she thought "That can't be right, because then it would shine as bright in gamma-rays as the unidentified...EGRET........sources...."
oops. No, that couldn't be right.
But just to make sure, she went and tried including this extra gas in the model of the gamma-ray background radiation, expecting it wouldn't fit the data well at all.
It fit very, very well...
And so, although still preliminary, she is working on a revised catalog of sources which is going to be much smaller than the previous one, that was supposed to be the final one. Well, at least she was right in one respect. When she suggested that the sources seemed to be somehow associated with the regions of gas and dust that make up the Gould belt, she was right.
It just didn't occur to her that they WERE the Gould belt.
So first, let us go to the banks of the Rhein to a little town called Bad Honnef. This is a little German resort town that is ridiculously quaint. A tower of a ruined medieval castle watches over the town and river from a steep crag. White houses with those crazy brown wooden beamed criss-crosses and well tended gardens line narrow windy streets, just waiting for some crazed Hansel to come wandering out of the creepy woods on the surrounding hills and start nibbling on the gingerbread. In this town there is a massive stone building built in the
early 1900s but harking back to the middle ages called the physikzentrum. In back, there is a modern lecture hall with a glassed in foyer, where a bunch of posters were hanging up. One, by Isabelle Grenier and collaborators, was talking about a 4th EGRET catalog.
Now, from the last post you will remember that EGRET was a gamma-ray telescope with the majority of the sources it detected being unidentified. These were cataloged in the brilliantly titled 3rd EGRET Catalog. To understand this story, I need to talk a bit about data analysis.
NO, GET AWAY FROM THAT MOUSE, IT IS NOT GOING TO BE BORING!
(Doest I protesteth too much?)
Ahem. Anyway. EGRET data analysis. Two things you need to know. One is that gamma-ray telescopes have terrible resolution, so all the sources get confused with each other and something called liklihood analysis is used to distinguish sources from each other and from the background glow. A key component of this type of analysis is knowing what to include in your model. For example, if there are two sources near each other, but you assume there is only one source in that region of the sky, the liklihood analysis will cheerfully tell you that there is one source located right between the two real sources. A rather painful iterative process is used where you make a guess of a source location, let the liklihood analysis tweak it up for you, then look at where the data says there is some unmodelled excess, try adding another source to your model, let the liklihood analysis tweak again, then see if what you ended up with is any better fit to the data than what you started with.
A key component of the analysis is the diffuse background glow. Look at this map of the sky.
What you are looking at is an image of the entire sky in all directions. It is orientated such that the direction to the center of our Galaxy is in the center of the image, and the thin disk of our Galaxy (the Galactic plane) is horizontal across the middle. This type of image is commonly used when talking about the entire sky. In this particular case, you are looking at what the sky looks like in high-energy gamma-rays. That bright streak across the center is due to very high-energy cosmic-rays interacting with gas and dust in our Galaxy. In order to see discrete, compact sources in this image, you have to have a model of this diffuse emission. We have a pretty good idea of what the Cosmic rays are like, since we are in the plane of the Galaxy and they are hitting the Earth all the time. But what about the distribution of gas and dust?
Well, for that, we have maps of the radio sky. In particular, single, neutral hydrogen atoms (HI) emit light with a wavelength of 21cm, which almost all radio telescopes are designed to be able to detect.
However, much of the time, hydrogen gas is in the form of a molecule with 2 hydrogen atoms (HII) which does not emit at this wavelength. We think, though, that where there is hydrogen molecular gas there is also carbon monoxide (CO), which also emits radio emission. So, using HI and CO radio maps of the sky, we can model that gamma-ray image of the sky as this:
+
And of all of those sources, this is how they were identified:
Now one group of these sources which were particularly mysterious appear to be associated with our Galaxy, but not in the thin disk where most of the stars, gas, and dust hang out. These are the smaller green dots that are mostly between the lines above and below the center line in the figure. Now Isabelle has been staring at those dots for a long time, trying
to figure out what they could be. And what she noticed maybe 8 or 9 years ago was that they are distributed around the sky in a way similar to a bunch of relatively nearby regions of gas and dust. These gas regions are where alot of young, giant stars were fairly recently born and are collectively known as the Gould Belt. Her study led to all sorts of work on how these gamma-ray sources may be a bunch of nearby pulsars, or maybe even black holes, which were born from the most massive of these stars exploding (massive stars, like SUVs, burn up their fuel very quickly, and die very young). Proponents of the two main opposing theories of how gamma-rays are produced by pulsars each came up with ways in which their pet theory could explain these proposed Gould Belt pulsars. Pulsar searches were conducted towards these sources. All sorts of fun.
And then, a year or two ago, Isabelle was watching a presentation of new results from cosmic microwave background instruments. And one of the things that was being discussed was how to get rid of all the foreground radiation from our Galaxy and how alot of extra gas was required in their models. Looking at their models of the gas, Isabelle felt that she had seen them somewhere before. And when she learned how much extra gas was required, she thought "That can't be right, because then it would shine as bright in gamma-rays as the unidentified...EGRET........sources...."
oops. No, that couldn't be right.
But just to make sure, she went and tried including this extra gas in the model of the gamma-ray background radiation, expecting it wouldn't fit the data well at all.
It fit very, very well...
And so, although still preliminary, she is working on a revised catalog of sources which is going to be much smaller than the previous one, that was supposed to be the final one. Well, at least she was right in one respect. When she suggested that the sources seemed to be somehow associated with the regions of gas and dust that make up the Gould belt, she was right.
It just didn't occur to her that they WERE the Gould belt.
