Text and pictures © 2004-2013 Guillaume Dargaud
Last updated on 2012/12/13
"For astronomy, Dome C has seeing conditions 4 times better than the previously best known site." — Nature, 16 September 2004.
At the preparatory meeting for the 2005 winter-over in Dome C, Karim Agabi and Eric Fossat gave a presentation about some experiments they've been running with an australian team to measure the 'seeing' conditions at Dome C. The 'seeing' is the quality of the sky for an astronomer. Already measured with success in summer from 2000 to 2004, they ran an automated experiment throughout the 2004 winter. It failed in difficult circumstances, but not before revealing that Dome C is 4 times better than the previously known best site on Earth, Mt Paranal, site of the European Very Large Telescope. The astronomers are drooling all over the table talking about building a 100 meter diameter telescope in Dome C within 15 years. ONE HUNDRED METERS !?!
Besides the obvious advantage of the night lasting several long months (for very long exposures or for running correlations), one of the best type of astronomy possible in Antarctica is actually done in broad daylight: 24/7 monitoring of the sun in summer. The first solar astronomy mission conducted at South Pole in 1979 lead to the new science of solar seismology. But apparently Dome C does it better !
The only potentially better site on the planet is Dome Argus but nobody has ever set foot there as it's near the inaccessibility pole (the chinese plan to reach it in 2005). Its higher altitude may be better for infrared astronomy but satellite measurements seem to indicate worse wind conditions. To say nothing of its probably insane winter temperature. This stays in my brain and before I catch my train I drop by an astronomy store where I purchase 3000$ worth of telescope equipment in half an hour, not having the slightest idea how it works or whether it'll freeze at once. When I get home, I jump onto internet where I get flamed, with good reason, for trying to figure out how it works in the 4 weeks I have left before departure: "Take your time to learn astronomy in a club, start with a pair of binoculars, etc..." I wish I had the time... Incidentally I learn that I currently live in the best area of France for astronomy observation, but of course, 2 days after I receive the scope, the weather turns sour for a week.
Right: This is the view that makes astronomers drool about Dome C. In early afternoon the light from the backlit sun hidden behind the flag is not scattered at all: no humidity in the atmosphere, no ice crystals. In other words, excellent viewing conditions, you can even see stars in the middle of the day. Also visible are our instruments: the Meteoflux (left) and Sodar (right).
So what astronomical goals can I set for my coming Dome C winter ?
And what problems do I expect to run into and what possible solutions ?
So what do I know about astronomy ? When I was nine I dreamed about having a telescope and I started saving money for one. After 2 years of careful accumulation, I chose the model I wanted, but instead of going to the store to pick it up, I spent all the money on an early personal computer ! This was in 1981 and I pretty much wrote my future as a computer engineer that day. I never looked back, except a few times while on vacation in the south-west desert of the United States: excellent observation conditions most night. We used to bivy right off the car, taking turns watching the sky with a pair of binoculars. We could even see the phases of Venus. I took quite a few star rotation pictures at that time. As for the technical aspect of astronomy, I once built an equatorial platform for a friend, but never used it myself as I was more interested in the mathematics involved than in actually getting up at night.
Right: Sun setting behind the AASTINO container.
I arrive in Dome C on December 4th but my telescope and the rest of my photographic gear will arrive only after the Italica reaches Terra Nova Bay and is then flown here, in mid-january. I'm surprised to find Karim already all setup with 4 or 5 11" telescopes in full daylight. The sky is so pure that he has no difficulty seeing some of the brightest stars, but his goals are not really to do observations but to evaluate the quality of the sky. He spends most of his time fine tuning the station of his scopes. He has two scopes on the ground and two scopes on a five meter platform, this way he can evaluate the difference in turbulence between ground and higher up. Some of his scopes are mounted with a cache with two holes, forming a double image of Canopus and allowing for other kinds of turbulence evaluations, namely by looking at the difference between both images on the same line of the video.
Karim and his summer-campaign colleague Eric perform the first fringe interferometry ever done in Antarctica but a week of hazy days hinder their progress. Sometimes he does some solar observation and plans on doing helioseismology later on, but most of his experiments are geared towards evaluating the turbulence by various means. Later on Eric is replaced by Tatiana, an astronomer from Ouzbekistan.
Around the 15th of January, one night we hear some weird beeping followed by unnatural silence: it was the last scream of the dying Sodar. Apparently AASTINO (Automated Astronomical Site-Testing International Observatory) died from a generator failure: one of the Stirling engines stopped and the other one kept trying to start. Nobody here knows how to fix them so we'll just wait for instructions from their masters...
Left: John inside the AASTINO container.
I try for the 4th time this season to do a 24 hour sun sequence. I'm planning it this way: have two cameras each on a tripod, one the Nikon F100 with an 8mm fisheye lens, pointing vertically; the other a digital camera I rotate by 15 degrees every hour to keep the sun in the middle of a vertical frame. The fisheye images I will scan and overlay in PhotoShop; the vertical shoots I will just assemble into a panorama showing the curve of the sun during the day. But things are not going well. I need to have a full day of perfect weather to do this and the weather is not collaborating. I start by placing the tripods in the evening in the middle of the summer camp, hoping that careless crane drivers won't just run them over. Then I get up a 4 in the morning, place the F100 pointing towards the sky, take a picture trying not to be on the image myself, and remove the battery leaving the camera in place. I then take an image with a borrowed digital camera but I remove it from the tripod, otherwise it will just freeze to death, unlike the F100 which doesn't seem to mind the -45°C night temperature at all. Unfortunately, each time I do that the sky covers up between 6 and 10 in the morning. Not as bad as the previous years mind you. In 1993 I tried to do that picture on my last day in DdU. At picture 20 I found the roll finished and I didn't have any spare. Upon processing it months later, I discovered that some as***les had taken pictures of their butts with it, ruining the sequence. Then in 2000 in Dome C everything went perfectly and I shipped all my rolls to my photo agency which subsequently... lost them all. So here I am, still trying to finish a picture I started 12 years ago. And here's the end of the tale.
On January 23rd I finally receive my winter equipment. After fighting with the computer all afternoon to get it to boot with a bad disk, I spend part of the evening trying to remember how my telescope works. I have no idea how to station it in broad daylight since there aren't many stars visible with the naked eye, so I can't use the Celestron program. For tonight I just want to do some solar observation to check things out. I get the scope outside next to the lab, connect the power supply to it, noticing that the power supply cables harden immediately. The scope works but after 2 minutes I can't read the LCD display anymore. I have a resistor network but it's way too big to just wrap around the GOTO control. I just follow the sun manually for a while while a few friends come to have a look at the two clearly visible sunspots. I try some of the accessories (the various eyepieces, the barlow, T-ring for the Nikon...).
Left: I also take the time to take a picture for Eugene of LinuxAstronomy.org who's given me plenty of advice.
After 2 hours outside, the motors of the scopes don't seem to mind the cold. It's a very nice evening, without any wind, but the temperature is still around -35°C. Plans for the next observation session: learn how to do daylight stationing, prepare a small resistor network for the GOTO, change the power supply cables to Teflon, prepare a 15° wedge to use the scope in polar mode (as if the weather is not polar enough here !)... I also test the Meade imager on an old laptop I just reformatted but the Pentium II with 64Mb is crawling like a slug pulling a sled through sastrugi.
In the meanwhile I start playing with astronomy programs. First let's compute some ephemerids. The sun will go below the horizon for the 1st time on Feb 11th 2005 and will disappear completely on May 4th. That is without taking into account the refraction of the lower atmosphere that can show it a bit longer. Then it will be seen again on August 10th and will be permanent on november 1st. That's 82 days of alternation followed by 98 days of permanent condition. What about the moon ? It's always close to the horizon, going from a few days permanently above the horizon, a few days alternating, and a few days below. In summer the moon is not very visible, as when it's full it's opposite the sun and thus below the horizon; it's only visible in its smaller quarters, but then it's during the day. But as soon as we'll get nights, the full moon will be above the horizon, lighting our nights.
Right: Karim taking pictures of the moon on the roof of Concordia.
[This is a copy/paste from my blog] On Feb 23rd, as I walk back to Concordia I notice that the sun is setting, already cut in half by the horizon, and is just perfectly clear. I speed up to Concordia, remembering my last words of the interview: "Here we enjoy the purest horizon in the world". It's the perfect day for the Green Ray. And as I open the window of my 3rd floor laboratory and point the 400mm lens, yes, this exceptional phenomenon is fully visible, one layer of orange sunlight right on the horizon, and another separate layer a fraction of a degree above it, a stupendous bright green. I take a picture but then notice that the parameters of the camera are all wrong. After removing the flash, the spot metering, the manual mode, the underexposure and others I put the camera back to my eye... and the green ray is gone. Raging, I stay at the window another 30 minutes, while Karim is at his and others in their rooms are watching the show as well. The sun takes forever to set, a thin strip of orange light lingers on the horizon, slowly moving left and several times we see green blotches on it, but nothing as bright as that first flash. There's a spot of light on the ground moving gradually away from us towards the sun and I understand it's the area of maximum intensity of the light reflected by the atmosphere's inversion layer, in other words it's the place where we should be to see the green ray. I take lots of pictures but you'll need to come back next year to see the processed slides... The next day the conditions are similar so I pull out my telescope and ready it by the window, unfortunately there are thin clouds on the horizon and no green ray, even though the sunset is still stunning.
The next day Karim and Christophe spot a strange object on the horizon right after sunset: it looks like a pyramidal mound, whitish and grows steadily. After a few minutes it comes out fully and we recognize it as the full moon, but strangely deformed by the lower atmosphere, it's flattened to less than a third of its normal diameter.
On the 25th I try to film the Green Ray with the Meade DSI mounted on the scope with the solar filter. It doesn't work well at all as the integration time of the DSI is so long that it mixes all the turbulence into one blurry mess. It's understandable in the sense that the DSI is meant for long pauses, but still it would be nice to be able to use it for high-speed photography as well. The correlation time of the turbulence is of the order of a millisecond, so I should be able to get decent images with my traditional camera. I move the telescope from the frustrating green ray search to point it at the raising moon, but the turbulence is horrible. Jupiter, right of the moon and also close to the horizon, is just a blurred spot. Karim confirms that in order to have good seeing I need to point 'up'. So I'll need to get out of the excellent protection of the window... Then he went out to take pictures of the moon rising behind Concordia and came back with blue fingers, crying and swearing like never before.
Left: Astronomy session on the roof of Concordia: -55°C temperature and too much wind.
On the 27th I decide to stop working early and go for a book in my room around 22:00. While I'm in the bathroom (read: 'taking a leak outside') I see Jupiter emerging from behind the moon on an already fairly dark sky. I run up to Karim's lab where he's all: 'Yeah, I just saw it too, I didn't know'. So we setup my telescope in the radio room and point it out the window, to the same results as a few days before: horrible turbulence. We then dress up and go try it on the roof. The Teflon cable I use as an 12V power supply works fine, but not the connector itself which hardens and slips out. A bent with the Swiss Army Knife and then it holds. I want to station the telescope properly as we have a great view, but I barely have time to press two buttons on the GoTo before its display goes completely blank: damn LCD, can't they use LEDs for this kind of thing ? Fortunately the manual control still works and we get a good view of the moon, but Jupiter is already out of the range of the 40mm eyepiece. The wind on the roof of the building is quite high, of the order of 6m/s and shakes the telescope too much to make for interesting observation. The temperature is -55°C and since I need to remove my mask to see properly through the eyepiece, my eyelashes cover with heavy ice quickly. Karim takes a few pictures with his 300mm but I don't even bother. After only 10 minutes we go back inside with numb fingers and the firm decisions to make some modifications to the equipment: a heated transparent box for the GoTo, a replacement cable for the GoTo (longer and insensitive to cold), a better location (probably behind some containers)
March 3rd — The nights are now getting quite dark and plenty of stars are visible. It's time to get the equipment ready. First let's replace the cable of the GoTo as the little twisted phone cord that comes with the scope hardens immediately in the cold. Not only it's too short but I'm also afraid of breaking it. I have to cut the little plug inside the goto and solder the tiny connectors individually under a magnifying glass. Then the heating of the LCD: I cut a resistor network to the dimensions of the GoTo and glue it underneath. Its power supply is 240V with a small thermostat that I don't know where to put. What else ? A control box into which to put the GoTo (and my hand) made of a cardboard box with a transparent plastic bag glued on top.
Right: Observing session from the warmth of Concordia to see the Green Ray of the sun. The trash bag is here to avoid airflow.
March 18th — Today the wind dropped suddenly to zero in the afternoon and the temperature raised by 10°C in a few minutes. The sky is pure as can be so we suspect we might see the green ray of the sun again. Last time the sun set right in front of the window of my lab, but now it's from Karim's lab. I move my telescope to one of his windows, moving his potted plants to the middle of the room. It takes me a long time to get the scope started (it's hard to do an alignment from inside a room !), all the accessories sorted and myself dressed up as the cold air gets into the room right from that window. Karim at the window just 3 meters away is only wearing a T-shirt without trouble. There's a thin cloud on the horizon that make us doubt we'll see it but we stand watch for more than half an hour and, even though we never see it fully, we see small green areas lit up briefly as part of the light that makes a line on the horizon. The problem is the heavy turbulence, mainly coming from the temperature contrast at the window. I shoot a lot of images, needing to add a weight to counterbalance the heavy camera on the back, but I need to adjust the focus continuously and all the images are actually out of focus. When we turn around to go for dinner, we notice a casualty: Karim's plants are all frozen on the floor of the room...
Left: The infamous and elusive green ray of the sun.
Next day, at around 6 in the afternoon the weather conditions are identical to the previous day: no clouds, wind around 3m/s, -58°C and same wind direction. We decide to attempt the green ray observation again. The only difference is that the turbulence, as measured by Karim's experiments, i 3 times calmer than yesterday. In order to avoid turbulence we know we should get on the top of one of his observatory stations, but I'm not too keen on freezing by buts off yet. Let's try something else first: in cut open a trash bag, punch a hole in the middle which I put around the tube of my Celestron, and tape the sides of the bag to the open window. It works better than expected: very little turbulence escapes from the window and we don't freeze. Too bad his plants are already dead. Karim puts his camera on my scope and connects it to his computer so we can see the pictures immediately. Observing conditions are much better than yesterday as there's very little turbulence and sharp focus. I take hundreds of pictures as the green ray comes and goes in stripes and sometimes I see something I've never even heard about: blue colors and even an entire spectrum. It's not surprising since this phenomenon comes from a refraction effect identical to a prism, but theoreticians claim that the blue or purple gets diffused by the atmosphere so you can't see it. Well, they're wrong, all you need to do is come to Dome C to freeze your butts off during the winter... Let me assert that this picture has received no post-processing whatsoever. Strike one off my to-do list.
Right: This is an enlargement of part of the Green Ray phenomenon, the image having been enhanced for contrast, and it shows a full spectrum.
March 22nd — The turbulence is beginning to be beyond any astronomer's dream now that the night is long enough to let the air stabilizes. It's autumn today, so, just like everywhere else on the planet, we get 12 hours of darkness. Yesterday Karim measured a incredible turbulence length, a lot better than the previous world record. And it's just the beginning. Now that the rule is that we cannot go out at night alone, we take turns going with him to his observatory shelter. Tonight I'm the one going and it's an opportunity to take some pictures while he's calibrating his telescopes, looking for stars that have moved out of the field and aligning his interference masks. In the distance we can see the lights of Emanuele and Roberto, gone to do some snow sampling in a hellish frozen snow hole a km away from the station. After a while bending our necks at the sky we just lay down on the hard snow, looking at the sky and letting the cold seep slowly through our clothing. I remember many such nights in the Utah desert, but Jenny was a bit more sexy than Karim and we were a lot warmer under the sleeping bag. I also remember the quality of the Utah sky, but here it's even better: the first thing one notices is that the stars don't twinkle at all; the sky looks static and dead. Even with the chart we have a hard time figuring out the constellations, so much for having an astronomer handy...
March 25th — Today again the conditions seemed good for another sighting of the green flash of the sun. So I station the telescope by the window of Karim's lab well in advance and spend some time trying to image the current sunspots with the Meade Deep Sky Imager camera. The quality is horrible, both from my lack of practice with this hardware but also from a high level of turbulence. When the sun sets, I can guess that there's some green and blue colors, but the turbulence keeps it all impossible to focus. After the sun sets I give up and head back to my lab. After 5 minutes Karim calls me all excited: the sun is visible again ! Some large areas on the ground are acting as reflectors and giving us a mirage of the sun lasting well over half an hour.
Left: Karim and my telescope stationed at the window to observe the sunset.
Right: Mirage of the sun, half an hour after sunset. The top dark line is actually the real horizon.
After the sunset and the balloon launch we tried to do some observations of the stars and M42 from the window, to test all the modifications I've done to my scope lately. First of all it's very hard to station a scope properly from the inside of a building ! Then the outflow of warm air from the window causes condensation on the refractor of the scope, which is exposed to the outside cold behind the gown I put between the scope and the windowsill. And this outflow also causes the turbulence to be horrible, like all astronomy manuals will tell you. The gown somehow minimizes this but it's still pretty bad.
Left: Karim observing from one of his telescopes.
March 29th — Today the seeing conditions are excellent, the wind non-existent and my equipment is ready for its first test outside. For the first test I decide to install the scope on the roof, although it's probably not perfect with regard to stability and thermal plumes. There's a very narrow spiral staircase leading on the roof, with only a wooden board as a horizontal trapdoor. There's barely room for my portable computer on top of the stairs and the mess of cables. Let's see: one cable from the computer to the GoTo which is inside a small cardboard box with a hole for the hand and a plastic window on top. One power cable also to that box to the heating resistor (I hope I won't get shocked). One 12V cable to the scope. One power cable to the other resistor I placed underneath the electronics of the scope. One cable to the Meade DSI camera installed on the back of the flip mirror. Once everything is connected I launch the scope in manual mode and start the alignment procedure. Something I notice quickly is that the alignment is wrong. I end up using Jupiter and the nearly full moon just to be sure I'm not the one screwing up. If I tell the scope to go to the moon immediately after the alignment procedure, it's off by a good 10 degrees. I have no idea why, but after a while I notice that the vertical motor has some trouble catching up, probably with the cold. I place a lead on the front of the scope to try to balance it and indeed it works better.
Right: The ConcordiAstro platform covered with snow in the cold of autumn.
I'm now ready to move to the warmth of the stairs and put the board back in place with the idea of doing everything by computer. First problem, I cannot get the remote GoTo to connect, which is strange since it's a simple cable and the (real) goto is warm inside its cardboard box outside. Then the camera displays only garbage: horizontal white lines. After a while I give up with the 'puter and since it was very nice outside before I just step outside for a 'classic' observing session. I still have trouble with the alignment but the seeing conditions are quite good and the only problem caused by the cold (-63°C) is that if I happen to breath on my eyepiece it freezes instantly. Just when I begin to be familiar enough with the sky the wind stops completely and the smoke from the power plant stays right here, turning the sky all hazy. After a while I give up, the only things I manage to observe were the moon, Jupiter and a hazy Mars, very low on the horizon.
Left: Halo phenomenon around the moon, above the summer camp.
April 29th — I've been meaning to test the new modifications of my scope for now more than a month. First we were very busy, especially with the balloon launches happening in the evening. Then I had other excuses. Then there was a week of cloudy weather. Tonight I finally I decided to test the webcam without the scope. I build a Nikon mount for the Meade DSI out of a plastic tube and a lens back-cover. The problem is that with wide angle lenses the precision of the distance between the CCD and the lens must be in the order of a couple hundredths of mm. Just passing the extremity of the tube with sandpaper is enough to go from not enough to too much. Finally I did some tests with a computer on top of the emergency stairs, the DSI laying flat on the roof of Concordia and the lens on top. After some difficulty focusing I got some basic images.
May 2nd — Now that darkness has come and will last for the next 3 months, I'm eager to get my scope out. In the afternoon I test all the equipment. The hand control of the scope begins to act weird: it doesn't respond, or moves the scope by 20 degrees, or just disappears, leaving only the red background light. Then at a certain point it starts getting really hot and I lose the display. I suspect a short circuit in the connector, so I cut it off and rewire it. After that nothing, no display, no light, no control. No more scope.
Right: A moonrise with some mirage phenomenon.
May 3rd — Never mind the scope (argh!), I'm sure I can get at least some images with my redesigned DSI: it's attached to a Sigma 8mm fisheye lens (or any other Nikon lens), and stuck inside a ring collar allowing the use of a tripod. I point it more or less at random on the Milky Way and fiddle with the settings for a while. I first use the default options of the stacking software (the Autostar DSI), but it follows only one star for the stacking, so all the other stars are in apparent rotation around it and the result is useless. So I have to save all the images and process them in some other software. I first save basic BMP images but the various image processing softwares I try won't open this format (they takes only FITS files). So FITS it is. A first problem is that although the image I have on the Live screen is OK, the FITS image is all dark because it's just a raw image, making it harder to manipulate in the program.
I first try to save as Uncombined Fits Int but the acquisition software just gives me a very helpful error 142 when trying to save. There's no help file. Then I try Uncombined Fits3P, which fails opening in the Autostar IP (Invalid Fits Image: GetLastError returned 0, which happen only if I try to load them as a list, not individually). Finally I use standard uncombined FITS. I can open those in the AutoStar Image Processing suite and thanks to ChuckR's advice I manage to align them properly.
Now I have a bunch of aligned FITS files and I consider it a big step forward. I merge them using [Group][Merge][Combine][Sum], but I have to be careful to keep the channels separate, otherwise it merges everything into a single B&W file and what I want in the end is a color image. Something else I cannot manage to do is a simple histogram adjustment: [Tools][Histogram] displays a fine graph, but doesn't allow me to change it. [Process][Stretch Image Data] looks promising but no matter what I select the resulting image is totally black or white, and the graph is empty.
Right: The Milky Way seen from Dome C, my very first image (so please be indulgent!) with the Meade DSI mounted on an 8mm Sigma lens. Stacking of about 15 images with 15s exposure each.
I fiddle with the program to try to combine them into an RGB image with [Color][RBG Merge]. It either crashes before even the menu (Image Buffers not allocated) or crashes while processing the 3 files I select in the popup window. I'd like to try combining those 3 files into another processing suite, but PixInsight gives errors if I try to open them... FITs files are apparently not even compatible with themselves.
I'm very quickly learning to hate the Meade programs. They look like they've been written in QuickBasic for Windows 3.11. The file popups never remember which directory was last opened. No drag'n'drop either. Hardly any window can be resized. You can press F1 as much as you want, there's no help coming. Plenty of image processing functions don't have a preview. The many boxes that contain file names cannot be enlarged nor scrolled, so you can't see the end of the file name; something particularly annoying since the Autostar IP adds plenty of tags to the file name each time you call an image processing function. And the killer bug: there's no UNDO; in an image processing program where you keep trying function after function&!?!
I want to thank the members of the Nexstar and Deep-Sky-Imager mailing lists for their kind assistance. As well as the authors of Registax who answered immediately to my query.
Left: Internal modifications to my Celestron Nexstar 5i: 2 resistor networks added on the inside, one near the electronics and Az axis, the other near the Alt step motor.
Right: Observing session from the trapdoor of the roof of Concordia: the computer barely fits for Scope control and DSI image acquisition.
May 5th — Good conditions tonight, little wind, temperature about -55°C. Let's go out and look at something. Since I want to operate the scope remotely with the USB webcam I cannot be very far. One of the most convenient setup is the roof, with the computer located below the trapdoor. I prepare everything inside the narrow emergency stairs: the portable computer on a narrow plank, the scope on its tripod tied temporarily with the velcro to avoid having it falling down the stairs, the DSI prepared for wide field imaging piggyback on the scope. I plug everything in: the DSI into the computer, the power of the scope, the resistors that I installed inside the scope, etc... I go back to my room to dress warmly and when I get back to the top of the emergency stairs... there's smoke coming out of the base of the scope !!! In a panic I tear the plugs off everything and bring the scope back down. I unscrew the cover of the base of the scope under a smell of burnt plastic and a sense of dread: one of the two resistors networks I'd placed inside has melted all the wires connecting the electronics to the step motors... I had tested my setup inside and it was working fine, so what just happened ? With 200W of resistor power, I had to use a thermostat, but apparently the thermostat would trigger only at 15°C and it never did during the tests. When on top of the stairs it did trigger, with an error in the wiring, and it heated too much. All the cables are merged into an ugly mass of plastic and there's not much I can do besides replace all of them. Out of laziness I just give it another try, I plug the scope back, power it up and send a motion order... it still works !!! I guess the metal parts of the cables don't touch each other.
Left: Karim taking moon shots in the back while Emanuele checks on us from the trapdoor of Concordia.
So let's go back outside to start this observing session. While I was down checking on the scope the wind has picked up and the windchill is tremendous: it fills like my face is being dipped in liquid nitrogen, even with the neoprene mask. I do an approximate (aptly named) polar station in a hurry and go back down. I'm still fiddling with the settings of the imager when the power goes off. Apparently the scope triggered the breakers. There's still something wrong in my newly redone electrical wiring of the heating resistors. It's not my day since after bringing back the scope inside I break the power plug.
Right: A slightly iced up Meade DSI operating at -70°C, mounted on a Sigma 8mm lens via a homemade and very imprecise mount. You can see a moving southern cross on the left.
In the next few days I experiment with the DSI on a static mount. I try to use it to capture auroras, but the big night with maximum activity I'm in my bed and nobody comes to wake me up. The problem with observing just outside the roof trapdoor is that there's a lot of condensation on the lens: some humid air escapes from the building, cools of quickly and deposits its load of humidity on whatever surface is available. The building antennas already look like Xmas trees loaded with snow. I try to add a heating fan to the contraption but the first tests don't show much improvement.
Left: Karim working indoors on one of his telescopes.
July — Karim is experimenting. He brought back one of his heavy telescopes from his observation platform to put it up on the roof of Concordia. The idea is to check if the observation quality improves with the height above ground. According to his (and some of my) measurements, the atmospheric perturbation is concentrated on the first 30 meters above ground. By putting the telescope on the roof 22m above ground, most of this turbulence should be eliminated, but there are other issues cropping up: it's more windy up there so there's more vibration on the scope. The installation is not easy as the roof is not meant to carry equipment. Jeff needs to build a small metal platform, but welding with windchills in the -100°C range proves next to impossible. After several days of work, the test is operational and concludes successfully at the end of the winter.
Right: Eric showing a spot of frozen condensation inside a newly installed telescope, just brought out from indoors. After a few hours in the sun it will be gone.
During the 2005 winter there are 3 main astronomy experiments, all geared towards site testing and evaluation. I've talked on other pages of the balloon launches, carrying special thermometers able to measure a 1/1000th of a °C thermal variation on two distances (20cm and 70cm), used to record the thermal turbulence directly, but only twice a week on a vertical profile.
The DIMM experiment, a telescope on an platform 8m above the ground used to measure the seeing, monitoring it every 2 minutes. It continuously follows a star with a mask with two holes on the input. One hole is just a parallel glass, while the other is a 1 arc minute prism used to deviate slightly the incoming light. The camera records the star and its deviated double and the difference between both images is integrated to give the variance of the motion resulting in a direct measurement of the seeing.
Then there are also two twin telescopes part of the GSM experiment, one on top of the platform and the other on the ground pointing the same star with the same light splitter as above used to correlate the 4 images 2 by 2. A computational model with 3 or 4 parameters can forecast the results depending on the observations and deduce all the important parameters: seeing, isoplane angle, coherence time and external scale. This last parameter is important for building large telescopes, it determines the numbers of actuators necessary for correcting the image when using active optics. The good thing is that above a certain value it's not necessary to add any more actuators.
Right: Eric, the new astronomer, in front of his new C14 and C16 scopes.
This was 2005 and there are several new astronomy projects for the 2006 winterover. First the continuation of the site testing with a Mig16 (406mm/16") telescope used to replace the balloon launches for turbulence measurements. This telescope acts as a Scidar by measuring the stratification of the turbulence, focusing at various heights to return a vertical profile of the turbulence. I has a 1km resolution in its upper reach (20km) and about 100m at its lowest range. It follows a star in an unfocused way and the camera itself is moved about the focal plane in order to focus on closer ranges. This should confirm the measurements of 2005, that is to say that most of the turbulence lays in the first 30 meters, with a seeing of 0.2" to 0.4" above that. It seems like a waste to use such a large scope to point at an out of focus star, but that's the way it works.Then there are some other experiments beginning with real sky observations. A Meade C14 (14") telescope will do some stellar coronography. A special lens divided in 4 quarters adds some phase to the incoming light (0, pi, 2pi, 3pi), then goes into a collimation lens and a Liot stop in order to remove the light from the central star by a factor x100. This can be used easily to image faint double stars but also large exoplanets. It's a prototype used for the first time in Antarctic conditions so we'll see how things go. The lens itself is a prototype where the wavelength-thick layers of glass are glued using molecular adhesion technology.
Another general purpose new experiment is Airbus. Nothing to do with the airplanes of the same name, this experiment measures the infrared background of the sky. With the very low humidity and high altitude of Dome C some IR windows are open for observations which are normally impossible in other places of the world. The K-band around 2µm is about 0.4µm large at Dome C and allows for the observation of young objects such as star cocoons, dust clouds, etc... The observable spectrum is much wider than at other sites, with more spectral bands allowing for more chemical elements to fit within the window. This experiment will measure the atmospheric absorption in various bands with a liquid nitrogen cooled camera mounted on a 10cm glass, pointing at random angles in the sky for some periods of time. Yes, even in the middle of Antarctica we still need to cool down some equipment ! And if you compare with the successor of Hubble, the future James Webb Space Telescope, which will do observations mainly in the infrared for a price tag of $4.5 billion (!), you'll understand why it's important to have cheaper observatories. And earlier ones too, as the completion for the JWST is still years ahead.
In 2006 the turbulence balloon launches won't continue and will be replaced by the same ultra-sensitive thermometers fixed on the 35 meter mast west of Concordia. There's no point in continuing to measure high altitude turbulence when most of it is so close to the ground that the balloons go through too quickly to record it precisely. There will be 4 sets of the same 4 receptors as on the balloons, placed at different heights on the mast to measure the Cn2 continuously. From the balloon measurements of 2005 it is estimated that 87% of the turbulence lays in the first 30 meters above the ground. The purpose of this mast experiment is to determine how the thickness of this turbulent layer varies with the time. In order to build a large telescope at Dome C one must first determine at which height to place it. Should it be 25 meters ? 35 meters ?
The turbulence is created both by the strong inversion gradient (about 20°C in a few tens of meters) and the wind speed gradient (wind shear). So building a snow butte is not necessarily the best way to raise a large scope as the wind (and the turbulence) would just follow the artificial slope and reach the scope. In South Pole the turbulence layer is 220m thick on average and thus the seeing (1.7") is no better than in the average urban backyard. It is of note that during the Dome C summer the thermal gradient is much less than in winter, sometimes disappearing completely in the afternoon. There are periods of one to 3 hours when the air column above our heads is totally stable and I've confirmed this with my own radiometer measurements.
Eric Aristidi, the astronomer in charge of the 2006 winterover also plans to take some deep sky exposures, thanks to the extreme darkness of the sky here. But it's more of a hobby than a real professional project. BTW, he just started a blog of his own winterover where you'll find a lot more Antarctic astronomy info.
Left: Animation of the green flash of the sun, lasting about 15 minutes, photographed in august 2005.
On a longer time scale there are several astronomy projects interested in Dome C. A sign of the times is the foundation of ARENA, a European group dedicated to Antarctic Astronomy with multiple backers (the European Space Agency and others). Their main goal is to determine the best way to build an observatory in Antarctica, in respect with other places of the world or outer space. Which experiments should be selected ? Where should they be placed ? Is Dome C really the best observatory on the planet ? Well, it's a difficult question as few places of the Antarctic plateau have been tested in winter. Both Vostok (Russia) and South Pole (USA) suffer from winds, clouds and thick turbulent layers. Dome A (China) is potentially very interesting for its high altitude but it's been reached for the first time only in 2005 and it will be a while before it can support winterovers. The cost of an observatory must take into account the cost of all the logistics involved in building a permanent station and Concordia now has the large pro of existing. Also Dome C has one of the best weather on the continent (and maybe on the planet): between february and june 2005 the weather was clear 96% of the time, allowing for very long windows of continuous observation or exposure.
Right: Milky Way above the glaciology shelter.
Among future projects (15~20 years) is a large interferometer. The only way to improve the results from telescopes is to increase their diameters as it increases both the resolution and the incoming brightness. In the '50s the largest diameters were around 4 meters (Mt Palomar, Kit Peak...). In the '80s it was about 8m (VLT, Gemini, Keck...). By 2015 astronomers expect to reach 30 to 50m diameter (ELT, OWL...) with either plain or segmented mirrors. But there is another cheaper and easier way: building an interferometer. Take an array of several small scopes and correlate their lights to turn them into a scope of equivalent diameter. The resolution is just as good but not the luminosity; it's a tradeoff.
Astronomers currently know how to interfere 4 telescopes into a common image. The Keops projects of the Nice University plans no less than 36 telescopes of 1.5 meters organized in 3 concentric circles, the outer ring having a diameter of 1km. With a diameter of 1.5m and the excellent quality of the seeing at Dome C there's no need for adaptative optics on each individual scopes. This setting would be able to resolve the Sun-Earth distance at 1kparsec (the thickness of the Milky Way). The principal application is to detect not only exoplanets (we already know several hundreds larger than Jupiter) but still undetected exo-Earths with a 90% probability. This kind of direct optical observation relies mainly on long exposure times to increase the signal to noise ratio and here in Dome C during the long winter nights it's possible to have exposure times above a month of duration with a 3 month window of perfect observing conditions ! Such a long exposure would bring the quality of the S/N to values never dreamed of before.
Left: Milky Way above the glaciology shelter.
Another project closer to us in time is stellar seismology: study of the vibrations of stars, starquakes. In 1976 Eric Fossat was the first to perform it on the sun... right here at Dome C. When you consider a star, it is a big ball of fluid that oscillates in various modes which change slightly its diameter (and its brightness). For instance the sun oscillates with a 5 minutes period with many harmonics due to compression waves. But this kind of optical observation relies on long periods of continuous observation in order to be able to extract useful frequencies and their harmonics. So this experiment thrives on the 3 months of continuous darkness present here (or the 3 months of permanent light in summer in the case of our sun). Month-long exposure times would be perfect to determine the internal structure of stars as well as their densities and other characteristics. This experiment was supposed to be tested some time ago already and is still a couple years late in the making.
In order to detect exoplanets three methods have been tried so far: finding the Doppler effect of a star rotating around the common gravity center with a large planet; direct optical observation using a coronograph to hide the bright star and try to see the dimmer planet; and detection of eclipses when a small planet passes in front of the star and dims its light temporarily. There's the project A-STEP based on such transit detection. If you are lucky enough to have a planet's orbit passing directly between us and the star this experiment should detect it. An additional condition is that the orbit must be fast in order to merge several transits during a single observation session, meaning that the planet is close to the star (and thus probably small). Here again Dome C conditions excel by allowing several months of continuous observation, something impossible in the rest of the world. From the shape of the light variation during the transit one can compute the size of the planet. So far the only exoplanets discovered by the various methods indicated above were at least Jupiter-sized, but this method should be able to detect planets smaller and closer to their stars than Mercury.
So in summary the strong points of Dome C for astronomy are: transparency of the atmosphere in infrared, long observation periods (3 months continuous), very low turbulence (best in the world), very good weather (best in Antarctica), very few auroras (which are difficult to filter out as they are not monochromatic). Expected results cover detection of Earth-like planets and Vulcain-style planets (Vulcain is the name given to an hypothetical and never detected 'zeroth' planet closer to the sun than Mercury).
I have taken many long exposure pictures during the winter, thanks to a Nikon EM I purchased on eBay before departure and which worked flawlessly even when doing 6-hour exposures below -70°C. As of august 2006 I'm done with the scanning and processing of all the resulting slides. I'm currently adding the new pics to the site. Thanks for visiting.
Right: The Milky Way seen through over the 180° horizon of the Dome C sky. The camera equipped with an 8mm fisheye lens was tied to a telescope to allow for a few minutes of exposure. The telescope is visible on the upper left of the image and the band next to it is a faint aurora. Then the two white blurs are the two Magellanic clouds. The dark area on the Milky way is called the Coal Sack and is a typical southern hemisphere feature, quite visible to the naked eye. The reddish lower part is light from the sun, still way bellow the horizon but shining nonetheless. More info about fisheye photography.