Focus On: Site Testing
  Gary Sanders & Matthias Schöck

For a telescope as ambitious as TMT, a great site has to be selected. The data we are collecting on the astronomy performance of our candidate sites is now good enough that we are learning a great deal from it. An internal quarterly review of site testing data, just conducted, now shows how good our five candidate sites are. They are all remarkably good, but they are different in significant ways.

Choosing a site for TMT is actually quite complex. We are learning this as we analyze the data. You might guess that you can specify in advance how to use the site testing data to rank the sites. It is not so simple. There are several qualities related to astronomy observing, and some sites are excellent in some, and offer less in others. The challenge is to choose a site that will facilitate all of the kinds of science observations that TMT can perform over several decades.

The TMT Web site has several earlier articles about our site testing and site selection activities. One is a travelogue that describes a visit to the sites in Chile and tells a bit about the technical requirements on the sites. Another chronicles visits and contacts with our site hosts, governments and institutions with roles in astronomy, and how TMT will solicit proposals from these hosts for TMT to be placed on their sites. More notably, this article tells how important the cultural and community setting is for each of these mountains. It asks whether we astronomers can be sensitive enough to balance our goals for a mountaintop site with local traditions and ceremonies, and with the mountain’s meaning to the native people of the area.

Foremost, however, is assessing the quality of how a given mountain would support astronomical observing. Astronomers record digital images, often in exposures lasting minutes or more. So the atmosphere above a mountain must leave the image as sharp as possible. Astronomers also record spectra of these imaged sources, and they want spectra from each little bit of the visual field so they can discern structure and dynamics. The atmosphere must leave the image sharp for this as well, and the light collection of a big telescope then makes it possible to slice and dice the visual field and still have enough light for each spectrum. In addition, astronomers want to measure brightnesses very accurately. This requires cloudless nights for long periods to obtain accurate measurements. Wind can shake the telescope, smearing the image. Yet that same wind can be directed to cool the air above the mirror to eliminate the heat-driven twinkling that we must avoid.

TMT has placed a suite of instruments on each of the five candidate mountaintops. These include a robotic telescope system to measure the optical “seeing” above the site. This is the quality that leads to sharpness. The figure shows one of our observatories on Mauna Kea. It is placed on cement blocks so as to avoid disturbing the ceremonially important mountaintop. Other TMT installations can be seen in the figures of the travelogue article mentioned above.

The robotic telescope system measures seeing quality from the ground up, and also from a few hundred meters above the ground and then upwards. This tells us how the ground layer, warm and shimmering, twinkles our stars. Other instruments record temperature, wind direction and velocity, turbulent layers close to the ground (see the Technology Nugget in this Newscast on the SODAR instrument), the fraction of clear skies (observed by a time-lapse all-sky camera), the amount of dust in the air, and the amount of water vapor above the site. The latter is important for astronomy at wavelengths in the mid-infrared, which may be absorbed by water molecules.

Wouldn’t it be nice if all of the mountains were excellent in all quantities? Then we could select the site on other considerations. This would be a bit like finding an identical house in different neighborhoods—we could weigh local factors. But we knew that would not happen.

That is why TMT used a global satellite survey to choose a suite of excellent but diverse sites. Coastal mountains bathed in ocean-mediated even temperature airflows are joined by inland, much higher sites with drier but more varied airflows. We knew that we would have a multidimensional set of data, and that the ranking and choosing would be complex. We knew that periodic reviews of the data would teach us the right questions to ask. We knew that each quarterly review would make our questions somewhat sharper. And that is happening.

The good news is that all the sites appear to be very good. With more data, that conclusion may well change as more seasons and statistics strengthen the message of the data. The bad news, though expected, is that now we have to consider each of our science thrusts and estimate how each of the diverse sites can be best used to carry these out. The combined observing model for each site is the meeting of the science program and site characteristics with the telescope capabilities. The hard work begins now that we have good data.

The even better news is that we can see that the data has a strong story to tell, and a couple more years of data and analysis will provide strong guidance in choosing a home for TMT.