Why I Lie Awake at Night

August 1, 1997

I'm still having nightmares. I think they are my subconscious telling me "Hey pal, these are all the things you need to be getting done in the next 42 days before we go into orbit around Mars." For five years now I've had a running list of the top 20 things that need to get done, but usually there are 40 things listed!

With about six weeks left before Mars Global Surveyor intercepts Mars on September 11, we on the Thermal Emission Spectrometer (TES) team are frantically worrying about finishing the absolutely crucial things on that list. The main thing we're struggling with now is the science analysis software and how we're going to handle the huge volume of data (about 100,000 spectra a day!). There are two problems. First, we've never flown an instrument anything like TES before and we've spent the last five to 10 years trying to figure out how to analyze the data and what we would do with it. In the old days we would've had the data and played with it for a year before going public, and after six months to a year we would've started to write science papers. In the current environment there'll only be three or four days before someone will come to us and ask where should we land the next rover? Where are the neat rocks? Where should we look for life?

The very enthusiasm about Mars has upped the ante in terms of the real-time analysis of data. The Pathfinder team is experiencing this first hand. "You've had the picture now for 20 minutes, what can you tell us about the history of Mars?" Now that everything's up on the Internet almost as fast as we receive it, you've got people calling in saying, "Hey, I see a rock over there on the horizon that looks kind of rounded. Why?" And the poor science team that's been up for 24 or 48 hours hasn't even looked at that picture yet! The pressure to do "instant science" is inevitable.

Back in November TES took its first test image of Earth. It wasn't really an image as Earth was smaller than one pixel in our instrument, but we got a beautiful spectrum of the Earth (that we've already written a paper about)! TES detected ozone and carbon dioxide in water vapor, which if you were searching for planets that had life, you would hone in on Earth. But in terms of further developing TES's science analysis system software techniques, it didn't help much because Earth was so tiny and the total number of spectra that we got with Earth in them were 15, so I processed them by hand. When we get into orbit around Mars, we'll get 100,000 spectra a day... that's why I lay awake at night.

In the spectroscopy world a grad student goes into the lab and takes 20 spectra of 20 rocks and ponders them for a month. The spectra are looked at one at a time, while noting the subtleties and details. No one has ever had to deal with 100,000 spectra a day. If I can process one spectrum at a time by hand, there's no conceptual reason why a computer can't process 100,000! The hard part is believing the numbers the computer is throwing back at you, and none of us wants to do that! We want to be left alone for a year to digest all the spectra and figure out the errors and the problems before we have to get up and announce our results!

Once Mars Global Surveyor begins transmitting TES data back to Earth, we're going to have to go public almost instantly. Our strategy will probably be to take our data, pick three of the wavelength bands and process them into a color image. Our brains are really good at identifying a blue pixel, a yellow pixel and a blob of orange pixels because they all look different. We'll then pull up the 10 spectra for each of those areas and then we'll be back in the old mode of doing 10 at a time by hand! But now the problem is, we're throwing away 999,990 while we focus on the 10!

If you go back through history and the literature, virtually every discovery that's been announced using spectroscopy has been wrong and had to be retracted because people rushed into things! For example, back in the Mariner 6 & 7 days the principal investigator who took the spectra of Mars announced at a press conference a week later, that methane had been found on Mars. The only thing that produces methane in a planet is life. Wow! Life on Mars! Well it turned out that a young guy had just published his Ph.D. dissertation at Caltech, with those exact spectra lines in it, which the principal investigator thought was methane. It turns out that the spectra were some previously unknown absorption bands due to carbon monoxide ice, which is a far less exciting thing to discover. So the principal investigator got very excited and announced things far too early, because after all, for discovering life on Mars you go down in history but for discovering carbon monoxide you don't! But you do go down in history for a royal blunder and that is hard to recover from.

It's easy to see why people have been discovering the wrong things. Every substance under the sun, including you and me, has a beautiful characteristic infrared spectrum because we're all made up of molecules in crystal structures and cells that vibrate. These vibrations are absolutely diagnostic of whether you're a mineral, carbon dioxide ice, water ice or gas. The TES instrument can detect everything because everything has an infrared spectrum. The bad news is that everything has an infrared spectrum and when you're trying to look at the minerals on the ground, you have to look through gases in the atmosphere, dust in the atmosphere, and water ice and clouds in the atmosphere. This is where you have to be incredibly careful not to confuse a mineral on the ground that's got an absorption feature that matches some mineral, but also matches carbon monoxide in the atmosphere. You have to be very cautious about what you announce. We're going to try and concentrate on what we hope are very strong obvious features that we can detect conclusively at first.

Today I'm working on two proposals that are due at the end of August, for instruments on the 2001 lander and orbiter going to Mars. I really want to participate in those missions! We've got an excellent team and good ideas so I'm spending a fair amount of my summer writing proposals for the next set of missions.

I'm incredibly psyched now. A month ago I was burned out but the Pathfinder lander has gotten me excited again. My biggest fear all along was that we'd fly the most spectacular instrument you could imagine to Mars and every spectrum would look the same--that Mars is made up of one type of rock, namely basalt, and there are none of these are evaporites and lake bed or hot spring deposits. But looking at the landscape and seeing that the rocks in the color pictures look different, has got me totally pumped!