When I describe to people how we search for meteorites out here – eight people on skidoos, looking with our eyes and collecting with our hands – a natural question that comes up is, why don’t we use robots for that? After all, we send some pretty sophisticated robots to do the work of humans in other inhospitable climates, like mines, ocean floors, and Mars. I’ve been giving this some thought this season. It turns out humans are actually pretty efficient at searching for meteorites.
One reason humans are better is that it’s difficult to come up with absolute criteria for positively identifying meteorites in the field. They typically have a fusion crust, except when it is weathered away. They are typically black, except when they are brown because the interior is exposed. They typically have a lustrous or iridescent appearance, except some don’t. They typically are neither squared off nor round, but some are. They typically contain metal, except the achondrites. And so on – you get the idea. Yet, ANSMET teams typically return only a few percent of “meteor-wrongs,” or terrestrial rocks that aren’t meteorites. This is because the human brain is really good at telling “same” from “different.” Once we each build a mental library of the kinds of terrestrial rocks that are out here, we are good at spotting things that are different, even if we have only a subconscious reason for spotting it rather than explicit decision rules.
Another reason humans are best is that meteorites in the wild aren’t big rocks sitting on a plain background. They are cupped in the ice, partially buried in snow, or behind terrestrial rocks in a moraine. We use our ability to look at different angles to tell shadows from rocks, to use the sun to see crystals glint, and our ability to pick up the rock and really look at it up close. This is an enormous advantage over rovers. I am a member of the Mars Exploration Rovers team (yes, Opportunity is still going strong at 10 years!) and we struggle every day with evaluating images of rocks from single vantage points or sun angles. I am fond of saying that a geologist gets an enormous amount of information from a rock simply by picking it up (density or heft), gripping it (is it coherent or crumbly), and turning it in the sunlight (does it have crystals and holes? what shape and size are the crystals?). As of yet, rovers can’t do this.
Even with all the logistical problems of getting a bunch of meatbags, all our food, and our fuel out here, we’re still a cost-efficient and effective “sample return” mission. We bring back about 1700x(!) as much material in a typical year as a robotic sample return mission could, for a fraction of the cost. Don’t misunderstand me – we need both humans and robots exploring the solar system together. But after 30 years of honing, ANSMET is optimized for grabbing these readily available, but remote, Antarctic samples right now.
-posted by….. Barb
Note from editor (rph). I get asked this informally dozens of times a year, and very seriously (by funding agencies) every year or so. Barb’s hit the nail on the head here. My shorter version of an answer is “intuition”. The human eye-brain system has evolved over millions of years to be superb at taking a little bit of visual information and spinning it into a complete story- is that waving stalk of grass over there a lion ready to eat me, or did a bird just take off? We’re good at looking at broad, ordinary scenes and immediately noticing what’s different (think “Where’s Waldo?”). Robots are not good at this…… yet. From an economics standpoint, the upcoming Osiris REx mission barb mentions will recover about 60 grams from an asteroid. that sample in front of her is at least 10x that; one sample. And the Osiris-REx budget would cover our costs for almost 1000 years (about 970 to be precise). For what we do, human searches win on the “science per dollar” and “science per minute” charts.