HiWish fulfilled (or, be careful what you HiWish for)

Yesterday the Jet Propulsion Laboratory issued a press release announcing the first eight image releases that resulted from HiWish suggestions. That's the program by which members of the public can request that the Mars Reconnaissance Orbiter's highest resolution HiRISE camera take photos of specific spots on Mars. And guess what -- one of the eight was one of my suggestions! Woo hoo! In fact, it was the observation that I described in detail in my earlier post about HiWish. Here's the screen cap I posted then, showing the area of the image that I requested, and the science justification I wrote for the observation request:

And here is the area of the actual image that was taken, which actually covers even more of the rim of the crater than I'd requested:

If it strikes you as a little more than a coincidence that one of the first HiWish images selected was from a blogger who might publicize the program, HiRISE principal investigator Alfred McEwen addressed that directly in the caption he wrote for the released image:

This image covers some high-standing topography just outside the rim of an impact crater about 30 kilometers (19 miles) in diameter near a Martian hill named Zephyria Tholus. What formed this hill? Could it be a volcano? That was hypothesized to be the case in a paper published in 2001, and this suggestion was entered to test that idea, perhaps from seeing internal layering exposed by the crater. This is an example of what we regard as an excellent science justification for a target suggestion: following up on a testable hypothesis with specific observational goals. It was entered by eminent blogospherian Emily Lakdawalla of the Planetary Society. If you're thinking we chose to acquire this image early because we want her to blog nice things about HiRISE and HiWISH, you're right! But it's also a good suggestion.

(Hey! "Eminent blogospherian!" That's a new one. Thanks, Alfred!)

So, what do I think of the image, which you can see here? Well, um, how do I say this? ...it's really, really dusty. In fact, it's so dusty that they changed the title from my proposed "domical feature near Zephyria Tholus" to "Dust-mantled topography near Zephyria Tholus." Ouch. That may make this image most educational by the cautionary tale that title change tells: that there are some kinds of features that really won't benefit from being seen at the astounding resolution of HiRISE. In fact, I think an argument can be made that of the eight images released yesterday, this is the least spectacular. The "Lobate debris apron in Deuteronilus Mensae" is much more instantly beautiful, and the caption to "Boulder strewn plain in northern Utopia Planitia" is really thought-provoking.

But HiRISE wasn't sent to Mars to get pictures that are pretty; that's just a side benefit (albeit a significant and valuable one). It was sent to Mars to get images for science. With that in mind, let's pore over my incredibly dusty view of a dust-mantled domical feature near Zephyria Tholus and see if there is anything I can learn about putative volcanism in the Zephyria region from it.

Checking out the context image, it looks like the center of the image crosses the feature I was interested in. That's important because it's only the center 20% of the pixels in any HiRISE observation that contain color information. I'd like to look at the color data, since it's prettier, and also because color may help guide me to less dusty (less red) areas, but I needed to make sure that the color data was going to contain features I was interested in. Also, looking at the context image tells me that the image's center is at the northern edge of the feature I identified as being a possible volcano. So I'm going to direct most of my attention to the southern half of the image.

So I open the color version of the image, and here's what it looks like:

Wait, did I click the wrong button? This is supposed to be in color, right? Yes, in fact, it is in color. Then why does it look so gray? The short answer is that it looks gray because there is dust everywhere.

But wait, you may be saying, isn't Mars' dust red? Yes, Mars' dust is red. But when the images are displayed for viewing on the HiRISE website, they are not shown as perfectly calibrated color, because calibrated color (as you've seen in the amateur-produced, carefully colored images from the Spirit and Opportunity sites on Mars) appears very washed out; there's little color detail visible in "true" colored views. Remember, we're not here for pretty pictures. We're here for science. And to be able to observe what's really going on in the picture, scientists zoom in on subtle color variations by stretching the contrast in images to emphasize differences between different areas. The quickest way to do this is to set the brightest pixel values in an image to be white, and the darkest pixel values to be black. If you do this independently for each channel in a red, green, blue color composite, it can produce very striking images from Mars, where the soil is bright red and the bedrock appears dark blue.

But if there is only one kind of material in an image -- if every spot has the same relative brightness from red, to green, to blue channels -- then this kind of contrast stretch makes everything appear gray.

Gray is bad news. That pretty much means there's no bedrock to find. (On the other hand, gray tells me it's a waste to just look at the color image; I really ought to be checking out the entire grayscale image.)

I can hold out some hope to learn something from the photo, however, for two reasons. First, the places where I am most likely to find bedrock are the places where a later impact crater cut the mountain. That impact crater is on the right; it cut a slope that faces east. Since the Sun is coming from the west (from the left), the sloping wall of the crater is in shadow. Therefore, the bedrock I'm looking for could be hiding in the shadows, where it wouldn't have been able to contribute much to the image's color anyway. Second, even if every single visible surface is mantled by dust, differential weathering of any layers that may be in the bedrock could result in topographic expressions that I should still be able to see. Wishful thinking, probably, but we have the data; it'd be a waste not to look!

So, I need to dive into those shadows and see what I can find. The way I look at the data in its highest resolution is to use the IAS Viewer, handily linked to on the right side of every single HiRISE image. It's the best way I know to view the data but it doesn't do one thing I really really wish it could do: allow me to save to disk any chunk of the data at its full resolution. Instead, if I want to grab full-resolution images, I have to save what's on my screen, one screenful at a time, and reassemble the tiles in Photoshop. It's a bit tedious, but it doesn't actually take all that long. Here's what that eastern flank of the hill looks like. The version below is at only half the full resolution, or 1 meter per pixel.

There's a lot of texture in that flank, enough to convince me that we are, in fact, seeing bedrock (though it's likely still got a blanket of dust on top). Do I see layering? Nope. If it's there, it's not obvious. Rather, I'm struck by a round, bouldery texture to the bedrock. Now, that's not necessarily inconsistent with volcanism; I need to do more reading on what volcanic landforms look like from space.

I will note that, seen at full HiRISE resolution, even the dust is beautiful. It makes such a smooth, blank canvas in some places, here and there striated by dunes that look like fingerprints. Where craters are fresh, there are beautiful splashy patterns around them, where the dust flowed like a liquid as it cascaded outward from the site of the impact. In some places those ejecta patterns are controlled by topography, making butterfly patterns that are elongated downhill. In other places the craters are almost totally filled in. Here and there, a boulder has rolled across the dust, leaving a bouncing trail behind it. In some places, some event has triggered a dust slide, and the dust fans out down slope into lovely feathery patterns.

How would I go about addressing the scientific questions I identified when I requested the image? I need to convince myself that I can tell the difference between rock that looks like it was deposited by volcanism and rock that has been shattered by billions of years of impacts. I should compare the texture I've found here to the texture of the rocks that make up Zephyria tholus itself. Another good thing to do would be to look at crater walls in the region that are not cutting these putative volcanic features, to see if the texture of the rock looks any different there. I should crawl around Google Earth looking for eroded volcanic constructs in desert regions to see what the texture looks like in those. It might even be interesting to compare to Lunar Reconnaissance Orbiter photo from the Moon of a similar spot, a place where a large-ish crater has cut through a topographic construct of some sort.

This would be a lot of work. Most of the time, you don't get scientific results from a single glance at an image. It really does take a lot of work, which is why I used to spend from 6 am to 7 pm in my office as a grad student. You have to study the image, then ask yourself questions. You create multiple working hypotheses -- this could be a volcano, or it could just be a chance hill built from the ridges of overlapping ancient craters. Figuring out which hypothesis is closer to true will take more observations and comparisons to other places on Mars and even other planets. It also takes a lot of reading about what other researchers have concluded based upon their work on the same ideas, and hopefully also talking to researchers actively doing work in the field.

Hopefully your questions can be answered with existing data, but a lot of the time, they can't, and you have to take the time to make more requests for more data, wait to see if the team thinks your request has merit, and, if they so, sequence an observation, uplink the sequence to the spacecraft, downlink and process the data. You might need to get some real help from another scientist who knows more about one aspect of the problem; if they're sufficiently interested, they may make time to work together with you on your puzzle, and they'll become a coauthor on whatever conference abstract or paper you eventually publish about the work.

If you publish. A nonzero proportion of the time, research winds up in a dead end. You wind up not being able to answer a question, or you simply decide that the problem isn't worth your limited time, and you move on to other things. If I were a researcher, I might follow up on this problem. It's a fairly contained question, so it'd be good for someone at my level of capability, which is to say, someone at about the third- or fourth-year grad student level. But I'm not a researcher, I'm a part-time blogger and full-time mom, and there are other fish to fry! (And a baby who is refusing to take a nap. Oy.)

This won't stop me from making more HiWish suggestions. I've already heard via Twitter that there's another picture on the ground from one of my HiWish suggestions, and the comment on that was that it, too, is dusty. Drat. I need to do a better job of finding places for imaging that are likely to be less dusty. Fortunately, there are scientists out there (led by Steve Ruff) who have already laid the groundwork for solving that problem: there is a global map of Mars derived from Mars Global Surveyor Thermal Emission Spectrometer data (the "TES Dust Cover Index") that can tell you where it's dusty and where it's not. I'll be consulting the TES dust map when I make my HiWish suggestions from now on!