PDA

View Full Version : Fluid flow on the walls of Endurance crater?



MicroKid
2004-Sep-13, 08:04 AM
Taken on Sol 168

http://img.photobucket.com/albums/v376/MicroKid/FluidFlow16.gif

http://img.photobucket.com/albums/v376/MicroKid/FluidFlow15.gif

Note the area on the left (in FluidFlow15) which seems to show two fluid flows running over the top of the rock slab and down the crater wall a short way.

MicroKid

EFossa
2004-Sep-13, 09:38 AM
Hi, very interesting stuff. Another possibilty is that it's perhaps a very small rock fall?

Could you post a link to the original images for that sol?

jt-3d
2004-Sep-13, 09:45 AM
I'm going to go with 'What are elongating shadows due to a setting sun?' - for $200, Alex.

MicroKid
2004-Sep-13, 10:13 AM
The photos are pam cam Sol 168 shots. Oppy did a long series. Maybe Nasa knew what they were looking for.

I'm going to go with 'What are elongating shadows due to a setting sun?' - for $200, Alex.

Cute but not correct. The photos are about 40 seconds apart. Notice any other shadows getting longer? No? Lots of rocks around to compare. With respect next time look first.

As for a rock fall, what rocks moving sideways along the crack between two slabs?

Next............

MicroKid

EFossa
2004-Sep-13, 11:25 AM
Yes I know they are from sol 168, but there are 68 pictures on that page, can you link to the specific images you refer to in your original post.

jt-3d
2004-Sep-13, 11:43 AM
Yes, please do. Link us to the originals next time instead of a couple of blurry, home made gifs. And as a matter of fact I do see more shaodws getting longer. They're all over the place. It's just that the other shadows are already there so they're not as easy to see unless you're a paid NASA disinfomation agent like me. Eh, it's a living.

MicroKid
2004-Sep-13, 12:10 PM
Another shot of a weepy rock slab on the walls of Enduraance crater.

http://img.photobucket.com/albums/v376/MicroKid/FluidFlow18.jpg

MicroKid

MicroKid
2004-Sep-13, 12:20 PM
One pair are:

http://marsrovers.jpl.nasa.gov/gallery/all/1/p/168/1P143098335ESF3221P2596R1M1.HTML

http://marsrovers.jpl.nasa.gov/gallery/all/1/p/168/1P143098375ESF3221P2596R1M1.HTML (this has the big flow)

and the other pair are:

http://marsrovers.jpl.nasa.gov/gallery/all/1/p/168/1P143098492ESF3221P2596R1M1.HTML

http://marsrovers.jpl.nasa.gov/gallery/all/1/p/168/1P143098529ESF3221P2596R1M1.HTML

I you can see rock shadow changes in this time span (37 - 40 seconds) you need help.

MicroKid

ToSeek
2004-Sep-13, 03:13 PM
What are the odds that this is some sort of JPEG artifact or somesuch? I note that the "big flow" is only in the one image, and an image taken 40 seconds later ( http://marsrovers.jpl.nasa.gov/gallery/all/1/p/168/1P143098414ESF3221P2596R1M1.JPG ) looks little different from the image taken 40 seconds before the "big flow" image.

MicroKid
2004-Sep-13, 10:45 PM
I've been involved with digital and analogue image systems for a long time and I sure it is a real event. However the disapperance of the main spot 40 seconds later is also real.

Anyone care to guess how long say 200ml of muddy water would last on the surface before it formed a ice crust and changed it's albedo from a darkish object (which the camera picked up) to whiter and disappeared?

Note the flow seems to have moved along the crack between the slabs which is again evidence against a pixel error.

Also in the second blinker two of the weep spots are the same as the big event. Then there is that flow event on the second blinker with the left most site suggestive of a short flow along and down the surface.

What did you think of the apparent flow from under the tilted slab? Notice there are several high water marks on the down crater deposit bed and that close examination suggests a muddy flow rolled over the adjacent berry bed. You can see the rolled leading edge.

MicroKid

jt-3d
2004-Sep-14, 05:02 AM
Well, I made this (http://home.satx.rr.com/jt3d/hmmm.gif). I admit that animated, it doesn't really look like a shadow. However, in the still, it does. I think this is due to the 256 color restriction of .gif format.

Or maybe it's just an artifact like the hundred of other dancing pixels in those shots. In particular the white on on the bottom left of the dark pile of dirt. Now I think that's a martian albino crab but that's just me.

The results don't look nearly as dramatic if you leave the contrast and size alone. But if you blow up that 'big flow' you can see that it's really just one really dark gray pixel and four or five lighter ones. I'm not opening a spa up there just yet.

MicroKid
2004-Sep-14, 06:05 AM
Hi It-3d,

Except there is evidence of other dark patches which follow the slab crack left and down crater from the bigger dark event. Then there is the other blinker which also shows darkening in the same spots of the slab to slab crack.

Now throw in the dark streaks in the second blinker which are suggestive of another flow event on the left side.

By the way if you still think it's a shadow which lasted for about 40 seconds, what caused it? Then figure out the sun angle and where an object would need to be to cast a shadow to look like the flow event. Can't be done.

I suggest it was a dark muddy flow whos surface quickly froze and altered it's albedo / reflective qualities as it went whiter.

Any comments on the flow from under the tilted slab?

MicroKid

Bamf
2004-Sep-14, 06:15 AM
Unlike most of the "anomalies" people come up with in MER images, this one is somewhat interesting. Here's some additional information:

This is a 15 image series (http://www.mars.asu.edu/~gorelick/s168.gif), taken over 10 minutes of time, about 40 seconds apart, with 13 second exposures. The "anomaly" only occurs in frame 9. It is, by far, the largest change among any two adjacent frames. That said, it's still only a couple of pixels in size (http://www.mars.asu.edu/~gorelick/s168_diff.png), putting it at the low end of believeable.

The downlink telemetry indicates that it was compressed with LOCO (the lossless compressor), there were no lost pixels, no corrupted blocks, and the compression segment boundaries are no where near those pixels.

The coregistered animation makes it pretty clearly that it's not a change in solar illumination, or a dust slide. However fluid ejection wouldn't have been one of my first choices. My current guess is transient dust on the lens. If I come up with any additional information, I'll post it.

MicroKid
2004-Sep-14, 06:31 AM
Hi Bamf,

Dust on the lens would be so far out of focus the only effect is to slightly reduce contrast due to scattered light. If you have a SLR or digital camera try it and see for yourself.

That darkenings occured on at least two occasions at the same spot on the slab to slab crack strongly suggest the darkening is real. Please do another comparision between frames 7 & 9 as frame 8 shows darking along the slab to slab crack. I think this will revel a stronger flow along the slab to slab crack.

Blow your blinker up to 300% and watch the other darkenings along that slab to slab crack. Your photos have a different contrast ratio that those on the Nasa site. Did you process them or are they from the .img files? If so I take it you have inside access?

By the way despite what some here have said, this event it NOT a single pixel. Here is the un resampled pixel map:

http://img.photobucket.com/albums/v376/MicroKid/FlowEventcrop.jpg

Even here the flow to the the left and down is clear.

What do you think about this shot which strongly suggests a muddy flow from under the rock slab, pooling into a muddy shelf with at least two high water marks and the leading / right edge of the muddy shelf rolling / flowing over the adjacent berry bed.

http://img.photobucket.com/albums/v376/MicroKid/FluidFlow18.jpg

Thanks for the heads up on the photo status. Where did you get that information if I may ask?

MicroKid

Astrophil
2004-Sep-14, 12:34 PM
Truly the human imagination is a treacherous thing. Looking at jt-3d's .gif, you find yourself reading it like a sequence of film and tracing falling bits of imaginary dirt from frame to frame to frame, even though in reality it's only two cycling images.

Just wanted to say, though, that regardless of what this blob is, the images that people have made to examine it, especially Bamf's big sequence, are among the most vivid I've seen for giving you the illusion you're somehow on Mars. So thanks guys!!!

Bamf
2004-Sep-14, 05:44 PM
Hi Bamf,

That darkenings occured on at least two occasions at the same spot on the slab to slab crack strongly suggest the darkening is real. Please do another comparision between frames 7 & 9 as frame 8 shows darking along the slab to slab crack. I think this will revel a stronger flow along the slab to slab crack.

There are no variations between frames 7 and 8 (or any other frames) that are above the noise level. Here's a plot (http://www.mars.asu.edu/~gorelick/s168_plot.png) of the pixels in question. The X-axis is frame number, the y-axis is unscaled radiance. This is for the 1 dark pixel, plus one above it, and the strip of 4 beneath it.



Even here the flow to the the left and down is clear.


I don't think you can say that anything about this 4 pixel event is "clear".



What do you think about this shot which strongly suggests a muddy flow from under the rock slab, pooling into a muddy shelf with at least two high water marks and the leading / right edge of the muddy shelf rolling / flowing over the adjacent berry bed.

I think you have a vivid imagination. I'm not (yet) saying you're wrong, but that much interpretation seems excessive at this point.


Did you process them or are they from the .img files? If so I take it you have inside access?

Thanks for the heads up on the photo status. Where did you get that information if I may ask?

And here, I just assumed everyone knew who I was by now.

I work for the Mars Space Flight Facility at Arizona State University. We have 4 instruments on 4 different spacecraft currently at or on Mars that I'm responsible for, including the mini-TES spectrometers on the rovers. So yeah, I have a full copy of everything for MER, and yes, I processed everything myself.

Bamf
2004-Sep-14, 09:18 PM
I had previously ruled out a cosmic-ray event because the anomaly is darker than the surroundings, instead of brigther as would be expected, but being suspcious, I asked the Pancam PI about it. He has this to say:


I love a good mystery... but this ain't one. What you're seeing is indeed the result of a cosmic ray hit. For each image we also acquire a zero-exposure "shutter subtraction" image which is subtracted onboard to remove the bias and frame transfer smear. Apparently there was a cosmic ray hit in the shutter image that came after image 9 in the sequence, yielding a bright cosmic ray streak in the zero exposure frame (I assume--that frame is not saved), and when that bright streak was subtracted from the darker scene at those pixel locations, the result is a dark smudge seen in the downlinked product. We've seen this occasionally in other sequences too.

01101001
2004-Sep-14, 09:45 PM
I love a good mystery... but this ain't one. What you're seeing is indeed the result of a cosmic ray hit. For each image we also acquire a zero-exposure "shutter subtraction" image which is subtracted onboard to remove the bias and frame transfer smear. Apparently there was a cosmic ray hit in the shutter image that came after image 9 in the sequence, yielding a bright cosmic ray streak in the zero exposure frame (I assume--that frame is not saved), and when that bright streak was subtracted from the darker scene at those pixel locations, the result is a dark smudge seen in the downlinked product. We've seen this occasionally in other sequences too.
No wonder it looked so straight. I thought ray right away, but then remembered that the examples I had seen were white. The shutter-subtraction frame explains it well. It's good to know.

Thanks for digging, Bamf, and welcome to the BABB.

freddo
2004-Sep-15, 12:18 AM
Straight from the horses mouth, Thanks Noel!

MicroKid
2004-Sep-15, 12:56 AM
Hi Bamf,

Thanks for the feedback. Is there a magnetic field around the CCDs? Maybe your Pam PI can plot the straight cosmic ray path in the below pixel difference map. To me it looks like it would need to curve.

http://img.photobucket.com/albums/v376/MicroKid/Flow_s168_diffcrop.png

Can I ask you to do another pixel brightness plot for a 3 x 3 array centred on the bright pixel, across the same shot base you used before?

Then there are the darkenings in the second blinker, some at the same spot as the first. Someone got a cosmic ray gun and firing at Oppy?

http://img.photobucket.com/albums/v376/MicroKid/FluidFlow15.gif

MicroKid

Bamf
2004-Sep-15, 02:56 AM
Hi Bamf,

Thanks for the feedback. Is there a magnetic field around the CCDs? Maybe your Pam PI can plot the straight cosmic ray path in the below pixel difference map. To me it looks like it would need to curve.

Cosmic rays aren't likely to intersect directly normal to the CCD (at least no more likely than from any other direction). They come from a random direction and cruise right on through the camera housing and the 2-D CCD array at an oblique, 3-D angle. That L-shaped pattern of brighter pixels is completely consistant with a ray that entered from about the 8:30 direction, through the front face of the CCD, only just grazing the left-most of the 4 pixels, influencing the photon-wells more as it penetrates deeper into the CCD while moving to the right, and finally hitting the darkest pixel hard before exiting the back of the CCD.


Can I ask you to do another pixel brightness plot for a 3 x 3 array centred on the bright pixel, across the same shot base you used before?

There's no reason you couldn't make this plot, but since the data is still handy, here (http://www.mars.asu.edu/~gorelick/s168_plot2.png).



Then there are the darkenings in the second blinker, some at the same spot as the first. Someone got a cosmic ray gun and firing at Oppy?

As I mentioned before, these "darker" pixels are well within the noise level.

The Bad Astronomer
2004-Sep-15, 04:36 AM
Well, CR streaks are pretty common. Sometimes they are truly cosmic, but you get the same thing from radioactive decay in the metals of the spacecraft. "Noisy dewars", we called 'em, because of the metal in the dewar that holds the coolant.

And before anyone asks, the "zero exposure" image can still get a cosmic ray in it, because it takes time to read out the chip. It can take seconds to minutes to dump the charge from a chip, and in that time you can get a hit.

MicroKid
2004-Sep-15, 08:19 AM
Hi Bamf,

As an electronics engineer I do understand how CCDs work, that cosmic rays can come from any angle and that they can form photoelectrons. Most examples I have seen are quite strong as generally CRs are a lot more energetic that photons.

You said:
"That L-shaped pattern of brighter pixels is completely consistant with a ray that entered from about the 8:30 direction, through the front face of the CCD, only just grazing the left-most of the 4 pixels, influencing the photon-wells more as it penetrates deeper into the CCD while moving to the right, and finally hitting the darkest pixel hard before exiting the back of the CCD."

And it did this without generating any photoelectrons in the pixel just to the left of the major hit site?

I just put some childs blocks on the desk and maybe you can show me the pathway cause I just can't see it. Now if the CR hit the left most pixel first (at the front top edge) then would I not expect to see a reducing effect on the pixels to the right as it angled further and further upward from the horizontal row moving toward the big hit pixel but being careful not to hit the pixel just to the left.

I also find it interesting that a contrast and gamma enhanced blinker (a 60 x 60 array @ 5ms per frame based on your blinker, thank you) centred on the big event site clearly shows the crater wall and slab to slab crack features. Also interesting is the fact that the darkening event follows the slab features and the "L" shape seems to be due to a slight downward protrusion on the upper slab edge just to the left of the "Big Event" site. Also there is the "Stuff" below the four pixels which on the rapid blinker suggests it ran down slope.

Thanks for the 3 x 3 data. Some interesting observations:

1) What happened on shot 1 to cause the brightening on shot 2? Did you use the right photo?
2) Almost all pixels from shot 10 to 11 show brightening.
3) Pixel 296,100 shows a delayed dark to light transition on shots 3 to 4 and on shots 9 to 10 to 11.
4) Pixel 297,100 shows a much sharper dark to light curve than pixel 296,100
5) Pixel 298,100 shows a sharp brightening peak the next shot after each pixel 297,100 dark peak.
6) Pixel 297,99 also shows a sharp brightening peak the next shot after each 297,100 dark peak.

This doesn't look like noise to me but a relationship.

Can you link an excel file for the 3 x 3 data so I can show the effects more clearly? I appreciate your assistance and your time. Wish I had your tools!

Thanks,
MicroKid

Bamf
2004-Sep-15, 09:56 PM
And it did this without generating any photoelectrons in the pixel just to the left of the major hit site? I just put some childs blocks on the desk and maybe you can show me the pathway cause I just can't see it.
There's your first problem. The important parts of a CCD are not shaped or arranged like blocks. The actual photon wells are tiny dots, usually fitted with a lens that is much, much larger than the actual accumulator. So it's more like an array of pennies, each sitting on its own pad the size of a floppy disk, with a silver dollar lens on top of it. And more importantly, once these penny sized detectors are exposed, their charge is shifted into another tiny register that can be located anywhere. In many CCDs the transfer registers are in alternating strips. So you have a strip of detectors, next to a strip of transfer registers, next to another strip of transfer registers, followd by another strip of detectors. And within a strip, it's not uncommon for the transfer registers to be grouped, rather than linerarly spaced, to make manufacturing easier. Since it's most likely that the ray struct while the data was in the transfer register, and since we don't have detailed specs on the pancam chip and don't actually know how things are arranged, your children's blocks aren't going to help you.


I also find it interesting that a contrast and gamma enhanced blinker (a 60 x 60 array @ 5ms per frame based on your blinker, thank you) centred on the big event site clearly shows the crater wall and slab to slab crack features. Also interesting is the fact that the darkening event follows the slab features and the "L" shape seems to be due to a slight downward protrusion on the upper slab edge just to the left of the "Big Event" site. Also there is the "Stuff" below the four pixels which on the rapid blinker suggests it ran down slope.
As I previously indicated, I think your overzealous interpretation of these 5 pixels is unwarrented and silly. Especially since you're using someone elses (quickly hacked together) data to do it.


Can you link an excel file for the 3 x 3 data so I can show the effects more clearly? I appreciate your assistance and your time. Wish I had your tools!

Actually, I'm pretty much through with this one. You had an interesting mystery so I looked into it, but I'm comfortable with the answer. If you're not, the data and the tools to work with it are all freely available for download.

MicroKid
2004-Sep-16, 12:14 AM
Hi Bamf,

"Actually, I'm pretty much through with this one. You had an interesting mystery so I looked into it, but I'm comfortable with the answer. If you're not, the data and the tools to work with it are all freely available for download."

No matter how you lay out the physical structure of the CCD cells I still fail to see how a CR could produce the result seen but hey that is just my view.

So where can I get access to the 12 bit data and tools that you are using? I only have access to 8 bit .jpg data. Like I said before I do appreciate your time and effort but do appreciate level playing fields.

MicroKid

Bamf
2004-Sep-16, 01:41 AM
So where can I get access to the 12 bit data and tools that you are using? I only have access to 8 bit .jpg data. Like I said before I do appreciate your time and effort but do appreciate level playing fields.

I write my own tools. You can get them from here (http://davinci.asu.edu).
You've got a point about the data though. The PDS files are only available for the first 30 days so far, so I guess you're stuck with the raw images until the next PDS dump.