Hello!

I am wondering why the SAR on Cassini is producing segments of the radar path, in most cases four parallel stripes per flyby. Does anyone know how the instrumentsworks exactly?

I'm sure people better informed than myself will be able to tell you - but as I understand it - the spacecraft essentially 'paints' that pattern onto the surface by slewing left and right in 4 stripes during closest approach.

Doug

To perform Synthetic Aperture Radar imaging, the spacecraft points its radar dish to one side of its track as it flies over Titan. To capture a SAR swath it broadcasts a burst from five different antenna beams (one at a time) and listens for the echo from the surface. The echo data is transmitted to Earth. On Earth, each burst echo is modeled in "delay-Doppler space". This is tough to describe -- but basically, the echo is returned from different parts of Titan's surface at different times depending on how far away it was from the spacecraft (that's the delay part) and the echo's frequency is shifted higher or lower depending on whether it was in front of the spacecraft's path, and hence Dopplered to higher frequency, or to the rear of the spacecraft, and Dopplered to lower frequency (the Doppler part). For this processing to work the speed and altitude of Cassini need to be known to extremely high accuracy. The resulting images are in delay-Doppler space. I've learned from the team that they always reproject their images to a resolution of 256 pixels per degree, giving a pixel size of around 176 meters per pixel. But the effective resolution is actually lower than that, ranging from around 350 to 720 meters per pixel. The Magellan instrument worked almost exactly the same way except it had only one beam in contrast to Cassini's 5 beams.

Here's a little something I wrote up for the Society website:
http://www.planetary.org/explore/topics/ca...ment_radar.html
And attached is a technical paper on the instrument.

--Emily

This is also why the long map strips produced on each SAR pass have little narrower tags at each end -- the SAR can pick up audible return echoes from the central down-looking track of its radar beam for greater distances from Titan than it can for the other SAR tracks aimed off to the sides of its flyby path. (By the way, it's only the fact that Cassini's SAR proved a bit more sensitive than expected -- and thus able to operate at greater distances from Titan -- that allowed it to get a SAR map of the Huygens landing point on this last pass.)

The pattern of the five beams, is as follows. extracted from the frames kernel (the red dot marking Huygens):



This may be why the predicted swaths show jagged ends?
As with:

I think your albedo map is showing the opposite side of Titan than Huygen's landing spot.

Anyway, cool illustration of the radar beams.

Oops, used wrong map (wrong center longitude)--fixed it. Thanks.

One small correction to this -- the center beam is only down-looking in altimetric mode. In SAR mode, the spacecraft always looks completely to one side of the spacecraft (there's an illustration in the technical report attached to my previous comment). That's because if the spacecraft were pointed downward and consequently looking to both sides, it would get echoes with the same delay time from both sides of its ground track. By looking completely off to one side, delay time translates one-for-one into cross-track distance from the spacecraft.

This is also why the RADAR swaths often have a curly tail on one end, from the center beam: as a SAR pass ends the spacecraft rotates to point nadir to do an altimetric observation, and they keep using the center beam in SAR mode right up to the last moment. The curly tail shows you the spacecraft's rotation.

The angle at which the spacecraft is pointed off its track is called the "look angle." Magellan, which was in a polar orbit, acquired most of its global imagery in a left-looking geometry, but late in the mission they changed modes to right-looking and got a significant amount of coverage that way, which allows some pretty accurate photoclinometry on topographic features. Presumably they will eventually be able to do the same with Cassini once they finally get some overlapping coverage on Titan.

One interesting result of looking off to one side is that all the RADAR passes planned in the mission for either right- or left-looking geometry could be flipped so that Cassini looks the other way. They've got it planned out for the maximum coverage, but if there is something particularly interesting that they might otherwise miss that they can catch by flipping the planned RADAR swath to the other side of the ground track, they'll do that.

--Emily

Thanks a lot! That was exactly I was looking for!

This probably a really obvious question, but I'm a humble humanities student !

Is there any prospect of multiple (>2) overlapping passes for improving the characterization of particularly interesting sites; or for set of various typical terrains and features encountered across the other SAR passes? [And therefore providing an improved understanding of the single-pass data.]

Something for the extended mission perhaps? Would there be much to be gained perhaps half a dozen of the Huygens site, perhaps coupled with a lower altitude ceiling? Edit: Looks as if the altitude won't be lowered anytime soon, though this does leave the possibility open in the future I think, as variations in the atmosphere become better known (http://www.planetary.org/blog/article/00000707/).

The RADAR seems to be the instrument that has the most flexibility and scope for improvement as the mission goes on.

Roly

I don't think there's much point to having MANY passes of the same site (n>2), but there is a lot of value, potentially in having exactly n=2. That's not coincidentally why just about every advanced animal has two eyes but basically none have 3. You get stereopsis with 2, and 3 is redundant. With SAR, it's not quite stereopsis we're after, but the principle is analogous: We want to distinguish bright=rough vs. bright=sloped towards the spacecraft. Two perspectives provide the discrimination in most cases, although pathological cases could thwart that.

Unfortunately, Cassini has no chance of an extended mission that so egregiously exceeds the main mission as we've seen with the MERs. I think the best case scenario for additional RADAR passes in the extended mission will be no more than about the same number in the main mission, but that is far on the optimistic side. We might get as few as 6, even assuming that Titan is the sole target (and it will be a major one, by necessity).

Overlap of some stripes is basically unavoidable as mapping continues, so we will get at least some terrain that enjoys that level of scrutiny. However, we're not going to get to map the whole globe once, so overlap comes at the expense of "once-over" coverage. Also, targeted coverage comes at the expense of pragmatics (near-nadir pointing provides best resolution, and orbits must be designed to bring Cassini back to Titan). I think we'll end up seeing Titan covered about 25-30% (?) with a small fraction of that SAR'd twice.

It has been reported (though I can't remember where -- possibly in a conference abstract) that some small amount of overlap was already achieved. Specifically, over the "sand dunes". A surprise at that was that although the radar illumination was from a different direction, the dunes didn't change their appearance. Bringing into question their vertical relief and topographic shading assumptions.

Thanks for the responses, they were most helpful, especially regarding the intrinsic properties of the surface vs. its orientation toward the beam.

I vaguely thought that performing more than two passes might allow for improvements in effective resolution through multi-sampling (maybe improving SNR), or some other kind of clever superresolution trick, but what works for optical images is probably wildly different from SAR. With such a brief extended mission (I was more optimistic), the trade between mapping and repeat coverage seems to weigh against deliberately targeted second passes (c.f. overlap.)