Hacker News new | past | comments | ask | show | jobs | submit login

It is certainly pretty. But I'm still not clear what it means.

OK, so the radar is measuring distance between the satellite and reflecting surface. They're compairing data from July 8, 2019 and April 8, 2018. I'm guessing that the two images look pretty much the same. Especially given limited spatial resolution.

But ELI5, what does the interference pattern show? I mean, are there 12 cm amplitude waves of vertical surface displacement? Something like frozen S waves?




An earthquake is rocks sliding past each other. What you're looking at here is a measure of how much they moved. (It doesn't "slide back" afterwards -- the motion is permanent.) For an earthquake of this size, the motion will be on the order of a few meters.

------------

In a bit more detail, the radar can't measure distance precisely enough to detect the movement. The distance measured before and after by radar is the same, within error. However, there's another part of the radar signal beyond just how long it takes to travel. That second part is the phase of the returned signal. Imagine the first time we imaged a small area, we got a return waveform that looked like this (zero phase):

    /\  /\  /\   
      \/  \/  \/
but then the next time we got back a slightly different result (270 degree phase):

    \  /\  /\  / 
     \/  \/  \/
The difference is shape of the returning signal is a phase shift. The radar wave is shifted slightly

We know that it moved at least three quarters wavelength in the ascii art example above. However, we'd get the same result if it moved ten and three quarters, though. We can measure part of the change very precisely, but the bulk of the motion looks the same to us. We're looking at that fined-grained part of the motion (phase difference) not the overall motion itself.

In programming terms, we're looking at the result of a modulo operator.


> we'd get the same result if it moved ten and three quarters

And also the same if it moved one quarter wavelength in the opposite direction.


Thanks.

I see that the ALOS-2 SAR uses L band, which seems to mean 1-2 GHz (30-15 cm). So maybe the ALOS-2 actually uses 12 cm?

So does that mean that the pattern shows something like contour lines?


Yep! You can think of the bands as contour lines of deformation.


Hey, thanks.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: