What does 'Apply Earth Gravitational Model' tick box do when using an External DEM for Range Doppler Terrain Correction?

I am using an external DEM (EGM08 geotiff) for terrain correction. Can anyone explain what the optional tick box ‘Apply Earth Gravitational Model’ does and why this would be necessary if we are already specifying a DEM?

I processed the same image with and without it, which gives a shift of ~30m to the west when you have this box ticked.

Does it effectively apply the correction twice resulting in the shift? Without it the corrected image lines up with the results of using ASTER or SRTM.

Anyone know whats going on here?

Because you tie the DEM to EGM that is in use, you will to better its precisions.

I’m sure I don’t need this option considering I’m using EGM08 (and accuracy gets worse if I do use it) - but does anyone have an example of a DEM where this is needed?

I wonder why you want to use EGM08 for Terrain Correction as it only defines the height of the Geoid. I don’t think it is suitable as an input DEM because terrain is not included but only gravitational fields.
In general, applying a graviational model as an addition to absolute heights ca make sense, especially for altimetry where you retrieve absolute heights from space but SAR data of medium resolution (30 meter or better) surely needs more precise information on topography.

Image source


Thanks for your reply - I’m using EGM08 at the moment because I’m only interested in the oceans - terrain is not needed so it’s good enough for this. I was using SRTM until now (by virtue of it being the default) but I do need to process images worldwide so looking for an option that works outside of +/- 60deg.

It’s part of a script using the snappy functions so ideally would like a one-size-fits all. Suspect ASTER would be fine also. :relaxed:

Wondering what the gravitational model built in is? EGM08 or 96 or something else?

alright, I didn’t think of ocean surfaces. Here EGM can be better than nothing. But for data is located over oceans you could as well apply the Ellipsoid Correction instead which doesn’t use topographic information at all and it should be sufficiently correct in terms of geometry.

To answer my own question above - DTED would be one - all ocean elevations are set to 0.

@ABraun I think using terrain correction with a global DEM that incorporates a gravitational model will be generally better at sea than correcting only to the ellipsoid - I’ll run an image through Ellipsoid Correction though and see what difference it makes (although maybe not a lot!)

Looks like others found the same in this thread.

Good to hear. If you could provide a screenshot of the comparison it could help others in the future

Terrain correction working a lot better than ellipsoid for me, here are some offshore platforms of known location (green points). EGM08 lines up nicely, Ellipsoid Correction shifted east ~320m (and took 30 minutes compared to 18 for TC on my laptop).


great, thank you for the comparison!

So will you opt for the Apply EGM option twice.
Once during terrain flattening and the second time during RDTC (both with external DEM). And are you trying to say that applying EGM will make a huge difference?