SAR interferometry measures the relative height of a surface with the difference in phase between two SAR images taken from different locations (either two satellites or two antennas on one satellite or airplane). The heights are relative within the SAR interferometry image until you can calibrate the absolute phase constant from some additional information such as a ground control point of known location and elevation.

Dear Eric,

Thanks for your interest

how can I calibrate the SAR DEM with ground control points ?

Is there  a specific software does this calibration?

of course I agree with Eric and I want to add a notation that SAR data stored at different instants and at different exposure stations and so, it reform the earth surface relative depending on distances from the exposure station (after removing the effect of polarization...etc)...and so it can model the earth relatively to absoultly because of the lack of dependency on refrence

I don't know the details of the NEST Toolbox. I believe that the DORIS software can use ground control points to calibrate InSAR DEMs from some satellites.

An additional source of error in InSAR DEMs is the atmosphere, if you are using SAR images acquired at different times. Changes in the water content of the atmosphere between the two dates will change the phase and cause errors in the elevations, even with a ground control point, so it is better to use several ground control points.

"Elevations from SAR are relative" means that they are not absolute... ;-)

It means that these elevations are consistent with each other, but they are not consistent if you want to compare them with elevation from an other source.

In that case, you need to project your SAR elevations in a common reference frame, for example using a ground control point of known elevation as said Eric. 

Simply, it means that SAT data does is not related to a particular coordinate system. The relative measurements are correct up to scale factor and they do not have a global origin and orientation. To convert the SAR data to a global coordinate system you can use 3D similarity transformation to compute the 3D translation vector as well as 3D rotation angles.

LiDAR and SAR have in common that measurements between points is accurate to the limits of the sensor. It is known as relative accuracy, a confusing term, that could be better expressed as inter-point accuracy. For modern LiDAR sensors this might be 5 cm. SAR sensor accuracy depends on many elements in the system.

Absolute measurements mean that whatever the value might have been in raw data, the point location has been corrected by projection to a coordinate system. Measuring the elevation of the single point on a feature by painstaking survey will tell you what the difference is between the two measurements of the same point. This is known as absolute accuracy because a single point is measured by two different methods.