Just to clarify:
The SAR measurement is a complex number (real and imaginary parts). To generate a detected product, the complex number is converted to intensity (i.e. i.i + q.q). To fit the dynamic range of the intensity data into the 16 bits of a product, the square root of the intensity values is applied - this is amplitude. So amplitude is just a convenient way of storing detected intensity data. For any measurements from detected data, the amplitude must be converted to intensity first. Also care should be taken when averaging over a region (e.g. a field) - convert to intensity first and then average.