As explained here by @lveci: TOPSAR Coregistration
Range and Azimuth shift apply the enhanced spectral diversity (see references).
Basically it takes advantage of the overlap between bursts to fine tune the coregistration. Generally the backgeocoding is accurate and the ESD can provide a pixel or sub-pixel shift to improve the coregistration.
P. Prats-Iraola, R. Scheiber, L. Marotti, S. Wollstadt, and A. Reigber, “TOPS interferometry with TerraSAR-X,” IEEE Trans. Geosci. Remote Sens., vol. 50, no. 8, pp. 3179–3188, 2012
R. Scheiber and A. Moreira, “Coregistration of interferometric SAR images using spectral diversity,” IEEE Trans. Geosci. Remote Sensing, vol. 38, no. 5, pp. 2179–2191, July 2000
A clear sentence can be found in the abstract of Yuxiao Qin, D. Perissin and Jing Bai:
saying: “In Sentinel-1 TOPS mode, the antenna sweeps in the azimuth direction for the purpose of illuminating the targets with the entire azimuth antenna pattern (AAP). This azimuth sweeping introduces an extra high-frequency Doppler term into the impulse response function (IRF), which poses a more strict coregistration accuracy for the interferometric purpose. A 1/1000 pixel coregistration accuracy is required for the interferometric phase error to be negligible, and the enhanced spectral diversity (ESD) method is applied for achieving such accuracy. However, since ESD derives miscoregistration from cross-interferometric phase, and phase is always wrapped to [ − π , π ) , an initial coregistration method with enough accuracy is required to resolve the phase ambiguity in ESD.”
You can go more in detail with it so will get a deeper understanding.
I hope this clarifies your doubts…