Dear Nitish,

I would suggest you to consider soil site PGA because, liquefaction phenomena completely related to soils. In some literature (Kotoda 1988; Zhu 2017) I found compare to PGA, PGV gives better results for liquefaction analysis, refer attached papers.

Regards,

Ramesh.

As you have performed PSHA, that means you have bed rock PGA, you can deaggreagte that PGA and select earthquake scenario (M and R combination), select real time site class B recorded ground motion (PEER website), perform site specfic ground response analysis (DEEPSOIL) using these selected ground motions to calculate response at the surface, use that PGA to calculate CSR.

This is a flow chart from report " RECOMMENDED GUIDELINES FOR LIQUEFACTION EVALUATIONS USING GROUND MOTIONS FROM PROBABILISTIC SEISMIC HAZARD ANALYSES ". you can refer for more clearification

You should perform site response analysis using the ground motion at the bedrock to determine PGA at surface. If you use the conventional method, you can simply put PGA at surface as the parameter to calculate CSR.

You should ofcourse use soil site PGA for liquefaction assessment.