PGA attenuation with depth?

Dear researchers; I have been working on Seismic Hazard Analysis for some time. The subject of my question is about the dynamic analysis of underground structures.

At one point, the seismic bedrock continues to the ground surface, which means we don't have soil layer or alluvium. PGA obtained from Site (PSHA method) with a return period of 2475 years is equal to 0.5. In international standards (ASHTO, FHWA, etc.) and many reports of engineering companies, for deep underground structures (deeper than 30 meters), the dynamic analysis is omitted or usually when PGA is transferred to the depth, a reduction factor is multiplied in PGA. However, this reduction factor is probably applied when there is some layers of soil or alluvium on the bedrock. Is this statement true?

The next question is that if the underground structure is completely in the seismic bedrock and at a depth of 30 to 100 meters (without soil layers or alluvium on the bedrock up to the ground surface), is it still necessary to use the same reduction coefficients to reduce PGA in depth?

Are there other methods to reduce the PGA in depth? Does "PGA" decline with increasing depth in bedrock?!! Otherwise, what will be the way of using PGA in the depth of the earth when the structure located in the seismic bedrock?

Currently, I am dealing with the dynamic analysis of tunnels near the surface of the earth using prevalent methods. However, for depths higher than 30 meters and assuming the reduction of the earthquake effect on deep structures, I have doubts in using the above method.

If there is a solution (and related software) to use PGA in the depth of ground, particular when the structure is located in seismic bedrock, Please give a brief guide. As mentioned above, I have obtained the PGA at the surface of seismic bedrock through the PSHA method. The ground surface and the seismic bedrock surface in the studied site are the same (without soil layers or alluvium on the bedrock).

Many thanks