The damping that you're referring to is not structural damping but that which was used to derive the response spectra.  Structural damping is an altogether different item.

It seems to me that the 5% damping allowed in seismic analysis (for typical structures) is the result of soil friction at the abutments and pier foundations.  I also remember (from some seminar) that most structures will exhibit damping characteristics greater than 5% (the 5% being a somewhat lower limit).  The damping allowed is due more to physical observations of several in-service structures than from theoretical analysis.

If the 5% damping allowed is due to soil-structure interaction, the superstructure type would not influence it significantly.

I'm not sure why "structural damping" would be a different item unless it refers solely to mechanical means.  Even if you add mechanical dampeners, it would be additive to the soil-structure friction that provides "natural" damping effects and you might consider some "effective" damping percentage based upon both.

I'm getting over my head pretty quick.  It's an interesting discussion and hope they'll be more responses.

Structural damping is useful in all dynamic analysis (seismic or traffic induced vibrations).
The damping ratio (fraction of the critical damping, value of no vibration) depends on the energy dissipation due to elastic viscosity of materials, friction, etc.

The traditionally considered value of 5% applies to concrete buildings and short bridges. Large bridges presents lower values.

Typically you can consider 3% for reinforced concrete bridges, 2% for prestressed ones and 0,5% for steel or composite bridges.

Of course it depends on a lot of parameters, including material, length and deck typology, bearing systems, non structural elements, abutments and pier dimensions, foundation, etc. Only physical observation of real structures can give a exact value for each one, as Rowe pointed.

Hope this helps.

If memory serves correct the UBC also uses a 5% damping.  The 5% is based on empirical observations including blast and shock induced ground motion.  

Caltrans states the 5% damping is "for ordinary standard concrete bridges...10% can be justified for bridges heavily influenced by energy dissipation at the abutments and are expected to respond like single DOF systems."

After the response of the structure, based on the ARS curve and elasticity is complete, some factors to account for the ductility of the particular structure are used.


Rambling:
The ARS curves work because higher frequencies will have smaller amplitudes.  The earth does not ring, higher frequencies will not propogate.  Current analysis has not moved much beyond a static analysis.  The only significant change is that current analysis discourages irregular geometries that might have passed with a strictly static analysis.  The precision these people attain for an indefinate answer is mind boggling.  After speaking with Caltrans engineers (on the posted subject), I'd rather take my tax dollars and light a fire.