When it comes to standards, I would first turn to ASME Section II Part D. Poisson's ratio and densities are provided in Table PRD. The introduction to Subpart 2 (of II-D, which holds these properties) says all the properties are typical - neither average nor minimum. It goes on to say that these properties are for information only and, if you have data supporting the use of different values, you may use them.
Table PRD shows a density of 0.28 lbf/in^3 for all the materials you mention. In your units this is 7750 kgf/m^3. That sits between the CAESAR II and AutoPipe numbers.
While I cannot say why we offer a different density for P91, I will note that we are talking about a 2.5% difference between the two CAESAR II densities that you mention. I understand that in certain systems this difference may be significant but I would then also ask how certain are we of many other assumptions used in these analyses.
If you do not accept the CAESAR II material values you can update the material database you use in CAESAR II.
_________________________
Dave Diehl