Glenn,
I thought we had been looking at this before, but I cannot find the issue now. Seems to still be a problem. This case is very high resolution, with lots of meshes (2000). The HRRPUV is noticeably different on different meshes. I guess this has something to do with how we compress on the FDS side, but again I thought we had dealt with this.

The case is here on spark:

mnv/FireModels_fork_home/fds/Research/SootDeposition/Thermo_Only_0p2cm_DNS_2000