Wan deals with physics based climate models, related to atmospheric and global weather models.
Data she deals with typically is "heavily averaged" with many variables
Abhi produces one-off apps to provide insight into specific problems she is working on. Pan3D could potentially reduce the need for customized apps.
Side-by-side comparison visualizations are often needed.
Typical workflow involves discussing need/requirement with Berk/Abhi, prototyping in paraview, then producing customized app. Could Pan3D replace the paraview step?
There were comments (from Berk?) that workflows might leverage temporal parallelism more so than spatial parallelism.
Hui usually works with low-resolution data. There are much higher levels, of course (e.g. "high" and "super" resolution)
There echnical issues sometimes relate to dealing with different models using different grids.
There are research efforts to develop regional/weather models and wind prediction models for energy
Also a "new atmospheric" model is being developed. Might be in conflict with existing models at ERF (?)
Notes