Scalable Lagrangian-based Visual Analysis on Multivariate Ensemble Simulations
Visualizing and analyzing large-scale and complex flow field data with scalability is a major challenge in various application domains. In this talk, I will present our recent efforts to tackle this challenge by two projects on multivariate and ensemble flow data visualization and analysis with Lagrangian-based methods.
The first project we introduce is LASP (Lagrangian-based Attribute Space Projection), which tightly couples the multivariate analysis and flow advection for large-scale datasets. This technique makes it possible to discover and explain complicated transport phenomena in various studies. The second project we present is eFLAA (ensemble Flow Line Adevection and Analysis), which provides comparisons across ensemble constituents.
The differences between the constituents are measured with the Lagrangian-based distance metric, instead of fixed locations with the Eulerian specification. This prototype system manages the big input data, the massive pathline computing and storage, and generates relatively small subsets of pathlines for interactive visualization and analysis. It also shows good scalability in National Supercomputer Center in Jinan, Shandong, China.