Visualization and Analysis Platform for Ocean, Atmosphere, and Solar Research (VAPOR)
The VAPOR project is an Open Source software development effort aimed at improving the ability of researchers in the Earth sciences to analyze and interpret results from some of the largest numerical simulation outputs. The genesis of this work was an NSF-ITR research grant that supported CISL as well as collaborators at the University of California, Davis and the Ohio State University. VAPOR has transitioned from a research project to a maturing product, and with the expiration of our ITR grant late in FY2008, VAPOR is presently supported entirely through NSF Core funding.
Though VAPOR's origins are strongly rooted in geophysical turbulence, recent work on VAPOR has focused on expanding its capabilities to support the needs of the broader Earth sciences community. Development of VAPOR is closely guided by a steering committee comprised of geoscientists from around the world that sets development priorities, dictates software requirements, and serves as friendly users for testing and evaluating new software features.
This figure shows vortices in the eye
wall of a WRF hurricane simulation, produced by Yongsheng Chen, NCAR.
Illustrated are two new capabilities added to VAPOR this year: 2D dense
texture flow visualization and support for WRF data. Though still in its
infancy, VAPOR's WRF model support is fulfilling a critical unmet need in
numerical weather research: providing tools for interactive, 3D visual
data analysis.
Decades of continual advancements in microprocessor technologies have led to numerical simulations of Earth sciences phenomena computed at unprecedented scales. These simulations generate extraordinary amounts of data. But our ability to manage, analyze, and gain insights from these data has not kept pace with our ability to generate them. For many numerical modelers, the greatest challenge in the discovery process begins after the simulation has completed when the analysis process begins. The VAPOR project focuses on the problem of large-data exploration with an intelligent approach: by exploiting multiresolution data representation coupled with advanced interactive visualization and quantitative analysis capabilities, VAPOR provides a comprehensive desktop environment suitable for exploring terascale-size -- and soon petascale-size -- data sets.
Thus VAPOR supports two of NCAR's strategic priorities: "Developing and providing advanced services and tools" and "Creating an Earth system knowledge environment."
Work in FY2008 focused on broadening VAPOR's capabilities to better meet the needs of a more general Earth and space sciences computational fluid dynamics community. In particular, much of the past year's development has been aimed at users of the Weather Research Forecast (WRF) model. Interest from WRF researchers is strong: following an initial release of the software offering rudimentary WRF support, over 500 copies of the package were downloaded in a single week.
Noteworthy events in FY2008 include:
- Two new major versions of the software were completed and released. Major new capabilities include: 2D dense texture flow visualization, mapped isosurfaces, user preferences support, and support for importing and visualizing WRF data sets.
- Over 6,000 copies of VAPOR have been download by the scientific community (~1,000 of the most current version, released in June 2008).
- VAPOR was cited by 10 refereed, scholarly scientific publications in 2008, and imagery produced by VAPOR was featured on two journal covers.
- The VAPOR team continued its education efforts, hosting Ph.D. interns from UCLA and the Colorado School of Mines.
- A second invited paper on VAPOR was accepted for publication by the New Journal of Physics.
- VAPOR was the cornerstone the TOY geophysical turbulence summer school, attended by ~30 PhD students from around the globe.
Work in FY2009 will continue to focus on expanding VAPOR's feature set to attract a wider geosciences user base. A survey of current and potential future users was conducted in January 2008; the results will guide much of the year's development. Significant areas of planned development include:
- Integration with CISL's NCL package to further VAPOR's geo-referenced data capabilities
- Improving the VAPOR data model and tools to better prepare for petascale simulations
- Improving our support structure to better handle the growing VAPOR user base
This project was originally made possible through support from NSF's Information Technology Research for National Priorities (ITR) program, and is now entirely supported by NSF Core funds.