Welcome to the FY2013 CISL Annual Report. This collection of highlights from our service, science, and education portfolios includes an overview of CISL’s broader impacts on the research community.
To start the fiscal year, CISL officially opened the NCAR-Wyoming Supercomputing Center (NWSC) in Cheyenne, Wyoming on 15 October 2012. To commemorate the grand opening with a ribbon-cutting ceremony, UCAR President Dr. Tom Bogdan was joined by National Science Foundation Director Dr. Subra Suresh, Wyoming Governor Matt Mead, former Wyoming Governor Dave Freudenthal, and University of Wyoming Vice President Dr. Bill Gern. This event marked the official unveiling of the Yellowstone supercomputer, which began operations with Accelerated Scientific Discovery (ASD) projects. An important feature of CISL’s services for many years, ASD provides very large allocations of computational resources to a small number of experienced scientists to boost their potential for discoveries through high-capability simulations. The NWSC provides petascale computing, data analysis, and visualization resources combined with exascale data management capabilities to support greater model resolution, increased model complexity, better statistics, longer simulation times, and more predictive power. This facility is carefully designed to continue serving the scientific community at the highest level for decades.
CISL is gratified that the arrival of NWSC and Yellowstone were instrumental in the support for climate scientists’ ambitious agenda of Community Earth System Model (CESM) and Community Climate System Model (CCSM) simulations for the Intergovernmental Panel on Climate Change (IPCC) Fifth Assessment Report (AR5). These researchers’ coordinated climate model experiments were completed in time to address pressing climate change issues for world leaders, and CISL now provides high-performance access, end-user support, accurate data use metrics, and helps publish the resulting data products. At a rate averaging 75 terabytes per month in FY2013, more than 2,000 scientists downloaded these data products from the Earth System Grid (ESG) science gateway at NCAR.
The Regional Integrated Science Collective (RISC) in CISL’s Institute for Mathematics Applied to Geosciences (IMAGe) made some groundbreaking progress and produced numerous other important advances in regional climate modeling. RISC pioneered a new data bias correction method using distribution-mapping techniques to develop a daily observational data product that includes uncertainty estimates. This new product improves on existing data products by providing not only gridded daily observations but also estimates the uncertainty associated with interpolated fields. This product, called MICA, is being developed in a transparent, modular, and reproducible form to facilitate adoption throughout the community.
The Data Assimilation Research Testbed (DART) project in IMAGe made a swift transition to petascale computing this year. DART was carefully prepared and tested in advance, and it ran on Yellowstone without modification for all applications. Considerable effort was made to find the most efficient ways to use DART on Yellowstone for large models, and a novel multiple-component assimilation interface to the coupled CESM models was completed and tested. This CESM coupler advances ensembles of all five component models using separate instances of DART for the atmosphere, land, and ocean models to assimilate observations into the ensemble. This fully coupled multiple-component assimilation allows scientists to better understand the relationships between the different coupled model components. Finally, a new interpolation mechanism was developed and tested to reduce memory usage for assimilations in ever-larger models.
Among CISL’s FY2013 advances in education, I want to call your attention to a new type of “data analytics” training program that multiplies our ability to broaden participation in the Earth System sciences. Developed in collaboration with a professor of statistics at the University of Wyoming, “Data analytics for the Geosciences using R” is a new type of workshop where experts train statistics instructors who can spark the interest of future scientists and engineers. Targeting math, science, and statistics professors, the workshop particularly encouraged attendance by faculty at community colleges who teach introductory statistics and teachers of Advanced Placement statistics at regional high schools. Breaking the tradition of teaching introductory courses using simple methods and small data sets, this new approach encourages instructors to immerse students in expansive real-world datasets with uncertain values and missing data. Using data sets that are rich in clues about how nature works and relevant to students’ interests can stimulate the curiosity and imagination of future scientists and engineers.
This annual report offers many more highlights of the breadth and excellence of CISL’s programs. These highlights include external recognition of the NWSC supercomputing facility, as it won two national awards in 2013: the Uptime Institute’s Green Enterprise IT first-place award for Facility Design Implementation, and the Datacenter Dynamics North American ‘Green’ Data Center award for demonstrated sustainability in the design and operation of facilities.
CISL provides far more than balanced, easy-to-use computational and data environments designed for the evolving requirements of the Earth System sciences. CISL also develops and delivers high-quality science and education programs to help secure the future of our scientific enterprise and the communities we serve. As you read this report, I hope you share our excitement about our recent progress. It is my pleasure to present our FY2013 CISL Annual Report.