Data-Centric Science at the NASA Center for Climate Simulation

Overview

From international climate assessments to groundbreaking global simulations, field campaigns, and spacecraft, the NASA Center for Climate Simulation (NCCS) data-centric computing environment is advancing NASA Earth Science missions.

Project Details

The NCCS centerpiece is the Discover supercomputer, which now contains more than 43,000 Intel Xeon processor cores as well as GPUs and Xeon Phi coprocessors. The Discover supercomputer provides 1.121 petaflops peak performance along with tens of petabytes of online disk and mass storage. NCCS is partnering with its parent organization, the Computational and Information Sciences and Technology Office, on Climate Model Data Services (CDS) to provide access to classic technologies such as OPeNDAP, emerging technologies such as the Earth System Grid Federation (ESGF), and advanced technologies such as the CDS Application Programming Interface (API) that allows developers to plug their software into several CDS services.

Results and Impact

The Intergovernmental Panel on Climate Change (IPCC) Working Group 1 released its Fifth Assessment Report in September 2013. NCCS served the assessment by hosting modeling contributions from NASA's Goddard Institute for Space Studies (GISS) and Global Modeling and Assimilation Office (GMAO). Coupled Model Intercomparison Project Phase 5 (CMIP5) simulation data submitted for the IPCC assessment plus observation and reanalysis data are available on the NCCS ESGF Gateway. Worldwide downloads from the gateway totaled more than 260 terabytes over the past year. CDS is also providing access to data from ongoing GMAO simulations for the U.S. National Climate Assessment.

In addition to studying phenomena as varied as the dust plume from the Chelyabinsk meteor and ozone's effects on climate over decades, current GMAO efforts at NCCS include Nature Run 2. This 2-year global climate simulation is using finer (7.5-kilometer) resolution than its predecessor as well as non-hydrostatic techniques and updated physics to explicitly resolve atmospheric convection and potentially better represent hurricanes. GMAO's new 50-year Ocean Data Analysis System (ODAS) will help initialize decadal simulations updating IPCC/CMIP5 contributions. Data services based on the integrated Rule-Oriented Data System (iRODS) automate ODAS monitoring and enable data annotation and publication.

Among multiple field campaigns supported by NCCS was the Iowa Flood Study (IFloodS) ground validation for the Global Precipitation Measurement (GPM) satellite being launched in February 2014. Daily forecasts from coupled weather and land-surface models helped field scientists prepare for precipitation events in order to protect staff and equipment from severe weather. NCCS also served campaigns focusing on hurricanes, sand storms, dust, and other aerosols. Beyond Earth studies, NCCS supplies extensive data processing for NASA's twin Gravity Recovery and Interior Laboratory (GRAIL) spacecraft that orbited and then were intentionally crashed into the moon.

Role of High-End Computing Resources

NASA missions push computing and data resources to their present limits. Nature Run 2 alone requires 7,680 Intel Xeon Sandy Bridge processor cores for months and will produce an estimated 4 petabytes of data.