High-Performance Computing and GFDL Research

John Sheldon
Geophysical Fluid Dynamics Laboratory (GFDL), OAR

NOAA's ability to conduct leading edge research in climate, oceanography, and weather has benefited tremendously from recent improvements in GFDL's High-Performance Computing
System. The HPCS, consisting of over 2400 CPU's, provides balanced, comprehensive computing capabilities, meaning that in addition to high-performance computing, it includes capabilities for data management and archiving, analysis and visualization of model results, and networking and telecommunications in order to advance the lab's research programs and support research collaborations within NOAA and with other government agencies. This system enables GFDL to achieve it's goal of expanding the scientific understanding of the physical processes that govern the behavior of the atmosphere and the oceans as complex fluid systems. The systems are modeled mathematically and studied using complex computer simulations.

GFDL has also undertaken a major effort to develop a shared software infrastructure for the construction of climate models and model components for vector and parallel computers. The result of this software modernization effort is the GFDL Flexible Modeling System (FMS). Because vgorous external collaborations are so fundamental to the scientific success of GFDL, the external dissemination of models and numerical modeling tools that are developed at GFDL is an essential element of GFDL's current and future mission. For these reasons, GFDL has made the software and documentation in the FMS repository publicly available.

Among recent research results made possible with these new resources is a high-resolution, Global Mesoscale Cirulation Model (GMCM) simulation. This is the first GMCM run at GFDL and, since it is non-hydrostatic with explicit convection, it may be the first of its kind anywhere.

Other research objectives for which the HPCS and FMS are being used are the continuing development of a more advanced GFDL Hurricane Prediction System to improve track forecasting accuracy, improved prediction of wind and precipitation fields, storm surge, and changes in storm intensity. In addition, they will be used to improve regional projections of climate change, investigate the effects of deep ocean circulation on model behavior, and analyze of the processes controlling El-Niño-Southern-Oscillation events.