Due to the lapse in federal government funding, NASA is not updating this website. We sincerely regret this inconvenience.
A Glimpse at the Future of Global Weather Prediction and Analysis at NASA
Dr. William Putman
Overview
The Goddard Earth Observing System (GEOS) model provides a peek into the future of global weather prediction and analysis. Using the non-hydrostatic “finite-volume cubed-sphere” (FV3) dynamics solver and advanced physics schemes, GEOS has pushed the limits of high-performance computing at NASA to execute simulations representative of the projected evolution of global weather prediction capabilities over the next decade.
Innovative numerical methods in combination with increased resolution and efficient computing power have revived a new exploration of the advanced simulation capabilities for global modeling systems like GEOS. These GEOS simulations provide the first glimpse at the potential for what are known as “global convection-allowing models” of the Earth’s atmosphere.
Project Details
In addition to the demand to predict weather and climate at fine, localized scales, high-resolution global modeling has two main applications at NASA. First, it is essential for NASA to be able to simulate existing and future types of observations with a resolution close to the footprint size of existing and planned satellite instruments. This applies to representations of weather—such as winds, clouds, and precipitation—as well as to the distributions of atmospheric pollutants. Second, as increased computing power routinely allows using models with finer spatial resolution and physical complexity, NASA is pioneering the use of ultra-high-resolution global models to improve our understanding of atmospheric processes within the Earth system. Testing the fidelity of these models today is one step towards evaluating the systems that will be in routine use for weather prediction and climate simulation over the next decade.
Results and Impact
GEOS model development consumes millions of computing hours on the NASA Center for Climate Simulation (NCCS) Discover supercomputer. This work includes the day-to-day routine development to implement new schemes and evaluate model performance against existing reanalyses and observations. GEOS production systems operate daily to produce high-quality analyses and forecasts from days to weeks in support of NASA missions and field campaigns. Experimental simulations with advanced numerics and physics serve the observing system community by providing a platform to evaluate new observations and assimilation techniques.
Why HPC Matters
High-end computing (HEC) at NASA, specifically the NCCS, provides an essential resource for preparing our modeling systems for emerging HEC systems and exploring new science developments. While current GEOS configurations require thousands of computational cores, configurations over the coming decade will require scalability approaching millions of cores. Exposing additional parallelism within GEOS becomes essential for reaching these scalability targets. Partnerships with the NCCS support our exploration of traditional hybrid parallelism approaches within GEOS today, and prepare our models for future HEC platforms by exposing new depths of parallelism within our application components.