The NOAA CPO Modeling, Analysis, Predictions, and Projections (MAPP) program hosted a webinar on the topic of Unified Weather-Climate Modeling: Challenges and Opportunities on Thursday, September 29, 2016. The announcement is provided below.
September 29, 2016
9:30 AM - 11:00 AM ET
The Met Office Unified Model: seamless development for weather and climate prediction
For over 25 years, the Met Office have developed and used a Unified Model for global and regional prediction of the Earth’s weather and climate systems. The original motivation for this approach was to consolidate the technical infrastructure of our modelling codes. Since then, we have increasingly moved towards more scientifically unified systems so that we can take advantage of the synergies between weather and climate modelling. This talk will present our experience of the unified modelling approach and how this helps us achieve our mission. It will include examples of the scientific, organisational and technical benefits this brings, as well as some of the challenges it presents and our approaches to address them.
Software infrastructure for a unified modeling system
ESMF-based software infrastructure is used in modeling systems that span predictive time scales from weather to climate. ESMF coupling infrastructure supports specialized approached for conservative interpolation, such as the exchange grid, explicit approaches that use a central mediator, and simple connectors that transfer fields directly between components. The construction of a unified modeling system that spans weather to climate scales requires careful consideration of the applicability of these and other coupling options, and the scientific and computational trade-offs that may arise from using the same model components and/or coupling techniques for all timescales. It also calls into question what exactly is meant by "unified". The NOAA Environment Modeling System (NEMS) aspires to be a unified modeling system, and participants in development are in the process of identifying open questions, and, with community partners, formulating strategies to address them. This talk discusses the current status and plans.
A new radiation code for atmospheric models across scales
Radiation is the fuel for the atmospheric heat engine and the fundamental driver for circulations at all scales. Thus all but the most idealized models require a radiation parameterization and errors can have systematic impacts forecast quality. As a modeling capability, radiation parameterizations are unique in being conceptually isolated, have clear benchmark results, and applicable essentially unchanged across two or more orders of grid resolution. A well-designed radiation parameterization is crucial since radiative transfer is often a significant portion of an atmospheric model’s computational cost.
In this presentation I’ll briefly introduce RRTMGP, the Rapid Radiative Transfer Model for General Parallel applications, a new radiation parameterization with roots in the well-regarded and widely-used RRTMG parameterization. RRTMGP increases accuracy relative to RRTMG by introducing new spectroscopic information but the majority of changes are aimed at making the model more efficient, on the one hand, and more flexible and easily adapted to use across scales and other contexts on the other hand. The flexibility is achieved through the use of Fortran 2003 that users may extend to efficiently and correctly couple the parameterization to their host model without modifying the underlying code. Efficiency is achieved by isolating computational tasks into simple “kernels” that may be highly optimized or replaced with architecture-specific implementations.
Dr. Annarita Mariotti
MAPP Program Director
Dr. Daniel Barrie
MAPP Program Manager
MAPP Program Specialist
MAPP Program Assistant
Oceanic and Atmospheric Research (OAR)
National Oceanic and Atmospheric Administration (NOAA)
Department of Commerce
Climate Program Office
1315 East-West Hwy, Suite 1100 Silver Spring, MD 20910
Copyright 2018 by NOAA
NOAA Privacy Statement|
Web Accessibility Statement|
Disclaimer for External Links|
U.S. Department of Commerce|