ECMWF and ORNL complete the first seasonal scale global simulation with 1 km resolution

On June 22, 2020, a collaborative research team of scientists from the European Center for Medium-Range Weather Forecasts (ECMWF) and the Oak Ridge National Laboratory (ORNL) in the United States completed the world’s first 1 Seasonal-scale global atmospheric simulations run on km grid spacing. The research team completed the simulation using the ECMWF Integrated Forecast System (IFS), running on the world's fastest computer (as of November 2019) - the Oak Ridge National Laboratory's Summit supercomputer.

Currently, ECMWF’s operational numerical weather prediction (NWP) operates on a 9 km grid spacing. The simulation data implemented on a 1 km grid spacing provide a basis for evaluating the advantages and disadvantages of operational numerical weather forecasting. refer to. These data will support the planning of future satellite missions, as new satellite tools can be used in the evaluation of global atmospheric simulations. Simulations on a 1 km grid spacing can be viewed as a “digital twin” prototype of the Earth. The simulation results show that the continuously improved IFS static numerical weather prediction model performs well under an average grid spacing of 1 km. This appears to challenge the common view in dynamical meteorology that non-hydrostatic equations will be required at the 1 km resolution level, and the results of this run provide a benchmark against which future non-hydrostatic simulations can be measured.

Although ECMWF's IFS has accumulated more than 40 years of experience in numerical weather prediction, simulations still face unknown territory. Obstacles include whether the model can remain stable over a season, whether there are problems with explicitly resolving convection, or problems due to steep terrain. The height of the Himalayas simulated at a grid spacing of 9 km reaches about 6000 m, while using IFS at a grid spacing of 1 km this height can be accurately determined to 8172 m. Although there is still a slight error from the actual situation, it is very close.

Simulations of the Earth's atmosphere with 1 km grid spacing enable realistic simulation of global mean circulation and improved advection through analytical feedback from deep convection and topography, as well as associated Rossby waves and inertial gravity waves layer of expression. The simulation results also provide the first direct indication of extreme weather, such as the likelihood of tornadoes. The resulting reference data set will directly support model development efforts and help estimate the impact of future observing systems. The researchers found that global energy redistribution was similar in simulations with 1 km grid spacing (explicit deep convection) and simulations with 9 km grid spacing (parameterized deep convection). It is worth noting that these features are not available without deep convection parameterization at coarser resolutions.

Please indicate the source and author when reprinting this article: "Earth Science Dynamic Monitoring Express" 2020, Issue 14, Lanzhou Documentation and Information Center, Chinese Academy of Sciences, compiled by Liu Yanfei.