Press Release
Japan Agency for Marine-Earth Science and Technology (JAMSTEC)

April 28, 2005
Japan Agency for Marine-Earth Science and Technology

Development of a Global Cloud-Resolving Atmosphere Model with 3.5km Grid Size
Using the Earth Simulator

Outline:

Dr. Masaki Sato, Sub-Leader, Dr. Hirofumi Tomita, Researcher and their group members of Next-Generation Model Development Group in the Global Environment Modeling Research Program of Frontier Research Center for Global Change (FRCGC), Japan Agency for Marine-Earth Science and Technology (JAMSTEC) come to the worldfs first success of a global cloud-resolving simulation of the atmospheric circulation with 3.5km grid size (figure 1). They have been developing a new global atmospheric model based on an icosahedral grid system called NICAM (Nonhydrostatic Icosahedral Atmospheric Model) since year of 2000, to be run on the Earth Simulator that had been under development at that time. NICAM has resolved the ambiguity caused by indirect handling of convective clouds, and lead to a possibility of increasing accuracy of the numerical weather prediction and the climate change prediction. Especially, the improvement of the prediction of summer in Japan that is controlled by Pacific high and generation and strength of typhoons will be expected.

This research result was presented at the Aqua-planet Experiment Program (notes 1) Workshop starting from 20 April 2005, in Reading, UK and the European Geosciences Union General Assembly 2005 in Vienna, Austria. It will also be presented at the Spring Meeting of Meteorological Society of Japan, Japan.

Background:

In the temperate region, there are high and low pressure systems and fronts, and they bring about weather change, but there is no such clear structure in the tropics. According to the picture from the weather satellite (figure 2), there is no system like in the temperate region but cloud clusters (notes 2) instead. Cloud clusters are the big bank of thunderclouds with 100~500km extents, but its core, an upward motion of air has only 10~20km size. Existing climate models such as numerical weather prediction models and climate change prediction models are with 50~300km grid size. Therefore, it is impossible to calculate and reproduce cloud clusters in the models directly. Instead, clouds were treated by parameterization (notes 3) in the models. However, small but unique characteristic of cloud clusters can be simulated only with less than 5km grid size by the calculation applying the equation which controls generation and change of clouds. This fact has been recognized as the greatest difficulty of the climate models.


Result:

The greatest achievement of the experiment of this time is to calculate convective clouds, especially cloud clusters scattered in the tropical area directly by the model with less than 5km grid -3.5km grid interval, and to simulate their behavior such as formation of groups of clouds and their disintegration realistically just as if we are looking at the cloud image sent from satellite. This result has not been attained without the Earth Simulator, the worldfs biggest super computer. NICAM opens a new era of the 50-year-long history of atmospheric general circulation models and numerical weather prediction and increases a possibility of more accurate numerical weather prediction and climate change prediction by calculating the behavior of convective clouds in the tropics directly following the equation.


Future plans:

Progressing further from a hypothetical Aqua-planet Experiment of this time, more realistic experiments introducing ocean-land distributions and mountains into the model will be conducted. This will enable us to make highly reliable simulation of the mechanisms of convective precipitation related to unique topography in the various regions. At the same time, simulation of the process of the generation of typhoons from convective clouds will also be expected on more reliable basis.


Contact:

Japan Agency for Marine-Earth Science and Technology

Mr. Katsuhiko Masuda, Director
Frontier Research Promotion Department,
Frontier Research Center for Global Change
Tel:045-778-5746 Fax:045-778-5497
URL: http://www.jamstec.go.jp/frcgc/jp/

Mr. Kenichi Takahashi, Manager
Public Relations Division, Administration Department
Tel:046-867-9050 Fax: 046-867-9055
URL: http://www.jamstec.go.jp/

Notes 1: Aqua Planet Experiment:
As a test of the basic performance of an atmosphere model, numerical experiments with a hypothetical condition that the whole earth is covered by ocean whose meridional temperature distribution is similar to the observation are often carried out. Those experiments are called Aqua-planet experiment.

Notes 2: Cloud Cluster
In the tropics strong convective clouds (cumulonimbi) develop one after another in a relatively small area (~100km) to form a cluster. Occasionally many cloud clusters aggregate to form gsuper cloud clusterh.

Notes 3: Parameterize
In numerical models of the atmosphere and ocean, their motion, pressure, temperature and so on are calculated at each grid point based on the laws of physics. But those phenomena whose scale is smaller than the grid box cannot be calculated by directly applying physical laws. Instead whose phenomena and their effects are incorporated in the model though some empirical relations expressed by parameters directly calculated by the grid. This procedure is called parameterization.