• Subscribe

Powering climate models with grid computing

Development of global climate modeling. Image courtesy Global Climate Model. Click to view larger image.

Meteorological models tell us what the temperature will be like tomorrow whereas climate models show a range of possibilities for the future. An update of knowledge on the scientific, technical and socio-economic aspects of climate and how it changes is well underway, as preparations for the Intergovernmental Panel on Climate Change (IPCC) 5th Assessment Report (AR5) enter their final stages. So, how have climate models evolved in the last 40 years and how is grid computing helping to project future climate? Eleni Katragkou, a climate scientist from the Aristotle University of Thessaloniki (AUTH) in Greece, answered many of these questions during a plenary session at the recent EGI Community Forum 2013.

In the 1960s, climate models focused only on the atmosphere (i.e. clouds, water vapor etc.), but by the early 2000s models started to mirror the reality of the entire Earth-atmosphere system more and more. Nowadays, five different climatic modeling components can be coupled together: biosphere (life), lithosphere (rock), cryosphere (cold), hydrosphere (water) and human activity. Spatial resolution of models has also vastly improved from 500km in the early 1990s to 100km scales in 2010s.

During her talk, Katragkou described the challenges the earth modeling community currently faces, and demonstrated how indispensable grid computing has become in powering progressively more sophisticated climate model simulations.

Relative importance of climate drivers to current global warming as determined by the 4th Assessment of the IPCC. The IPCC quantifies the relative impact, in terms of its contribution to climate change, of a range of natural and human factors (collectively known as 'forcing factors') . This is based on radiative forcing (the influence that each forcing factor has in altering the balance of energy in the Earth-atmosphere system). The IPCC also gauges the current level of scientific understanding (LOSU) of each driving force. Image courtesy IPCC. Click to view larger image.

Modeling the climate system

General circulation models (GCMs) use complex algorithms, the laws of physics, and 3-dimensional grid systems, as well as differential equations, to relate fundamental physical processes (temperature, pressure, specific humidity etc.) to each other. Each equation is solved at discrete grid points on the earth's surface, at a fixed time interval and across several vertical layers. The more grid cells the higher the resolution, and the more calculations that must be computed. "Every time one doubles the resolution it requires ten times more CPU," explains Katragkou.

Parameterizing the complex interactions

"What really complicates the modeling process is the fact that all climatic components (i.e. atmosphere, cryosphere etc.) are interconnected and interact with each other. Parameterizing those interactions and even more the related climatic feedbacks requires sophisticated modeling algorithms and computational demanding climatic simulations", explains Katragkou.

Additionally, for specific applications a 100km spatial resolution is insufficient to resolve important sub-grid scale features (e.g. vegetation distribution, coastlines and mountain topography), which all impact significantly on local climate change. Dynamical downscaling is a method for obtaining higher resolution climate information from relatively coarse-resolution GCMs. "By nesting a Regional climate model (RCM) into an existing GCM we can simulate climate features at resolutions of even 10-50km, improving the spatial detail considerably", says Katragkou.

Map of the predicted changes in ozone concentration for the 2090s decade with respect to the 1990s. The shade of red is proportionate to the increase. The map shows that by the 2090s, south-west Europe will experience an increase of ozone concentration at ground level of about six parts per billion. Image courtesy of Katragkou et al., 2011. Click to view larger image.

Simulating the present and future of surface ozone concentrations

In 2010, Katragkou used the HellasGrid computing resources (720 hours x 4 cores = 2,880 CPU hours) from the AUTh computing centre to simulate the effect of climate change on ozone concentrations in Europe over the next 100 years. Using a past decade (1990s) as a control, and the IPCC future emission scenario A1B (typified by rapid economic growth), Katragkou et al., 2011, suggested that the concentration of ground level ozone is likely to increase towards the end of the century due to climate change, especially in south-west Europe by 6.2 parts per billion [ppb]. The atmospheric models applied took into account chemistry and climate interactions, assessing the impact of key meteorological parameters (temperature, solar radiation, wind, cloudiness) on atmospheric composition. Their model suggests that the median summer near surface temperature for the whole Europe will be 2.7 degrees celcius higher at the end of the 21st century.

Simulating the present and future climate over south-east Europe

Being able to visualize what might happen on a map unquestionably aids clarity for non-experts. The aim of another ongoing project - the GEOCLIMA project, is to develop an integrated Geographic Information System allowing the user to visualize, manage and analyze the information which is directly or indirectly related to climate and its future projections over south-eastern Europe. "We need high resolution geographical information systems to allow analysis and dissemination of credible climate information to the public. The GEOCLIMA database will be soon available for impact assessment studies and climate-related adaptation strategies aiming to reduce vulnerability and risks of climate change," explains Katragkou.

Join the conversation

Do you have story ideas or something to contribute? Let us know!

Copyright © 2022 Science Node ™  |  Privacy Notice  |  Sitemap

Disclaimer: While Science Node ™ does its best to provide complete and up-to-date information, it does not warrant that the information is error-free and disclaims all liability with respect to results from the use of the information.


We encourage you to republish this article online and in print, it’s free under our creative commons attribution license, but please follow some simple guidelines:
  1. You have to credit our authors.
  2. You have to credit ScienceNode.org — where possible include our logo with a link back to the original article.
  3. You can simply run the first few lines of the article and then add: “Read the full article on ScienceNode.org” containing a link back to the original article.
  4. The easiest way to get the article on your site is to embed the code below.