Climate models – large and complex computer codes – require hardware with high capacity both in terms of computational speed and storage. High performance computing facilities of this kind are part of the Swedish research infrastructure on the national level, and even larger facilities are available within Europe. The Bolin Centre modelling coordination team’s primary task is to ascertain that adequate resources are available for the researchers within the Bolin Centre.

Climate models are important tools for improving our understanding and predictability of climate behavior on seasonal, annual, decadal, centennial and millennial time scales. The Bolin Centre uses the climate models EC-Earth and NorESM to understand how the Earth’s climate varies in time and space, and how the physical and biogeochemical cycles, including human activities, interact with the climate system.

So what kind of climate modelling is performed and what has been observed? Bolin researcher Qiong Zhang and colleagues have used EC-Earth to simulate the climate response to a greening of Sahara that happened during mid-Holocene (6000 years ago) and observed globe-wide climate changes such as a warmer Arctic, weakening of the climate phenomenon El Niño – Southern Oscillation (ENSO), and more tropical Atlantic cyclones. Similar modelling strategies are applied to evaluate the climate consequences to today’s clean energy proposal on deployment of massive Sahara solar farms.

Industrial activities generate not only greenhouse gases, but also nano- to micrometer-sized particles in the air. These particles affect climate at the same time as they are harmful for air quality and health. A Bolin Centre team led by Annica Ekman and Hans-Christen Hansson, have used NorESM to examine how future emission paths affect particle concentrations in the air as well as climate – in particular in the Arctic. The simulations show that future decreases in particle emissions may enhance mid-century Arctic warming by about 0.4 oC, unless strong compensating reductions are made in greenhouse gas emissions.

Looking ahead, the Achilles heel of climate models are clouds since these are mostly too small to be simulated with the approximately 100 by 100 kilometer grid cells used in regular climate models. Cloud processes determine both the strength of global warming through cloud feedbacks, and the character of climate change such as precipitation extremes and droughts. With the increasing computational power available an international team led by Thorsten Mauritsen are working to bring the next-generation ICON climate model to a new supercomputer named LUMI that is being built in Finland. They hope to be able to run with a global grid of 1 by 1 kilometers, permitting most clouds to be simulated.

Bolin Centre researchers also use more detailed models than climate models to better understand different processes and interactions within the climate system, for example models covering only a column of the atmosphere and ocean, ice sheets models, hydrological models, and large-eddy simulation.

A climate model is a three-dimensional representation of the atmosphere that is coupled with the Earth’s surface and the seas. In short, the atmosphere and the seas are divided into a series of boxes, or grid cells. In each cell of the grid, values of the temperature, humidity, air pressure, wind speed, sea ice and vegetation are calculated. After the climate’s condition in each cell has been calculated, a step forward in time is taken, and all values are calculated again. Illustration: SMHI