The issue of scale in climate models and its resolution
Guest post by Callum Munday
The simulation of climate by numeric models is emergent from the fundamental physics which govern the climate system. However, limits on computer power leave most general circulation models (GCMs) unable to resolve this physics at scales finer than 100km. Climatic processes occurring at these scales therefore must be approximated. This is achieved through the aggregation of sub-grid processes across the larger (~100km) grid box in model parameterisation schemes.
These parameterisation schemes take resolved physics at grid scale and use a physically-informed statistical model to estimate the mean magnitude of a particular process. For example, convective parameterisation schemes use a simple metric between convective updraft, convective downdraft and compensating subsidence to estimate mean cloud formation and precipitation. Despite the crudeness of such schemes, climate models are remarkably effective in reproducing large scale atmospheric patterns.
However, as approximations of unresolved processes, parameterisations represent a significant source of error in climate model simulations. This is especially relevant in relation to cloud processes and cloud-aerosol interactions, which are thought to contribute the largest uncertainty in estimates of The Earth’s changing radiative balance (Boucher et al., 2013). The issue of scale is therefore one that needs to be tackled ahead of accurate predictions of future climate change.
Attempts to resolve the scale issue have traditionally tended to focus on two goals. Firstly, to improve computer resources and secondly to develop better parameterisation schemes. The first goal has been met with some success, with a steady improvement in the average grid resolution of the ensembles used in successive Intergovernmental Panel on Climate Change (IPCC) reports from ~180km in the Third Assessment Report (TAR) in 2001 to ~100km in the latest 2013 report (AR5). However, even the highest resolution GCM with grid resolution of 40km (the Japanese MIROC4h model) is unable to explicitly resolve convection. This places a greater emphasis on the second goal, to improve parameterisations.
On this front, there has also been progress, led by improvements in the understanding of fundamental physics of cloud formation. In part, this has been achieved through targeted observational campaigns, such as NASA’s CERES remote sensing project. However, incorporating this understanding into parameterisation schemes is fraught with difficulty, because of the coarse information which is resolved and available at the grid box scale. This brings us back to the problem of how to resolve finer scale processes.
A New Approach
A new approach, set out in a 2014 paper authored by Tim Palmer attempts to move beyond the divide between resolved processes and sub-grid parameterisations (Palmer, 2014). Rather than the resolution of the model being truncated at a scale limited by computer power, the simulation is degraded progressively from the meso-scales (100km and less). In this way, the simulation of processes at a meso-scale is less computationally expensive, allowing for the explicit resolution of processes at a cloud scale. Resolving cloud processes explicitly could avoid the uncertainties associated with parameterisation and subsequently lead to a more accurate climate predictions. Consequently, the practical application of this theory is a promising area for climate research.
Callum Munday is a research student in the 2014 cohort.
Boucher, O., D. Randall, P. Artaxo, C. Bretherton, G. Feingold, P. Forster, V.-M. Kerminen, Y. Kondo, H. Liao, U. Lohmann, P. Rasch, S.K. Satheesh, S. Sherwood, B. Stevens and X.Y. Zhang, 2013: Clouds and Aerosols. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Stocker, T.F., D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.
Palmer, TN, 2014, More reliable forecasts with less precise computations: a fast-track route to cloud-resolved weather and climate simulators? Philos Trans R Soc A 372(2018):20130391.