The most advanced climate models describe the main physical processes in the climate system (atmosphere, ocean, and land surface) and are based on a set of equations for energy, momentum, and mass conservation. Even after being considerably simplified, the governing equations of climate models can only be solved numerically and thereby the development and broad applications of climate models began after the advent of sufficiently powerful computers. Climate models simulate a large set of physical characteristics of the climate system, such as atmospheric and ocean circulation, radiative fluxes, temperature, cloudiness, precipitation, snow and sea-ice cover. Although fundamental physical processes in the atmosphere, the ocean, and on land surface can be described by well-known laws of physics, the enormous complexity of the climate system and limitations of modern computers do not allow to design climate models solely based on the first principles. For example, micro-physics of individual water droplets in clouds is well understood but not only individual water droplets, also individual clouds, cannot be resolved in climate models since their spatial resolution is about 100 km at best. Thereby, the description of many important processes in the climate system is based not on the physical laws but on the so-called parametrizations, that is, rather simple, often semiempirical, submodels of individual processes. With the progress in geosciences and growing computer power, these parametrizations become more sophisticated and realistic but the use of different para-metrizations in climate models still remains the major source of uncertainties in climate predictions.
Was this article helpful?