## Types of Climate Models

There are several types of models used at present for climate research (see also Table 1). The most complex and realistic climate models are the so-called coupled general circulation models (GCMs) of the atmosphere and the ocean. These models are based on the most comprehensive set of dynamical and thermodynamical equations and they describe a large set of relevant processes in the ocean and the atmosphere. The name 'general circulation' reflects the fact that these models, unlike more simple models, simulate three-dimensional circulation of the atmosphere (wind speed) and the ocean (current velocity). Climate GCMs originated from the weather prediction models and some models can be used both for weather and climate predictions. The main difference is how the models are used. The weather prediction is aimed on simulations of temporal development of individual weather systems, such as cyclones and anticyclones. Due to chaotic nature of weathers, an accurate prediction of meteorological conditions is only possible on the timescale order of 1 week. On a longer timescale, even small differences in the initial conditions

 Types of climate or Spatial aggregation and Resolved Number of Earth system models resolution timescales variables Examples Simple models Box type, 1-Da or2-D 1-109 yr 1-10 Budyko-Sellers energy balance model, Stommel ocean box model EMICs 2.5-D or 3-D spatial 1 day-106 yr 10-50 CLIMBER-2, LOVECLIM, UVic resolution: 500- 1000 km GCMs 3-D spatial resolution: 1 min-103 yr >100 CCM3, ECHAM-5 HadCM3 100-300 km

an-D - n-dimensional model.

an-D - n-dimensional model.

result in large differences in simulated fields. The aim of climate modeling is to simulate climate, that is, the averaged weather conditions. To obtain a sufficiently accurate climate state (i.e., statistics of weather), an averaging of simulated meteorological conditions over at least several decades is required. Current generation of coupled GCMs employs spatial discretization with the resolution of about several hundred kilometers and both the atmosphere and the ocean are divided in vertical direction by several dozens of unevenly spaced levels. Since atmosphere and the ocean are characterized by a number of fast processes, a relatively short time step of numerical integrations - from minutes to hours - is required to guarantee numerical stability and accuracy. This makes coupled GCMs extremely computationally expensive tools which require the use of the most powerful computers.

Another extreme in the spectrum of climate models is represented by simple climate models. Such models describe only a very limited subset of the processes in the climate system and, usually, they employ a very coarse spatial resolution. A prominent example of simple climate models is an energy balance model developed in the mid-1960s. This model is based on one equation for the energy balance of the climate system and it simulates only atmospheric temperature. Atmospheric circulation in this model is parametrized as a large-scale horizontal diffusion. In spite of their simplicity, this class of climate models still remains a useful tool for the analysis and better understanding of some important aspects of climate dynamics, especially related to its nonlinear aspects.

At last, a new type of climate models, the so-called models of intermediate complexity, emerged in the recent decade. These models are aiming on closing a wide gap between simple climate models and GCMs. Design of the models of intermediate complexity represents a compromise between a high degree of complexity required to realistically simulate climate and the necessity to reduce computational cost to perform long-term simulations. Unlike simple climate models, models of intermediate

complexity are able to simulate a much large set of climate characteristics, often comparable with GCMs, but due to considerable simplifications of the governing equations and, usually, a much coarse spatial resolution, models of intermediate complexity are suitable for much longer simulations than GCMs. This makes models of intermediate complexity very useful for the study of past climate changes and for long-term (thousand years and longer) future climate predictions.