Before application to future climate predictions, it is important to test model ability to simulate observed climate state and past climate variability. Current generation of climate models demonstrates a considerable skill in simulation of different atmospheric and oceanic characteristics as well as their interannual and interdeca-dal variability, such as tropical climate variability associated with ENSO. This is, however, only the first step in validation of the climate models. Another important step in climate models validation is testing of their ability to simulate different climates and climate change known from observations. For example, a pronounced global warming trend during the twentieth century is successfully simulated by climate models when changes in all major climate forcings, both natural and anthropogenic, are prescribed. Paleoclimate reconstructions of past climates present another important opportunity to test models under climate conditions different from the present one. For example, climate of the last glacial maximum about 21 000 years ago is relatively well studied and all necessary boundary conditions and atmospheric composition are known for this time with sufficient accuracy. Comparison of model simulations of the last glacial maximum with numerous available paleo-climate reconstructions shows that climate models are able to reproduce major aspects of the glacial climate reasonably well, although some discrepancies between data and model simulations remain and still have to be explained.
Was this article helpful?