By ‘Evaluate’ we mean comparing the model simulations in describing observed climate properties and statistics (observations usually refer to a specific climate data set). The type of evaluation varies with context and needs to address the question of how suitable the models are in providing me answer(s) to my question. Typically, an evaluation involves some type of metrics (a number or a score), but can also include the use of statistical analysis or description.

Typical types of evaluations include bias (systematic differences such as being too warm/cold or having too pronounced variations), the mean annual cycle (a prerequisite for reliability is that the models are able to reproduce well-known features), the Taylor diagrams, etc.