Climate model data

The user should be aware that grid box values are not directly comparable to station data. Climate models attempt to represent the full climate system from first principles on large scales. Physical "parameterizations" are used to approximate the effects of unresolved small scale processes because it is not economically feasible to include detailed representations of these processes in present day models. Caution is therefore needed when comparing climate model output with observations or analyses on spatial scales shorter than several grid lengths (hundreds of km), or when using model output to study the impacts of climate variability and change.

The user is further cautioned that estimates of climate variability and change obtained from climate model results are subject to sampling variability. This uncertainty arises from the natural variability that is part of the observed climate system and is generally well simulated by the climate models.