Global general circulation models are the dominant tool in the effort to forecast the effects of climate change. Given the expansive scope of these models, some simplifications need to be made when representing smaller-scale processes, such as the effects of regional topography. To compensate, regional climate models are sometimes used to incorporate local influences and, in theory, improve the accuracy of projections of regional climate change. In testing the abilities of a regional model and a global model to represent historically observed climate change for the continental United States, however, Racherla et al. found that the regional model provided only a small increase in model accuracy and in some cases actually made the forecasts worse.