A heat wave of unprecedented magnitude occurred over parts of Western North America (WNA) during June 2021, with severe societal and ecological impacts. Although this event was predicted by weather forecast models, questions have been raised as to whether such extraordinarily strong events can be predicted by global climate models (GCMs), which operate at lower grid resolutions, and suffer from biases in their simulated climate.

Factors affecting the extreme heat waves in the WNA region include the large-scale high-pressure ridge, small-scale regional topography, and climate change. A high-resolution weather model can capture all these effects to make a good forecast, but that does not help quantify the potential role of climate change. Attribution studies are needed for that purpose, and these studies often rely upon lower resolution GCMs to simulate the statistics of large-scale atmospheric flow patterns.

We analyzed sets of historical and future GCM simulations from 1950 to 2050, and identified extreme heat waves of comparable amplitude in the WNA region, in both high- (horizontal resolution of 0.25^o for atmosphere) and low-resolution (~1^o) runs. This underscores the possibility of predicting the statistics of such extreme events using GCMs. We further carry out different simulations initialized from observations using the Energy Exascale Earth System Model (E3SM) atmospheric component (EAM) at forecast lead times of 2 weeks, 1 week, 5 days and 3 days at horizontal resolutions of 1^o and 0.25^o, and compare the properties of simulated extreme heats with the observed one. This study aims to assess the ability of GCMs to simulate extreme heat waves, and assess the potential impact of model resolution and bias.