Analysis of Daily Timescale Surface Temperature Variability in the Community Atmosphere Model

Extreme temperature events take thousands of lives every year and are a threat to agriculture and the biosphere. Therefore, it is essential that climate models be able to accurately reproduce the statistics of these extreme events and provide accurate predictions for how the probability of extreme events is expected to change in the future. This study analyzed the accuracy of the Community Atmosphere Model 5 (CAM5) in modeling the distribution of extreme temperature events through statistical analysis and comparisons with the MERRA-2 reanalysis. In regions where CAM was deficient in its representation of daily temperature variability, the dynamical processes that underlie the variability in these regions were contrasted and compared between CAM5 and MERRA-2 in order to understand the reason for these deficiencies. In particular, it was found that the variance of daily timescale surface temperatures in the winter season (December/January/February) over northwest North America in CAM were considerably higher than in MERRA-2. This large variance seems to be due, in large part, to strong circulation anomalies in the model. This leads to a stronger temperature gradient from north to south in the region of interest, which results in increased advection of cold air from the north into the northwestern part of North America.