Diagnosing track feature bias of extratropical cyclones over North America in the GFDL SPEAR model

The simulated track features of extratropical cyclones (ETCs) in three different atmospheric resolution versions of GFDL’s Seamless System for Prediction and Earth System Research (SPEAR) (e.g., SPEAR-LO (1°), SPEAR-MED (0.5°), and SPEAR-HI (0.25°)), are evaluated using ERA5 data. A cyclone feature tracking is applied to the 6-hourly sea level pressure field of SPEAR’s historical simulations for 30 years (1991–2020). SPEAR models generally simulate more frequent and stronger ETCs over the U.S. West and Southeast than observations. To investigate impact of the sea surface temperature (SST) bias on the ETC bias, we contrast the historical simulations and the atmosphere-only simulation (AMIP) of SPEAR models. The remote SST biases over the North Pacific and the Atlantic Coast in SPEAR partially account for ETC biases in the U.S. west coast and Southeast. For instance, local cold SST biases over the North Pacific induce the strengthening of upper-level wind in the south flank of the cold bias center through thermal wind balance, leading to overestimations of baroclinicity, and causing more cyclogenesis over the North Pacific. Further cyclone source tracking analysis demonstrates that the overestimation of ETC in the West Coast comes from more frequent cyclogenesis over the North Pacific. To some extent, the warm biases in the Atlantic coastal zone tend to cause ETC biases over the U.S. Southeast Coast. The possible connection between these ETC biases and precipitation biases over North America will also be discussed in this talk.