Bias Analysis in the Simulation of the Western North Pacific Tropical Cyclone Characteristics by Two High-Resolution Global Atmospheric Models

This study compares the atmosphere-only HighResMIP simulations from FGOALS-f3-H (FGOALS) and MRI-AGCM3-2-S (MRI) with respect to tropical cyclone (TC) characteristics over the Western North Pacific (WNP) for the July–October months of 1985–2014. The focus is on investigating the role of the tropical easterly jet over the Western Pacific (WP_TEJ) in modulating the simulation biases in terms of their climatological distribution and interannual variability of WNP TC genesis frequency (TCGF) based on the analysis of the genesis potential index (GPI). Results show that the two models reasonably capture the main TC genesis location, the maximum center of frequency, and track density; however, their biases mainly lie in simulating the intense TCs and TCGF distributions. The MRI better simulates the wind-pressure relationship (WPR) but overestimates the proportion of super typhoons (SSTYs). At the same time, FGOALS underestimates the WPR and the proportion of SSTYs but better simulates the total WNP TC precipitation. In particular, FGOALS overestimates the TCGF in the northeastern WNP, which is strongly tied to an overestimated WP_TEJ and the enhanced vertical circulation to the north of its entrance region. In contrast, the MRI simulates a weaker WP_TEJ and vertical circulation, leading to a negative TCGF bias in most of the WNP. Both models exhibit comparable capability in simulating the interannual variability of WP_TEJ intensity, but the composite difference of large-scale atmospheric factors between strong and weak WP_TEJ years is overestimated, resulting in larger interannual anomalies of WNP TCGF, especially for FGOALS. Therefore, accurate simulations of the WP_TEJ and the associated oceanic and atmospheric factors are crucial to further improving WNP TC simulations for both models.