Statistical model for seasonal prediction of tropical cyclone frequency in the mid-latitudes of East Asia

This study constructed a multiple linear regression model (MLRM) for the seasonal prediction of summer tropical cyclone (TC) frequency in the mid-latitudes of East Asia and then analyzed its validity using large-scale environments. The 850-hPa geopotential heights of the preceding April in the open ocean east of the Philippines and in the Bering Sea were used as independent variables. In the low-frequency years predicted by the MLRM, there was a larger amount of sea ice around the Sea of Okhotsk during the preceding spring and its cooling effect continued into the summer. In addition, topographic and geographic effects around the Sea of Okhotsk that results in the easy formation of cold air created an anomalous cold high over this region in the summer. As a result, the northerlies from an anomalous cold high around the Sea of Okhotsk caused cold surface air temperature anomalies in the mid-latitudes of East Asia, which played an important role in preventing a western Pacific subtropical high from advancing toward the mid-latitudes of East Asia. Eventually, these environments led to a reduced summer TC frequency in the mid-latitudes of East Asia.