Abstract:Using the subjective forecasts, and the observation data of China's automatic weather station (AWS) observations, the European Centre for Medium-Range Weather Forecasts (ECMWF) reanalysis data, the Characteristics and Forecasting difficulties of Typhoon Gaemi are analyzed. The results as following: (1) Gaemi gradually intensified after forming over east sea of the Philippines and made landfall in Yilan, Taiwan and Putian, Fujian successively, bringing widespread and intense wind and rain impacts. Daily precipitation at 14 national meteorological observatories broke historical records, demonstrating strong extremity. (2) During the formation stage of Gaemi, the atmospheric circulation was in a dynamic adjustment stage, and it formed concurrently with Prapiroon. The binary interaction between the two typhoon vortices further added the difficulty of track forecasting. Although subjective forecasts have positive skill corrections, significant deviations still existed in the typhoon track forecasts. (3) The complex terrain of Taiwan caused high uncertainty into the track forecasting of Typhoon Gaemi before and after it crossed the Taiwan island, greatly increasing the challenge of accurate landing forecasts. However, high-resolution numerical models have certain advantages in forecasting small-scale fluctuations (such as track loops) in the track under complex terrain conditions. (4) Gaemi suddenly turned westward in northern Jiangxi, leading to a significant deviation in the track forecast This was because Gaemi was in a complex saddle-shaped environment, where its movement was extremely sensitive to minor changes in the surrounding circulation systems, greatly reducing the predictability of the typhoon's later track. Nevertheless, AI-models such as Fengqing demonstrated unique advantages in forecasting the westward turn of the track. In the future, we will enhance our understanding and application of AI meteorological models. The factors influencing the track of typhoon over land are more and complex, making forecasting much more challenging than that over the sea. In the future, it is necessary to conduct in-depth research on the mechanisms of typhoon track changes over land to enhance our understanding of them.