Tropical cyclones (TCs) in the western North Pacific (WNP) basin pose a significant threat to densely populated coastal regions of East Asia. While TC intensity (maximum sustained surface wind) is a widely used metric for assessing threat, the damage caused also depends heavily on forecasting accuracy, which is often low for rapidly intensifying TCs. This study addresses this limitation by introducing the concept of TC potential threat (PT), incorporating both TC intensity and intensification rate (IR). The authors hypothesize that a combined assessment of intensity and IR provides a more comprehensive measure of TC threat than intensity alone. The research aims to objectively categorize TCs based on PT levels using a clustering algorithm, analyze long-term fluctuations in PT, and investigate the influence of environmental factors, particularly ocean temperature, on these trends. The study’s importance lies in offering a more nuanced understanding of TC risk and its potential escalation due to climate change. The western North Pacific is chosen for this study due to the high frequency and intensity of TCs in this region.

Previous research primarily focused on TC intensity as the primary indicator of threat, neglecting the crucial role of forecasting accuracy. Studies have shown slow progress in TC intensity forecasting, particularly for TCs with high intensification rates. Existing research highlights the importance of various factors in TC destruction, including precipitation, minimum sea-level pressure, outer size, and storm surge. However, the challenge of accurately forecasting rapid intensification (RI) remains a significant obstacle. While some studies have examined increases in the number of major TCs and their intensification rates, the combined effect of intensity and IR on TC threat and the associated environmental drivers hasn't been comprehensively investigated. Existing research on the impact of atmospheric and oceanic conditions (humidity, wind shear, ocean temperature) on TC activity primarily focuses on individual factors, without fully considering the joint effects on PT. This study builds on this existing literature by integrating intensity and IR, providing a more complete picture of TC threat and its relationship to climate change.

The study utilizes the International Best Track Archive for Climate Stewardship (IBTRACS) dataset from 1980 to 2020 for TC best-track observations. Data from the Joint Typhoon Warning Center (JTWC) is primarily used, complemented by data from the China Meteorological Administration (CMA), Hong Kong Observatory (HKO), and Japan Meteorological Agency (JMA) for comparison and validation. The K-means clustering algorithm is employed to categorize TCs into distinct groups based on lifetime maximum intensity (LMI), average intensity change within 24 hours (AV), intensification change within 24 hours before LMI (AV24), and lifetime maximum intensification rate (LMIR). The silhouette and Davies-Bouldin indices were used to determine the optimal number of clusters. The European Centre for Medium-Range Weather Forecasts (ECMWF) prediction errors, obtained from the THORPEX Interactive Grand Global Ensemble (TIGGE) dataset, were analyzed to assess forecast skill for different TC groups. Long-term fluctuations in PT and associated environmental conditions were analyzed using composite analysis. High-resolution reanalysis data (ERA5 for atmospheric variables and GLORYS for oceanic variables) were used to characterize the environmental factors associated with different PT levels. Long-term trends in PT were investigated using linear regression, considering the influence of global warming and internal climate variability (ENSO, PDO, AMO). A bootstrap experiment was conducted to assess the robustness of the long-term trend in TC counts. The study used two-tailed t-tests with p-values of 0.01 to test significance.

The K-means clustering resulted in four groups of TCs. Groups A and B, designated as severe TCs, showed significantly higher LMI and IR compared to Groups C and D. Group A TCs, while not necessarily the most intense, displayed significantly higher forecast errors due to their rapid intensification. The study found a significant increase (22% per decade) in the annual number of Group A TCs over the 41-year period, while the overall number of TCs showed only a slight decrease. Composite analysis revealed that subsurface ocean temperature, specifically Tropical Cyclone Heat Potential (TCHP), was the most significant environmental factor differentiating Group A and B TCs. Group A TCs had considerably higher TCHP values. The long-term trend analysis showed a significant increase in TCHP in the primary TC basin (5–20°N, 120–180°E) over the past four decades, consistent with the observed increase in Group A TCs. The increase in Group A TC activity was found to be more closely related to global SST warming than to PDO and AMO indices. After removing the signal of global SST, the trend of the residual count of TCs in Group A was no longer statistically significant. In contrast, the remaining TC count still increased significantly after removing PDO, Niño3, or AMO signals.

The findings demonstrate that incorporating both intensity and intensification rate in assessing TC threat offers a more comprehensive understanding of TC risk than using intensity alone. The significant increase in the number of high-PT TCs, especially those in Group A, highlights the increasing risk of TC-related hazards in a warming climate. The dominance of TCHP as the environmental factor influencing TC intensification underscores the importance of monitoring subsurface ocean temperatures for improved TC prediction. The strong correlation between the increase in high-PT TCs and global SST warming suggests that global warming is a significant driver of this increase. The study contributes to a more refined approach to TC risk assessment and emphasizes the need for improved forecasting and mitigation strategies.

This study introduces the novel concept of TC potential threat (PT), integrating intensity and intensification rate, providing a more comprehensive assessment of TC risk. The observed increase in high-PT TCs is strongly linked to rising subsurface ocean temperatures, highlighting the impact of global warming. The findings emphasize the need for improved TC forecasting and preparedness measures to mitigate the increasing risk of TC-related hazards. Future research should focus on expanding the PT concept to include other factors affecting TC damage (size, precipitation, storm surge) and exploring the underlying physical mechanisms connecting subsurface ocean warming and TC intensification using higher-resolution numerical models.

The study is limited by the availability of reliable best-track data, restricting the analysis to the period 1980–2020. The impact of spatial and seasonal distributions of ocean temperature anomalies was not fully explored. While the bootstrap experiment adds robustness to the findings, a longer observational period would strengthen the long-term trend conclusions. The study focused primarily on TC intensity, neglecting other factors contributing to TC damage. Further, the specific K-means clustering method employed might influence the results; investigating other clustering models would enhance the study's generalizability.