Search

An air-sea interaction field campaign was conducted in September of 2017 within the warm and cold eddy region of the western Northwest Pacific (WNP) (17.5˚-20.5˚N, 127.5˚E-133.5˚E). Both near-surface oceanic and atmospheric conditions in addition to ocean heat content (OHC) were examined to better understand the mechanisms governing high heat flux magnitudes responsible for rapidly intensifying tropical cyclones. Observations from this experiment indicate that the latent heat flux (LHF) under modest wind conditions reached 400 W m-2 within the vicinity of a warm eddy, with OHC higher than 100 kJ cm-2 of warm eddy region being 2~3 times as large as that of cold eddy region. These high OHC and a resultant high LHF in the warm eddy, comparable to the magnitude of LHF in the North Equatorial Current, may explain the mechanism of why tropical cyclones over a warm eddy in eddy-rich zones often rapidly intensify in the WNP. A month later typhoon Lan rapidly intensified into a super typhoon, while passing over the boundary region of warm and cold eddies during the observation period. Results from this study illustrate that both the wind-normalized LHF and the difference (Qs-Qa) between the specific humidity at air (Qa) and at the sea surface (Qs), closely correlate with OHC patterns, which suggests that the ocean looks likely to control the spatial pattern of LHF. Overall, both the ocean and weather conditions govern the pattern of specific humidity difference between the air-sea interface, with large OHC over the warm eddy controlling higher Qs and the pattern of Qa depending on the pattern of wind direction. Qa as a factor impacting LHF magnitude is strongly linked to wind direction in the experimental area, that is, the drier northwesterly flow and southeasterly moist wind, resulting in the enhanced contrast of specific humidity at cold eddy region.

In this study, a machine learning (ML)-based Tropical Cyclones (TCs) Rapid Intensification (RI) prediction model has been developed by using the Net Energy Gain Rate Index (NGR). This index realistically captures the energy exchanges between the ocean and the atmosphere during the intensification of TCs. It does so by incorporating the thermal conditions of the upper ocean and using an accurate parameterization for sea surface roughness. To evaluate the effectiveness of NGR in enhancing prediction accuracy, five distinct ML algorithms were utilized: Decision Tree, Logistic Regression, Support Vector Machine, K-Nearest Neighbors, and Feed-forward Neural Network. Two sets of experiments were performed for each algorithm. The first set used only traditional predictors, while the second set incorporated NGR. The outcomes revealed that models trained with the inclusion of NGR exhibited superior performance compared to those that only used traditional predictors. Additionally, an ensemble model was developed by utilizing a hard-voting method, combining the predictions of all five individual algorithms. This ensemble approach showed a noteworthy improvement of approximately 10% in the skill score of RI prediction when NGR was included. The findings of this study emphasize the potential of NGR in refining TC intensity prediction and underline the effectiveness of ensemble ML models in RI event detection.

The air-sea drag coefficient (Cd) is closely related to tropical cyclone (TC) intensification. Several recent studies suggested that the Cd decreases in winds greater than 33 m s−1. The effects of Cd reduction in high winds on TC intensity, especially rapid intensification (RI) and the lifetime maximum intensity (LMI) distribution, were investigated by analyzing the wind-dependent Cd-based ocean vertical mixing and the energy budget. In addition, to consider the uncertainty of the Cd in extreme winds (above 50 m s−1), three types of Cd fitting that decrease after 33 m s−1, which show different trends after 50 m s−1 (increase, flat, and decrease), were adopted. The results were then compared with those for the control fitting (saturated after 33 m s−1) and show that the reduced Cd in high winds drives an increase in net energy by reducing frictional dissipation and suppressing sea surface cooling. This extra energy prevents the TC from achieving a steady-state, causing the bimodality of simulated maximum potential intensity. The observed steady-state probability (intensification rate and RI probability) in the Cd reduced wind range were significantly lower (higher) than in the others. These results suggest that Cd reduction might potentially induce the RI and LMI bimodality.

The East Sea (ES; Sea of Japan) meridional overturning circulation (MOC) serves as a crucial mechanism for the transportation of dissolved, colloidal, and suspended particulate matters, including pollutants, on the surface to deep waters via thermohaline circulation. Therefore, understanding the structure of the ES MOC is critical for characterizing its temporal and spatial distribution. Numerous studies have estimated these parameters indirectly using chemical tracers, severely limiting the accuracy of the results. In this study, we provide a method for directly estimating the turnover times of the ES MOC using the stream functions calculated from HYbrid Coordinate Ocean Model (HYCOM) reanalysis data by averaging the flow pattern in the meridional 2-D plane. Because the flow pattern is not consistent but various over time, three cases of stream function fields were computed over a 20-year period. The turnover time was estimated by calculating the time required for water particles to circulate along the streamlines. In the cases of multiple (two or three) convection cells, we considered all possible scenarios of the exchange of water particles between adjacent cells, so that they circulated over those cells until finally returning to the original position and completing the journey on the ES MOC. Three different cell cases were tested, and each case had different water particle exchange scenarios. The resulting turnover times were 17.91–58.59 years, 26.41–37.28 years, and 8.68–45.44 years for the mean, deep, and shallow convection cases, respectively. The maximum turnover time, namely 58.59 years, was obtained when circulating the water particle over all three cells, and it was approximately half of that estimated by the chemical tracers in previous studies (∼100 years). This underestimation arose because the streamlines and water particle movement were not calculated in the shallow (3,000 m) in this study. Regardless, the results of this study provide insight into the ES MOC dynamics and indicate that the traditional chemical turnover time represents only one of the various turnover scenarios that could exist in the ES.