Search

This paper proposes a spatiotemporal attention convolutional network (STAC-Pred) that leverages deep learning techniques to model the spatiotemporal features of tropical cyclones (TCs) and enable real-time prediction of their intensity. The proposed model employs dual branches to concurrently extract and integrate features from intensity heatmaps and satellite cloud imagery. Additionally, a residual attention (RA) module is integrated into the three-channel cloud imagery convolution process to automatically respond to high wind speed regions. TC's longitude, latitude, and radius of winds are injected into the multi-timepoint prediction model to assist in the prediction task. Furthermore, a rolling mechanism (RM) is employed to smooth the fluctuation of losses, achieving accurate prediction of TC intensity. We use several TC records to evaluate and validate the universality and effectiveness of the model. The results indicate that STAC-Pred achieves satisfactory performance. Specifically, the STAC-Pred model improves prediction performance by 47.69% and 28.26% compared to the baseline (official institutions) at 3- and 6-h intervals, respectively. Significance Statement: Tropical cyclones are one of the most deadly and damaging natural disasters in coastal areas worldwide. Early prediction can significantly reduce casualties and property losses. This study innovatively conducts dimensionality augmentation on one-dimensional intensity numerical sequences and proposes a new network model for rolling forecast of their future intensity. The proposed prototype model (not yet incorporating any atmospheric conditions) shows promising results for 3- and 6-h advance forecasts, providing valuable guidance for forecasters regarding real-time operational predictions of short-term tropical cyclone intensity. [ABSTRACT FROM AUTHOR]

The upper ocean mixed‐layer depth (MLD) has been recognized as a crucial factor in understanding the impact of climate change on tropical cyclone (TC) intensity, while the specific MLD variations in the western North Pacific (WNP) basin have received comparatively little attention. Using two ocean reanalysis datasets and the output of 28 CMIP6 models, the climatological mean MLD and its trend in the TC active region are examined. Compared to the ORAS4 and EN4 datasets, the spatial pattern of the climatological mean MLD can be well simulated in the multi‐model ensemble (MME). The mean MLD decreases with increasing latitude and maintains the maximum in the equatorial region. The spatial patterns of MLD long‐term (1958–2017) and short‐term (1979–2017) trends are deepening in the TC active region, with the relatively large positive trends occurring to the southeastern WNP primarily between 5°N and 15°N. However, the simulated trend in the MME is dominant with shoaling MLD, and the shoaling of the MLD is also projected under the SSP5‐8.5 scenario. It is suggested that the bias on the simulated shoaling of the MLD in CMIP6 models may result from the overestimated enhancement of the upper‐ocean stratification. Thus, there is a discrepancy between the observed and simulated MLD trends in the WNP basin, which may have implications for TC intensity change under global warming scenarios. [ABSTRACT FROM AUTHOR]

The North Pacific monsoon trough (NPMT) is an imperative large‐scale circulation pattern influencing tropical cyclone (TC) activity in the western North Pacific (WNP), thus its future change has great implications for the WNP TC activity. Here future change in the NPMT and its uncertainty are examined by using 35 climate models from Phase six of Coupled Model Intercomparison Project (CMIP6). Due to the El Niño‐like sea surface temperature (SST) pattern, the multi‐model ensemble (MME) mean projects an eastward extension of the NPMT in the latter half of the 21st century. However, considerable inter‐model uncertainty exists in the projection, which is associated with the diversity in the zonal SST gradient across the tropical Pacific. This diversity in the projected SST is largely rooted in the simulated precipitation biases in the tropical Pacific in the historical experiments of CMIP6 models, which can modulate future Pacific SST distribution through the shortwave‐SST feedback. This linkage between biases in historical climate and future climate allows us to constrain the projection by using observational data sets. Emergent constraint using observational precipitation data set reduces the projection uncertainty by 48% and projects an eastward extension of the north Pacific NPMT by 6.2°. This eastward extension of the NPMT indicates an eastward migration of TC genesis location, and elongated TC track and thus strengthened TC intensity, suggesting an increased TC‐related disaster for residents near the WNP. Plain Language Summary: Tropical cyclones (TCs) in the western North Pacific (WNP) greatly threaten the residents in the Pacific coastal regions due to the TC‐related enormous disasters. How TCs will change in a warmer climate is of great scientific and societal concern. However, current state‐of‐the‐art climate models are still insufficient to resolve TCs. Here we take an alternative way to examine the future change in a large‐scale circulation important for TC activity (i.e., North Pacific monsoon trough (NPMT)) to infer the future WNP TC activity. It is found that large inter‐model uncertainty exists in the projected NPMT, which is linked to simulated precipitation biases of climate models. After constraining the bias by using observational data, the projection shows a robust eastward extension of the NPMT in warming scenarios, indicating an increased TC‐related disaster for residents near the WNP in the future. The results highlight the crucial role of climate models' skills in simulating the current climate in the future climate projection and have important implications for projecting future TC activity. Key Points: The large inter‐model uncertainty in projecting the North Pacific monsoon trough is linked to the diverse zonal Pacific SST gradientThis diversity is rooted in precipitation biases of CIMP6 models, which can modulate future Pacific SST through the shortwave‐SST feedbackEmergent constraint projects an eastward extension of the North Pacific monsoon trough and reduces the projection uncertainty by 48% [ABSTRACT FROM AUTHOR]

Attaining skillful decadal predictions for the western North Pacific tropical cyclone (TC) emerges as a formidable challenge, mainly stemming from the limited prediction skills of Pacific Decadal Variability (PDV) within the state‐of‐the‐art models. Assessing sixth Coupled Model Intercomparison Project models' retrospective predictions finds that the predictability of PDV transcends the expectations set by raw forecasts featured by constrained temporal skills and low signal‐to‐noise ratio. Employing a refined approach, we selectively identify the models that capture the diverse phases of PDV and subsequently adjust their variances. This tailored approach yields a compelling concordance between the predicted PDV and observation in phases and variances. Anchored in the heightened prediction skill of PDV, we establish a sophisticated statistical model adept at predicting the latitude of TC's lifetime maximum intensity (LMI). The near‐term prediction indicates a sustained poleward migration of LMI latitude by 1.53° during 2020–2027, increasing subtropical East Asia's TC‐related disaster vulnerability in the coming decade. Plain Language Summary: Tropical cyclones (TCs) account for one‐third of the deaths and economic losses from weather‐, climate‐ and water‐related disasters. This ratio tends to be much greater in the East Asia region due to its neighborhood to the busiest TC basin in the world‐the western North Pacific (WNP). However, it is particularly challenging to make skillful decadal predictions for the WNP TC activity due to the limited prediction skill of Pacific Decadal Variability (PDV), the foremost driver for the TC decadal variability. Here, we find that the capacity of the current state‐of‐the‐art climate model predicting PDV can be substantially improved. The improved prediction indicates that the latitude of the TCs' lifetime maximum intensity will migrate poleward in 2020–2027, making subtropical East Asia more vulnerable to TC disasters. The finding has profound implications for disaster prevention and mitigation strategies in subtropical East Asia, offering pivotal insights that can significantly enhance preparedness measures. Key Points: Near‐term prediction of tropical cyclone (TC) activity depends on Models' skills to predict Pacific Decadal Variability (PDV), which is limitedWe substantially improved PDV's hindcast skill and the refined approach predicts a PDV negative phase in the near‐term (2020–2027)The predicted PDV phase implies a poleward shift of location of TC's lifetime maximum intensity, increasing subtropical East Asia's TC‐related risk [ABSTRACT FROM AUTHOR]