Lightning Tames Typhoon Intensity Forecasting

Typhoons regularly drench densely populated western Pacific regions, but lightning could forecast intensity more than a day before a storm’s strength peaks.

Decision-Tree-Based Classification of Lifetime Maximum Intensity of Tropical Cyclones in the Tropical Western North Pacific

The National Typhoon Center of the Korea Meteorological Administration developed a statistical–dynamical typhoon intensity prediction model for the western North Pacific, the CSTIPS-DAT, using a track-pattern clustering technique. The model led to significant improvements in the prediction of the intensity of tropical cyclones (TCs). However, relatively large errors have been found in a cluster located in the tropical western North Pacific (TWNP), mainly because of the large predictand variance. In this study, a decision-tree algorithm was employed to reduce the predictand variance for TCs in the TWNP. The tree predicts the likelihood of a TC reaching a maximum lifetime intensity greater than 70 knots at its genesis. The developed four rules suggest that the pre-existing ocean thermal structures along the track and the latitude of a TC’s position play significant roles in the determination of its intensity. The developed decision-tree classification exhibited 90.0% and 80.5% accuracy in the training and test periods, respectively. These results suggest that intensity prediction with the CSTIPS-DAT can be further improved by developing independent statistical models for TC groups classified by the present algorithm.

Domain knowledge integration into deep learning for typhoon intensity classification

In this report, we propose a deep learning technique for high-accuracy estimation of the intensity class of a typhoon from a single satellite image, by incorporating meteorological domain knowledge. By using the Visual Geometric Group’s model, VGG-16, with images preprocessed with fisheye distortion, which enhances a typhoon’s eye, eyewall, and cloud distribution, we achieved much higher classification accuracy than that of a previous study, even with sequential-split validation. Through comparison of t-distributed stochastic neighbor embedding (t-SNE) plots for the feature maps of VGG with the original satellite images, we also verified that the fisheye preprocessing facilitated cluster formation, suggesting that our model could successfully extract image features related to the typhoon intensity class. Moreover, gradient-weighted class activation mapping (Grad-CAM) was applied to highlight the eye and the cloud distributions surrounding the eye, which are important regions for intensity classification; the results suggest that our model qualitatively gained a viewpoint similar to that of domain experts. A series of analyses revealed that the data-driven approach using only deep learning has limitations, and the integration of domain knowledge could bring new breakthroughs.

Can a Warm Ocean Feature Cause a Typhoon to Intensify Rapidly?

The hypothesis that a warm ocean feature (WOF) such as a warm eddy may cause a passing typhoon to undergo rapid intensification (RI), that is, the storm’s maximum 1-min wind speed at 10-m height increases by more than 15.4 m/s in 1 day, is of interest to forecasters. Testing the hypothesis is a challenge, however. Besides the storm’s internal dynamics, typhoon intensity depends on other environmental factors such as vertical wind shear and storm translation. Here we designed numerical experiments that exclude these other factors, retaining only the WOF’s influence on the storm’s intensity change. We use a storm’s translation speed Uh = 5 m/s when surface cooling is predominantly due to 1D vertical mixing. Observations have shown that the vast majority (70%) of RI events occur in storms that translate between 3 to 7 m/s. We conducted a large ensemble of twin experiments with and without ocean feedback and with and without the WOF to estimate model uncertainty due to internal variability. The results show that the WOF increases surface enthalpy flux and moisture convergence in the storm’s core, resulting in stronger updrafts and intensity. However, the intensification rate is, in general, insufficiently rapid. Consequently, the number of RIs is not statistically significantly different between simulations with and without the WOF. An analytical coupled model supports the numerical findings. Furthermore, it shows that WOF-induced RI can develop only over eddies and ambient waters that are a few °C warmer than presently observed in the ocean.

Deep learning for typhoon intensity classification using satellite cloud images

Tropical cyclone, also known as typhoon, is one of the most destructive weather phenomena. Its intense cyclonic eddy circulations often cause serious damages to coastal areas. Accurate classification or prediction for typhoon intensity is crucial to the disaster warning and mitigation management. But typhoon intensity-related feature extraction is a challenging task as it requires significant pre-processing and human intervention for analysis, and its recognition rate is poor due to various physical factors such as tropical disturbance. In this study, we built a Typhoon-CNNs framework, an automatic classifier for typhoon intensity based on convolutional neural network (CNN). Typhoon-CNNs framework utilized a cyclical convolution strategy supplemented with dropout zero-set, which extracted sensitive features of existing spiral cloud band (SCB) more effectively and reduces over-fitting phenomenon. To further optimize the performance of Typhoon-CNNs, we also proposed the improved activation function (T-ReLU) and the loss function (CE-FMCE). The improved Typhoon-CNNs was trained and validated using more than 10,000 multiple sensor satellite cloud images of National Institute of Informatics. The classification accuracy reached to 88.74%. Compared with other deep learning methods, the accuracy of our improved Typhoon-CNNs was 7.43% higher than ResNet50, 10.27% higher than InceptionV3 and 14.71% higher than VGG16. Finally, by visualizing hierarchic feature maps derived from Typhoon-CNNs, we can easily identify the sensitive characteristics such as typhoon eyes, dense-shadowing cloud areas and SCBs, which facilitates classify and forecast typhoon intensity.

A Tale of Two Rapidly-Intensifying Supertyphoons: Hagibis (2019) and Haiyan (2013)

Devastating Japan in October 2019, Supertyphoon (STY) Hagibis was an important typhoon in the history of the Pacific. A striking feature of Hagibis was its explosive RI (rapid intensification). In 24 h, Hagibis intensified by 100 kt, making it one of the fastest-intensifying typhoons ever observed. After RI, Hagibis’s intensification stalled. Using the current typhoon intensity record holder, i.e., STY Haiyan (2013), as a benchmark, this work explores the intensity evolution differences of these 2 high-impact STYs.We found that the extremely high pre-storm sea surface temperature reaching 30.5°C, deep/warm pre-storm ocean heat content reaching 160 kJ cm−2, fast forward storm motion of ~8 ms−1, small during-storm ocean cooling effect of ~ 0.5C, significant thunderstorm activity at its center, and rapid eyewall contraction were all important contributors to Hagibis’s impressive intensification. There was 36% more air-sea flux for Hagibis’s RI than for Haiyan’s.After its spectacular RI, Hagibis’s intensification stopped, despite favorable environments. Haiyan, by contrast, continued to intensify, reaching its record-breaking intensity of 170 kt. A key finding here is the multiple pathways that storm size affected the intensity evolution for both typhoons. After RI, Hagibis experienced a major size expansion, becoming the largest typhoon on record in the Pacific. This size enlargement, combined with a reduction in storm translational speed, induced stronger ocean cooling that reduced ocean flux and hindered intensification. The large storm size also contributed to slower eyewall replacement cycles (ERCs), which prolonged the negative impact of the ERC on intensification.

Impacts of Air-Sea Energy Transfer on Typhoon Modelling

The Coupled Ocean-Atmosphere-Wave-Sediment Transport model has been used to simulate Super Typhoon Yutu (2018). The impacts of four momentum transfer parameterization schemes (COARE, TY, OT, and DN) and three heat transfer parameterization schemes (COARE, GR, and ZK) on typhoon modelling have been studied by means of the track, intensity, and radial structure of typhoon. The results show that the track of Yutu is not sensitive to the choice of parameterization scheme, while the combinations of different parameterization schemes affect the intensity of Yutu. Among the four momentum flux parameterization schemes, three wave-state-based schemes (TY, OT, and DN) provide better intensity results than the wind-speed-based COARE scheme, but the differences between the three wave-state-based schemes are not obvious. Among the three heat flux parameterization schemes, the results of the GR scheme are slightly better than those of the COARE scheme, and both the GR and COARE schemes are significantly better than the ZK scheme, from which the intensity of Yutu is underpredicted obviously. The influence of the combination of different parameterization schemes on the intensity of the typhoon is also reflected in the radial structure of the typhoon, and the radial structure of the typhoon simulated by experiments with stronger typhoon intensity also develops faster. Differences of intensity between experiments are due mainly to the differences in sea surface heat flux, the enthalpy transferred from sea surface to the atmosphere has a significant impact on the bottom atmosphere wind field, and there is a strong correspondence between the distribution of enthalpy flux and the bottom wind field.

Coordinated observation system for extreme weathers consisting of AWS network with lightning sensor and micro-satellites

<p>In order to predict the intensity and location of extreme weathers, such as torrential rainfall by individual thunderstorm or typhoon, we are developing the new methodology of weather monitoring using a ground AWS network with lightning sensors and micro-satellites weighting about 50kg, which will realize quasi-real-time thunderstorm monitoring with broad coverage. Based on the AWS network data, we plan to operate micro-satellites in nearly real-time, manipulating the attitude of satellite for capturing the most dangerous or important cloud images for 3D reconstruction. We have developed and launched several micro-satellites and been improving the target pointing operation for this decade. We succeeded in obtaining the images of the typhoon center at a resolution of 60-100 m for Typhoon Trami in 2018 and Typhoon Maysak in 2020. Using 4 or a few 10s images captured from different angles by one micro-satellite when it passed over the typhoon area, 3D models of typhoon eye were reconstructed, which have a ground resolution of ~100 m. Due to the unusual temperature profile around typhoon eye, it’s very difficult to estimate the heigh distribution of cloud top only with a thermal infrared image at a resolution of 2 km taken by geostationary meteorological satellite. This is one of the biggest limitations in estimating the precise intensity of typhoons, namely, the center pressure or the maximum wind velocity. The on-demand flexible operation of micro-satellite will achieve the high accuracy estimation of typhoon intensity as well as the speed estimation of individual thunderstorm development, which can be applied to disaster management. This research was conducted by a mixed team of Japan and the Philippines, supported by Science and Technology Research Partnership for Sustainable Development (SATREPS), which is funded by Japan Science and Technology Agency (JST) / Japan International Cooperation Agency (JICA).</p>

Typhoon Intensity Forecasting Based on LSTM Using the Rolling Forecast Method

A typhoon is an extreme weather event with strong destructive force, which can bring huge losses of life and economic damage to people. Thus, it is meaningful to reduce the prediction errors of typhoon intensity forecasting. Artificial and deep neural networks have recently become widely used for typhoon forecasting in order to ensure typhoon intensity forecasting is accurate and timely. Typhoon intensity forecasting models based on long short-term memory (LSTM) are proposed herein, which forecast typhoon intensity as a time series problem based on historical typhoon data. First, the typhoon intensity forecasting models are trained and tested with processed typhoon data from 2000 to 2014 to find the optimal prediction factors. Then, the models are validated using the optimal prediction factors compared to a feed-forward neural network (FNN). As per the results of the model applied for typhoons Chan-hom and Soudelor in 2015, the model based on LSTM using the optimal prediction factors shows the best performance and lowest prediction errors. Thus, the model based on LSTM is practical and meaningful for predicting typhoon intensity within 120 h.