Tropical cyclone track prediction by a high resolution limited area model using synthetic observations

ABSTRACT. Results of tropical cyclone track prediction experiments in die Indian seas by a high resolution limited area numerical weather prediction model (1° × 1° lat./long. grid) are presented. As the tropical cyclones form in data sparse regions of tropical oceans, and are, therefore, not well analysed in die initial fields, a scheme has been developed for generation of synthetic observations -based on die empirical structure of tropical cyclones, and their assimilation into the objective analysis, for preparing initial fields for running a forecast model. Experiments on track prediction have beat : conducted for die cyclones forming in the Bay of Bengal and Arabian Sea during the period 1990-95. Forecast errors of the model for 24 hr and 48 hr forecasts have been computed. A sensitivity experiment has been carried out to demonstrate the importance of initial humidity field on forecast model performance. The experiment brings out crucial important of the initial humidity field prescription in accurate track prediction by die forecast model.

A Novel Deep Learning Model by BiGRU with Attention Mechanism for Tropical Cyclone Track Prediction in Northwest Pacific

Tropical cyclones are amongst the most powerful and destructive meteorological systems on earth. In this paper, we propose a novel deep learning model for tropical cyclone track prediction method. Specifically, the track task is regarded as a time series predicting challenge, and then a deep learning framework by Bi-directional Gate Recurrent Unit network (BiGRU) with attention mechanism is developed for track prediction. This proposed model can excavate the effective information of the historical track in a deeper and more accurate way. Data exepriments are conducted on tropical cyclone best track data provided by the Joint Typhoon Warning Center (JTWC) from 1988 to 2017 in the Northwest Pacific. As results, our model performs well in tracks of 6h, 12h, 24h, 48h and 72h in the future. The prediction results show that our proposed combined model are superior to state-of-the-art deep learning models, include Recurrent Neural Network (RNN), Long Short-Term Memory neural network (LSTM), Gate Recurrent Unit network (GRU) and BiGRU without the use of attention mechanism. Compared with China Meteorological Administration (CMA), Japan Meteorological Agency (JMA) and Joint Typhoon Warning Center (JTWC), our method has obvious advantage in the mid- to long-term track forecasting, especially in the next 72 hours.

WRF-model sensitivity test and assimilation studies of Cempaka tropical cyclone

The WRF model was used to forecast the most intensive stage of Cempaka Tropical Cyclone (TC) on 27 - 29 November 2017. This study evaluates the combination of cumulus and microphysics parameterization and the efficiency of assimilation method to predict pressure values at the center of the cyclone, maximum wind speed, and cyclone track. This study tested 18 combinations of cumulus and microphysics parameterization schemes to obtain the best combination of both parameterization schemes which later on called as control model (CTL). Afterward, assimilation schemes using 3DVAR cycles of 1, 3, 6 hours, and 4DVAR, namely RUC01, RUC03, RUC06, and 4DV, were evaluated for two domains with grid size of each 30 and 10 km. GFS data of 0.25-degree and the Yogyakarta Doppler Radar data were used as the initial data and assimilation data input, respectively. The result of the parameterization test shows that there is no combination of parameterization schemes that constantly outperform all variables. However, the combination of Kain-Fritsch and Thompson can produce the best prediction of tropical cyclone track compared to other combinations. While, the RUC03 assimilation scheme was noted as the most efficient method based on the accuracy of track prediction and duration of model time integration.

Rapid prediction of peak storm surge from tropical cyclone track time series using machine learning

Rapid and accurate prediction of peak storm surges across an extensive coastal region is necessary to inform assessments used to design the systems that protect coastal communities’ life and property. Significant advances in high-fidelity, physics-based numerical models have been made in recent years, but use of these models for probabilistic forecasting and probabilistic hazard assessment is computationally intensive. Several surrogate modeling approaches based on existing databases of high-fidelity synthetic storm surge simulations have been recently suggested to reduce computational burden without substantial loss of accuracy. In these previous studies, however, the surrogate modeling approaches relied on a tropical cyclone condition at one moment (usually at or near landfall), which is not always most correlated with the peak storm surge. In this study, a new one-dimensional convolutional neural network model combined with principal component analysis and a k-means clustering (C1PKNet) is presented that can rapidly predict peak storm surge across an extensive coastal region from time-series of tropical cyclone conditions, namely the storm track. The C1PKNet model was trained and cross-validated for the Chesapeake Bay area of the United States using existing database of 1031 high-fidelity storm surge simulations, including both landfalling and bypassing storms. Moreover, the performance of the C1PKNet model was evaluated based on observations from three historical hurricanes (Hurricane Isabel in 2003, Hurricane Irene in 2011, and Hurricane Sandy in 2012). The results indicate that the C1PKNet model is computationally e cient and can predict peak storm surges from realistic tropical cyclone track time-series. We believe that this new surrogate model can enhance coastal resilience by providing rapid storm surge predictions.