Ensemble forecast of tropical cyclone tracks based on deep neural networks

doi:10.1007/s11707-021-0931-8

Front. Earth Sci.

2022, Vol. 16

Issue (3) : 671-677 https://doi.org/10.1007/s11707-021-0931-8

RESEARCH ARTICLE

Ensemble forecast of tropical cyclone tracks based on deep neural networks

Chong WANG¹, Qing XU²(

), Yongcun CHENG^3,⁴, Yi PAN⁵, Hong LI⁶

¹. Key Laboratory of Marine Hazards Forecasting (Ministry of Natural Resources), Hohai University, Nanjing 210098, China
². College of Marine Technology, Faculty of Information Science and Engineering, Ocean University of China, Qingdao 266100, China
³. Southern Marine Science and Engineering Guangdong Laboratory (Guangzhou), Guangzhou 511458, China
⁴. PIESAT Information Technology Co. Ltd., Beijing 100195, China
⁵. College of Harbor, Coastal and Offshore Engineering, Hohai University, Nanjing 210098, China
⁶. Shanghai Typhoon Institute, China Meteorological Administration, Shanghai 200030, China

Download: PDF(2139 KB) HTML
Export: BibTeX | EndNote | Reference Manager | ProCite | RefWorks

Abstract

A nonlinear artificial intelligence ensemble forecast model has been developed in this paper for predicting tropical cyclone (TC) tracks based on the deep neural network (DNN) by using the 24-h forecast data from the China Meteorological Administration (CMA), Japan Meteorological Agency (JMA) and Joint Typhoon Warning Center (JTWC). Data from a total of 287 TC cases over the Northwest Pacific Ocean from 2004 to 2015 were used to train and validate the DNN based ensemble forecast (DNNEF) model. The comparison of model results with Best Track data of TCs shows that the DNNEF model has a higher accuracy than any individual forecast center or the traditional ensemble forecast model. The average 24-h forecast error of 82 TCs from 2016 to 2018 is 63 km, which has been reduced by 17.1%, 16.0%, 20.3%, and 4.6%, respectively, compared with that of CMA, JMA, JTWC, and the error-estimation based ensemble method. The results indicate that the nonlinear DNNEF model has the capability of adjusting the model parameter dynamically and automatically, thus improving the accuracy and stability of TC prediction.

Keywords tropical cyclone track deep neural network ensemble forecast

Corresponding Author(s): Qing XU

Online First Date: 26 April 2022 Issue Date: 29 December 2022

Cite this article:

Chong WANG,Qing XU,Yongcun CHENG, et al. Ensemble forecast of tropical cyclone tracks based on deep neural networks[J]. Front. Earth Sci., 2022, 16(3): 671-677.

URL:

https://academic.hep.com.cn/fesci/EN/10.1007/s11707-021-0931-8
https://academic.hep.com.cn/fesci/EN/Y2022/V16/I3/671

Fig.1 (a) Tracks of TCs over the Northwest Pacific Ocean between 2016 and 2018. The colored dot represents the Saffir-Simpson TC intensity scale. (b) TCs with sudden track changes. The blue dot indicates where the TC suddenly changes its direction.

Tab.1 The average 24-h TC track forecast error (km) of the DNNEF-CJJ model with different schemes between 2016 and 2018

Fig.2 Architecture of our DNNEF model for 24-h prediction of longitude or latitude of TC centers.

Fig.3 Loss value in the training process.

Tab.2 24-h TC track forecast error (km) of different forecast centers and models. The number in bracket denotes the accuracy improvement (%) of the DNNEF-CJ model compared with that of each forecast center or EEF method

Fig.4 24-h forecast error (km) of TC tracks in different years.

Tab.3 24-h TC center latitude and longitude forecast error (° ) of different forecast centers and models. Positive/negative values denote northward/southward or eastward/westward bias of the predicted TC track

Fig.5 24-h forecast error (° ) of TC center latitude and longitude in different years.

Fig.6 Comparison of 24-h forecast results of TC tracks from different models with the observation from the Best Track data every 6 h: (a) Lionrock in August 2016, (b) Ampil in July 2018. The value in the parentheses is the forecast error in km.

Fig.7 Improvement (%) in accuracy of our DNNEF-CJJ model for track forecast of Lionrock with time. The colored dot on the curves represents the maximum wind speed (m/s) of Lionrock at the corresponding time with a time interval of 6 h.

1	C Buckingham, T Marchok, I Ginis, L Rothstein, D Rowe. ( 2010). Short- and medium-range prediction of tropical and transitioning cyclone tracks within the NCEP global ensemble forecasting system. Weather Forecast, 25( 6): 1736– 1754 https://doi.org/10.1175/2010WAF2222398.1
2	R Buizza, M Milleer, T N Palmer. ( 1999). Stochastic representation of model uncertainties in the ECMWF ensemble prediction system. Q J R Meteorol Soc, 125( 560): 2887– 2908 https://doi.org/10.1002/qj.49712556006
3	G M Chen, H Yu, Q Cao, Z Zeng. ( 2013). The performance of global models in TC track forecasting over the Western North Pacific from 2010 to 2012. Trop Cyclone Res Rev, 2( 3): 149– 158
4	L Deng J Li J T Huang K Yao D Yu F Seide M Seltzer G Zweig X He J Williams Y Gong A (2013) Acero. Recent advances in deep learning for speech research at Microsoft. In: Proceedings of International Conference on Acoustics, Speech, and Signal. Vancouver: 8604– 8608
5	R L Elsberry. ( 1995). Recent advancements in dynamical tropical cyclone track predictions. Meteorol Atmos Phys, 56( 1−2): 81– 99 https://doi.org/10.1007/BF01022522
6	A Esteva, B Kuprel, R A Novoa, J Ko, S M Swetter, H M Blau, S Thrun. ( 2017). Dermatologist-level classification of skin cancer with deep neural networks. Nature, 542( 7639): 115– 118 https://doi.org/10.1038/nature21056
7	E S Epstein. ( 1969). Stochastic dynamic prediction. Tellus, 21( 6): 739– 759 https://doi.org/10.3402/tellusa.v21i6.10143
8	S Gao, P Zhao, B Pan, Y Li, M Zhou, J Xu, S Zhong, Z Shi. ( 2018). A nowcasting model for the prediction of typhoon tracks based on a long short term memory neural network. Acta Oceanol Sin, 37( 5): 8– 12 https://doi.org/10.1007/s13131-018-1219-z
9	T M Hamill, J S Whitaker, M Fiorino, S G Benjamin. ( 2011). Global ensemble predictions of 2009’s tropical cyclones initialized with an ensemble kalman filter. Mon Weather Rev, 139( 2): 668– 688 https://doi.org/10.1175/2010MWR3456.1
10	R A Jeffries C R Sampson L E Carr J (1993) Chu. Tropical cyclone forecasters reference guide numerical track forecast guidance. Tech. Rep. No. NRL/PU/7515–93–0011
11	J Jin, M Li, L Jin. ( 2015). Data normalization to accelerate training for linear neural net to predict tropical cyclone tracks. Math Probl Eng, 2015: 931629 https://doi.org/10.1155/2015/931629
12	J A Knaff, C R Sampson, M DeMaria, T P Marchok, J M Gross, C J McAdie. ( 2007). Statistical tropical cyclone wind radii prediction using climatology and persistence. Weather Forecast, 22( 4): 781– 791 https://doi.org/10.1175/WAF1026.1
13	T N Krishnamurti, C M Kishtawal, T E LaRow, D R Bachiochi, Z Zhang, C E Williford, S Gadgil, S Surendran. ( 1999). Improved weather and seasonal climate forecasts from multimodel superensemble. Science, 285( 5433): 1548– 1550 https://doi.org/10.1126/science.285.5433.1548
14	A Krizhevsky, I Sutskever, G E Hinton. ( 2012). ImageNet classification with deep convolutional neural networks. In: Proceedings of the 25th International Conference on Neural Information Processing Systems, 1097– 1105
15	C W Landsea J P Cangialosi ( 2018). Have we reached the limits of predictability for tropical cyclone track forecasting? Bull Am Meteorol Soc, 99(11): 2237–2243
16	C E Leith. ( 1974). Theoretical skill of Monte Carlo forecasts. Mon Weather Rev, 102( 6): 409– 418 https://doi.org/10.1175/1520-0493(1974)102<0409:TSOMCF>2.0.CO;2
17	M Leutbecher, T N Palmer. ( 2008). Ensemble forecasting. J Comput Phys, 227( 7): 3515– 3539 https://doi.org/10.1016/j.jcp.2007.02.014
18	H Li, J Luo, M Xu. ( 2019). Ensemble data assimilation and prediction of typhoon and associated hazards using TEDAPS: Evaluation for 2015–18 Seasons. Front Earth Sci, 13( 4): 733– 743 https://doi.org/10.1007/s11707-019-0794-4
19	C J Neumann, M B Lawrence. ( 1975). An operational experiment in the satistical-dynamical prediction of tropical cyclone motion. Mon Weather Rev, 103( 8): 665– 673 https://doi.org/10.1175/1520-0493(1975)103<0665:AOEITS>2.0.CO;2
20	X Peng, J Fei, X Huang, X Cheng. ( 2017). Evaluation and error analysis of official forecasts of tropical cyclones during 2005–14 over the Western North Pacific. Part I: storm Tracks. Weather Forecast, 32( 2): 689– 712 https://doi.org/10.1175/WAF-D-16-0043.1
21	M Plu. ( 2011). A new assessment of the predictability of tropical cyclone tracks. Mon Weather Rev, 139( 11): 3600– 3608 https://doi.org/10.1175/2011MWR3627.1
22	B (2012) Raddaway. Newsletter No.130-Winter 2011/12. Available at ECMWF website
23	E N Rappaport, J L Franklin, L A Avila, S R Baig, J L II Beven, E S Blake, C A Burr, J G Jiing, C A Juckins, R D Knabb, C W Landsea, M Mainelli, M Mayfield, C J McAdie, R J Pasch, C Sisko, S R Stewart, A N Tribble. ( 2009). Advances and challenges at the National Hurricane Center. Weather Forecast, 24( 2): 395– 419 https://doi.org/10.1175/2008WAF2222128.1
24	S Saha, S Moorthi, X Wu, J Wang, S Nadiga, P Tripp, D Behringer, Y T Hou, H Chuang, M Iredell, M Ek, J Meng, R Yang, M P Mendez, H van den Dool, Q Zhang, W Wang, M Chen, E Becker. ( 2014). The NCEP climate forecast system version2. J Clim, 27( 6): 2185– 2208 https://doi.org/10.1175/JCLI-D-12-00823.1
25	D Silver, A Huang, C J Maddison, A Guez, L Sifre, G van den Driessche, J Schrittwieser, I Antonoglou, V Panneershelvam, M Lanctot, S Dieleman, D Grewe, J Nham, N Kalchbrenner, I Sutskever, T Lillicrap, M Leach, K Kavukcuoglu, T Graepel, D Hassabis. ( 2016). Mastering the game of Go with deep neural networks and tree search. Nature, 529( 7587): 484– 489 https://doi.org/10.1038/nature16961
26	V Sze, Y Chen, T Yang, J S Emer. ( 2017). Efficient processing of deep neural networks: a tutorial and survey. Proc IEEE, 105( 12): 2295– 2329 https://doi.org/10.1109/JPROC.2017.2761740
27	Z Toth, E Kalnay. ( 1997). Ensemble forecasting at NCEP and the breeding method. Mon Weather Rev, 125( 12): 3297– 3319 https://doi.org/10.1175/1520-0493(1997)125<3297:EFANAT>2.0.CO;2
28	K W (1996) Veigas. The development of statistical-physical hurricane prediction model. Final Report, U.S.W.B. Contract Cwb 10966, Travellers Weather Research Center, Hartford, CT, 1996. 19
29	L Wu, H Zong, J Liang. ( 2011). Observational analysis of sudden tropical cyclone track changes in the vicinity of the East China Sea. J Atmos Sci, 68( 12): 3012– 3031 https://doi.org/10.1175/2010JAS3559.1
30	M Ying, W Zhang, H Yu, X Lu, J Feng, Y Fan, Y Zhu, D Chen. ( 2014). An overview of the China Meteorological Administration tropical cyclone database. J Atmos Ocean Technol, 31( 2): 287– 301 https://doi.org/10.1175/JTECH-D-12-00119.1
31	J Yuan Y Chen Y Pan J Dong Y (2017) Luo. Improvement of ensemble forecast of typhoon track in the Northwestern Pacific. Marine Forecasts, 34(2): 37– 42 (in Chinese)
32	R H Zhang, X S Shen. ( 2008). On the development of GRAPES-A new generation of the national operational NWP system in China. Chin Sci Bull, 53: 3429– 3432
33	X Zhi, L Zhang, Y Bai. ( 2011). Application of the multi-model ensemble forecast in the QPF. In: Proceedings of International Conference on Information Science and Technology, 657– 660
34	L Zhu, J Jin, A J Cannon, W W Hsieh. ( 2016). Bayesian neural networks based bootstrap aggregating for tropical cyclone tracks prediction in South China Sea. In: Proceeding of International Conference on Neural Information Processing, 475– 482

[1]	Yixuan ZHONG, Shenglian GUO, Feng XIONG, Dedi LIU, Huanhuan BA, Xushu WU. Probabilistic forecasting based on ensemble forecasts and EMOS method for TGR inflow[J]. Front. Earth Sci., 2020, 14(1): 188-200.
[2]	Hong LI, Jingyao LUO, Mengting XU. Ensemble data assimilation and prediction of typhoon and associated hazards using TEDAPS: evaluation for 2015–2018 seasons[J]. Front. Earth Sci., 2019, 13(4): 733-743.

Viewed

Full text

Abstract

Cited

Shared

Discussed