Authors

Nguyen Gia Trong 1, 2 , Nguyen Xuan Hien 3 , Vu Duc Manh 4 , Tran Duc Vinh 5*

Affiliations

1 Hanoi University of Mining and Geology; nguyengiatrong@humg.edu.vn    

2 Geodesy and Environment research group, Hanoi University of Mining and Geology 

3 Vietnam Meteorological and Hydrological Administration; nguyenxuanhien79@gmail.com

4 Hanoi University of Natural Resources and Environment; vdmanh@hunre.edu.vn

5 Viet Nam’s people naval hydrographic and oceanogtaphic department; vinhtduc@gmail.com

*Corresponding author: vinhtduc@gmail.com; Tel.: +84–984215679

Abstracts

Building or predicting the trajectory of drifting objects is significant in maritime studies and search and rescue operations. The trajectory of a drifting object can be determined using traditional tools with marine dynamic models or through artificial intelligence models. From the drifting buoy data collected between December 19 and December 28, 2003, the research team employed the CNN (Conv1D) model for analysis. The analysis results indicated that by using the Adam optimizer, the Huber loss function, and 256 filters in the hidden layer, the characteristic parameters for the model’s performance were determined as RMSE = 0.04004, MAE = 0.032304 degrees, and R² = 98%. When using the SGD optimizer and the mean squared error (MSE) loss function, both RMSE and MAE values decreased by up to four times compared to the previous case, while the R² value reached 99.9% with 64 filters in the hidden layer. When the number of filters in the hidden layer was increased to 128, the performance of the CNN (Conv1D) model improved by up to 20%, with RMSE = 0.007863deg and MAE = 0.006653deg. The R² value when predicting the trajectory of drifting buoys using the CNN (Conv1D) model with the SGD optimizer and the MSE loss function approached approximately 100%, indicating that this model is suitable for the input data in predicting the trajectory of drifting buoys. Increasing the number of filters in the model's hidden layer from 128 to 256 did not change the model's predictive performance, demonstrating that the optimal number of filters for this case is 128. However, the RMSE result achieved in this study is still relatively large (0.87 km), possibly due to the limited input data. Future work should continue to experiment with drifting buoy data analysis using a larger input dataset.

Keywords

Drifting buoy data; Artificial intelligence; Deep learning; Time series data.

Cite this paper

Trong, N.G.; Hien, N.X.; Manh, V.D.; Vinh, T.D. Spatiotemporal data analysis using deep learning models: A case study with drifting buoy data. J. Hydro-Meteorol. 2025, 22, 1–9.

References

1. Zeng, F.; Ou, H.; Wu, Q. Short-term drift prediction of multi-functional buoys in inland rivers based on deep learning. Sensors 2022, 22(14), 5120.

2. Ha, S.Y.; Yoon, H.S.; Kim, Y.T. A study on the prediction of the surface drifter trajectories in the Korean Strait. J. Korean Soc. Coastal Ocean Eng. 2022, 34(1), 11–18.

3. Spaulding, M.L.; Hewlett, E. Application of SARMAP to estimate probable search area for objects lost at sea. Mar. Technol. Soc. J. 1996, 30(2), 17–25.

4. Han, D.; Kim, Y.J.; Lee, S.; Lee, Y.; Kim, H.C. The estimation of arctic air temperature in summer based on Machine Learning approaches using IABP Buoy and AMSR2 satellite data. Korean J. Remote Sens. 2018, 34(6_2),1261–1272.

5. Song, M.; Hu, W.; Liu, S.; Chen, S.; Fu, X.; Zhang, J.; Li, W.; Xu, Y. Developing an artificial intelligence-based method for predicting the trajectory of surface drifting buoys using a hybrid multi-layer neural network model. J. Mar. Sci. Eng. 2024, 12(6), 958.

6. Aksamit, N.O.; Sapsis, T.P.; Haller, G. Machine-learning ocean dynamics from lagrangian drifter trajectories. arXiv preprint arXiv 2019, 12895.

7. Baidai, Y.; Dagorn, L.; Amande, M.J.; Gaertner, D.; Capello, M. Machine learning for characterizing tropical tuna aggregations under drifting fish aggregating devices (DFADs) from commercial echosounder buoys data. Fish. Res. 2020, 229, 105613.

8. Diamant, R. Prediction of water current using a swarm of submerged drifters. IEEE Sensors J. 2020, 20(19), 11598–11607.

9. Apel, H.; Hung, N.G.; Thoss, H.; Schöne, T. GPS buoys for stage monitoring of large rivers. J. Hydrol. 2012, 412, 182–192.

10. Häfner, D.; Gemmrich, J.; Jochum, M. FOWD: A free ocean wave dataset for data mining and machine learning. J. Atmos. Ocean Sci. 2021, 38(7), 1305–1322.

11. Pokhrel, P. Machine Learning in weakly nonlinear systems: A case study on Significant wave heights. arXiv preprint arXiv 2021, 08583.

12. Mozo, A.; Morón-López, J.; Vakaruk, S.; Pompa-Pernía, Á.G.; González-Prieto, Á.; Aguilar, J.A.P.; Gómez-Canaval, S.; Ortiz, J.M. Chlorophyll soft-sensor based on machine learning models for algal bloom predictions. Sci. Rep. 2022, 12(1), 13529.

13. Bhaganagar, K.; Kolar, P.; Faruqui, S.H.A.; Bhattacharjee, D.; Alaeddini, A.; Subbarao, K. A novel machine-learning framework with a moving platform for maritime drift calculations. Front. Mar. Sci. 2022, 9, 831501.

14. Chen, J.; Ashton, I.G.C.; Steele, E.C.C.; Pillai, A.C. A real-time spatiotemporal machine learning framework for the prediction of nearshore wave conditions. Artif. Intell. Earth Syst. 2023, 2(1), 220033.

15. Kim, J.C.; Yu, D.H.; Sim, J.E.; Son, Y.T.; Bang, K.Y.; Shin, S. Validation of opendrift-based drifter trajectory prediction technique for maritime search and rescue. J. Ocean. Eng. Technol. 2023, 37(4), 145–157.

16. Trinh, N.Q.; Huan, N.M.; Hieu, P.D.; Toan, D.V. Simulation of drifting object motion in the east sea using numerical methods. Proceeding of the 14th Asian Congress of Fluid Mechanics - 14ACFM, October 15-19, 2013, Hanoi and Halong, Vietnam, 2013, pp. 1144–1151.

17. Trinh, N.Q.; Cham, D.D.; Binh, H.T.; Thao, D.T.; Hanh, L.D.; Son, N.T.; Hung, L.T.; Giang, H.H.; Vinh, N.Q. Application of 2D models for the problem of plastic waste transmission to seasonal characteristics in the Da Nang–Quang Nam sea. J. Hydro-Meteorol. 2022, 743, 36–51.

18. Trong, N.G.; Quang, P.N.; Cuong, N.V.; Le, H.A.; Nguyen, H.L.; Tien Bui, D. Spatial prediction of fluvial flood in high-frequency tropical cyclone area using TensorFlow 1D-convolution neural networks and geospatial data. Remote Sens. 2023, 15(22), 5429.

19. Tinh, L.D.; Thao, D.T.P.; Bui, D.T.; Trong, N.G. Integrating Harris Hawks optimization and TensorFlow deep learning for flash flood susceptibility mapping using geospatial data. Earth Sci. Inf. 2024, 17, 3397–3412.

20. Phong, D.V.; Trong, N.G.; Chien, N.V.; Thanh, N.H.; Ha, L.L.; Quan, N.V.; Quang, P.N. Analysis of land vertical movement using ANN function from the results of processing GNSS time series data. J. Hydro-Meteorol. 2023, 752, 41–50.

21. Tinh, L.D.; Thao, D.T.P.; Thang, T.D.; Hop, D.T.; Trong, N.G. Evaluating the Performance of CNN (Conv1D) and CNN (Conv3D) Models in GNSS Data Analysis. J. Hydro-Meteorol. 2024, 771, 55–64.

22. NOAA. Online available: https://www.ndbc.noaa.gov/.

23. ESA. Copernicus marine service. Online available: https://marine.copernicus.eu/.

24. Lawhead, J. Learning geospatial analysis with Python: Understand GIS fundamentals and perform remote sensing data analysis using Python 3.7. Packt Publishing Ltd, 2019.

25. Campesato, O. Python 3 for machine learning. Mercury learning and information, 2020.

26. Chen, D.Y. Pandas for everyone: Python data analysis. Addison-Wesley Professional, 2017.

27. Raschka, S.; Mirjalili, V. Python machine learning: Machine learning and deep learning with Python, scikit-learn, and TensorFlow 2. Packt publishing Ltd, 2019.