Novosibirsk State University Journal of Information Technologies|
ISSN 2410-0420 (Online), ISSN 1818-7900 (Print)
Forecasting of dynamic processes on the earth surface based on data assimilation
Artem Sergeyevich Feoktistov, Evgeny Semenovich Nezhevenko
A method of modeling a dynamic process on the Earth surface, such as wildfires, floods, mudslides, oil spills, in the conditions of a priori uncertainty is proposed. This method is based on the data assimilation approach. Data assimilation is realized using recurrent neural networks and Kalman filtering. The training process of the neural network is described and also we propose a method for accelerating neural network training through the using of a Kalman filter. The efficiency of its application is analyzed.
computer simulation, wildfire, recurrent neural network, data assimilation, Kalman filter, 519.6
How to cite:
Feoktistov A. S., Nezhevenko E. S. Forecasting of dynamic processes on the earth surface based on data assimilation // Vestnik NSU Series: Information Technologies. - 2015. - Volume 13, Issue No 2. - P. 226-231. - ISSN 1818-7900. (in Russian).
Full Text in Russian
Available in PDF
1. Kozik V. I., Nezhevenko E. S., Feoktistov A. S. Adaptive prediction of dynamic processes on the surface of the earth using a recurrent neural networks. International Conference «Mathematical
and Informational Technologies, MIT-2011», 2012, p. 226–231.
2. Kozik V. I., Nezhevenko E. S., Feoktistov A. S. Neural network method for predicting the behavior of dynamic processes on the surface of the Earth. Abstracts XIV Russian conference with
international participation «Distributed information and computational resources» (DICR-2012) (Novosibirsk, Russia, 26–30 November 2012). Novosibirsk, 2012.
3. Kozik V. I., Nezhevenko E. S., Feokitstov A. S. Adaptive prediction of forest fire behavior on the basis of recurrent neural networks. Optoelectronics, Instrumentation and Data Processing, May
2013, vol. 49, iss. 3, p. 250–259.
4. Ipatova V. M., Shutyaev V. P. Algorithms and data assimilation problem for models of the dynamics of the atmosphere and ocean. Dolgoprudny, 2013
5. Penenko V. V. Variational data assimilation in real time. Computational Technologies, 2005, vol. 10, part 1.
6. Haykin S. Kalman filtering and neural networks. Awiley, Intrescience publication, 2001.
7. Dorrer G. A. The dynamics of forest fires. Novosibirsk, SB RAS, 2008.
8. Haykin S. Neural networks. N. Y., Prentice Hall, 1999.
9. Kozik V. I., Nezhevenko E. S., Feokitstov A. S. Studying the Method of Adaptive Prediction of Forest Fire Evolution on the Basis of Recurrent Neural Networks. Optoelectronics, Instrumentation and Data Processing, 2014, vol. 50, iss. 4, p. 1–7.
10. Dottori F., Todini E. A 2d flood inundation model based on cellular automata approach. XVIII International Conference on Water Resources CMWR. J. Carrera (ed.). Barcelona, 2010.
11. Korotenko K. A., Mamedov R. M. Modeling the propagation of oil spills in the coastal zone of the Caspian Sea. Oceanology, 2001, vol. 41 (6), p. 34–45.
Main title Vestnik NSU Series: Information Technologies, Volume 13, Issue No 2 (2015).
Parallel title: Novosibirsk State University Journal of Information Technologies Volume 13, Issue No 2 (2015).
Key title: Vestnik Novosibirskogo gosudarstvennogo universiteta. Seriâ: Informacionnye tehnologii
Abbreviated key title: Vestn. Novosib. Gos. Univ., Ser.: Inf. Tehnol.
Variant title: Vestnik NGU. Seriâ: Informacionnye tehnologii
Year of Publication: 2015
ISSN: 1818-7900 (Print), ISSN 2410-0420 (Online)
Publisher: Novosibirsk State University Press
© 2006-2017, Novosibirsk State University.