|
Kozhevnikova I.A., Shveykina V.I. - Analysis of Non-linear and Linear Models of Fluctuations of Caspian Sea Level |
|
|
|
|
Written by Erkin Turdibaev
|
|
Thursday, 10 August 2017 |
|
The article is devoted to the simulation of fluctuations of the Caspian Sea level. At the beginning of the article there is given a substantiation of the use of nonlinear methods which relies on the thermal physical achievements of the modern science. The main conclusion of this part is that for the Caspian sea not only the external unpredictability created by climatic changes is characteristic but also internal due to the unstable nonlinear dynamics of the water balance. Then follows a nonlinear stochastic model of sea level fluctuations based on the nonlinear autonomous differential equation. A thorough analysis of the residual sequence was performed using the methods of spectral analysis resulted in determination and identification of polyharmonic components. This resulted in the stochastic model which adequacy to the initial data is proved by means of modern statistical methods. The simulated trajectories obtained using the constructed model move from one stable level to another without going beyond the maximum and minimum values of the initial series. Further in the article there is given an example of fluctuations simulation of the Caspian sea level using a linear model based on auto regression of the 1st order. The simulated in such a way trajectories gradually in the course of time can take any large values and have only one stable level equal to the initial value. The forecast on the basis of such models will give exactly the value with which modeling starts. Such models can be used to predict only for 1 year ahead. The main conclusion of the article is as follows. For more long-term forecasts of the Caspian sea level it is necessary to use nonlinear models such as those presented in the article.
|
