HURST STATISTICS (R/S-ANALYSIS) INTHESTUDY OF CLIMATIC VARIABLES

Introduction. An R/S analysis of the persistence of trends in climatic variables was carried out in the article using the normalized range method, which is one of the nonparametric approaches for studying series that do not satisfy all the conditions of standard Gaussian statistics. To study the stability of a system behaving not as a random variable, but passing a longer path (shifted Brownian motion with the presence of a trend), the Hurst indicator Н was used. Materials and methods of research. The trend stability (persistence) of air temperature changes was assessed using the normalized range method (R/S analysis). The method for determining the Hurst indicator H is based on in order to analyze the range of the parameter (the largest and the smallest values on the period of the segment) and the standard deviation and its dependence on the period of the studied time T, more or less than the current one. The use of long-term data of average, maximum and minimum surface air temperature of 20 meteorological stations of different climatic southern Russia (according to the state observational network of Roshydromet of the North Caucasian Directorate of the Hydrometeorological Service) is used. Results of the study and their discussion. When analyzing the climate, the initial data are time series containing the values of certain climatic indicators (temperature, precipitation, humidity, etc.) for a certain period. Traditionally, trends are used to analyze data for a number of climatic parameters. This solves the problem of predicting future values of the series. At the same time, the trend does not say anything about how stable the series is. Thus, classical methods of analysis are not very informative and have many methodological limitations for their application. The paper presents the results of time series analysis using the method of the normalized range R/S. It was found that the indicators of stability H characterize the stability and long-term changes in the time series of annual and summer average temperatures (H = 0.80), as well as autumn average temperatures (H = 0.73). Series of annual, summer (H = 0.75) and autumn maximum temperatures (H = 0.70), as well as spring minimum temperatures (H = 0.72) also had stable trends. Conclusions. The results of the R/S-analysis showed that the temperature series is not an ideal Poisson process (without memory), on the contrary, there is some long-term correlation between the last events and the initial ones. Change in climatic variables. as a phenomenon, it bears the dual characteristics of randomness and regularity, and the more the Hurst indicator H deviates from 0.5, the more regularity appears in the time series, and vice versa.

R/S-анализ

1. Исаев, А. А. Статистика в метеорологии и климатологии: учебник / А. А. Исаев. Москва: Издательство МГУ, 1988. 244 с.

2. Вероятностные разделы математики / под ред. Ю. Д. Максимова. СПб. : Иван Фёдоров, 2001. 592 с.

3. Hurst, H. Long Term Storage Capacity of Reservoirs / Hurst, H.; Transactions of the American Society of Civil Engineers, 1951. Vol. 116. P. 770-799.

4. Мандельброт, Б. Фрактальная геометрия природы: учеб. пособие / Б. Мандельброт; Институт компьютерных исследований. М., 2002. 666 с.

5. Rubalcaba, J. J. O., Fractal Analysis of Climatic Data: Annual Precipitation Records in Spain /j. J. O. Rubalcaba // Theoretical and Applied Climatology, 1997. Vol. 56. P. 83-87.

6. Carvalho, L. M.V. Antipersistence in the global temperature anomaly field / L. M.V. Carvalho, A. A. Tsonis, C. Jones, H. R. Rocha, P. S. Polito // Nonlinear Processes in Geophysics, 2007 Vol. 14. P. 723-733.

7. Солнцев, Л. А. Фрактальный анализ векового хода средней температуры воздуха в г. Нижнем Новгороде / Л. А. Солнцев, Д. И. Иудин, М. С. Снегирева // Вестник Нижегородского университета им Н И Лобачевского Нижний Новгород, 2007. № 4. С. 88-91.

8. Калуш Ю.А., Логинов В.М. Показатель Херста и его скрытые свойства // Сибирский журнал индустриальной математики - 2002. Т. 5. Вып. 4. С. 29-37.

9. Кроновер, Р. М. Фракталы и хаос в динамических системах. Основы теории: учеб. пособие / Р. М. Кроновер. Москва: Постмаркет, 2000. 352 с.

10. Федер Е. Фракталы / пер. с англ. М.: Мир, 1991. 254 с.