The Research And Application Of The Abrupt Detecting Methods In Dynamical Structures

This paper presents three new methods which can be used to detect the abrupt changes in dynamical structures.These methods are mainly based on the power law and detrended fluctuation analysis method.We have proved the reliability of these methods through some ideal tests and studied the influences of all kinds of trends existed in the data and nonstationarity in the observation data because of the observation errors and observation scarcity.We use the DFA to analyze the long-term correlations of the climate factors and apply the MC-DFA to analyze the abrupt detecting of the air pressure and temperature through which we find out the climates abrupt changes property and space time distributing characters which happened in 1960s and from the end of 1970s to the early of 1980s,at the same time,we study the abnormity of the solar activity.In the last,we do the scale analysis to the climatic proxy and prove the feasibility of these methods in the application of abrupt changes detecting in the dynamical structures.The main results are just as follows:(1)The ideal tests show that the three kind of dynamical detecting methods all can detect the abrupt changes exactly and the results don't depend on the lengths of subseries, what's more,these methods have a strong ability to prevent against the influence of noise.We also find MDFA and SA-DFA are suitable for those situations when the scale exponent is lager than 1.5;however,MC-DFA is suitable for those time series which own a correlation.(2)The periodicity trend of the time series cause a big influence to the detecting results when we use the three new methods to detect abrupt changes,but iinearity trend and multinomial trend don't influence the detecting results.(3)Approximate entropy can't find out the abrupt changes exactly,however,it can only reflect the abrupt changes' information in some extend which existed in the time series and show a probable abrupt time instance,but the instance is related with the length of time series and has a large deviation with the true abrupt changes.Many traditional detecting methods such as moving t-test,Cramer method,Yamamoto method and Mann-Kendall method can't vividly identify the abrupt changes in the dynamical structures.(4)The records of temperature,air pressure,average wind speed and precipitation all exhibit a long persistence and this persistence has an obvious area distributing.Averagely, temperature and average wind speed have a strong persistence while air pressure and precipitation have a weak persistence.(5)Air pressure has an obvious abrupt change in the dynamical structures during the middle-late 1960s and in the end of 1970s.The abrupt change time of Southern China is 1968 which is later than 1964 of Northern China,but the abrupt change time in the most Western China is later than Eastern China except the northern Xinjiang(1965),what's more,there is no abrupt changes of the air pressure in the southern Xinjiang in the 1960s.In the 1970s,the air pressure in many areas has an abrupt change except the northern areas of northeast,Sichuan basin and its neighboring areas and the most areas of the Qingzang altiplano.In Eastern China, the abrupt change happened in 1976 while in Western China,the time is one year later,the latest time is in 1979 happened in Hetao area.In the early 1980s,there are only a few stations which detect the abrupt changes of the air pressure in the dynamical structures.(6)We find the temperature had an abrupt change in the most areas of china in the 1960s and the late 1970s to the early 1980s,but there are still some areas don't have abrupt changes such as in 1960s,the west north in northeast and the most areas in the south of the Yangtse Rive according to the lowest temperature detecting.The results of the highest temperature detecting are similar with the results of the lowest temperature detecting;the biggest difference between them is that there are abrupt changes in the west south area of the south of the Yangtse Rive and the middle north according to the highest temperature detecting. In the late 1970s to the early 1980s,the highest daily temperature detecting shows there are no abrupt changes in the most areas of the east north,but according the results of the lowest daily temperature detecting,only in the west north of the east north can find the abrupt changes.The another obvious difference between the highest daily temperature detecting and the lowest daily temperature detecting is there is abrupt change in the east north area of the Qingzang altiplano by the highest daily temperature detecting,conversely,there is no abrupt changes by the lowest daily temperature detecting.In the early 2000s,there are obvious abrupt changes in the most north areas.The highest daily temperature detecting show there are abrupt changes in the most area of the southeast,but the lowest daily temperature detecting don't fred any obviously abrupt changes.(7)The evolvement of the sunspots has a scaling character and a long term correlation in some certain time scale.That means the solar activity can last for a long time,if solar activity is strong in this year and in the next year,the solar activity is always strong too.The solar activities had abrupt changes in the 1946 to 1961 and in the end of 1970s to the early of 1980s and the climate changes happened in the 1960s and in the end of 1970s to the early of 1980s.So,we can see that the climate changes have a close relationship with the solar activity; however,there is no relationship between the solar activity and the abrupt changes happened in the northeast and the southeast in the early 21th century.We will make a further research to investigate the cause.The change of the earth's rotation speed is similar with the Brownian motion.Researches show the earth's rotation speed had an abrupt change in 1901,1957 and 1980,what's more,1957 and 1980 are very close to the abrupt changes time of the solar activity.Thus,the change of the earth's rotation speed may have a close relation with the abrupt change of the solar activity.(8)A lot of scale analyses of climatic proxy all show a strong long term correlation such as the tree annual ring of the Northern Hemisphere,the pine's annual ring width of the mountain Hua,the cypress's annual ring width of the Qilian,the stalagmite of the Beijing shihua hole.There are many climate change information in these dates and they show the long term persistence of climate change which may last for more than 200 years,however,the Guliya ire don't have a correlation.It's very difficult to analyze the abrupt change in dynamical structures of the substitutive dates with the MC-DFA because the uncertainty of these dates destroys the true evolvement information.