Analysis Of Regional Drought Characteristics Based On Information Diffusion Theory And Multiple Fractal Theory

Drought is one of the most common and most affected climate disaster in China, more over in recent years, the drought happened frequently, which cause serious influence on China’s agriculture. The climate general situation of China every year published on Meteorological Monthly refer that during the nature disasters happened in China the drought event never absent for every year, besides the affected population and crops to the affected area caused by drought are most serious, so it is likely to be the focus and core for the drought science researchers to predict and monitor the drought event effectually and assess the influence caused by drought quantificationally. Given this, the occurrence of drought are predicted in nonlinear probability with recurrence period of drought as main reference factor based on the theory of multifractal in the paper. What’s more, the semiquantitative assessment of influence caused by different levels of drought event can be got combining the agricultural losses caused by drought disaster with the different levels of drought intensity. By this, more science references of forecast monitoring and assessment of drought can be offered from different views, thus the influence of China’s agricultual economy caused by drought disaster can be reduced. As Northeast China Region is the most granary of China and recent years the Southwest China suffered from drought, this two regions are choosed to analyze the characteristics of drought in the paper. One of difficult problems of the disaster assessment is the sample size of data. For this reason, the information distribution and diffusion theory is used to diffuse the data of simple size in the paper, with which, although it cannot eliminate the influence of limited data radically, it can reduce the influence. The drought vulnerability characteristics from 1971 to 2012 in Northeast China is analysed and the risk in Northeast China is calculated based on the theory of information distribution and diffusion, combining with the Standardized Precipitation Index (SPI) and the drought damage index (DDI) in Northeast China, considering the meteorological factors and social factors.The probability density function (PDF) of DSI is estimated by using the method of information distribution. The vulnerability of the relationship between DSI and DDI is constructed by using the method of two dimensional normal information diffusion.Then the average risk can be obtained by calculating the sum (discrete distribution) or integral (continuous distribution) of the product of the PDF of DSI and the vulnerability curve.The results show that, the introduction of information distribution and diffusion method to analyze the problem of vulnerability for the situation of small sample could get the PDF of DSI smoother.What’s more, the vulnerability of the relationship between DSI and DDI through the DSI to get the DDI has clearer physical meaning.Therefor the results could be realistic. More significantly, similar results can be got with different simple length which means the method is insensitive to the simple length and it can overcome the instability of analysis of the small sample.The nonlinear characteristics of recurrence of drought in Southwest China was analyzed combined the multifractal detrended fluctuation analysis(MFDFA) and the Standardized precipitation index(SPI) calculated with the monthly precipitation data of 303 stations in southwest China during 1961-2012 reported by the China meteprplogical administration, based on the similar characteristics of the theory of multifractal. The result shows that the drought index in Southwest China belong to multifractal sets and are provided with character of long-term correlation and the property is uncorrelated with time scales; A power-law decay can be shown both in the probability density function and autocorrelation of return intervals of drought events, whch can indicate the fact that the return intervals of drought events exist memorability and the memory decline gradually; On the basis of further, a drought risk index based on the character of multifractal and long-term correlation was put forward. The index we mentioned not only reflects the probability distribution and character of memory of the return intevals of drought events but also can depict the drought risk in the future. In a word, the drought risk index we got has certain instruction significance on judging the drought happen or not in the future, what’s more provides the basis for preventing disasters and reducing damages.