| For the past few years,dramatic changes have taken place in the distribution of hydrological element series,drought in Henan has found an increasingly wide and became more frequent,the ecological environment,which the human survival relies on suffer from a serious threat.Therefore,in order to against the drought initiatively and scientifically,it seems that it is urge to evaluate the risk of drought,this paper take Henan as an example,base on the monthly data from 1951 to 2013 and monthly hydrological data from 1956 to 2010,this paper build the risk assessment index system in two ways,they are Meteorological drought and Hydrological drought,this risk assessment index system base on Wavelet Analysis,Entropy,Neural networks,Multivariate Copula function and other methods to conduct risk assessment about drought in Henan Province.The results are as follows:(1)The space-time rules of drought in Henan Province.In order to know the feature of the space-time rules of drought in Henan Province precisely,we analyzed it precisely with Mann-Kendall(M-K)trend analysis,Wavelet Analysis and their's drought characteristic map system:from the point of Meteorological drought,it shows a trend of shifting from north to south and from east to west,from the point of Hydrological drought,annual run off from 1956 to 1988 shows an increasing trend,but it has an decreasing trend from 1988 to 2010;as a whole,The arid tendency in Henan Province will more and more obvious in the future,both the meteorological and hydrological characteristics are show a periodic variation of2-4a,According to cross wavelet analysis,we know that Hydrological drought usually lag behind Meteorological drought for 1-2 months.(2)Drought prediction of Henan Province.For improving forecast accuracy,firstly,this paper selects the main factors that affecting drought,because it has error accumulation in edge and joint distribution between different variables,we make use of Copulas entropy to selecte the factors of drought,temperature,water vapor pressure and relative humidity are the factors that have great effect on drought among the many factors that affecting drought,BP neural network and LSTM neural network are used by us to predict,In the future,the drought situation in Henan Province will be center on light drought and medium drought,of which the possibility of occurrence of light drought in the future in Henan province is higher,but there is a tendency of severe drought.(3)The risk assessment of drought in Henan Province.in this test,we extract three characteristic factors of drought,they are drought duration,drought severity and drought peak intensity.we extract them base on run-length theory in order to reduce the information loss of the single variable's marginal distribution in estimates,implement precise estimate of drought risk in Henan Province,building a high-dimensional Copula function risk assessment model which is based on drought characteristic variables.As a result:the probability of light drought is higher in Henan,the period of time is 4 months,The probability of severe drought happening is low in Henan,the cycle of severe drought occurrence is 2-4 years;For the three-dimensional Copula function risk assessment model,the Gumbel-Hougaard Copula function has the best imitative effect on random variables,When the probability of any two together increase to a certain degree of drought duration,drought severity and drought peak intensity,as the density of drought increase,severe droughts occur.Through the research and analysis of the space-time rules of drought,drought prediction,risk assessment of drought in Henan,this paper is set multi-angle,multi-level,multi-dimensional research and analysis methods in one,the drought condition in Henan province in recent decades is disscussed comprehensively and systematically,the drought development mechanism and risk assessment results of this test can help government departments of Henan,they can put forward the measures to reduce drought based on this paper rusult. |
PDF Full Text Request