Grey Modeling Method For Regional Agricultural Drought Disaster Prediction And Assessment

Drought disaster is a worldwide meteorological natural disaster,which has caused great harm to agricultural production and human society.Drought disaster is affected by both natural environment and human social environment.It has the characteristics of multiple factors and complex correlation between factors.At present,there is a lack of clear understanding of the formation mechanism of drought disaster.Therefore,the formation mechanism of agricultural drought disaster,the prediction and assessment of drought disaster have always been the focus.Based on the viewpoint of system science,this paper takes the regional agricultural drought disaster system as the research object,and the grey system theory and method is used to study the grey modeling methods of quantitative identification of influencing factors,evolution trend prediction and disaster assessment by analyzing its system structure,constituent elements,operation behavior and interaction relationship between elements.The main research contents and results are as follows:(1)The formation process and mechanism of agricultural drought are deeply analyzed,and the basic model of regional agricultural drought research is proposed.By analyzing the differences and relations between the basic concepts of agricultural drought,agricultural drought disaster and conditions of agricultural disaster,the connotation and extension of the concept of agricultural drought disaster are further clarified.By analyzing the interaction among disaster pregnant environment,disaster causing factors,disaster bearing body and the condition of a disaster,the formation mechanism of agricultural drought disaster is revealed,and it is pointed out that the agricultural drought disaster system has the characteristics of high dimension,nonlinearity,complex correlation and uncertainty,and a basic research model of regional agricultural drought prediction and evaluation based on the regional agricultural drought disaster system is proposed.The grey information processing technology and data sequence processing method in grey system theory are studied,which provides a basic method for factor identification,evolution trend prediction and evaluation of regional agricultural drought disaster.(2)With the problem of identifying the influencing factors of regional agricultural drought system,a multi-dimensional quantitative identification method based on grey correlation theory is proposed.According to the characteristics of regional agricultural drought system,the grey correlation model is proposed to identify the complex correlation relationship of drought influencing factors,the high-dimensional characteristics of drought data are represented by multi-dimensional panel data,and a multi-dimensional grey correlation model based on panel data is constructed.By comparing the absolute gap,and dynamic gap between index matrices,the key influencing factors are identified from the static and dynamic aspects;With the problem of the relationship between univariate and multivariate in the system,a multivariate grey correlation model is constructed.The dimension of the index of multivariate sequence is reduced through the spatial mapping relationship,the correlation analysis of univariate sequence of multivariate sequence is realized,and the contribution rate function is designed to select the influencing factors of the system quantitatively.(3)With prediction of the evolution trend of regional agricultural drought disaster,a grey prediction method based on dynamic analysis of sequence characteristics is proposed.By decomposing the drought characteristic sequence,it is found that it usually has the fluctuation characteristics of trend,periodicity and randomness.A method of constructing the prediction model according to the sequence characteristics and the principle of grey modeling is proposed.Aiming at the multi period characteristics in the time series,the multi period characteristics are fitted by Fourier series,and combined with the grey prediction method,a grey prediction model based on multi period time series is constructed;According to the sequence characteristics of both seasonal periodic fluctuation and annual random fluctuation,a seasonal factor with annual effect and interval grey number seasonal factor are designed.The structure of the grey prediction model is changed by the seasonal factor,and A prediction model based on multiple fluctuation characteristic is constructed.By analyzing the time-delay effect of multivariable in regional agricultural drought system,an improved multivariable prediction model is constructed by using time-delay parameters and polynomial to modify the structure of model,which solves the problem of multivariable prediction in agricultural drought disaster system.(4)With the problem of regional agricultural drought disaster assessment,a multi-dimensional grey assessment method based on system characteristics is proposed.The evaluation of regional agricultural drought disaster is realized through the comprehensive evaluation of drought disaster system.According to the high-dimensional and uncertain characteristics of drought disaster system,panel data is proposed to represent the temporal and spatial characteristics of the evaluation object,and interval grey number is proposed to represent the uncertainty of the evaluation object and evaluation standard.Combined with grey correlation and cluster evaluation theory,a multi-dimensional grey evaluation method based on the characteristics of regional agricultural drought system is proposed,which solves the evaluation problem of regional agricultural drought disaster.With the problems of influencing factor identification,characteristic sequence prediction and disaster evaluation of regional agricultural drought,grey system theory as the main line,a multi-dimensional grey correlation model for key factor identification,a prediction model based on sequence characteristics and a grey interval whitening weight clustering evaluation model based on Panel Data are constructed.The models are applied to the empirical analysis of disaster prediction and evaluation in various regions of Henan Province.The results show that: the agricultural drought in Henan has obvious regional specific characteristics,and the risk of drought increases from south to north.The exposed areas are lower in the western region,the highest in the southeastern region,the higher sensitivity in the southwest,the lower in other areas,the higher drought resistance ability in the northern part,and the lower drought resistance ability in other regions.On the whole,the disaster risk of rain fed agriculture in Henan Province is high,and the irrigation agriculture is relatively low.Finally,according to the evaluation results of agricultural drought disaster in Henan Province,the policy suggestions for drought prevention and control are put forward.