Study On Assessment Model Of Agricultural Catastrophe Risk

Agricultural catastrophe risk assessment is an imperative need to resolve problem for agricultural steady development under the background of global climate change, meanwhile, it is a scientific problem urgently need to solve in agricultural risk assessment discipline. At present, the academic sector has mainly undertaken researches on theory and method of agricultural catastrophe risk assessment from different viewing angles based on risk factors, risk loss, and risk mechanisms, but problems of the loss degree of agricultural production caused by extreme climate events and the probability distribution of agricultural catastrophe loss did not be given to a definite answer. Aimed at these problems, this paper has constructed the basic framework of agricultural catastrophe risk assessment based on modern risk analysis and evaluation theory, applied the basic analysis means of quantile regression, Monte Carlo Simulation, peak over threshold, Value at risk, etc., and gained breakthroughs in the following two aspects: the effect of extreme climate events based on quantile regression model, and the probability distribution of agricultural catastrophe loss based on extreme value theory model. On this basis, this paper has unfolded the application research on agricultural catastrophe disaster prediction and agricultural catastrophe insurance rating.This paper has explored further in theory, methodology and application agricultural catastrophe risk assessment models.As in theoretical research, this paper has defined agricultural catastrophe as the event that have the characteristics of "low frequency and high loss" from the perspective of quantitative and qualitative perspectives. Agricultural catastrophe can be measured through conditional expectation, while the risk is a kind of future scenarios related to unfavorable events. On this basis, this paper has introduced the theoretical connotation of agricultural catastrophe risk assessment aiming to solve the difficult problems of crop production loss caused by extreme climate events and the probability distribution of agricultural catastrophe loss; furthermore, two kinds of paradigm of agricultural catastrophe risk assessment have been constructed in this paper.In terms of research methods, this paper has constructed assessment models of agricultural catastrophe risk from the perspective of quantile regression theory and extreme value theory.(1) The assessment model of agricultural catastrophe risk based on quantile regression theory aims at resolving the problem of what degree of losses extreme climate events accounting for the poor crop. It is consisted mainly of hazard model and vulnerability model of agricultural catastrophe risk. In the construction of hazard model of agricultural catastrophe risk, this paper has adopted parameter approach to fitting the probability distribution functions of agricultural catastrophe risk of hazard‐formative factors. Via calculating the probability of occurrence or recurrence interval of specific extreme metrological disaster events in the research region, certain time period, and this paper has realized the purpose of evaluating agricultural catastrophe risk. In the aspect of constructing vulnerability model of agricultural catastrophe risk, this paper holds the view that vulnerability of agricultural catastrophe risk is a result of resultant action of agricultural system external impact mechanism and internal stability mechanism. Viewing external impact mechanism, agricultural catastrophe risk of hazard‐formative factors will produce random shocks through system internal exposure and vulnerability transmission, and its result is demonstrated in hectares covered by agricultural disaster. Using quantile regression approach to analyze the effect of agricultural catastrophe risk of hazard‐formative factors (illumination, temperature and precipitation) on hectares covered by disaster, this paper has separated hectares covered by agricultural conventional disaster risk and by agricultural catastrophe risk, respectively. Hence we can describe the relation between agricultural catastrophe events and hectares covered by agricultural catastrophe located in the tail of distribution curve. Viewing the internal stability mechanism, through production function this paper has further described the relation between the yield and the hectares covered by agricultural disaster under the internal stability mechanism of agricultural natural resources endowment, agricultural production inputs, and agricultural anti‐disaster ability. This paper has taken drought catastrophe in Jilin province of China as the empirical case and assesses grain‐production risk using agricultural catastrophe risk assessment model based on quantile regression theory.(2) The assessment model of agricultural catastrophe risk based on extreme value theory aims at resolving the problem of what probability distribution agricultural catastrophe loss is submitted to. This paper has applied indirect indicators concerning hectares covered by natural disasters, hectares affected by natural disasters, and hectares destroyed by natural disasters to gain agricultural disaster loss data, while Monte‐Carlo simulation technology was used to expand sample spaces so as to solve the problems of greater agricultural catastrophe risk assessment errors caused by fewer data. Peak over threshold (POT) was used to fitting the tail distribution of agricultural disaster losses effectively, and this paper has gained the generalized Pareto distribution (GPD) of agricultural catastrophe losses, which effectively overcome the shortcoming of traditional statistical approaches in fitting catastrophe risk. On this basis, this paper has introduced the Value at risk (VaR) to realize precision measurement of agricultural catastrophe losses. This paper has cited drought catastrophe risk in Henan Province as the empirical case and assesses grain‐production risk using agricultural catastrophe risk assessment model based on extreme value theory.In terms of applied research, this paper has expanded and extended the agricultural catastrophe risk assessment model based on quantile regression theory and extreme value theory from the perspective of agricultural catastrophe disaster prediction and agricultural catastrophe insurance rating. In the aspect of agricultural catastrophe disaster prediction, this paper has added the prediction models of agricultural catastrophe risks affected by external shock and internal stability. Through applying random walk model and Monte‐Carlo simulation technique, this paper grasped the predictive value of external shocks factors of agricultural catastrophe risk, which have been categorized as "violent fluctuation type" variables. While this paper have taken internal stability factor as "smooth fluctuations type" variables, the prediction value of internal stability factors was realized by comprehensively application of Holt‐Winters no season model, ARIMA model, Grey GM (1,1) model, combined forecasting model. In the aspect of agricultural catastrophe insurance rating, under the precondition of certain security level, probability density function of agricultural catastrophe loss becomes the key of rating agricultural catastrophe insurance. Given the scenario of summer drought catastrophe in Jilin Province, the prediction result of grain production is very similar as the actual situation, while the results of rating pure premium of drought catastrophe insurance in the region of13major grain‐producing provinces (autonomous regions, municipalities) also coincide with the actual situation. It has fully demonstrated the rationality of agricultural catastrophe risk assessment model constructed by this paper, and its importance in agricultural production practice.