Research On Regional Agricultural Drought Risk Assessment And Temporal And Spatial Variation

As a universal natural disaster,drought hazard had a profound impact on agriculture,water resources,environment and human society,and drought hazard had become one of the bottleneck factors which restricted the sustainable development of economy.Agriculture was the biggest industry affected by drought hazard,and regional food security was threatened by drought seriously.This paper takes Heilongjiang Province as an example,which is the largest cultivated land area and the highest grain yield area in China.Based on the concepts of drought,vulnerability and drought hazard,Regional agricultural drought,drought vulnerability and drought risk were evaluated and spatio-temporal analyzed by improved SPEI drought index,evaluation method of agricultural drought vulnerability under the perspective of grain production and improved projection pursuit model.The main research results were concluded:(1)The theory of agricultural drought risk has been deeply analyzed,and the definition and connotation of agricultural drought,agricultural drought vulnerability and agricultural drought risk are analyzed.?Agricultural drought is related to soil water storage and crop physiological factors,as precipitation decreases or evaporation increases,water deficit continued below the critical water required for the growth of crops,and then threaten the growth and development of crops.?Agricultural drought hazard vulnerability is a condition of disaster bearing body,specifically defined as the extent to which agricultural disaster bearing bodies are lost when threatened by drought,includeing exposure,sensitivity and adaptive capacity.Exposure reflects the magnitude to which disaster bearing bodies are affected by disaster factors,which can be represented as the quantity,value and distribution.Sensitivity reflects the extent to which agricultural production is affected by the hazard factor.The adaptive capacity is the ability to overcome the effects of disaster factors and reduce potential losses,and it is related to social and economic factors,scientific and technological factors and drought resistance.? The drought hazard risk process involves the transformation of the vulnerability of a disaster bearing body into a drought disaster due to various hazard factors,and which is a comprehensive of agricultural drought and agricultural drought hazard vulnerability.(2)Based on the idea of SPEI,the Standardized Precipitation Crop Evapotranspiration Index(SPCEI)was constructed,and SPCEI was applied to the evaluation of agricultural drought and its temporal and spatial variability in Qigihar of Heilongjiang Province.?The advantages and limitations of SPEI are analyzed.The advantage is that SPEI takes into account the impact of evapotranspiration on drought and SPEI has a variety of time scales.The limitation is that the effects of different crop cover and underlying surface water holding capacity on evapotranspiration are not considered.? Considering the limitations of SPEI,the effects of crop cover and underlying surface water holding capacity on evapotranspiration were considered,A Standardized Precipitation Crop Evapotranspiration Index(SPCEI)was constructed.And from the point of view of index construction,the differences between SPEI and SPCEI are analyzed,and the applicable conditions of SPCEI are explained.? SPCEI was applied to the assessment of agricultural drought and temporal and spatial variability in the Qigihar which is the semi-arid region of I leilongjiang Province.Taking the city Keshan Station,Tsitsihar Station and TaiLai Station as the research point,taking maize and soybean as an example,four growth cycles of two crops were taken as the research object.The temporal and spatial variation characteristics of agricultural drought in different crop angles and growth cycles in Qigihar were analyzed.In the analysis process,firstly,the fitted distribution function is selected according to the actual data,and the similarities and differences between SPCEI and SPEI are analyzed.Secondly,the temporal and spatial variation characteristics of drought were analyzed.The results show that the trend is not obvious from the time change,similar to the result of SPEI.From a spatial point of view,the northern region is relatively humid,the central and southern parts are relatively dry,and the drought in the South has further intensified.(2)The evaluation method of agricultural drought vulnerability was constructed from the perspective of grain yield,and it was applied to the study of agricultural drought vulnerability assessment and temporal and spatial variability.?According to the theory of regional disaster system,the sensitive yield were separated from the grain yield using a regression method,and the sensitive yield represent the sensitivity.Additionally,the trend component of the grain yield represents the adaptive capacity,and the crop planting area represents exposure.Based on these definitions,a model of agricultural drought vulnerability is constructed.?Further,the concept of unit drought vulnerability and regional drought vulnerability is proposed,this method is suitable for areas where the grain yield exhibits a certain tendency and is sensitive to climatic factors.?Taking the semi-arid areas of Qigihar,Daqing,Heihe and Suihua in Heilongjiang Province as an example,the spatial distribution characteristics and temporal variation characteristics of regional agricultural drought vulnerability are analyzed.The results show that,from the time point of view,the unit vulnerability shows a trend of decline first and then rise.From a spatial point of view,the unit drought vulnerability of Qigihar and Daqing is higher,and the regional drought vulnerability in Qigihar and Suihua is higher.The main reasons for these differences are the improvement of adaptability and the difference of exposure.(4)based on Friedman-Tukey projection pursuit model,a projection pursuit model based on information entropy and a projection pursuit model based on cluster analysis are constructed.The projection pursuit model based on cluster analysis was applied to the study of agricultural drought risk assessment and temporal and spatial variability in 13 prefecture level cities(regions)of Heilongjiang province.?Projection pursuit model based on information entropy:The concept of the overall dispersion and local aggregation of projection points was proposed by Friedman and Tukey,and improvements to the projection pursuit model are proposed here by measuring discrete projection points according to information entropy.Information entropy was shown to provide improved measurements in the data discreteness relative to standard deviations,and the cutoff radius was defined between 0 and ln2,thus allowing the use of the exhaustion method to determine the cutoff radius.?Projection pursuit model based on cluster analysis:The shortcomings of the method for determining the cutoff radius in the projection pursuit model are analyzed.The standard deviation Sz is used to measure the overall scatter of projection points.Cluster analysis method was used to construct local density Dz,The high dimensional samples are classified by k-means cluster,The clustering results are represented by clustering matrix,and local density Dz is constructed based on clustering matrix.This method avoided the cutoff radius,therefore,the problem that the cutoff radius is not easy to be determined is solved.An improved projection pursuit model is proposed by the combination of cluster analysis and projection pursuit model,and the advantage of classification in clustering analysis and the advantage of assessment in projection pursuit are used in the improved model.The improved model enriches the theory of projection pursuit and has certain theoretical significance.?The two models are compared with the original model by the iris data.The results show that the two models constructed in this paper are slightly better than the original model.?In the case study,the projection pursuit model based on cluster is applied to the evaluation of agricultural drought risk in Heilongjiang Province,and the temporal variation and spatial difference of agricultural drought risk in Heilongjiang province are analyzed.The spatial distribution results indicated that Suihua,Qigihar,Daqing,Harbin,Jiamusi are located in high-risk areas,Daxing'anling and Yichun are located in low-risk areas,and the differences among the regions were primarily caused by the aspects exposure and resistance capacity.The temporal variation results indicated that the risk of agricultural drought in most areas presented an initially increasing and then decreasing trend.A higher value for the exposure aspect increased the risk of drought,whereas higher value for the resistance capacity aspect reduced the risk of drought.Over the long term,the exposure level of the region presented limited increases,whereas the resistance capacity presented considerable increases.Therefore,the risk of agricultural drought in Heilongjiang Province will continue to exhibit a decreasing trend.