Multimodal Spafiotemporal Evolution Characteristics And Mechanisms Of Natural Hazards In The Middle And Upper Reaches Of The Yellow River

Through data cleaning,calibrating,grading,and rasterization,a spatio-temporal big database of eight natural disasters in the middle and upper reaches of the Yellow River in the past 2000a was established for drought flood,hail,freeze,wind,earthquake,insect,and epidemic.Based on multimodal theory,utilizing spatio-temporal big data detection technology,correlation analysis,moving correlation analysis,spatial bias correlation analysis and GeoDetector to analyze multimodal characteristics and mechanisms of natural disasters,such as multi-types,multi-scale,multi-associations,multi-dimensional dynamics,and multi-reference systems,with important theoretical and practical significance for the construction of modem disaster prevention and mitigation system,ecological protection and high-quality socio-economic development.The research results revealed that:(1)The multi-type characteristics of natural disastersA total of 7229 natural disasters occurred in the study area in the past 2000a,among which the number of droughts,floods,hailstorms,insects,frosts,wind and sand,plagues and earthquakes decreased in order.However,after the 17th century,the proportion of droughts decreased significantly and the proportion of floods and hailstorms increased significantly.The frequency of extreme natural disasters was highest in the 20th century,followed by the 17th century.The proportion of various natural disasters was close and the frequency was the highest in the 16th and 20th centuries,which was the period of various natural disaster clusters.Summer was the period of frequent natural disasters,with high frequency of droughts,floods,hailstorms,pests and plagues in summer,and high frequency of frosts and sandstorms in spring.(2)The multi-scale and multi-association characteristics of natural disastersThe spatial kernel density values of natural hazards were highest in the Guanzhong Plain.The spatial kernel density values of drought,flood,hail,frost,pest and plague hazards were the highest in the Guanzhong Plain,while those of hail and frost hazards were the highest in Guyuan City,and those of frost,sand and earthquake hazards were the highest in northern Ningxia.The interdecadal spatial distribution of natural hazards showed a trend of contraction-expansion-recontraction,and the center of gravity of hazards moved in the direction of south-southeast-east-northwest The spatial center of gravity of drought hazards moved in the following directions:south-northwest-east-northwest,and the spatial center of gravity of flood hazards moved in the following directions:south-northwest-eastsouthwest.Drought hazards showed spatial correlation in Xianyang,Guyuan,Qingyang and Haidong,and flood hazards showed spatial correlation in Weinan,Guyuan,Pingliang and Xining.(3)The multi-dimensional dynamic characteristics of natural disastersNatural hazards in the study area over the past 2000a showed a significant increasing trend in the spatio-temporal dimension,with natural hazard hotspots first appearing in the Guanzhong plain,and then graduall spreading westward and then northward with time.The spatio-temporal hot spot pattern was concentrated in Guanzhong,Longdong,Longnan,Longzhong and northern Ningxia,and the cold spot pattern was concentrated around the 400 mm isohyet in northern Shaanxi,central Ningxia and Haidong of Qinghai.In the spatio-temporal pattern,all of the eight natural hazards showed oscillatory distribution;except for plague,other hazards showed a significant increasing trend;the spatio-temporal correlation of each hazard was the strongest in Guanzhong region in the last 200a.Additionally,the frequency of natural disasters in the central Guanzhong Plain and northern Ningxia will increase in the next 50 years.(4)Mechanisms of natural disaster in multiple reference systemsNatural disasters had a good correspondence with natural social factors in the past 2000a.In recent times,topography,climate,and vegetation factors were the dominant factors of natural disaster intensity in the whole study area,while natural disaster intensity in Ningxia,Longdong Loess Plateau,and Shaanbei Loess Plateau was dominated by GDP,and natural disaster intensity in Guanzhong Plain,Longzhong Loess Plateau,and Qinghai Haidong area was the result of the combined effect of GDP,temperature,and precipitation.