Analysis Of Dry And Wet Temporal-spatial Characteristics In Jiangsu Zhejiang Anhui And Shanghai Region During Ming And Qing Dynasties

Jiangsu,Zhejiang,Anhui and Shanghai are located in the Middle East of China,near the East China sea,and are the hotspots of droughts and floods since the ancient times.Experts and scholars from various disciplines accumulated the rich data of flood and drought disasters through data collection and data analysis.In this paper,taking this basis,modern climate analysis,data mining methods for space-time analysis of flood and drought disasters are introduced and temporal-spatial data mining are used for studying the drought and flood disasters,which aim at discovering the space-time characteristics hidden in the historical data of drought and flood.This paper expounds the least-square method,the moving average method,the cumulative departure method,sliding t-test,the wavelet analysis method and the method of DBSCAN,and does temporal-spatial data mining research on drought and flood disasters with these methods,which aim to find the space-time characteristics of drought and flood disasters,to help people understand the situation of drought and flood disaster during the Ming and Qing dynasties,and provide the basis for disaster prevention and mitigation.(1)”A compendium of Chinese meteorological records of the last 3,000 years" was put into the attribute database.Making use of "typical dictionary",the drought?flood and normal years records were extracted and the drought and flood grade series was obtained.Years of waterlogging were significantly more than drought years.(2)Analyse the wet and dry years and the frequency with the statistical method,flooding often occurred in 1551-1600 years,and drought often occurred in 1501-1550.Use the least squares method to fitting analyze the drought and flood disasters,the anomaly of waterlog was negative before the year 1550,after that was the positive;drought was positive before the year 1700,after that was the negative;use the moving average method and the cumulative departure method to analyze the stage of droughts and floods,the happening of drought and flood disasters can be divided into seven main stages;use t-test for mutation detection of the dry and wet disasters,and 12 obvious jump points or jump range exist;use the wavelet analysis method to analyze the periodicity of droughts and floods,find that drought and flood disasters have a quasi-periodicity of 2,8-11,18,30,50,75,100 and 150 years.(3)DBSCAN clustering method is used for clustering mining of the drought and flood level value of nine cold and warm period of the sites,it can be seen from the clustering results in different period that the spatial distribution of drought and flood disasters change over time.Drought,partial drought disaster,partial waterlogging disasters and floods in 1470-1529,1530-1569,1700-1809 and 1810-1889 respectively were the largest proportion.