Research Of Spatio-temporal Law Mining And Knowledge Expression Of Drought Propagation Of Yangtze River Basin

Drought disaster is a universal global natural disaster.It has aroused widespread concern.At present,the single and multivariate drought index based on sites or remote sensing data are widely used in drought monitoring and early warning system at domestic and overseas.It realizes the quantitative monitoring of drought state.However,there is a propagation process of drought from breeding to disaster.While the existing research lacks quantitative analysis of drought propagation process,expecially the full propagation from meteorological drought to hydrological drought.With the deepening of drought propagation research,a large amount of drought propagation information has been presented.How to associate and integrate the messy information,and to explore the laws of knowledge hidden behind the information has also become one of the current research hotspots.In this paper,Yangtze River Basin was selected as the experimental area.The drought propagation frome meteorological drought to hydrological drought,and from hydrological drought to soil moisture drought were studied.And the cross wavelet analysis method,k-means method,run theory,and regression analysis method were used to analyze the drought propagation.Meanwhile,the periodic characteristics,duration characteristics,magnitude characteristics and other structural data of drought propagation were extracted.What's more,KNN algorithm,thulac and Open NRE tool were used to extract unstructured data.All the structural and unstructured data were used to build the knowledge base of drought propagation.Finally,the knowledge graph system of drought propagation was realized,and the functional module of the system was displayed through operation examples.The main work and results of this paper are as follows:(1)A knowledge graph model of drought propagation was established.From the perspective of simple knowledge engineering,this research established a knowledge graph model based on drought events.The entities involved include drought time,drought location,drought impact,drought evaluation,propagation characteristics,and so on.Based on the propagation model,the data expression,extraction,storage,etc.of the drought propagation knowledge graph are determined.(2)The characteristics of the drought propagation period in the Yangtze River Basin was studied.From the experimental results,it can be seen that in terms of the distribution of periodic characteristics,the characteristics of periodic distribution in the upper,middle and lower reaches of the Yangtze River Basin are not the same,and the difference in the time scale of their period is mainly reflected in the relatively long time scale.In the analysis of the results of SPI-SRI and SRI-SSI,it was found that the maximum time scale in the downstream area is longer than that in the middle and upstream areas.(3)The characteristics of drought events in the Yangtze River Basin were analyzed.By identifying the drought events in various regions of the Yangtze River Basin,and obtaining the magnitude and duration characteristics of each event,it was found that the linear model has the best fitting effect on the drought characteristics.And the fitting results of drought duration characteristics are better than the magnitude.Meteorological drought to hydrological drought are more susceptible to differences in the Yangtze River basin in terms of magnitude.(4)The knowledge graph system of drought propagation was realized.Using python,neo4 j graph database,html,css and other technologies,four functions are realized.It included drought entity recognition,drought entity query,entity relationship query and drought event overview.Through this system,the knowledge of drought propagation can be systematically integrated and displayed.