| Plague is one of the existing class A virulent infectious diseases in China,which has been included in the border inspection category.Yunnan province,located in the hinterland of Central Asia and the southwest border area of China,is the foci of the plague of the historian and has an ancient plague epidemic history.In the border area of Dehong Prefecture,the plague epidemic among animals has been active over the years,which has been the concentration area of rodent plague such as Ratte flavipectus in the past years,posing a great threat to local and even national human health.It is of great significance to explore the temporal and spatial distribution of plague epidemic in domestic plague sources such as Dehong Prefecture,improve the plague surveillance index system,and guide the prevention and control of rodent plague in Yunnan province and even in southern China.This paper takes dehong Prefecture of Yunnan Province as the study area,and Ratte flavipectus as the representative host type of local rodents,to analyze the spatial and temporal pattern of plague in Ratte flavipectus in Dehong Prefecture,analyze the surveillance indexes,estimate the historical risk of plague in Ratte flavipectus and analyze the prevalence factors.The following is the main research content of this paper:(1)Spatial and temporal pattern of plague epidemic in Dehong from 1983 to2013.The results showed that the transmission path and spatial distribution pattern of human plague cases were consistent with that of plague in Ratte flavipectus.In terms of spatial distribution characteristics,the indicators of plague in Ratte flavipectus showed spatial correlation and aggregation distribution characteristics,which were the same as human plague cases,showing a cluster aggregation distribution centered on road intersections and administrative units.In terms of time distribution,the SARIMA model was used to analyze the monthly total positive values of Ratte flavipectus.The results showed that the monthly total positive values of Ratte flavipectus had the characteristics of periodicity(12 months),seasonality and time autocorrelation during1983-1987 and 2002-2007;(2)Rat density FRK spatiotemporal interpolation.The main transmission route of rodent plague was "rat-flea-rat".Based on the historical surveillance and investigation data of plague animals,the study selected the density of rodents that affected the occurrence and transmission of plague among rodents for monitoring index analysis.Based on 3S technology,this part creatively introduced the spatio-temporal fixed kriging interpolation(FRK)method to estimate the population density of Ratte flavipectus in different habitats.Based on the interpolation results,the indoor and outdoor density data of Ratte flavipectus from 1983 to 2013 were collected and analyzed.(3)Analysis and risk mapping of plague occurrence factors in Ratte flavipectus in Dehong Prefecture.Based on the results of literature review,three natural environmental factors,such as the indoor and outdoor densities of Ratte flavipectus and flea index of Xenopsylla cheopis on the body of Ratte flavipectus,including host and vector population factors,monthly bioclimatic factors,leaf area index,terrain,soil and land use type,were selected.And 31 indicators of human factors such as population density,per capita GDP and main crop yield.Based on the spatial analysis method and remote sensing information extraction,the plague spatio-temporal factor data set of Ratte flavipectus was collected based on plague surveillance sites among animals.Qualitative descriptive analysis and quantitative mechanism exploration were conducted with whether positive rodents were detected by bacteriological test at monitoring sites as response variables.The random sampling method was used to construct the random forest classification model,and 75% data were selected as the test set.The accuracy verification showed that the OOB value of model error rate was 4.73%,and the AUC value was 87.23%.The accuracy of the classification model met the basic requirements.In order to further explore the seasonal effects of plague occurrence in r.tanestris,all data sets were divided into drought and rain season layers according to precipitation,from November to April and from May to October,respectively.According to the overall climate characteristics,the layered random forest model is divided into four layers: spring(February to April),summer(May to July),autumn(August to October)and winter(November to January).The comparison of classification accuracy showed that the seasonal stratified model was more accurate to distinguish plague occurrence.According to the factor contribution size to determine the influence of Dehong prefecture plague in Ratte flavipectus occurred in the first 10 important factor,respectively is: monthly total numbers of Ratte flavipectus checked bacteria,indoor and outdoor rat densities of Ratte flavipectus and flea index of Xenopsylla cheopis on the body of Ratte flavipectus,monthly temperature difference,monthly relative humidity,monthly maximum temperature,monthly minimum temperature,monthly total precipitation,after four climate elements in the secondary factors.The composition and order of important factors are different in different seasonal layer models.Based on the results of seasonal stratified random forest model,the monthly plague risk probability(pi)of historical monitoring sites was predicted in spring,summer,autumn and winter,and the annual average plague risk(Pi)of historical monitoring sites in four seasons was calculated.The results showed that: in general,Pi values in each township were summer > autumn > winter > spring.It is suggested that the surveillance of Ratte flavipectus as the main host should be carried out in the later period,and the intensity should be strengthened in summer and autumn.In the central part of Dehong prefecture,the areas within 5km of each county,township center and provincial road are all historically high risk areas,so regular surveillance of plague among animals should be actively carried out. |
PDF Full Text Request