Research And Application On The Identification Of The Suitable Areas For Marmota Himalayan In Yushu City

Plague is one of the Class A infectious diseases in my country.The Qinghai-Tibet Plateau is the main plague foci in my country,and the Marmot himalayan(hereinafter referred to as "marmot")is the main host animals in this foci.Effective monitoring and identification of its suitable areas in Qinghai-Tibet plateau is important for active preventing and controlling of the marmot plague epidemic.At present,the identification and monitoring of marmot suitable areas are still mainly based on field investigation,the application of spatial information technology is insufficient,and the suitable area model considering natural and human activity factors is lacking,which leads to the fact that field investigation mainly relies on subjective experience and lacks appropriate global guidance,which is not conducive to comprehensive,efficient and scientific prevention and control work.To solve these problems,this paper took Yushu City of Qinghai Province as the study area,combed the factors affecting the distribution pattern of marmot suitable areas in this area,such as topography,soil,vegetation cover,climate conditions and human activities,and constructed a multi-dimensional evaluation index system;On this basis,combined with satellite remote sensing,UAV remote sensing,crowdsource data and other multi-source data,the above factors were quantified spatially,and the identification of marmot suitable areas based on multiple species distribution models and their integration was carried out;Combined with the consistency of multi-model research results and the actual needs of local plague prevention and control,the relevant reference routes focusing on prevention operations and improving the prediction of marmot suitable areas were put forward respectively,which provided a scientific basis for guiding the local marmot monitoring operations.The main research contents and results of this paper are as follows:(1)Based on the site survey data,remote sensing data and other thematic data,the spatial distribution characteristics of marmot suitable area elements in the study area were comprehensively analyzed from the aspects of topography,climate,soil,energy source and human disturbance,and a multi-dimensional evaluation index system for marmot suitable habitat areas was constructed.The results showed that there were two main problems in the existing marmot survey site data: the spatial distribution was unbalanced and the proportion of existing sites and non-existing sites was unbalanced,and the relevant treatment methods were proposed.Through further analysis of spatial characteristics,The characteristic parameters such as normalized difference temperature and humidity index,aspect and soil type were highly correlated with marmot,and 15 multi-dimensional indicators were selected for the evaluation and modeling of marmot suitable areas.(2)Based on the investigation of ten species distribution models,such as RF(Random Forest)and GBM(Generalized Boosting Model),the results,accuracy and variable importance of each model were compared according to different classification algorithms,and then the model screening and integration were carried out according to the AUC value,and the combination model of marmot suitable area identification in the study area was constructed.The results showed that the species distribution models had good prediction effect on the spatial distribution of marmot.The machine learning models(RF,GBM),the maximum entropy model(MAXENT)and the classification models(CTA,FDA)had higher accuracy and stability,and obtained a larger range of suitable areas for marmot;The accuracy of the regression model was slightly lower,and the range of marmot suitable habitat was smaller.Compared with the single species distribution model,the accuracy and stability of the ensemble model were significantly better.(3)The results of marmot suitable area identification were further applied to the plague investigation and prevention and control work in Yushu City,and the relevant reference routes focusing on prevention and control work and improving the prediction of marmot suitable area were designed respectively.The results showed that the former focused on the suitable areas for marmot determined by multiple species distribution models,while the latter focused on the areas with inconsistent output between species distribution models,both of which complemented each other and considered road factors to ensure the convenience of transportation along the line,which had important guiding significance for improving the level of plague prevention and control.