Study On Early Warning Of Grassland Rodent Based On Meteorological Factors

Grassland,as a precious umbrella of nature,is similar to cultivated land,forest,ocean and other natural resources,and is also an important strategic resource in China.It plays an indispensable role in protecting wild biodiversity,preventing soil erosion and purifying air.However,due to global temperature changes in recent years,grassland rodent in the Inner Mongolia Autonomous Region has occurred frequently,which not only destroyed the grassland environment,but also caused huge economic losses to herdsmen.Taking Horqin Right Front Banner(referred to as Ke You Front Banner)as the study area to predict the area of grassland rodent damage in this county,and establish a scientific and practical value of grassland rodent prediction model to predict the area of grassland rodent damage and to warn the occurrence of grassland rodent infestation.Two models,linear model and neural network model,were introduced to predict the occurrence area of rodent in grassland.By combining 8 sets of data such as average temperature,precipitation,maximum wind speed and sunshine hours with the area of grassland rodent infestation,the prediction of the area of grassland rodent infestation in Ke You Front Banner was carried out as follows:Firstly,through combing and analyzing the research progress on the area of grassland rodent damage and the dominant rodent species—Altai zokor in the study area,the relationship between the area of grassland rodent damage and meteorological factors is summarized.At the same time,the establishment of the study area,experimental methods,statistical methods for the area of grassland rodent damage,field experiment procedures,and the acquisition and sources of meteorological data are introduced in detail.Secondly,through the analysis and summary of the area of grassland rodent damage and the results of field experiments,it is found that the area of grassland rodent damage in the past year and human factors cannot have a decisive influence on the area of grassland rodent damage in the second year,while will have a decisive influence on the area of grassland rodent damage.Thirdly,through the correlation analysis between the area of grassland rodent damage and 8 sets of data including average temperature,precipitation,maximum wind speed,sunshine hours,etc.,the results showed that all the other factors except the average relative humidity had significant correlation with the rodent infestation area in Ke You Front Banner.Fourthly,the area of grassland rodent damage in Ke You Front Banner was predicted by using a linear model,a curve model,a multivariate linear model and a neural network prediction model,and the accuracy of the model was tested.The analysis showed that the above linear models passed the significance and variance analysis tests.Although the multivariate linear model is better than the uni-variate linear model and the uni-variate curve model,the multivariate linear model has the defect of model instability.The introduction of the neural network model not only overcomes the shortcomings of instability,but also has higher prediction accuracy.The prediction accuracy is more than 91%,so it meets the business requirements of predicting the area of grassland rodent damage.Lastly,summarizes the research content and analysis results in the paper,and also some problems and goals in the future research are explained and prospected.