Study On Plague Of Grassland Locust Data Reconfiguration Modeling In Ar Horqin Banner Based On History Record

Grassland,as the largest surface protection barrier in China,plays an extremely important and indispensable role in the protection of ecological environment and biodiversity.Especially in preventing land salinization,soil erosion and wind erosion aspect,other ecosystems are not as good as grassland.In recent years,with the global climate anomalies and the deterioration of ecological environment,plague of grassland locust occurred frequently in Inner Mongolia Autonomous Region,that destroyed a lot of pastures,caused serious economic losses to herdsmen,and affected the sustainable development of animal husbandry in the autonomous region and the quality of grassland ecological environment.Ar Horqin Banner is one of the plague of grassland locust hazard prone areas in Inner Mongolia Autonomous Region,and it has different degrees plague of grassland locust every year.So taking this as the study area,determine the relationship between meteorological factors with the locust and establish practical application value refactoring model of the grassland locust and population density.It not only provides data support for grassland locust areas and population density,but also provides scientific basis and model support for monitoring and early warning of Inner Mongolia plague of grassland locust.Linear regression mainly used to establish the plague of grassland locust predicting model,but few scholars have predicted history data,that predicting is refactoring.And most predicting models have some questions,that are the use of inadequate information and the model predicting accuracy is low.In this paper,seven meteorological factors are introduced into the refactoring model,with the multiple linear regression analysis and dimension reducing method are used to overcome the above shortcomings.This article based on the existing data,amied at grassland locust areas and population density reconsitution in Ar Horqin Banner,mainly done the following work:Firstly,progress on effects of grassland locust research at home and abroad were analyzed,the interaction between locust and meteorological factors is summarized,and the methods of refactoring grassland locust areas and population density with using meteorological factors are summarized.Secondly,statistical methods and standard of grassland locust areas and population density are introduced,the acquisition methods and processing methods of meteorological data are also described.Thirdly,by using the year 2004 to 2016 history records,the grassland locust areas,population density and seven meteorological factors of the correlation results show that: the grassland locust areas were positively correlated with the sunshine of late August last year,the temperature and ground temperature of late September last year,the ground temperature of early and late October last year,the wind of early November last year,the wind of middle June that year,the rainfall and relative humidity of early July that year.And it was negatively correlated with the relative humidity of October last year,the relative humidity of middle November last year,the rainfall of middle January that year,the temperature,ground temperature and sunshine of early July that year,the relative humidity of late July that year.The population density was positively correlated with the temperature of middle November last year,the sunshine of early and late December last year,the rainfall of early April that year,the temperature and ground temperature of late April that year,the sunshine of late May that year,the wind of early June that year,the relative humidity and rainfall of middle June that year.And it was negatively correlated with the wind of early and late October last year.But the pressure is not influence on grassland locust areas and population density.In the forth chapter,combining with the history data in 2004 to 2016,we use a linear regression,curve estimation and multiple linear regression method to establish refactoring models,and use the 1997,1998 and 2004 to 2016 history records to complete the validation of the reconfiguration model.According to the significant 0.05 t test and variance analysis showed that although the three models have all passed the test,but the multivariate linear model has more meteorological factors as independent variable characteristics,higher fitting degree,and from the results,we can accurate reconstruct grassland locusts history status.Lastly,the research results of this thesis are summarized and discussed,and the outlook for further works is pointed out.