Remote Sensing Extraction And Spatial-temporal Evolution Of Oriental Migratory Locust Potential Distribution

Oriental Migratory Locust is one of the major crop pests in China.With climate change,the potential distribution of locust(areas suitable for breeding and development of locusts)is constantly changing,posing a great challenge to agricultural pest monitoring and control.Therefore,there is an urgent need to realize extraction and spatial-temporal evolution analysis of the large scale oriental migratory locust potential distribution to facilitate the precise control of agricultural pest in China.Remote sensing,as a real-time,rapid and large-area earth observation technique,can provide new ideas and methods for the extraction and spatial-temporal evolution analysis of oriental migratory locust potential distribution.This study used multi-source remote sensing data as the main data source,combined with meteorological re-analysis data and locust regional survey data,taking the Tianjin north Dagang reservoir as an example to carry out the remote sensing extraction method of oriental migratory locust potential distribution with the relationship between host and environmental elements and locust occurrence.Taking into account the influence of the differences in locust fertility caused by the complex regional meteorological background on the host and environmental elements acquisition time,taking the middle and lower reaches of the Yellow River as an example,the remote sensing extraction method for oriental migratory locust potential distribution in large areas combining meteorological differences was carried out to realize the spatial extension of the remote sensing extraction method for locust potential distribution.Finally,the spatial-temporal evolution of oriental migratory locust potential distribution was analyzed in terms of its ecological control needs,based on an accurate extraction of the long time series of its potential distribution.The study not only improves the monitoring level of oriental migratory locust in China,but also provides scientific suggestions for the ecological control measures of oriental migratory locust,which in turn supports the precise and green control of oriental migratory locust and is of great significance to the food security production in China.The main research results are as follows:(1)To improve the accuracy of locust habitat extraction,the study used the Tianjin north Dagang reservoir as the study area,combined the relationship between locust occurrence and host and environmental elements,selected and extracted key habitat factors based on Landsat and MODIS remote sensing data to comprehensively characterize locust host and environmental elements,combined regional survey data and correlation analysis to determine the order of importance of habitat factors influencing the extraction of oriental migratory locust potential distribution,and used the analytic hierarchy process method to determine the weights of each habitat factor.Finally,a remote sensing extraction method(LHS)for oriental migratory locust potential distribution in small areas was established,by integrated expert knowledge and the moving window algorithm.The results showed that the four habitat factors of vegetation cover,land cover class,surface temperature and soil water content can be used to characterize the host and environmental factors affecting locust occurrence.The weighting of each habitat factor is highest for land cover class(0.28),followed by vegetation coverage(0.27)and land surface temperature(0.25),and lowest for soil water content(0.20).The LHS model,which combined host and environmental elements and added a moving window algorithm to quantify landscape structure,extracts oriental migratory locust potential distribution with 87.5%accuracy,which is10.0%better than the HS model without quantification of landscape structure.The above results indicated that the remote sensing method of extracting oriental migratory locust potential distribution considering host and environmental elements has potential for extracting oriental migratory locust potential distribution in small areas.(2)For the influence of a large regional complex meteorology on the life cycle of oriental migratory locust,the study introduced degree day model to simulate locust development in different provinces based on meteorological re-analysis data and regional survey data to determine the development time of locusts in each province and extract habitat factors by combining remote sensing data.On this basis,the habitat factors were further optimized and updated,and finally the spatial extension of the potential distribution extraction method in small areas combining host and environmental elements was achieved by combining the patch distribution of different land class to establish remote sensing extraction method for oriental migratory locust potential distribution in large areas(DD+ILHS).The results showed that there were significant differences in the life cycle period of locust in different provinces.After updating the weights of each factor,the weight of land cover class weight is 0.29,followed by vegetation coverage(0.28)and land surface temperature(0.22),soil water content was still the lowest(0.21);the R~2of the DD+ILHS model combining meteorological differences can reach 0.83,the highest among all models.The above results showed that the remote sensing extraction method of oriental migratory locust potential distribution can effectively solve the influence of the differences in locust life cycle caused by regional meteorological differences on the extraction of the potential distribution in large areas,and realize the spatial extension of the remote sensing extraction method of potential distribution,which provides a basis for the analysis of the spatial-temporal evolution of oriental migratory locust potential distribution.(3)In order to support the scientific control of oriental migratory locust,the study used remote sensing extraction of oriental migratory locust potential distribution with meteorological differences to extract locust potential distribution along the middle and lower reaches of the Yellow River in the 1990s,2000s,2010s and 2020s.A multi-temporal binary coding method was used to analyze the spatial-temporal dynamics of locust potential distribution in four periods,and on this basis,the correlation between the spatial-temporal evolution of locust potential distribution and land use/cover change was investigated in conjunction with land cover transfer matrix analysis.The results showed that the evolution of oriental migratory locust potential distribution can be classified into four categories:extinction,new,repetition and stable;the locust potential distribution along the middle and lower reaches of the Yellow River was generally stable over the four periods.Temporally,from 1990s to 2020s,the large extinction of the locust potential distribution along the middle and lower reaches of the Yellow River was accompanied by the new of a certain area,the area of the locust potential distribution showed a trend of increasing and then decreasing and reaches a peak in the2000s.Spatially,the extinct and stable areas were mainly distributed in the Yellow River estuary area,the repetition areas were mainly located along the Yellow River,and the newly areas were relatively less evenly distributed in the study area.The extinction of locust potential distribution along the middle and lower reaches of the Yellow River is closely related to the transformation of cropland and wetland into forest,artificial surface and water,or grassland into cropland and forest.The newly locust potential distribution was closely related to the transformation of forest,artificial surface,water and other land classes into cropland,or cropland into grassland.The overall change of land cover in stable areas is small.The results of the above study can provide a scientific basis for the ecological control of oriental migratory locust.