Analysis Of Land Use Change And Driving Factors Before And After The Earthquake Based On GEE

As the most intuitive manifestation of human and natural activities on the surface,land use change has become an important way to grasp the process of urbanization and understand the status of the ecological environment.As one of the most serious disasters to urban and ecological damage,the earthquake not only changes the surface morphology of the disaster area but also affects the change of land use.Ludian County is located in the south of Zhaotong City,Yunnan Province,with complex geological conditions and a fragile ecological environment.After the Ms6.5 earthquake occurred in Ludian County in August 2014,it had a certain impact on the current status of local land use.Scientific analysis of the law of land use change and its driving factors before and after the earthquake in Ludian County will help to grasp the post-earthquake recovery process,promote the rational development and utilization of land,and promote highquality economic development.This thesis takes Ludian County as the research area,and takes 2013,2014,2017 and 2020 as the research period.First,based on the Google Earth Engine(GEE)cloud platform,using the random forest classification algorithm obtains the land use classification data of Ludian County,and on this basis,the distribution characteristics of land use in Ludian County before and after the earthquake are clarified from the three dimensions of township scale,terrain gradient,and seismic intensity.And then,using the methods such as the land use dynamic degree,the land use degree composite index and the land use transfer matrix reveal the spatiotemporal change law of land use before and after the earthquake in Ludian County.Finally,considering natural,human and disaster factors comprehensively,the optimal parameter-based geographic detector model(OPGD)is used to screen the optimal parameters for the discretization of driving factors,using factor detectors and interactive detectors to analyze the driving factors of land use change in Ludian County before and after the earthquake,and put forward optimization suggestions for land use protection in Ludian County after the earthquake.The main conclusions are as follows:(1)Based on the GEE cloud platform,the overall accuracy of land use classification results in different periods in Ludian County obtained by using the random forest classification algorithm is greater than 85%,and the Kappa coefficient is greater than 0.82.And by selecting typical areas affected by earthquakes for verification,it is found that the classification results have a high consistency with the high-resolution historical image features of Google Earth Pro.The above results show that the classification results of this thesis are accurate and reliable.(2)This thesis analyzes the distribution characteristics of land use in Ludian County before and after the earthquake from the three dimensions of township scale,terrain gradient,and seismic intensity.It is found that the areas with large changes in land use types in Ludian County before and after the earthquake are mainly located in Longtoushan Town and Huodehong Town,the area with an elevation of 1400-1800 m,the area with a slope of less than 11° and greater than40°,and within the range of magnitude Ⅷ and Ⅸ seismic intensities.The main performance is that the proportion of unused land increases first and then gradually decreases,the proportion of forest,grassland and built-up showed a trend of decreasing first and then increasing,and the proportion of cropland was always in a decreasing trend.(3)The main land use types in Ludian County are forest,grassland and cropland,and the sum of the three areas accounts for more than 93% of the total area of Ludian County;the builtup has been expanding after the earthquake;the unused land has the largest change range,showing a trend of substantial increase first and then a continuous decrease,with a dynamic degree of 712.80% during 2013-2014.The land use degree composite index in Ludian County was 239.17,227.61,300.63 and 313.50 before and after the earthquake,respectively;the spatial variation of land use degree in Longtoushan Town was more obvious,showing a trend of first decreasing and then rising.The transfer of land use mainly occurred in Longtoushan Town and Huodehong Town,in 2013-2014,it was the transfer of grassland,cultivated land,construction land and forest land,and the transfer of unused land;in 2014-2017 and 2017-2020 Years are the transfer of unused land and cultivated land,and the transfer of construction land,grassland and forest land.(4)The OPGD can screen out the optimal parameters for the discretization of the driving factors,which improves the explanatory power of the driving factors.The factor detection results show that the land use change in Ludian County before and after the earthquake is mainly affected by population factors and economic factors;in 2014,2017,and 2020,the q value of the seismic intensity and the distance to the active fault reached above 0.3,the explanatory power is significantly improved compared with that in 2013.The results of interactive detection show that the interaction between each factor shows the characteristics of two-factor enhancement or non-linear enhancement.Moreover,the explanatory power of the interaction between seismic intensity or distance to active faults and other factors has been significantly improved,and seismic factors continue to affect the land use change in Ludian County after the earthquake.