Study On Dynamic Monitoring And Driving Mechanism Of Desertification In Arid Area Based On Deep Learning

The ecological environment in arid areas is closely related to water resources,and desertification is an important variable reflecting the relationship between them.Therefore,monitoring desertification based on remote sensing has become a hot spot of environmental monitoring in arid areas.With the increasing of remote sensing data sources,the monitoring method based on deep learning has been developed rapidly.This paper takes the lower reaches of Tarim River,near Daxihaizi reservoir(40°22′31″N～40°45′35″N,87°23′16″E～88°0′15″E)as the research area.Based on the Landsat medium resolution remote sensing satellite images of five periods(2000,2005,2010,2015 and 2018)in the study area,and based on ENVI 5.5the remote sensing image software processing platform used for image preprocessing,and then the deep learning model is used to classify the features of desert,water body,construction land and cultivated land in the study area in recent 19 years.The classification results satisfying the accuracy are obtained;furthermore,the annual change of desertification in the region is statistically analyzed by using single land dynamic degree and land use transfer matrix method And the mutual transfer between desert and other features.Based on the above classification results,the desert area was extracted to obtain the normalized vegetation index(NDVI),and then the vegetation coverage of the study area was obtained.Combined with the quantitative relationship between vegetation coverage and desertification,the classification results of desertification grades in different years in the study area were obtained,and the mutual transformation among severe,moderate and mild desertification areas was analyzed by using transfer matrix method.Combined with natural factors(annual precipitation,annual average temperature,annual average wind speed,annual maximum wind speed)and human factors(ecological water supply,water surface area,cultivated land area,population),the driving mechanism of desertification was discussed by using grey correlation degree method,and the development of desert was predicted by GM(1,1).Finally,some effective suggestions on desertification control and ecological restoration are put forward.The main conclusions are as follows.1.The accuracy analysis of 5 phases(2000,2005,2010,2015 and 2018)shows that the overall classification accuracy of the images is more than 85%,and the kappa coefficient is about 0.83.2.Desert is the main feature in the study area,accounting for more than 80%of the total area of the study area;in the past 19 years,the desert area has been decreasing continuously,with a total reduction of106.27 km~2;the proportion has been continuously reduced,with an average annual growth rate of-0.30%,and desertification has been continuously reversed.The overall ecological environment has improved.3.In recent 19 years,the area of severe desert decreased by 334.71 km~2;the area of moderate desert increased by 182.92 km~2;the area of light desert increased by 48.05 km~2.The average annual growth rates of severe desert area,moderate desert area and light desert area were-1.24%,2.24%and 5.68%respectively.The area of non desert is small and should be ignored.The area of unclassified land increased by 106.27 km~2,with an average annual growth rate of 2.52%.In general,the severe desertification is continuously reversed,and the area of moderate,mild desert and unclassified land is increasing,and the ecological environment is effectively improved.4.Through the grey correlation analysis of natural factors,human factors and desert and severe desert,the correlation degree of ecological water supply and desert and severe desert is the highest.The increase of ecological water supply in the study period is helpful to vegetation restoration and desert The reversion of transformation.The impact of human factors on desertification is decisive,especially the ecological water supply;it leads to the change of surface water area,which provides the water needed by crops for cultivated land reclamation;it provides water for vegetation growth in the study area;it provides necessary water for human life.5.The GM(1,1)model is used to forecast the desert and severe desert in 2025,2030,2035,2040 and2045 years;During the prediction period,the area of desert and severe desert showed a decreasing trend.The dynamic monitoring of desertification and the research on the driving mechanism of desertification in this paper can timely grasp the change of desertification in this area,and provide useful reference for the government and people to effectively control desertification.