| Land desertification is closely related to the living environment of human beings and the development of social economy.Because of its rapid development and serious hazards,it has attracted wide attention of scholars all over the world..Based on the dynamic change law and influencing factors of desertification,the dynamic prediction model is explored,which is of great significance for the prevention and control of desertification in the study area.This paper takes Jinghe County as the research target area,based on the Landsat series of remote sensing images of 1990,2000 and 2016,and the related maps and statistics are auxiliary data.According to the selection and setting of driving factors,the ANN-CA model is constructed.Based on the model correction and accuracy verification,the 2016 desertification is taken as the initial data to simulate the trend of land desertification in the Jinghe County in 2026,which is the future of the research area.Land management provides a reference basis.Research indicates:(1)Extracting the land area of desertification type from 1990 to 2016 and analyzing its dynamic change information.The results show that the area of??severe and moderate desertification is reduced,the desertification of species is reduced from 3855.49 km~2to 3457.7 km~2,and the moderate desertification is reduced from 2958.95 km~2to 1721.02 km~2;the area of mild and non-desertification continued to increase,the light desertification increased from 1755.75 km~2to2559.27 km~2,and the non-desertification land increased from 1676.03 km~2to2408.23 km~2.From the perspective of the spatial pattern of desertification,the desertification in the northeastern part of the Aibi Lake Conservation Area continues to increase,and the local area continues to deteriorate;the desertification around the southern mountainous area and the Ulandadangai Desert is weakening,and the ecological situation has improved.(2)Based on the ANN-CA model to predict the trend of land desertification in the study area in 2026,the simulation results show that under the natural trend scenario,the overall change of the area of different desertification types is not large,and the original level is basically maintained,but the non-desertification type is-5.96%.The rate of change is slightly higher than that of other land types;in the protection trend scenario,non-desertification increased from 2408.237 km~2to3168.857 km~2in the early stage of prediction,and other types of desertification area decreased by 0.61%,2.46%,and 25.18%,respectively;Under the trend scenario,the rate of change of desertification types increased significantly during the same period.Unless desertification land types were used,their desertification land was reversed by dynamic changes of 3.27%,34.6%and 5.47%respectively.(3)Land desertification in the study area is the result of the combination of nature and humanity.By selecting the three natural factor parameters of precipitation,soil fertility and temperature,the relationship between each parameter and the desertification difference index(DDI)shows that the correlation coefficients of precipitation,soil fertility and DDI are 0.784 and 0.801,respectively,which are highly correlated;The correlation between temperature and DDI is not obvious,only individual regions are relevant.The impact of population growth and socio-economic development on the land use pattern,surface temperature and vegetation cover in the human factor will increase the degree of land desertification in the study area.(4)The simulation results of ANN-CA model are highly consistent with the actual situation,indicating that the model has good adaptability in this study area. |
PDF Full Text Request