Study On The Dynamic Change Of Farmland Shelterbelt And The Benefit Of Windbreak And Sand Fixation In Ganzhou District

Farmland shelterbelt is an important part of the three North Shelterbelt System,which makes an important contribution to the ecological environment protection and grain production in the three north area.The spatial distribution of farmland shelterbelts and the dynamic estimation of windbreak and sand fixation are the important contents of the ecological engineering monitoring of the three North shelterbelts.The traditional way of artificial investigation to obtain farmland shelterbelts and estimate the amount of windbreak and sand fixation is time-consuming and inefficient.Remote sensing technology can make up for the shortcomings of traditional artificial investigation.In the study of spatial distribution of farmland shelterbelt,vegetation information can be extracted from single-stage high-resolution remote sensing image,but the phenomenon of "same spectrum foreign body" and "same object different spectrum" can not be solved.Therefore,it is of great significance to obtain the spatial distribution of farmland shelterbelt and vegetation coverage quickly and accurately,which can provide data support for the estimation of windbreak and sand fixation benefits,and then provide timely and effective information support for the modern management and service of agriculture and forestry.This paper takes Ganzhou District of Zhangye City as the research area,using stratified random sampling to obtain 462 sample plots of 30 m × 30 m.Using Landsat 5 and Landsat 8 time series images in 2000,2005,2010,2015 and 2018 as data sources,the dynamic changes of farmland shelterbelts and the benefits of windbreak and sand fixation were studied.Firstly,the phenological parameters of the study area are extracted from time series data,and the dynamic data of farmland shelterbelts and land use types are extracted by multi time series and multi feature hierarchical classification method.Secondly,extract remote sensing image and vegetation index factors,use Pearson correlation coefficient,variance expansion factor and importance analysis method to analyze the correlation between remote sensing variables and vegetation coverage,carry out collinear diagnosis between variables,and use linear stepwise regression method to screen variables.The endmembers was selected and purified from the pixels without land use change in the 2000-2018 remote sensing image.The vegetation coverage of the study area was inversed by using the mixed pixel decomposition and random forest model to monitor the vegetation coverage change and vegetation growth in Ganzhou area in 2000-2018.Finally,based on the dynamic data of vegetation coverage and land use type,the revised wind erosion equation(RWEQ)was selected to evaluate and analyze the wind and sand fixation in Ganzhou area from 2000 to 2018.The results are as follows:(1)NDVI time series data can extract the phenological parameters effectively.Using phenological parameters such as growth season length and growth amplitude can separate forest land and cultivated land information to extract farmland shelterbelt,which based on vegetation phenological information is widely used in crop subdivision.In this study,the phenological differences between forest and crops are used for reference to extract farmland shelterbelt in Ganzhou district.The conditions for distinguishing ground objects are not as strict as those for subdividing crops,and the classification accuracy is higher than that of traditional classification methods based on spectral characteristics differences.(2)Multi temporal and multi feature hierarchical classification method can effectively extract farmland shelterbelt information.Among 33 samples of farmland shelterbelts,29 were extracted correctly,and the classification accuracy was 87.8%.Most of the existing researches focus on decision tree classification,which used spectral information,shape index,etc.And few researches combine multi time series and multi feature methods to improve the classification accuracy of ground objects.(3)From 2000 to 2018,the ecosystem types of Ganzhou district changed significantly.Based on the analysis of the dynamic change of the ecosystem types in Ganzhou District,we found that the area of farmland and other ecosystem types increased significantly,149.35 km2 and 76.13 km2 respectively.The area of grassland and desert ecosystem decreased by 156.48km2 and 83.9 km2 respectively.(4)The inversion effect of random forest vegetation coverage is better than that of the limited least square mixed pixel decomposition method.In the estimation of vegetation coverage in 2018,the decision coefficient R2 of the estimation results of the random forest method is 0.753,and the root mean square error(RMSE)is 0.151,which is better than the results of the limited least square mixed pixel decomposition method with R2 0.689 and RMSE 0.173.(5)It is feasible to select the end element from the pixels without land use change during 2000-2018 to carry out vegetation coverage inversion.In the case of only the measured value of vegetation coverage in 2018,the method of endmembers selection and purification is used to determine the measured value of multi-phase image at the same time,and the random forest model is used to predict the vegetation coverage in Ganzhou District in 2000-2018.The vegetation coverage distribution map of Ganzhou District from 2000 to 2018 shows that from 2000 to 2018,the vegetation coverage of Ganzhou District has improved significantly,mainly reflected in the central region.The distribution of vegetation cover in Ganzhou area is almost the same for many years.It is high in the middle,low in the South and North,and the boundary is obvious.(6)From 2000 to 2018,the total amount of windbreak and sand fixation in Ganzhou district increased.Based on the vegetation coverage and combined with the relevant natural environment data,RWEQ model is used to quantitatively estimate the windbreak and sand fixation amount in Ganzhou District from 2000 to 2018.The results show that from 2000 to 2018,the annual total amount of windbreak and sand fixation materials in Ganzhou district is between 25500-36200 tons,and the overall amount of windbreak and sand fixation is increasing,among which the growth of forest land and grassland are the largest.