Vegetation Coverage Changes Of West Lakeside Oasis In Bosten Lake Based On Pixel Dividing Model

Vegetation coverage reflects the change of vegetation type,quantity and quality in terrestrial ecosystem to the certain extent,which is an important indicator of vegetation status.take the west bank of Bosten Lake as study area.1990?2000 and 2015 three different period remote sensing images as the basic data sources.combined with field research data used dimidiate pixel model to estimate the vegetation coverage for the study area in 1990?2000 and 2015,accorded the actual vegetation coverage to divided the vegetation coverage degree.used the matrix method to analyed the change of area of all vegetation cover types from 1990 to 2015.used the difference method to calculated the vegetation coverage,analyed the vegetation coverage in space changes from 1990 to 2015.Compared BP neural network model and LMBP neural network model,selected the high accuracy of the model to predict the vegetation coverage.selected the driving factors influencing the change of vegetation coverage from the aspects of natural factors and human factors.Through the correlation analysis,four kinds of coverage such as very low vegetation coverage,low vegetation coverage,medium vegetation coverage and high vegetation coverage were discussed type change main driving factor.The main conclusions are as follows:1.From 1990 to 2015,very low vegetation coverage,low vegetation coverage area is became smaller,medium vegetation coverage,and high vegetation coverage area continued to increase.Nearly 25 a,high vegetation coverage area significantly changed from 71.93km2 to 361.17km2,increased of nearly 5 times.Very low vegetation coverage area is reduced by 323.94km2,the overall decline of 81.28%.2.At 25 a,the transfer rate of low vegetation to medium vegetation was the highest,the transfer rate was 59.01%,and the transfer area was 156.41km2.second was the low vegetation vegetation coverage area toward the transfer,the transfer area 227.12km2,the transfer rate of 45.51%,third transferred from the vegetation of the very low vegetation to the vegetation coverage area,the transfer area reached 142.10km2,the transfer rate reached 28.50%.Finally,the high vegetation was transferred to the very low vegetation coverage area,the transfer rate was only 0.048%.3.From 1990 to 2015,the area of vegetation coverage in the study area accounted for 65.95% of the total area,with an area of 558.47km2.Significant improvement in the area of 243.94km2,the total area of 28.81%.The improvement area is mainly in the middle and eastern areas of the study area.The degraded areas are in the southern and western regions of the study area.Vegetation coverage in the past 25 years showed an improvement trend4.Constructed and compared the BP neural network model and LMBP neural network model.The results show that the LMBP neural network model has a higher accuracy than BP neural network model.The LMBP neural network model predicts that the correlation coefficient between the true value and the predicted value is 0.9400,which indicates that it is feasible to use the LMBP neural network to predict vegetation coverage in the study area The The vegetation coverage of the study area is forecasted in 2020,and the results show that the vegetation coverage is degraded.5.Vegetation coverage changed were influenced by natural factors and human factors.There was a significant negative correlation between very low vegetation coverage and annual precipitation(by p <0.05),and the correlation coefficient was-0.969.The correlation between the low vegetation coverage and the medium vegetation coverage area was-0.972 and 0.888,respectively.The area of high vegetation coverage area is related to the agricultural output value,the correlation coefficient is 0.988.