| Located in the middle reaches of the Yellow River,the Loess Plateau was one of the most serious soil erosion areas in the world.Yan ’an was the central part of the Loess Plateau,with serious soil erosion and fragile ecological environment.However,since the implementation of the Three North Shelterbelt,the protection of natural forest resources,and the conversion of farmland to forest and grassland,the ecological environment in this area has been improved obviously.In order to explore the vegetation change and influencing factors,this paper constructed NDVI time series based on Landsat TM/ETM+/OLI remote sensing images of the growing season during 2000-2019,detected the change time of vegetation with DBEST algorithm,verified the changes of different types based on GF-1PMS remote sensing images and proposed a multi-stage time trend analysis method based on breakpoint detection to reveal the spatiotemporal distribution characteristics of vegetation change of Yan ’an.Correlation and partial correlation analysis methods were used to explore the relationship between temperature and precipitation and NDVI.The main conclusions were as follows:(1)DBEST algorithm could well detect the change time and type of break point of NDVI time series in Yan ’an,with the accuracy of 91.67% and 81.67%,respectively.The results of the temporal and spatial distribution of the break points in Yan ’an showed that most of the years of vegetation change occurred before 2012,and the duration was basically between 1-4 years.There were few mutation areas,which were characterized by high variation range and a downward trend,mainly located in the urban areas of counties and along river roads.In most of the areas,there was a gradual change and a rising trend.The rise rate of Qiaoshan and Liangshan natural forests was lower than the northern areas and Luohe River valley.(2)Breakpoint detection results showed that 62.47% of Yan’an area had obvious trend changes during the study period,and the vegetation changes showed an overall improvement trend.A multi-stage trend analysis was carried out in this area.By dividing the time series with breakpoints and end time of the change,the longest time period was regarded as the first trend period,and a second long time period was regarded as the second trend period if it was greater than 5 years.The multi-stage trend of vegetation in Yan’an showed that the improvement was distributed in most areas and lasted for more than 16 years in the first trend.The north and central areas improved greatly.The degraded areas accounted for about 3.26%,and the southeast part of the degradation degree was relatively light.35.27% of the area had no second trend,and mainly distributed in the north and Luohe River Valley,and the vegetation remained stable for a long time;27.20% of the area had second trend.In the second trend,the areas where vegetation were basically unchanged and improved were mainly located in the central area of Yan’an;and the trend change mainly started in 2000 or 2014,lasting for 6-7 years;vegetation degradation started in 2000 or 2013,and distributed in Yichuan and Yanchang areas in the edge of Wuqi County,Northwest China,lasting for more than 7 years.(3)The results of multi-stage trend analysis could reveal the dynamic trend of vegetation temporal and spatial changes.For areas where only the first trend existed,there was little difference between the first trend and the overall trend.For areas where there was a second trend,the multi-stage trend analysis method could detect more trend declines,but the overall trend could only be improved or basically unchanged.In recent years,forest land,sparse forest land,shrub forest land,and unfforested land have all improved significantly,and the ecological construction project of Yan’an has achieved good results.(4)The yan ’an area precipitation in 2000-2015 and the annual average temperature factor to different regional difference is obvious,had the following change rules.The average annual temperature of Yan ’an was in a gentle state.The increasing trend of annual precipitation was not significant,but the fluctuation was large.Both of them decreased from southeast to northwest.Meanwhile,the precipitation in southern mountainous areas was significantly higher than that in surrounding areas with lower elevations.On the time scale,the maximum annual NDVI was correlated with annual mean temperature and annual precipitation,but the correlation coefficients were low.On the spatial scale,the effects of temperature and precipitation on vegetation in different regions showed different correlations.There were fewer regions where the maximum annual NDVI was significantly correlated with annual average temperature and annual precipitation,and the regions where there was an insignificant positive correlation were the largest.In general,the vegetation in Yan’an was more sensitive to precipitation. |