Vegetation Cover Change Process And Future Distribution Prediction On The Loess Plateau After“Grain To Green”

In order to explore the dynamic changes of vegetation cover and its driving factors after performing the program of “Grain to Green” in the Loess Plateau,we analyzed the seasonal and annual trends of NDVI based on MODIS/NDVI vegetation remote sensing data and ESA CCI-LC vegetation cover classification data,in the whole area of the Loess Plateau and different vegetation cover type zones,respectively.We use linear trend analysis,F test,Pearson correlation analysis,multiple linear regression model,and path analysis to analyze the mechanism of temperature,precipitation,relative humidity,sunshine hours and other factors affecting vegetation NDVI in the whole area and different vegetation cover types from annual scale and seasonal scale.At the same time,using the statistical yearbook data of each province(autonomous region),select the population density,per capita GDP,grain production per hectare and other indicators,analyze the impact of human activities on vegetation cover from the county and the regional scale.We use CA-Markov model to predict the distribution of different vegetation cover types in the Loess Plateau in the next 10 years,finally.The main conclusions are as follows:(1)From 2000 to 2015,the types and growth conditions of vegetation cover in the Loess Plateau have changed greatly.The areas with large changes in vegetation cover types are mainly distributed in the transition zone between grassland and desert in the northwest of the Loess Plateau,the southwestern region,the northern Shaanxi region and the central Shanxi region.In general,the area of forest and grass vegetation in the Loess Plateau is expanding,but the growth rate is slower than that of urbanization,the crop area is reduced,and the mosaic type of vegetation is also greatly reduced.The average value of NDVI in most areas of the region showed asignificant increase in the four seasons of spring,summer,autumn and winter,and the increase rate was lower than 0.02/a in most areas.The annual increase of NDVI was74.25%,and the NDVI of different vegetation cover types also showed a significant increase trend.The evergreen broad leaf forests and crop increased the most,with an average annual average of 0.012/a.Since the performing of “ Grain to Green”program,vegetation cover in the Loess Plateau has been greatly improved.(2)The response of different vegetation cover types NDVI to climatic factors on the annual scale can reveal the mechanism of climate change on the Loess Plateau.Vegetation cover types dominated by trees,such as evergreen broad leaf forests,deciduous broad leaf forests and evergreen needle leaf forests and mosaic trees were significantly affected by mean annual relative humidity and mean annual sunshine hours.While grass dominated types,such as grassland and mosaic grassland,were significantly affected by total annual precipitation.The influence of temperature on vegetation is not significant either on the whole regional scale or on the different vegetation type zones.(3)Precipitation in the whole area and different vegetation cover type zones were positively correlated with NDVI.It indicated that the vegetation in the Loess Plateau had a relatively consistent response to precipitation.Sunshine hours were negatively correlated with vegetation NDVI in the whole region,and only positively correlated with shrub NDVI in different vegetation cover type zones.The temperature of the area with low precipitation,low relative humidity and long sunshine hours were negatively correlated with NDVI.The temperature of the area with high precipitation,high relative humidity,and short sunshine hours were positively correlated with NDVI.The response of different vegetation cover types to relative humidity is completely opposite to the temperature.The relative humidity is high where the relationship between relative humidity and NDVI was negative;and the relative humidity is low where the relationship between relative humidity and NDVI was positive.It indicated that the range of relative humidity exceeds the tolerance limit of vegetation growth.(4)At the county scale,the annual NDVI of each county showed a significantincrease trend;the population density of most areas showed a significant increase trend;the per capita GDP of most counties in the whole district showed a significant increase trend.The grain production per hectare has been a significant increase since2001.Most of the regions with significant correlations among population density,per capita GDP,and grain production per hectare and NDVI with significantly positively correlated.At the regional scale,NDVI has a relatively consistent response to various socio-economic indicators.In the counties,per capita GDP,population density,and grain production per hectare were significantly positively correlated with NDVI.The regression equations between population density,per capita GDP,grain production per hectare and NDVI were extremely significant at p < 0.001.The regression equation coefficients were 3.2E-03,3.1E-06 and 0.079,indicating that NDVI is most sensitive to the changes of grain production per hectare,followed by population density,and the response weight to the changes of per capita GDP is lower.(5)The model prediction results showed that the distribution area of vegetation cover types such as crop,mosaic grassland,shrub,mosaic tree,evergreen broad leaf forests and deciduous broad leaf forests in the Loess Plateau is reduced in a small range compared with 2015.The amplitudes are all below 3%.The reduction in bare land is larger,occupying 11.7% in 2015.Compared with 2015,the grassland and evergreen needle leaf forests increased slightly,less than 1%.The distribution of deciduous needle leaf forests and city increased significantly,with the range of45.78% and 111% in 2015.