Spatio-temporal Evolution And Driving Force Analysis Of NDVI In Longmen Mountain Area

Vegetation is the "indicator" of ecological environment.Studying the trend and driving force of vegetation change is helpful to reflect the state of ecological environment.The relationship between vegetation and climate,socio-economic,topographic and other factors can be discerned by monitoring vegetation change,distribution and driving forces.In this paper,the change trend analysis method is used to analyze the spatial and temporal characteristics of MYD13Q1-NDVI changes in different vegetation and geomorphological types month by month,quarter by season and year by year from 2002 to 2019.Based on the change trend of NDVI in the growing season,the correlation analysis method is used to discuss the relationship between NDVI and climatic factors.Random forest regression model and geographically weighted regression model are used to analyze the relationship between climatic factors,topographic factors,social economy and NDVI.Combined with the results of the two models,the main driving forces of NDVI in Longmenshan area are analyzed comprehensively,in conclusion:(1)The monthly NDVI in Longmen Mountains fluctuates with time,and the change trend was basically the same,which was in the form of normal distribution curve.The seasonal mean NDVI was: growing season > summer > Autumn > Spring,the change of NDVI was similar in different years,but there was significant difference in local areas.The average annual and seasonal NDVI in the Longmen Mountains area showed a decreasing trend with the increase of elevation.(2)The interannual change of NDVI in different seasons showed a significant improvement trend.The proportion of significant and extremely significant improvement in spring was relatively large,and most areas of NDVI in the study area were at an increasing level,the proportion of significant improvement in summer was less than that in spring,and the distribution of extremely significant degradation areas was concentrated.In autumn,there was a marked difference between east and west,and vegetation improvement was distributed in the southeast of the study area,the vegetation degradation was mainly concentrated in the northwest of the study area,and the change of growing season was similar to that of summer.(3)The geomorphologic type showed a remarkable differentiation to the vegetation change,and the significance of different vegetation types was also different.Shallow cutting of low mountains had the highest vegetation improvement area ratio,followed by shallow cutting of Zhongshan and deep cutting of Zhongshan,the vegetation improvement area ratio of shallow cut low mountain was the highest,followed by shallow cut Zhongshan and deep cut Zhongshan,and the vegetation degradation of middle cut Zhongshan was the highest,followed by middle cut low mountain and very high mountain,and the vegetation improvement area ratio of high mountain was the highest,in the second place,there were grass and broad-leaved forest,and the grass had the highest rate of degradation,followed by alpine vegetation and shrub.(4)Based on the correlation analysis of 18-year NDVI,annual precipitation and annual average temperature in Longmen Mountains,the results showed that 1.11% of Longmen Mountains was driven by precipitation,6.77% was driven by temperature,and 5.19% was driven by precipitation and temperature.(5)Stochastic forest regression model and geographically weighted regression model were used to analyze the relationship between NDVI and 12 impact factors:(1)The results of stochastic forest regression model showed that the summation of the effects of natural factors presents the spatial pattern of east-west differentiation as a whole,and that of human factors presents the pattern of east-west differentiation as a whole,the ratio value of man-made natural factors showed the pattern of east-west differentiation.Among them,land use types had the greatest impact on NDVI,secondly,DEM,annual mean temperature,soil type,annual precipitation,nuclear density of geological hazards,roads euclidean metric,slope,night light rate,water euclidean metric,slope direction and slope position.(2)The results of geographically weighted regression model showed that,the influence of natural factors on the whole presents a pattern of east-west differentiation,the high value of human factors was concentrated in the southwest of the study area,the ratio of human factors was medium high,and the low value of northwest and southeast.Among them,the biggest impact on NDVI change was the core density of geological hazards,it was followed by land use type,soil type,annual mean temperature,annual precipitation,DEM,night light rate,road euclidean metric,water euclidean metric,slope,aspect and slope position.(6)Combining the results of geographically weighted regression model and stochastic forest regression model,the influence factors of different landforms and vegetation types were quite different.Deep-cut,medium-cut,shallow-cut,and very-high-mountain areas were less affected than other geomorphologic types,and grass,alpine vegetation,and meadows were less affected than other vegetation types.NDVI changes were not simply by a factor,but in a variety of driving factors under the joint effect.