Multiscale Spatial Relationship Between NDVI And Environmental Variables In The Upper Reaches Of The Yangtze River Basin

The ecological environment of the upper reaches of the Yangtze River basin is related to the ecological security of economically developed areas in the middle reaches,areas with rich biodiversity in the lower reaches,and areas with severe flood disasters in the Yangtze River Delta.In order to explore how climate,terrain,and soil jointly shape the spatial relationship of multi-scale vegetation in the upper reaches of the Yangtze River basin.Based on the background of vegetation types in the study area,this study selected the Normalized Difference Vegetation Index(NDVI)data from 2000 to 2019,combined with corresponding environmental variables,and uniformly divided them into 2000 to2009(D1)and 2010 to 2019(D2).A method framework combining two-dimensional discrete wavelet transform(2D-dWT)and geographic weighted regression(GWR)was constructed.Under this framework,we will explore the spatial relationship between NDVI and environmental variables at multiple scales in the upper reaches of the Yangtze River basin,and further divide the research area into six sub watersheds for key analysis to detect the impact of various environmental variables on NDVI.The following research conclusions were obtained:(1)The analysis of the multi-scale spatial distribution characteristics of NDVI shows that,in terms of approximate components,the annual average of NDVI is basically consistent with the spatial distribution pattern of elevation.Areas with elevations above4600 meters have the lowest annual average of NDVI in the previous ten years;In terms of detail components,there is a clear spatial distribution pattern,showing significant spatial similarity with terrain and temperature factors.In terms of approximate components,except for the edge and northwest of the Sichuan Basin,the annual average value of NDVI in the entire study area has significantly increased.In terms of detail components,the regions with significant spatial fluctuations in the NDVI annual average of D1 are concentrated in the southeastern and northwestern parts of the Sichuan Basin;During the D2 period,it was concentrated in the central Hengduan Mountains area of the study area.(2)The correlation analysis between multi-scale NDVI and environmental variables found that the dominant environmental factors are temperature,precipitation,and elevation.On the annual average,the first dominant factor in the previous and last ten years is controlled by temperature and precipitation,respectively.The correlation difference between the two on a scale is relatively small,which can be seen as having a strong impact on NDVI in the study area together.In terms of annual variation,the NDVI of D1 has the highest correlation with pH at the original scale and approximate components,reaching0.311;The NDVI of D2 has the highest correlation with slope,reaching 0.513.The NDVI impact of sand particles on D1 in this area is greater than that of clay particles,and the NDVI impact of clay particles on D2 in this area is greater than that of sand particles.The correlation of organic matter remained basically consistent over the past decade.In terms of detail components,NDVI follows a consistent pattern with terrain and climate.(3)Based on the constructed 2D-dWT and GWR research frameworks,the spatial relationship between multi-scale NDVI and environmental variables is mainly manifested as: in terms of annual average values,temperature,elevation,and pH are the most important factors affecting the spatial relationship of NDVI in the upper reaches of the Yangtze River basin.In terms of approximate components,the impact of environmental variables on NDVI did not significantly change with scale decomposition.In terms of detail components,the spatial fluctuation of D2 is much greater than that of D1.The positive impact of sand and clay particles on NDVI varies under different vegetation growth states.When vegetation is sparse,organic matter and pH have a greater impact on vegetation.In terms of annual variation,precipitation has a stronger control over vegetation growth than temperature,and elevation has a stronger impact on vegetation growth than slope.In ecologically fragile areas,the annual variation of NDVI in D2 has a strong positive impact compared to D1.As plant growth continues to improve,its negative correlation coefficient significantly decreases.(4)The impact of environmental variables on NDVI in each sub basin is different from that of the entire upper reaches of the Yangtze River basin.In terms of approximate components,the temperature and organic matter effects of D2 significantly increase,forming a situation of jointly controlling the annual average NDVI in each sub basin with elevation,precipitation,or slope.The influence of sand and clay particles is relatively small and only manifests on a large scale.The degree of influence of pH varies greatly among different sub watersheds.In terms of detail components,the scale effect of NDVI annual average and organic matter is significant in all four sub watersheds.The optimal scale effect of precipitation,temperature,and elevation is on a scale of 4 or 5.In small watersheds with complex terrain and significant terrain fluctuations,the most suitable scale decomposition is slightly larger than in flat areas.Overall,the optimal 2D-dWT decomposition scale for the NDVI annual average of D2 has decreased compared to D1.In terms of approximate components,the decomposition of most environmental factors and NDVI annual changes to a scale of 4 is sufficient.However,the decomposition of organic matter to a scale of 5 still exhibits good regularity.The larger the change in NDVI,the better the vegetation growth tends to be,and the vegetation growth in the watershed shifts from precipitation control to more temperature control.In terms of detail components,all environmental factors,except for organic matter,show good relationships when decomposed to scale 2,while the impact intensity gradually weakens with further decomposition of 2D-dWT.