Impacts Of Groundwater Depth On Spatio-temporal Variations In Vegetation Index

In arid and semi-arid areas,groundwater exerts significant effects on vegetation growth.This influence can be represented by the relationship between vegetation index and groundwater depth.Previous studies have revealed the statistical relationship between vegetation index and groundwater depth via remote sensing data,but there are still some shortcomings.How does groundwater depth affect the spatial variability and the seasonal and interannual variation characteristics of vegetation index?This paper will focus on these unsolved ecohydrogeological problems.The paper selects the Ordos Plateau,a typical arid and semi-arid region,as the study area,in which the Hailiutu River Basin,the Nalin River Basin,the plateau lake region and the Dusitu River Basin are across the Mu Us Sandy Land.The resolution of the grid data of vegetation index(EVI)and groundwater depth(GD)is 250 m.Linear filtering,geostatistical semi-variogram model,general likelihood uncertainty estimation(GLUE)and other statistical models were used to analyze the spatial variability characteristics of EVI.Then the impacts of GD on the spatial variability in EVI at pixel and landscape scale were discussed.As for seasonal and interannual variations of EVI,we determined the effects of GD on annual EVI peak(EVIp)and peak time(Tp)and proposed the concept of climate elasticity of vegetation index.The climate elasticity of EVI during multi-scale drying and wetting periods was calculated,and the role of GD on it was also analyzed.The results show that:(1)EVI is positively correlated with elevation and air temperature,and negatively correlated with drought index when GD>15 m.The correlation between EVI and climatic factors and elevation varies significantly when GD?15 m,reflecting the influence of groundwater.(2)The spatial variability of relative vegetation index(W_EVI)removing the background influence of climate and elevation is affected by groundwater.The parameters of the semi-variogram have a nonlinear relationship with GD.When GD?5 m,the base value in the semi-variogram of W_EVI is obviously higher than that of larger GD,and the nugget value of W_EVIis proportional to that in GD.(3)Both the EVIp and the growing-season mean EVIa show non-linear decreasing trends with GD.When GD gets smaller,Tp gets more dispersed and increases the probability in the middle of July,suggesting more vegetation diversity.On the contrary,Tp in large GD areas concentrates in the middle and late August,meaning the relatively homogeneous distribution of vegetation.(4)The relationship between climate elasticity of vegetation index and GD is non-linear,and it depends on different time scales.In general,the climate elasticity of EVI increases with GD(i.e.the sensitivity is enhanced).The suitable depth is 1-7 m,3-10 m and 3-7 respectively from west to east with the lowest climate elasticity of EVI,meaning the strongest ecological stability and insensitive to climate change.In this paper,the mechanism of the GD effect on the spatio-temporal variation in vegetation is preliminarily analyzed.It is believed that this is the mixing result of the hypoxia stress and water stress in sandy vegetation with GD,thus the detailed processes need to be further studied.