Study On The Scale Response Between NDVI And Environmental Factors In Wushen County Of Mu Us Sandland

Scale is the key to correctly understand the interaction between patterns and processes. Dependencies relationship between pattern, process and scale is one of the core issues in landscape ecology. The patterns and processes in response to disturbances in different scales is the main source of the inherent and complex heterogeneity of natural systems in time and space.Mu Us Sandland located in ecological transition zone from desert or desert steppe to forest steppe. The process of desertification plays a significant role in pattern changes of sand ecosystems in different time and space scale, which is an ideal site to carry out the research of desertification processes and its driving mechanism with different spatial and temporal scales. The landscape, vegetation patterns of spatial heterogeneity and its response to climate and the topographic factor of Mu Us Sandland under different spatial and temporal scales would be systematicly studied, which would help to understand and explore the progress of desertification process and driving mechanism, to scientificly monitor and evaluate the dynamic of desertification and to provide the basis for the formulation strategy of desertification prevention.Taken Wushen County in the hinterland of Mu Us Sandland as objects, screened out the suitable spatial interpolation method of climatic factors of local area and analyzed it’s periodic oscillation characteristics of interannual variability, based on the theory of ecological and applied mathematics and on the3S technology, set five re-sampling transects from north to south and from east to west (totally17328pixel), data about vegetation index (-1≤NDVI≤1) with vegetation coverage (NDVI>0) and the principal climate factor with basic topographic factor in1990,2005and2009in Mu Us Sandland was get. Using analysis methods of the detail coefficients of wavelet multi-scale correlation and stepwise multiple regression at original scale, spatial patterns of vegetation-climate-topography and response law at different scales were analyzed. The following conclusions were getting:(1) Under the conditions of the existing data density, data about precipitation and relative humidity of Inner Mongolia was appropriate to adopt the Ordinary Kriging with tetraspherical and rational quadratic as semivariogram model of the raw data for optimal spatial data interpolation method, Inverse Distance Weighting with p=3is optimal spatial interpolation method about average annual atmospheric temperature. (2) The annual average temperature of Mu Us Sandland in Inner Mongolia in the past41years increased by0.401℃/10a, the trends of annual precipitation and annual mean relative humidity are not obvious. Complex nested structure was existed in periodic oscillation of climate change at multiple time scales.(3) The characteristic scales of regularity in vegetation index (-1≤NDVI<1) with vegetation cover and the principal climate factors with basic topographic factors getting by the wavelet multi-resolution analysis was shown as follow: vegetation index (-1≤NDVI≤1), vegetation cover, the spatial heterogeneity of the major climatic factors and basic topographic factors increased with scale-up and tend to express at large-scale. The spatial patterns of the vegetation index (-1≤NDVI≤1) and vegetation cover (NDVI≥0) were of a complex non-linear characteristics. No characteristic scale was found with climatic factors and elevation of the study area in the research scales. The characteristic scale of the slope and aspect were complex in spatial distribution.(4) In the original scale, the results of multiple stepwise regression analysis of vegetation cover and climate with topographic factors showed that: in the year with high yield of precipitation, the effects of precipitation, relative humidity, temperature on the vegetation cover were stronger than the effect of the topographic factor. While in the year with low yield of rainfall, the effects of local topographic factors of slope and elevation on the vegetation cover were stronger than the effect of climatic factors.(5) There are complex characteristics in scale between vegetation index (-1≤NDVI≤1) with the vegetation cover and climate with topographic factors. The maximum correlation scale was often observed in middle and large scale. This indicates that the role of meteorological and topographic tend to show at the macroscopical level. The sensitive scale of temperature and vegetation index (-1≤NDVI≤1) with vegetation cover was small. Precipitation and vegetation index (-1≤NDVI≤1) with vegetation cover were significantly or extremely related with each other at multi-scale. Vegetation index (-1≤NDVI≤1) was related with slope and aspect at macro-and micro-scale. There are multiple relationships in scales between vegetation cover and elevation with aspect, which is helpful to reveal the patterns and processes in response to scales, to understand the role of various factors at different scales during the process of pattern formation.(6) The process of desertification was closely related with scale and pattern. There were differences in the process of desertification in spatial and temporal scales. This difference was indicated in the changes of desertification in sensitive scale and in the driving factors. To Mu Us Sandland, it is difficult to measure the desertification process and driving mechanism under a unified scale in fixed or semi-fixed sandland and bottomla- nd with the valley since which is impacted by wind or water erosion and salinization. So the desertification monitoring and evaluation should be done in multiple scales, which would provide a basis for drafting the control strategies scientifically and rationally.