Spatial Variation Of Aeolian Sediment Deposition In Patchy Vegetation Plot Of Desert-Oasis Ecotone In The Middle Reaches Of The Heihe River And Its Scale Effect

Soil texture coarsening, nutrient loss and productivity decline caused by wind erosion in the process of sand activities, will affect the vegetation growth in the desert-oasis transition zone. Once deteriorated, patchy vegetation will rarely restore and severely influence the ecological protective function of desert-oasis transition. In order to explore the principle of sand movement in the desert-oasis transition, combining classical statistics and geo-statistics method, we analyzed the spatial heterogeneity of aeolian sediment deposition from sample, belt transect to the region and its causes of space mutation. While we explored the influence of its coefficient of variation, its semivariable function and its Moran's I, when the sampling amplitude, sampling density and scale transformation between sample, belt transect to the region changed. We conclude that:?1? On the sample, belt transect and the region scale, Spatial variation characters of aeolian sediment deposition were different, especially in A belt transect, B belt transect and C belt transect. In general, the aeolian sediment deposition distributed within the low content, which accounts for 38.62% in the random variation and 54.38% in the autocorrelation variation. Its semivariable function model were Linear model, Gaussian model, Exponential model, Spherical model and Pure nugget model. Its spatial distribution was a grid in the random variation and a zonal in the autocorrelation variation.?2? Some factor changed with the season, such as air temperature difference, precipitation, wind etc. can cause the spatial variation of aeolian sediment deposition. Air temperature difference promote it in the random variation, but restrain in the autocorrelation variation. Precipitation restrain it in the random variation, but promote it in the autocorrelation variation. The effect of drift potential?DP? in the random variation was outstrip that in the autocorrelation variations to the spatial variation of aeolian sediment deposition. But resultant drift potential?RDP? was the opposite of drift potential?DP?. Resultant drift direction?RDD? was E-S in the random variation and W-N in the autocorrelation variation. Soil fine particle was the highest content in patchy vegetation plot of desert-oasis ecotone, and was also the main particle of aeolian sediment deposition. Vegetation, terrain and internal topography can form the grid unit, which have a important for the variation of aeolian sediment deposition.?4? In the random variation or autocorrelation variation, changes of sampling amplitude and sampling density can cause the scale effect of aeolian sediment deposition. The coefficient of variation?C_v? was in "M" type change as the sampling amplitude increases, which was middle variance in the 220m×220m. The coefficient of variation?C_v? was larger difference within the small density and kept stable within the large density. The semivariable mode and the nugget coefficient C₀/?C₀+C? have the directional trends that were opposite as the sampling amplitude increases. The semivariable mode and the nugget coefficient C₀/?C₀+C? kept stable as the sampling density increases. The ranges?A? were different when the sampling amplitude and sampling density changed. In the random variation, the Moran's I represented the "gather-scattered" change. While in the autocorrelation variation, the Moran's I represented the "gather – scattered- gather" change as the sampling amplitude increased, and the Moran's I represented the "gather – scattered- gather" change as the sampling density increased.?5? When the scale of sample, belt transect and the region transformed, the coefficient of variation?C_v?, the semivariable model, the nugget coefficient C₀/?C₀+C? and Moran's I were stable in the sample and region, and big difference.