| Surface roughness is one of the most important factors to affect soil erosion process,which is also the research hotspot in soil erosion science. Study the spatial heterogeneity ofsurface roughness of4typically loess tilled slopes during different erosive stages under amicro-scale (2cm×2cm) condition is significant for further understand the process of soilerosion. When slopes with4different gradients and tillage measures were prepared (artificialbackhoe, artificial digging, contour tillage and check), a erosive rainfall under a rainfallintensity of60mm/h was carried out. Soil surface roughness relative elevation data pointswere taken by a laser scanner and analyzed with both the semivariogram function and fractaldimension models. It can promote understandings and quantitative description of the couplingrelationship between the surface roughness and erosion, and it also can provide a theoreticalbasis to guide future tillage practices in the loess sloped regions(1) The results of classical statistical analysis showed that: The loess tillage slope surfaceroughness has a weak spatial variability. Which, the variability of the different slope was inthe order5°＞10°＞20°＞15°＞25°, With the increase of gradient, the spatial variability ofslope surface roughness tends to weaken; The variability of the different loess tilled was in theorder AD＞CT＞AB＞CK, it showed that AD slope has strong spatial variability than othercultivation measures slopes. In addition, different tillage measures slope for rainfall erosionhas certain inhibition, this inhibitory effect was in the order AD＞CT＞AB generally. Thus,AD farming slope has the good effect of soil and water conservation in loess area. In differentstages of erosion, the variability of surface roughness showed BR＞SpE≈ShE≈RE, and itindicated that rainfall erosion weakens the surface roughness.(2) The results of the semivariogram analysis indicated that the loess tilled slope surfaceroughness has a higher spatial autocorrelation, their spatial autocorrelation scale of surfaceroughness range from2.02m to3.59m. Which, the spatial autocorrelation of AD slopesurface in each rainfall erosion stage are characterized by strong spatial autocorrelationcharacteristic; the spatial autocorrelation of the different slope was in the order5°＞15°＞25°,but on the whole, the spatial autocorrelation of different rainfall stage did not show moreobvious regularity, this suggests that the artificial cultivation, the influence of the slope of thesurface roughness is greater than the influence of the rainfall. Different tillage measures, different slope slope surface from the relevant range is not thesame, seen from the point of different tillage measures, the spatial autocorrelation of AB slopeis3.59m, show that AB slope surface space continuity is better, the studies in2cm×2cmscales its spatial variation is not obvious, it needs to be studied more small scale. The effectiverange difference of AD slope and CK slope is small, effective, respectively2.33,2.21m,show that the ecological processes works on the same scale. The spatial autocorrelation of CTslope is1.54m, indicate that CT slope surface space continuity is worse, in future studiesshould take smaller sampling interval.From the point of different slope, The range of5°,10°,15°,20°,25°slope surfacewere2.05,2.52,3.82,1.68,2.02m. Overall, micro scale variation of terrain is2.42m.(3) The fractal dimension analysis results showed as follows: The loess tilled slopesurface roughness has good fractal features. In the research of scale range, it was notcompletely random, and it ranged from1.59to1.91. Its fractal features was in the orderCK(1.742)<AB(1.754)<AD(1.787)<CT(1.789), showed the variation of surface roughnessgradually enhanced on the small scale, its spatial heterogeneity tends to complex. From thepoint of slope, the fractal dimension of20°slope was the highest (1.81), followed by25°(1.79),15°(1.79),(1.76),10°(1.75),5°(1.73) slope reduced, With the increase ofgradient, the spatial distribution of slope surface roughness tends to complex, its spatialheterogeneity is more stronger. From the perspective of the fractal dimension value ofdifferent rainfall erosion phase, it was characterized by BR (1.78)> SpE (1.77)> ShE(1.76)> RE (1.75), suggested that along with the development of the erosion, the variation ofthe slope weakened on a small scale gradually, its spatial heterogeneity is more weaken, thespace configuration pattern of surface roughness is mainly controlled by variation in a largerscale range.(4) The anisotropy analysis results showed as follows: In the different process of watererosion, semivariance function of loess slope surface roughness has obvious directivity, showit is isotropic on the direction of90degrees and135degrees, the change of surface roughnessis consistent. On the direction of0degrees and45degrees, the semivariance function of loessslope surface roughness is anisotropic.(5) The spatial autocorrelation analysis showed that: From the point of different tillagemeasures, with the sampling spacing increasing, the spatial autocorrelation of the loess tilledslope surface roughness is plus or minus alternate phenomenon, the complexity of the spatialstructure of the slope was in the order CT＞AD＞AB＞CK, the results agree fairly well withthose obtained by the fractal dimension analysis. Overall, within the variation range of0to10cm, and the surface roughness show significantly positive spatial autocorrelation or negative spatial autocorrelation, this show that within the variation, The spatial heterogeneity ofsurface roughness was affected by random factors and structural factors work together, theresults agree fairly well with those obtained by the semivariagram analysis.(6) The results of kriging interpolation analysis showed that: The spatial distribution ofthe loess tilled slope surface roughness has the characteristics of obvious plaques orcontinuous distribution. Seen from the point of different slope, in the20degrees slope, thespatial distribution of the surface roughness is more complex, and present obvious patchshape distribution, and the difference is obvious; Seen from the point of different cultivationmeasures, in the CT slope, the spatial distribution of the surface roughness is more complex,and present obvious banded and patch shape distribution, shows that slope, cultivationmeasures will affect the change of surface roughness. |
PDF Full Text Request