| Soil erosion is now a global environmental problem and it is particularly so on LoessPlateau of China. Vegetation has long been an effective measure to control soil erosion, whichhas aroused universal concern of numerous scholars. In this study, indoor simulated rainfallexperiments were conducted to explore the effects of vegetation (Artemisia capillaris)patterns on soil erosion and hydrodynamic characteristics of slope flow. Runoff and sedimentreduction benefit of both above-and under-ground parts of Artemisia capillaris, as well as therelationships between vegetation parameters and runoff and sediment rate were furtheranalyzed to reveal the mechanism of vegetation reducing erosion. In addition, we studied thespatial variability of slope erosion coupled with geostatistics method. Our findings canprovide theoretical and data basis for optimal allocation of vegetation measures and thoroughunderstanding of the soil erosion processes and dynamic mechanism of slope flow in LoessPlateau. The main results are as follows:The patterning of Artemisia capillaris could effectively control soil erosion. Comparedwith the bare plot (CK), runoff and sediment rate in patterned plots were reduced by7~25%and50~92%, respectively, indicting that patterned Artemisia capillaris had stronger effect onreducing sediment than reducing runoff. The aboveground parts of Artemisia capillariscontributed more on runoff reduction, ranging from53%to66%, while the underground partshad more contributions on sediment reduction, about51%~71%. Regression analysis showedthat runoff rate linearly (p<0.05) correlated with the aboveground biomass (SD), stem height(SH), and stem number (SN), respectively. The erosion rate had very significantly exponentialrelationships (p<0.01) with root density (RD), root surface area density (RSAD), root lengthdensity (RLD), and root area ratio (RAR), respectively. However, no relationship wasdetected between erosion rate and root diameter. Path analysis demonstrated that RLD and SNwere the main factors affecting runoff rate, and RSAD and RD were responsible for erosionrate changing. Therefore, increasing sowing density properly in order to increase the plantnumbers and influence the root indices, ultimately conserving soil and water. The findingsshowed that pattern BP and SP were more appropriate in soil and water conservation.The classical statistical analysis found that the variability of slope erosion was strong. The geostatistical analysis revealed that the spatial variability of slope erosion was fitted wellby spherical or exponential models and it showed an intermediate or strong spatialdependence. Kriging interpolation directly reflected the erosion or deposition area and theirspatial distribution. Bare plot was mainly distributed by erosion plaque (△h>0) and thedeposition plaque (△h<0) only occupied a small part. Erosion area of patterns BP and SPwere smaller than pattern LP, which was in accordance with the runoff and sediment yieldingresults. After the aboveground parts were removed at soil surface, erosion area increased andthe plaques presented more continuous distribution for different patterned plots.Cross-validation showed that kriged mean prediction errors were close to0androot-mean-square standardized prediction errors were close to1, thus indicating the kriginginterpolation were reliable.Both vegetation patterns and rainfall intensities had significant effects on flow velocity.Mean surface flow velocity increased with the increase of rainfall intensity and elevated in theorder of SP (BP)<LP <CK. For the same patterned plots, the flow velocities of lower slopeswere higher than the upper ones. Moreover, flow velocities in grass covering section werelower than bare ones. Removing the aboveground parts significantly increased the flowvelocity. Artemisia capillaris (both canopy and root) could reduce the flow velocity. Therewas difference in contribution rate of velocity reduction between the canopy and root underdifferent rainfall intensity conditions. The canopy contributed more for rainfall intensity of60and90mm/h and ranged from53%to99.8%; nevertheless, the root had great contribution for120and150mm/h rainfall intensity and varied from51%to80%. Mean flow depth increasedwith increasing rainfall intensity and it declined after the aboveground parts were cleared. Thetemporal changes in flow velocity and flow depth were similar, both increasing initially andthen tending to be stable. Renold numbers (Re) for different treatments were25~80,indicating the slope flow was laminar. Furthermore, Re increased significantly with theincrease of rainfall intensity. There was no statistical difference in Re before and after clearingthe aboveground parts. According to the Froude numbers (Fr), flow in bare plot was criticalflow or jet, however, in Artemisia capillaris patterned plots it was subcritical flow. Fr of thelower slope was higher than the upper slope. Both Darcy-weisbach (f) and Manning (n)friction coefficient can reflect the flow resistance features. Besides, both had similar changelaw: ranked in the order of CK <LP <BP (SP), grass sections higher than the bare section,value of upper slope higher than the lower slope, and both decreased after removing theaboveground parts. Soil erosion rate was a power function of mean flow velocity and flowresistance, respectively, while it had no obvious relationship to flow depth. Compared to theflow Re, soil erosion rate could be well described by flow Fr. Comprehensively considering the measured and predicted values of each hydraulic parameter, it could be found flowvelocity was an ideal parameter to simulate and estimate soil erosion rate.Artemisia capillaris patterns had no significant effects on hydrodynamic factors, yetvalues of the hydrodynamic parameters increased with increasing rainfall intensity. Graycorrelation analysis showed that the correlation degree between unit stream power (P) andsoil erosion rate was the highest (0.767), stream power (ω) and unit flow energy (E) rankedsecond,0.697and0.659, respectively. The correlation degree between flow shear stress andsoil erosion rate was the lowest (0.618). The results suggested that stream power is closer tothe growing trend of soil erosion rate, and it might have more influence on soil erosion.Regression analysis also showed linear relationships between the erosion rate and (unit)stream power.
|
PDF Full Text Request