干旱区生态环境知识资源中心

Models of the two aeolian processes (saltation and sandblasting) that lead to emission of fine dust particles (PM20) by wind erosion in arid and semi-arid areas have been combined to form the so-called ’Dust Production Model’ (DPM). In this model, the size dependent binding energies of PM20 embedded within the wind-erodible loose soil aggregates or in the soil surface itself are key input parameters. Indeed, their values condition at the same time the intensity of emissions and their initial size distribution. Previous comparisons of vertical mass fluxes measured on-field with the model predictions suggest that these energies might be relatively independent of soil texture and also probably composition. Because this would greatly facilitate application of the DPM at regional or global scale, the objective of this work is to check experimentally the veracity of this result. The strategy that has been designed for this has involved selecting four natural soil samples collected in various source areas of the world and covering a wide range of textures and compositions. Then, these soil samples have been used to perform carefully controlled wind erosion simulations in a wind tunnel. During the experiments, which were carried out at different wind speeds with each soil, the horizontal flux (F-h) of saltating soil aggregates has been monitored. At the same time, number concentrations (C-i) of PM20 released by the sandblasting process were recorded in the 6 size classes of an optical size analyzer. Thus, the efficiency of the sandblasting process (defined as the ratio of C-i to F-h) could be determined for each of these size classes. Analysis of the results obtained in similar saltation conditions shows that for the four tested soils, and within the range of contents in clay and other components favoring aggregation (mostly organic matter and carbonates), the influence of soil composition and texture on binding energies of the PM20 particles within soil aggregates is at best a second order effect that can be neglected in large scale modeling of wind erosion by the DPM. (c) 2007 Elsevier B.V. All rights reserved.