Wind erosion prediction based on soil particle size characteristics and strong wind days

Abstract: Wind erosion is a severe environmental problem in arid, semi-arid region of the world, and it occurs frequently in northwest part of China and leads to soil degradation. Inner Mongolia is located in Northern China as a very important ecological barrier with various of landforms including desert, grassland and forest. In order to quantify the regional soil degradation degree and predict potential wind erosion risk, 5 typical sites of wind erosion, i.e. Jilantai Gebi, Ulan Buh desert, Mu Us sandy land, Wuchuan farmland and Xilamuren Grassland were chosen as experimental sites. Based on digital image technology, the pictures of these 5 sites were taken using Canon EF24-105 mm, and the grey scale information of the images was analyzed by Erdas-Image software; an area compensation equation was established to obtain the actual particle size, and then the residual amount of coarse surface was calculated. On the basis of wind erosion estimation model of Inner Mongolia Houshan Area, surface un-erodible particle (>0.84 mm) was chosen as key indicator combined with local strong wind days, and then an applicable soil erosion potential risk exponential equation was established and the soil potential wind erosion risk was evaluated. The results suggested that wind erosion was a physical process of erodible particle loss, as a result, serious degradation surface kept a high content of coarse grains. Jilantai Gebi and surface around shrub has experienced severe wind erosion, for there existed a large proportion of coarse particles, indicating that the area was in late period of the coarsening process with strong anti-erosion ability, and the average anti-erosion indices were 91.40% and 81.40% respectively. Soil particles of Ulan Buh Desert showed obvious spatial heterogeneous characteristics due to the effect of few shrub communities, and the soil anti-wind erosion ability was Artemisia xerophytica Krasch plot > Artemisia arenaria DC.plot > Nitraria tangutorum plot. In Mu Us Sandy land, aeolian material was the main component and the vegetation coverage was less than 5%; the area was dominated by barchan, and soil particle content showed intensive spatial heterogeneity with variation coefficient of 37%, which was mainly caused by different position of sand dunes. The anti-erosion index was only 13.40% indicating that wind erosion occurred easily in this site. In terms of typical farming land, surface fine material was maintained at a high level due to frequent plow activities, which led to a lower anti-erosion index of 29.40%; on the contrary, due to long-term wind erosion, much fine particle of abandoned farmland got lost, at last soil surface was covered by coarsening grains mostly, and the anti-wind erosion index was greater than that of farming land. Soil particle turned to show obvious spatial heterogeneity in deserted grassland, un-erodible particle decreased along with the distance far away from the tourist area, and the regional average anti-erosion index was 45.20%. Poor soil and strong wind were the main factors contributing to the sand dust in the tested areas. Based on the potential wind erosion risk equation, Jilantai Gebi was in the lowest potential wind erosion risk among the 5 sites, and it showed that soil surface reached a new stable balance. The potential wind erosion risk of shrub communities in Ulan Buh Desert was contrary to anti-wind erosion ability. However, based on the analysis effect, potential wind erosion risk was farmland > abandoned land > shift sand dune, which seemed to be a little bit abnormal. This was because farmland surface was looser than others, and there were different strong wind days among these areas. As a simple prediction equation, it was applicable in different sites of arid areas, especially in small scale predication, as was shown in the results. Farmland and grassland were in higher wind erosion risk and was becoming a new sand source, and human disturbance played an important role in this process.