Variations of sand transportation rates in sandy grasslands along a desertification gradient in northern China

To assess the variation of sand transportation rate in sandy grasslands at different levels of desertification, sand transportation rates and wind speeds were concurrently measured along a desertification gradient during two windstorm events. Some surface properties (vegetation cover, plant height, soil surface hardness and surface soil moisture) were also measured to quantify their influence on sand transportation rates. Rates increased linearly from the least desertified fixed sandy land to the severely desertified mobile sandy land on both measurement dates. The rate decreased with height by an exponential or power function. Most of the trapped sediments occurred within 12 cm above the surface, suggesting that sand transportation and deposition are near-surface processes. Dry particle size analyses showed that most particles were 0.25–0.1 mm in size, followed in decreasing order by particles of 0.5–0.25, 0.1–0.05, >0.5 and <0.05 mm. Differences in sand transportation rates among the four sandy lands were largely attributable to differences in their surface properties. Vegetation cover was most closely related to transportation rate, followed by soil surface hardness. Plant height and surface soil moisture content were relatively unimportant. Positive relationships between vegetation cover or soil surface hardness and surface roughness and negative relationships between vegetation cover or soil surface hardness and mean wind speed at 20-cm height suggest that the direct effect of vegetation cover or soil surface hardness is increased surface roughness. This resulted in more absorption of the wind momentum, lower surface wind speeds and less wind erosion. A multivariate predictive model based on measurements of vegetation cover, soil surface hardness and mean wind speed at 20-cm height was developed by regressing surface roughness length on vegetation cover, soil surface hardness and mean wind speed at 20-cm height. It had a high determination coefficient (R2) of 0.99. The model can be used for predicting the surface roughness length of the experimental sandy grasslands, and is also useful for testing the combined effects of vegetation cover, soil surface hardness and mean wind speed at 20-cm height on changes in surface roughness length.