The WIND-O-V field experiment: WIND erOsion in presence of sparse Vegetation

International audience; Wind erosion in semiarid areas is a major threat for the soil productivity as it impoverishes soil in organic matters and nutrients. Compared to desert regions, these regions are characterized by sparse seasonal vegetation that impacts the erosion process. Semiarid areas face two major evolutions that may modify their wind soil erosion in the future: (1) climate change, with a modification of the amplitude and frequency of precipitations, affecting the surface vegetation cover, and (2) population growth, generating a considerable human pressure on the land use. Characterizing wind erosion in such complex regions is, therefore, crucial and challenging. In order to better predict the amount and composition of emitted dust from semiarid areas, a novel field experiment named WIND-O-V (wind erosion in presence of sparse vegetation) has been performed in 2017 and 2018 in South Tunisia. The originality of this experiment is (1) to cover successively a plot without and with sparse vegetation, and (2) to combine detailed measurements of wind dynamics (including turbulence), size-resolved saltation and dust fluxes, and erosion-flux compositional fractionation along the soil-saltation-suspension continuum. The experiment took place from March to May in the experimental range of the Institut des Régions Arides (IRA) of Médenine (Dar Dhaoui, 25 km east of Médenine). The site approximates a flat half-circle plot of 150 m radius where measurements were performed at the center of the circle in order to ensure a fetch of at least 150 m. In 2017, the surface has been tilled with a disc plough and levelled with a wood board in order to meet the conditions of an ideal flat bare soil without soil crust or ridges. In 2018, sparse vegetation consisting of barley tufts have been grown on the plot with a 3.3 m wide regular arrangement. Three types of measures were carried out. Meteorology: on a 9 m high mast erected at the center of the plot, turbulent velocity components and air temperature fluctuations were measured simultaneously at 1.0, 1.9, 3.0, and 4.1 m height using four 3D sonic anemometers sampling at 60, 50, 50, and 20 Hz, respectively. On the same mast, 7 cup anemometers (0.2, 0.6, 1.3, 1.8, 3.0, 4.0, 5.2 m) and 4 thermocouples (0.4, 1.6, 3.7, 5.0 m) were also installed to measure simultaneously at 0.1 Hz the mean horizontal wind velocity and temperature profiles, respectively. Three additional 2D sonic anemometers were installed in 2018 around a barley tuft to characterize the wind around the vegetation. Roughness length of the surface and friction velocity were computed on the bare plot case by comparing the Law-of-the-wall and Eddy-Covariance methods. Saltation flux: one vertical array of 5 sediment traps like Big Spring Number Eight (BSNE) was deployed to quantify the saltation flux and its size distribution. The modified BSNE had a 5 times wider opening area to collect larger sediment quantities, allowing sequential (in time) sampling of individual erosive events and guarantying the possibility of applying size resolved analyses. Saltation flux measurements with a better temporal resolution were thus associated with more stable friction velocity conditions. In 2018, 5 MWAC masts were added to measure the spatial variability of the flux due to the sparse vegetation. A Saltiphone and a camera gave information on the beginning, end, and duration of erosive events. Dust flux: for the first time size-resolved dust fluxes were estimated from both the traditional flux-gradient approach and the eddy covariance approach. For the first approach, mass and size resolved number concentrations were measured at two levels (2 and 4 m). To that purpose, two TEOM microbalances and two optical particle counters (WELAS Promo 2300) were used. Both sensor-types were connected to omnidirectional air sampling inlets. The WELAS monitored at 1 Hz the dust concentrations per size class (32 classes between 0.3 and 17 μm). For the second approach, a third WELAS was coupled to the 3 m high sonic anemometer in order to correlate the size-resolved dust concentration and the vertical wind velocity fluctuations. Finally, the chemical composition of dust fluxes was estimated from the sequential sampling of dust particles at two levels (2 and 4 m) with online filters equipped with inlets of different size cutoffs (20, 10, 2.5 and 1 μm).