Search

International audience; Estimating accurately dust emission flux during aeolian erosion events is crucial for quantifying the amount of dust in the atmosphere. The rare existing field experiments quantifying such flux were mainly performed using the flux-gradient (FG) method. Here, we present the first intercomparison of the size-resolved dust fluxes estimated by both the FG and the eddy-covariance (EC) methods during several erosion events. Both methods were applied simultaneously during the WIND-O-V (WIND erOsion in presence of sparse Vegetation)'s 2017 field experiment over an isolated erodible bare plot in South Tunisia. Overall, both methods predict similar dust fluxes for particle smaller than about 4 μm. For coarser particles, the EC method predicts a smaller dust flux than the FG method. Factors explaining this difference are discussed such as the different sampling heads used by the dust particle counters of both methods, or the possible weakening of the constant dust flux layer at the location of the upper dust particle counter of the FG method. This intercomparison highlights the difficulties and advantages of each method as well as their complementarity.

Predicting the particle size distribution (PSD) of near-surface turbulent dust flux ( F wc ) is a key issue for estimating the size of atmospheric mineral dust. Existing dust emission schemes differ in their parametrization of the emitted dust ( F emi ) PSD, defining differently the surface inter-particle cohesive force and the influence of wind intensity. Moreover, these schemes have often been validated-fitted against field measurements, assuming PSD similarity between F wc and F emi . Here, we investigate numerically the main factors influencing F wc -PSD during erosion events. To this effect, we developed a 1D dust-dispersal model. After evaluating the model against published results, it is shown that F wc -PSD is influenced by both deposition and F emi -PSD. This latter one is shaped by the inter-particle cohesive bond exponent and the surface dust PSD. A time-to-space conversion of the dust flux variations reveals an increasing enrichment of F wc in small particles compared to F emi . This enrichment remains lower than a few percent of the total dust flux (in number) for fetch lower than 100 m, but it can rise to more than 10% for fetch longer than 1 km. This fetch dependence of F wc -PSD is explained by the slow deposition of particles having the lowest deposition velocities. Importantly, this difference between F wc and F emi PSDs is accentuated with wind intensity, with F emi -PSD dominated by particles with large deposition velocities, and in presence of a large-scale background dust concentration. The role played by the deposition process in shaping the F wc -PSD should be considered when evaluating dust emission schemes against near-surface field measurements.

Soil deposition is an essential pathway for tropospheric ozone (O3) removal, but its controlling factors remain unclear. Here, we explored the variability of soil O3 resistance in response to soil texture. To this aim, data of O3 deposition over bare soil obtained from micrometeorological measurements under contrasted meteorological conditions for five sites were used. The results obtained are twofold: (i) soil resistance (Rsoil) increased with soil surface relative humidity (RHsurf), but (ii) this relationship exhibited large site-by-site variability. Further analysis showed that the minimum soil resistance (corresponding to completely dry soil surface or RHsurf = 0%) and the increase of Rsoil with RHsurf are both linked to soil clay content. These results can be explained by (i) the soil surface available for O3 deposition at a microscopic scale which is a function of the soil specific surface area, and (ii) the capacity of a soil to adsorb water according to its clay content and therefore to reduce the surface active for O3 deposition. From these results, a new parameterization has been established to estimate Rsoil as a function of RHsurf and soil clay fraction.

International audience; During two growing seasons of a maritime pine stand, in 2014 and 2015, ozone (O-3) fluxes have been determined using the eddy covariance (EC) method and compared to the outputs of a big-leaf O-3 deposition model including stomatal, cuticular and soil pathways. The model developed in this study generally allowed to properly reproduce the measured ozone deposition. Ozone fluxes showed a strong reduction during two water stressed periods in September 2014 and July 2015. The model partly explain this fall due to the reduction of stomatal deposition. Despite this stomatal closure, measured O-3 fluxes presented systematically lower negative values than the model outputs, and sometimes even positive values around midday during periods marked by strong water stress. In other words, the difference between observed and modelled O-3 fluxes (hereinafter referred to as the residual O-3 flux) is systematically positive on daytime during these water-stressed periods. This positive residual flux traduced the existence of an O-3 source below the flux measurement level, responsible for positive fluxes that counterbalance deposition fluxes. We developed an O-3 production module based on a terpene emission algorithm and an OH concentration proxy, to try to explain the observed ozone production. As this parametrisation allowed us to reproduce well the daily and inter-daily dynamics of the residual O-3 flux, it confirms that the latter actually resulted from O-3 production processes. This ozone production is here highlighted for the first time using O-3 fluxes measurements by the EC method. The chemical reactions possibly involved in O-3 production processes in this maritime pine forest have been discussed and different mechanisms are proposed, based on peroxy radicals chemistry or stress-induced BVOCs.

Ozone is a potentially phyto-toxic air pollutant, which can cause leaf damage and drastically alter crop yields, causing serious economic losses around the world. The VULNOZ (VULNerability to OZone in Anthropised Ecosystems) project is a biology and modeling project that aims to understand how plants respond to the stress of high ozone concentrations, then use a set of models to (i) predict the impact of ozone on plant growth, (ii) represent ozone deposition fluxes to vegetation, and finally (iii) estimate the economic consequences of an increasing ozone background the future. In this work, as part of the VULNOZ project, an innovative representation of ozone deposition to vegetation was developed and implemented in the CHIMERE regional chemistry-transport model. This type of model calculates the average amount of ozone deposited on a parcel each hour, as well as the integrated amount of ozone deposited to the surface at the regional or country level. Our new approach was based on a refinement of the representation of crop types in the model and the use of empirical parameters specific to each crop category. The results obtained were compared with a conventional ozone deposition modeling approach, and evaluated against observations from several agricultural areas in France. They showed that a better representation of the distribution between stomatal and non-stomatal ozone fluxes was obtained in the empirical approach, and they allowed us to produce a new estimate of the total amount of ozone deposited on the subtypes of vegetation at the national level.

International audience; Improving the understanding of processes related to atmospheric particle sources is essential to better assess future climate. Especially, how biogenic volatile organic compounds (BVOCs) are involved in new particle formation (NPF) is still unclear, highlighting the need for field studies in sites that have not yet been explored. Weakly anthropised, mostly composed of maritime pines (known as strong monoterpene emitters), vast and under the influence of sea spray inputs, the Landes forest (located in the southwestern part of France) is a suitable ecosystem to explore these questions. The aim of the present work was to investigate for the first time NPF in the Landes forest, and to identify the conditions for NPF. During a field campaign conducted in July 2015, clear NPF was observed during nighttime, at a high frequency rate (37.5%), whereas only two daytime episodes were observed. Growth rates during NPF events were in the range 9.0-15.7 nm h(-1), and nucleation rates (J(10)) in the range 0.8-8 particles cm(3) s(-1), typically in the range of reported values from rural sites. Nocturnal NPF started at sunset, lagging the reductions of temperature and ozone concentration as well as the increase of relative humidity, atmospheric stability and monoterpene concentration. We established that NPF occurred during more stratified atmosphere episodes, reflecting that NPF is more influenced by local processes at the Landes forest site (Bilos). Concentration of the sum of monoterpenes, here mainly alpha- and beta-pinene, was observed to be maximal during NPF episodes. On the contrary, ozone concentration was lower, which may indicate a larger consumption during nights where NPF episodes occur. Results strongly suggest the contribution of BVOC oxidation to nocturnal NPF, in both nucleation and the growth stages.

National audience

Understanding how ozone is deposited on vegetation canopies is needed to perform tropospheric greenhouse gas budgets and evaluate the associated damage on vegetation. In this study, we propose a new multilayer scheme of ozone deposition on vegetation canopies that predicts stomatal, cuticular and soil deposition pathways separately. This mechanistic ozone deposition scheme is based on the multi-layer, multi-leaf mass and energy transfer model MuSICA. This model was chosen because it explicitly simulates the processes of rain interception, through fall and evaporation at different depths within the vegetation canopy, so that ozone deposition on wet leaf cuticles can be explicitly modelled with ozone dissolution, diffusion and chemical reaction inside the water films. The model was evaluated against a 3-year dataset of ozone, CO2 and evapotranspiration flux measurements over a winter wheat field near Paris, France (ICOS Fr-GRI). Only periods with fully developed canopies (including senescence) were considered to minimise the contribution of soil deposition to the total ozone flux. Before senescence, the model could reproduce the measured ozone deposition rates as well as the CO2 and water vapour fluxes. During senescence, large ozone deposition rates were observed under wet canopy conditions that could only be explained by first-order reaction rates in the water film of around 105 s−1. Such reaction rates are not compatible with the chemical composition of rainwater. We therefore hypothesise that, during senescence, the cell content leaks out of the leaves when they become wet, exposing anti-oxidants to ozone. These results provide for the first time a mechanistic explanation of the commonly observed increase in ozone deposition rates during rain or dew formation.

The eddy covariance (EC) method is widely considered as the reference method for heat and trace gas exchange flux measurements. However, for some species of interest, the aerodynamic gradient method (AG) is still a valuable method. Furthermore, some useful datasets are based on the AG method. In this study we compare the EC and the AG methods for latent (LE) and sensible (H) heat, carbon dioxide (Fc) fluxes and ozone deposition velocity ( V d O 3 ) over a maize field near Paris. The AG method gave roughly 15% smaller Fc and LE, similar H, and 40% larger V d O 3 than the EC method. The differences between the two methods are discussed. In particular, the effects of the displacement height and heights of measurements on the AG fluxes are explored and the similarity among heat, CO2, H2O and O3 is tested. Furthermore, the vertical divergence of the flux above the canopy is estimated with the AG method.

As forest edges are a major source of heterogeneity in fragmented landscapes, the atmospheric flow over forested areas is often under their influence. Understanding how far the upstream edge has an impact on the turbulent wind flow in a forest canopy is important, in particular for scalar flux measurement. In this study, edge and stand flows over a maritime pine forest characterized by a dense crown layer located above a deep and sparse trunk space are analysed in detail from in situ measurements and large-eddy simulations (LES). The LES model used here appears to simulate remarkably well most characteristics of the turbulent wind flow for this particular canopy structure. It is shown that the main characteristics of the edge flow in this case differ from those usually observed in forests with a more uniform vertical foliage distribution. The main differences are (i) the development of turbulence above the canopy occurring closer to the edge, (ii) the absence of a well-defined enhanced gust zone around the top of the canopy, (iii) the presence of a large secondary wind maximum within the trunk space, and (iv) the development of a positive momentum flux layer below the crown layer. Most of these differences are related to the presence of a substantial sub-canopy wind jet induced by the wind flow through the trunk space at the edge. The secondary velocity maximum induced by this wind jet differs from that observed in homogeneous stand conditions, where it seems to be related to the mesoscale pressure gradient. The wind jet appears to decrease very slowly with distance from the edge, so that edge effects are still significant at 9 h from the edge (where h is the mean canopy height). The length of the adjustment region is shown to be greater than 10–15 h , and to depend on the depth of the trunk space. In very fragmented forested areas with deep and sparse trunk space, within-canopy flow may always be under the influence of edges.

The variables controlling ozone deposition onto bare soil are still unknown and it is necessary to understand this pathway well, as it represents a significant sink for ozone. Eddy-covariance measurements of ozone (O3) fluxes were performed over bare soils in agricultural land. Three datasets with contrasted meteorological conditions and soil nitric oxide (NO) emissions were used to study the factors controlling soil deposition. It is considered that ozone deposition can be represented with an aerodynamic resistance (Ra), a quasi-laminar boundary layer resistance (Rb O3), and an additional resistance, named soil resistance (Rsoil). Although it is assumed in previous studies that soil resistance is a function of soil water content (SWC) and could be considered constant as variation of SWC at monthly scale are generally weak, the results of this study indicate that SWC is not the main factor controlling Rsoil which shows daily and hourly variations. The main factor controlling soil resistance is the surface relative humidity which is positively correlated with Rsoil, contrary to non stomatal resistance onto canopies which show a negative correlation with relative humidity. The relationship between Rsoil and the surface relative humidity is probably due to a decrease in the surface available for ozone deposition, due to an increasing adsorption of water molecules onto the ground with relative humidity. A new parameterisation of Rsoil was established, where Rsoil is a function of the surface relative humidity only (Rsoil = Rsoil min × e(k×RHsurf), and Rsoil min = 21 ± 1.01 s m−1 and k = 0.024 ± 0.001, mean ± SD). The measured and parameterised ozone deposition velocities agree well when soil NO emissions are negligible. However, when there are large soil NO emissions, the parameterised ozone deposition strongly underestimates the measured deposition velocity even if the chemical destruction of ozone by reaction with NO in the air column was evaluated to be negligible. This suggests that soil NO emissions enhance soil ozone deposition by chemical reaction at or near the soil surface. The new parameterisation allows a better estimation of soil deposition, especially during daytime when Rsoil is overestimated using previously published parameterisations. It is an important step towards a better parameterisation of the non-stomatal uptake of ozone.

Ozone (O-3) flux measurements, obtained by eddy-covariance technique, over a developed maize crop were used to partition the overall O-3 deposition between stomatal and non-stomatal uptakes. Data were analysed using a big-leaf model, which was developed from current knowledge on O-3 deposition. The classical parameters used in dry deposition models (i.e. the in-canopy aerodynamic resistance R-ac, the intrinsic ground resistance R-ig and the cuticular resistance R-cut) were determined for the maize crop from the relationship between the experimental non-stomatal conductance (g(ns)) and the friction velocity (u(*)) in dry conditions (relative humidity (RH) < 60%). g. was determined as the difference between the O-3 canopy conductance (g(c)) and the O-3 stomatal conductance (g(s)), where g(s) was estimated by a method which combines the Penman-Monteith approach and the use of the CO2 assimilation flux. Data analysis revealed that chemical reactions between O-3 and nitrogen monoxide (NO) between the canopy top and the O-3 flux measurement level (z(m)) could induce high values of the observed O-3 conductance, not representative of ozone deposition to the canopy. The actual O-3 canopy conductance was derived from the observed O-3 conductance by including a correction term function of z(m) and the NO concentration at this height, based on the previous studies on O-3 destruction above canopies. The estimations of R-ac, R-ig and R-cut given by the non-linear regression of g(ns) vs u(*) are in agreement with previously published results. Our analysis also confirms previous studies which have shown that the cuticular conductance (g(cut)) increases exponentially with RH, and we propose a new parameterization of g(cut) as a function of RH, based on experimental evidence. Using our model to partition the total O-3 deposition to the canopy, we showed that the relative contributions of stomatal and non-stomatal uptakes varied strongly with the physiological activity of the maize and the meteorological conditions. This point is of major importance for studies dedicated to the impact of ozone on plant physiology, since it emphasizes the necessity to determine accurately the amount of O-3 actually absorbed by the plants via their stomatal activity. (C) 2009 Elsevier B.V. All rights reserved.

International audience; Carbon dioxide, water vapour and energy fluxes were measured above and within a maritime pine forest during an atypical year with long-lasting reduced soil water availibility. Energy balance closure was adequately good at both levels. As compared with what is usually observed at this site the ecosystem dissipated less energy via latent heat flux and more via sensible heat flux. The understorey canopy was responsible for a variable, significant component of the whole canopy fluxes of water vapour and carbon dioxide. The annual contribution of the understorey was 38% (154 mm) of the overall evaporation (399 mm) and 32% (89 mm) of the overall sensible heat flux (274 mm). The participation of the understorey reached 45% of the overall evaporation and 30% of the daytime overall assimilation during significant soil water deficit periods in summertime. Even during winter, understorey photosynthesis was consistent as it compensated soil and understorey respiration. The ecosystem behaved as a sink of carbon, with a negative annual carbon budget (-57 gC m-2). However, due to high soil water deficit, the annual ecosystem GPP was 40% less than usually observed at this site. This budget resulted from a sink of -131 gC m-2 for the overstorey and a source of +74 gC m-2 for the understorey. Moreover, on an annual basis the overstorey layer contributed to almost two thirds of the ecosystem respiration. Finally, the effect of long-lasting soil water deficit on the maritime pine forest was found more important than the effect of the heat wave and drought of summer 2003.

Over the past 15 years atmospheric surface-layer experiments over heterogeneous canopies have shown that the vertical transfer of sensible heat and water vapour exhibit a strong dissimilarity. In particular, the sensible-heat-to-water-vapour transport efficiencies generally exceed unity. One of the main consequences is that evaporation (latent heat flux) computed by the flux-variance method is overestimated, as persistently demonstrated by comparisons with evaporation obtained with the eddy-correlation method. Various authors proposed to take into account the temperature–humidity dissimilarity to extend the applicability of the flux-variance method in order to compute evaporation from non-uniform surfaces. They attempted to connect the sensible-heat-to-water-vapour transport efficiency (λ) to the correlation coefficient between temperature and humidity turbulent fluctuations (R Tq ). This approach was found to be successful over ‘wet’ surfaces for which λ can be approximated by R Tq and ‘dry’ surfaces for which λ can be approximated by 1/R Tq . However, no solution has been proposed until now for intermediate hydrological conditions. We investigated this question using eddy-correlation measurements above and inside a pine forest canopy. For both levels, our data present a strong likeness with previously published results over heterogeneous surfaces. In particular, they confirm that λ is R Tq in wet conditions and 1/R Tq in dry conditions. Moreover, we defined the range of the Bowen ratio (Bo) values for which those two approximations are valid (below 0.1 and greater than 1, respectively) and established a relationship between λ, R Tq and Bo for the intermediate range of Bo. We are confident that this new parameterization will enlarge the applicability of the flux-variance method to all kinds of heterogeneous surfaces in various hydrological conditions

Quantification of sources and sinks of carbon at global and regional scales requires not only a good description of the land sources and sinks of carbon, but also of the synoptic and mesoscale meteorology. An experiment was performed in Les Landes, southwest France, during May¿June 2005, to determine the variability in concentration gradients and fluxes of CO2. The CarboEurope Regional Experiment Strategy (CERES; see also http://carboregional.mediasfrance.org/index) aimed to produce aggregated estimates of the carbon balance of a region that can be meaningfully compared to those obtained from the smallest downscaled information of atmospheric measurements and continental-scale inversions. We deployed several aircraft to concentration sample the CO2 and fluxes over the whole area, while fixed stations observed the fluxes and concentrations at high accuracy. Several (mesoscale) meteorological modeling tools were used to plan the experiment and flight patterns. Results show that at regional scale the relation between profiles and fluxes is not obvious, and is strongly influenced by airmass history and mesoscale flow patterns. In particular, we show from an analysis of data for a single day that taking either the concentration at several locations as representative of local fluxes or taking the flux measurements at those sites as representative of larger regions would lead to incorrect conclusions about the distribution of sources and sinks of carbon. Joint consideration of the synoptic and regional flow, fluxes, and land surface is required for a correct interpretation. This calls for an experimental and modeling strategy that takes into account the large spatial gradients in concentrations and the variability in sources and sinks that arise from different land use types. We briefly describe how such an analysis can be performed and evaluate the usefulness of the data for planning of future networks or longer campaigns with reduced experimental efforts.

International audience; Because most forests are multi-specific and vertically heterogeneous, understanding the functioning of such complex ecosystems remains a challenge for quantifying and modeling the fluxes of CO2 between the atmosphere and the vegetation at regional scale. In particular, there is lack of knowledge concerning flux partitioning between overstorey and understorey of forests across different climates and forest types. The main objective of this study is to compare the micrometeorology and the fluxes of CO2, water and sensible heat measured by the eddy-covariance method above and below the main canopy of different forests. The eight study sites are part of the Fluxnet network, and range through various climates and structural types. Across these different biomes, the leaf area index (LAI) of the overstorey is the main factor influencing flux footprint and turbulent mixing below the canopy, and flux partitioning between overstorey and understorey. Wind direction and rates of turbulent mixing generally differ between the two levels for close canopies and low wind speed in the boundary layer. Forcing of understorey turbulent mixing by overstorey conditions is stronger in case of low overstorey LAI, while trunk space flow tends to decouple from that in the boundary layer in closed canopies. Vertical light extinction is positively correlated with overstorey LAI. During the day, efflux of CO2 from the understorey is positively related to soil temperature. Uptake of CO2 by the understorey only occurs if a significant amount of light penetrates through the overstorey and reaches a significant amount of understorey leaf area. During the night, the probability that CO2 emitted from the soil reach the top of the canopy is negatively correlated with overstorey LAI. These results will help developing sound ecosystem modeling schemes for quantifying carbon budgets at regional scales, taking into account the wide variety of mixed and vertically complex forest canopies.

International audience; The physiological and physical processes controlling ozone dry deposition to vegetated surfaces are still not fully understood. In particular, the role of the understorey and the possible action of dew on ozone deposition have not received much attention so far. This paper presents the results of an experiment aimed at quantifying ozone dry deposition to a maritime pine forest in the “Les Landes” area in France. Ozone deposition fluxes were measured using the eddy-covariance technique above and within the canopy. We investigate the factors acting on ozone deposition in both dew-wetted and dry conditions. The values obtained for the ozone deposition velocity are well in the range of previously published measurements over coniferous forests. For the present forest, ozone uptake by the understorey is a significant portion of ozone deposition to the whole pine stand. The understorey contributes more to the overall ozone flux than to the other measured scalar fluxes (sensible heat and water vapour). During dry nights the surface conductance for ozone and the friction velocity are strongly correlated, showing that ozone deposition is largely controlled by dynamical processes. During the day, in dry conditions, the canopy stomatal conductance is the major parameter controlling ozone deposition. However, in winter, when the stomatal conductance is low, the influence of dynamical processes persists during day-time. It is also found that surface wetness associated with dew significantly enhances ozone deposition, during the night as well as in the morning.

Measurements of turbulent exchange in the understorey of a forest canopy are necessary to understand and model both the functioning of the lower stratum and its contribution to turbulent exchange at the whole canopy scale. Eddy covariance measurements of eddy fluxes just above the floor of a forest canopy must be thoroughly validated, given the particular conditions prevailing there. Our objective is two-fold: (i) check the overall quality of such eddy flux measurements through the analysis of the understorey energy balance closure and (ii) define quality criteria for each half-hourly sample, based on the residual term of the energy balance. A subset of the EUROFLUX data base, collected within a pine forest canopy in south-west France, was used for this purpose. During this experiment, all heat storage terms were carefully measured, which allowed the closure of the understorey energy balance to be rigorously tested. As in most experiments storage term measurements are not available, we also developed a method to estimate them, in order to apply the above-mentioned data selection method. The energy balance closure was found to be quite satisfactory (the slope of the sum of eddy fluxes and storage terms versus transmitted net radiation is 0.99, the intercept is less than 1 W m 2 , r 2 is 0.94, there is no deviation from a linear trend). The data selection procedure allows a fair description of the daily and day-to-day variation of turbulent fluxes while rejecting the most dubious data, whether experimental or estimated storage terms are used. This analysis proves the validity of eddy flux measurements in the lower part of the forest and offers tools for flux data selection, depending on the type of studies such data are intended for. © 2001 Elsevier Science B.V. All rights reserved.

International audience; Two experiments were performed in a confierous forest (maritime pine) in the southwest of France, one in 1994 and the other in 1995. Two sites were chosen, differing by age, height and structure of the trees, as well as the nature of the understorey. In both cases measurements of turbulent fluxes were made at two levels above and within the forest canopy, using sonic anemometers and open-path infrared CO2-H2O analysers. The flux differences derived from the two measurement levels allowed the Radiation and Water Use Efficiencies (RUE and WUE, respectively) to be evaluated for both canopy crowns. The results are based on the analysis of about ten days from each experiment. For both campaigns RUE is significantly larger during cloudy conditions when the fraction of diffuse radiation () increases. An empirical linear relation between RUE and is established for each site, with a smaller intercept and a larger slope for the older forest. In clear conditions (), RUE is about 30 % lower for this forest. Tree photosynthesis, estimated as the net CO2 flux of the foliated layer Fc, appears poorly correlated (r2 < 0.4) with transpiration (net water vapour flux E). This is shown to result from strong variations in the atmospheric saturation deficit D during both campaigns. At both sites WUE turns out to be a hyperbolic function of D (). The coefficient k is 50 % larger for the younger forest. This is in agreement with the values obtained for RUE, and indicates that photosynthetic rates decrease with the age of the trees.

International audience; An experiment was conducted in a pine forest in southwestern France during late spring 1992. The aim was fourfold: testing various flux measurement methodologies for chemically reactive species; quantifying the exchanges between the forest and the atmosphere; analyzing the involved mechanisms; and studying their influence on the chemistry of the surface boundary layer. This paper presents preliminary results obtained on the dry deposition of ozone and submicronic aerosol particles, measured using eddy correlation. Once properly normalized, the spectra and cospectra of all scalar species exhibit universal shapes over the whole frequency range. However, evidence is provided that under some meteorological conditions the time series of turbulent variables can be affected by nonstationary trends, or low-frequency fluctuations that do not contribute to vertical transfer but whose presence can induce large errors in the calculated fluxes. The time variations of the deposition velocities for ozone and aerosol particles are then presented over 2 days with different meteorological conditions. The deposition velocities are shown to be consistent with other reported studies. Dry deposition of ozone appears to be mainly governed by the stomatal resistance, whereas friction velocity and atmospheric instability in the boundary layer seem to govern the deposition of aerosol particles.

Atmospheric mineral dust resulting from aeolian soil erosion affects the Earth system. Their size-distribution (PSD) plays a key role on atmospheric radiation balance, cloud formation, atmospheric chemistry, and the productivity of terrestrial and marine ecosystems. However, climate models still fail to reproduce accurately the suspended dust PSD. This is explained by the poor representation of the dust emission mechanisms and the associated surface wind speed in these large-scale models. This is particularly true in the presence of surface roughnesses such as vegetation in semiarid regions. This thesis aims at improving the understanding of dust emission in semi-arid environments, characterized by heterogeneous surfaces with sparse seasonal vegetation. To this end, a combination of numerical and field experiments was employed, with investigations progressing from a bare erodible soil to surfaces with sparse vegetation.A review of the existing dust emission schemes showed ambiguities in the parametrization of the processes influencing the emitted dust. A sensitivity analysis, using a 1D dust dispersal model, demonstrated (i) the importance of surface dust PSD and inter-particle cohesive bondparametrization on the emitted dust PSD, and (ii) the importance of the deposition process on the net dust flux PSD. Based on this analysis, a new emission scheme was incorporated into a 3D erosion model, coupled with a Large Eddy Simulation (LES) airflow model, and evaluated first on a bare surface against the WIND-O-V’s 2017 field experiment in Tunisia. The model was able to reproduce the near-surface turbulent transport dissimilarity between dust and momentum observed during the experiment. This means that momentum and dust are not always transported by the same turbulent eddies. The model demonstrated that the main cause of this dissimilarity is the dust emission intermittency, which varies as a function of wind intensity and fetch. The role of sparse vegetation on the net emitted dust flux was then explored using the WINDO-V’s 2018 experiment, conducted at the same site as the 2017 experiment. The resulting field measurements were used to evaluate the 3D erosion model, including vegetation characteristics. A comparison between the 2017 and 2018 experiments confirmed that sparse vegetation reduces dust emission by increasing the erosion threshold friction velocity, which depends on vegetation characteristics and wind direction relative to the vegetation arrangement. During the 2018 experiment, the net emitted dust flux PSD varied continuously, unlike the 2017 experiment, with a progressive impoverishment in coarse particles (1.50 mm). This impoverishment was found independent of the vegetation, and resulted from the depletion of coarse particles at the surface due to longer emission periods in 2018 without surface tillage or precipitation. This non-influence of vegetation on the dust flux PSD was validated by the similarityof the dust flux PSD at the beginning of the 2018 experiment, when the vegetation was at its maximum height, with the one of the 2017 experiment without vegetation. It was further confirmed by the simulations that demonstrated (i) negligible re-deposition of coarse particles on to vegetation during emission events, and (ii) negligible effect of the turbulence induced by the vegetation on the PSD of the net emitted dust flux. Our 3D erosion model appears as a promising tool for characterizing dust emissions over heterogeneous surfaces typical of semi-arid regions and for deriving dust emission schemes for climate models as a function of surface roughness properties.; La poussière minérale atmosphérique résultant de l’érosion éolienne des sols affecte le système terrestre. Leur distribution en taille (PSD) joue un rôle clé dans le bilan radiatif et la chimie atmosphérique, la formation des nuages et la productivité des écosystèmes terrestres et marins. Néanmoins les modèles climatiques peinent à reproduire précisément la PSD de la poussière émise. Ceci vient de la mauvaise représentation dans ces modèles à grandes echelles des mécanismes d’émission de poussière et des vitesses de vent de surface associées. C’est particulièrement vrai en présence d’éléments de rugosité de surface comme la végétation en région semi-aride. Cette thèse vise à améliorer la compréhension de l’émission de poussière en environnement semi-aride, caractérisé par des surfaces hétérogènes liées à la végétation saisonnière éparse. A cette fin, une combinaison d’expériences numériques et de terrain a été employée, en partant d’un sol nu érodable, puis de surfaces couvertes en végétation éparse.Une revue des schémas existants a montré des ambiguïtés dans la paramétrisation des processus influençant l’émission de poussière. Une analyse de sensibilité utilisant un modèle 1D de dispersion de poussière a démontré (i) l’importance de la PSD de la poussière à la surface etde la paramétrisation de la cohésion entre les particules sur la PSD de la poussière émise, et (ii) l’importance du processus de dépôt sur la PSD du flux net de poussière. A partir de cette analyse, un nouveau schéma d’émission a été incorporé à un modèle 3D d’érosion, couplé à unmodèle turbulent Large Eddy Simulation (LES), et évalué sur une surface nue sur la base de l’expérimentation WIND-O-V 2017 en Tunisie. Le modèle a été capable de reproduire la dissimilarité entre les transports turbulents de la poussière et de la quantité de mouvement dans la couche de surface, telle que observée durant l’expérience. Cela signifie que la poussière et la quantité de mouvement ne sont pas toujours transportées par les mêmes tourbillons. Le modèle a démontré que la cause principale de cette dissimilarité est l’intermittence de l’émission des poussières, qui varie avec l’intensité du vent et le fetch.L’impact de la végétation éparse sur le flux net de poussière émis a ensuite été étudié sur la base de l’expérimentation WIND-O-V 2018, conduite sur le même site que celle de 2017. Les mesures ont été utilisées pour évaluer le modèle 3D d’érosion incluant les caractéristiques de la végétation. La comparaison entre les expérimentations 2017 et 2018 a confirmé que la végétation éparse réduit l’émission de poussière en augmentant la vitesse de frottement seuil de l’érosion, qui dépend des caractéristiques de la végétation et de la direction du vent par rapport à l’arrangement de la végétation. Nous avons observé que la PSD du flux net de poussière émis à varié au cours de l’expérimentation 2018, contrairement à 2017, avec un appauvrissement progressif en grosses particules (1.50 mm). Il s’est avéré que cet appauvrissement n’était pas lié à la présence de végétation, mais à l’épuisement du sol en grosses particules en raison de périodes d’émission plus longues en 2018 sans travail de la surface ou précipitations.Cette absence d’influence de la végétation sur la PSD du flux de poussière a été validée par la similarité entre la PSD du flux de poussière au début de l’expérimentation 2018, quand la végétation était à sa hauteur maximum, et celle de 2017 sans végétation. Elle a été aussi confirmée par les simulations qui ont montré (i) une re-déposition négligeable des grosses particules sur la végétation durant les émissions, et (ii) un effet négligeable de la turbulence induite par la végétation sur la PSD du flux net de poussière émis. Notre modèle 3D d’érosion apparaît comme un outil prometteur pour caractériser les émissions de poussière sur des surfaces hétérogènes représentatives des régions semi-arides et pour établir des schémas d’émission de poussière pour les modèles climatiques en fonction des propriétés de rugosité de la surface.

Atmospheric mineral dust resulting from aeolian soil erosion affects the Earth system. Their size-distribution (PSD) plays a key role on atmospheric radiation balance, cloud formation, atmospheric chemistry, and the productivity of terrestrial and marine ecosystems. However, climate models still fail to reproduce accurately the suspended dust PSD. This is explained by the poor representation of the dust emission mechanisms and the associated surface wind speed in these large-scale models. This is particularly true in the presence of surface roughnesses such as vegetation in semiarid regions. This thesis aims at improving the understanding of dust emission in semi-arid environments, characterized by heterogeneous surfaces with sparse seasonal vegetation. To this end, a combination of numerical and field experiments was employed, with investigations progressing from a bare erodible soil to surfaces with sparse vegetation.A review of the existing dust emission schemes showed ambiguities in the parametrization of the processes influencing the emitted dust. A sensitivity analysis, using a 1D dust dispersal model, demonstrated (i) the importance of surface dust PSD and inter-particle cohesive bond parametrization on the emitted dust PSD, and (ii) the importance of the deposition process on the net dust flux PSD. Based on this analysis, a new emission scheme was incorporated into a 3D erosion model, coupled with a Large Eddy Simulation (LES) airflow model, and evaluated first on a bare surface against the WIND-O-V’s 2017 field experiment in Tunisia. The model was able to reproduce the near-surface turbulent transport dissimilarity between dust and momentum observed during the experiment. This means that momentum and dust are not always transported by the same turbulent eddies. The model demonstrated that the main cause of this dissimilarity is the dust emission intermittency, which varies as a function of wind intensity and fetch.The role of sparse vegetation on the net emitted dust flux was then explored using the WIND-O-V’s 2018 experiment, conducted at the same site as the 2017 experiment. The resulting field measurements were used to evaluate the 3D erosion model, including vegetation characteristics. A comparison between the 2017 and 2018 experiments confirmed that sparse vegetation reduces dust emission by increasing the erosion threshold friction velocity, which depends on vegetation characteristics and wind direction relative to the vegetation arrangement. During the 2018 experiment, the net emitted dust flux PSD varied continuously, unlike the 2017 experiment, with a progressive impoverishment in coarse particles (1.50 μm). This impoverishment was found independent of the vegetation, and resulted from the depletion of coarse particles at the surface due to longer emission periods in 2018 without surface tillage or precipitation. This non-influence of vegetation on the dust flux PSD was validated by the similarity of the dust flux PSD at the beginning of the 2018 experiment, when the vegetation was at its maximum height, with the one of the 2017 experiment without vegetation. It was further confirmed by the simulations that demonstrated (i) negligible re-deposition of coarse particles on to vegetation during emission events, and (ii) negligible effect of the turbulence induced by the vegetation on the PSD of the net emitted dust flux.Our 3D erosion model appears as a promising tool for characterizing dust emissions over heterogeneous surfaces typical of semi-arid regions and for deriving dust emission schemes for climate models as a function of surface roughness properties.; La poussière minérale atmosphérique résultant de l’érosion éolienne des sols affecte le système terrestre. La distribution en taille (PSD) de cette poussière joue un rôle clé dans le bilan radiatif et la chimie atmosphérique, la formation des nuages et la productivité des écosystèmes terrestres et marins. Néanmoins les modèles climatiques peinent à reproduire précisément la PSD de la poussière émise. Ceci vient de la représentation imparfaite des mécanismes d’émission de poussières et des vitesses de vent de surface associées. C’est particulièrement vrai en présence d’éléments de rugosité de surface comme la végétation en régions semi-arides. Cette thèse vise à améliorer la compréhension de l’émission de poussière en environnements semi-arides, caractérisé par des surfaces hétérogènes liées à la végétation saisonnière éparse. A cette fin, une combinaison d’expériences numériques et de terrain a été employée, en partant d’un sol nu érodable à des surfaces couvertes de végétation éparse.Une revue des schémas existants a montré des ambiguïtés dans la paramétrisation des processus influençant l’émission de poussières. Une analyse de sensibilité utilisant un modèle 1D de dispersion de poussière a démontré l’importance (i) de la PSD de la poussière à la surface et de la paramétrisation de la cohésion interparticules qui affectent la PSD de la poussière émise, et (ii) des processus de dépôt qui influencent la PSD du flux net de poussière dans la couche de surface atmosphérique. A partir de cette analyse, un nouveau schéma d’émission a été incorporé à un modèle 3D d’érosion, couplé à un modèle turbulent Large Eddy Simulation (LES), et évalué d’abord sur une surface nue sur la base de l’expérimentation WIND-O-V 2017 en Tunisie. Le modèle a ainsi été capable de reproduire la dissimilarité entre les transports turbulents de poussière et de quantité de mouvement dans la couche de surface, observée durant l’expérience. Cela signifie que poussière et quantité de mouvement ne sont pas toujours transportées par les mêmes tourbillons. Le modèle a démontré que la cause principale de cette dissimilarité est l’intermittence de l’émission de poussières, qui varie avec l’intensité du vent et le fetch.L’impact de la végétation éparse sur le flux net de poussière émis a ensuite été étudié sur la base de l’expérimentation WIND-O-V 2018, conduite sur le même site que celle de 2017. Les mesures ont été utilisées pour évaluer le modèle 3D d’érosion incluant les caractéristiques de la végétation. La comparaison entre les expérimentations 2017 et 2018 a confirmé que la végétation éparse réduit l’émission en augmentant la vitesse de frottement seuil de l’érosion, qui dépend des caractéristiques de la végétation et de la direction du vent. Au cours de l’expérimentation 2018, nous avons observé que la PSD du flux net de poussière émis variait, contrairement à 2017, avec un appauvrissement progressif en grosses particules (1.50 µm). Il s’est avéré que cet appauvrissement n’était pas lié à la présence de végétation, mais à l'épuisement du sol en grosses particules en raison de périodes d’émission plus longues sans modification de la surface, comparé à 2017. Cette absence d’influence de la végétation a été validée par la similarité entre la PSD du flux de poussière au début de l’expérimentation 2018, quand la végétation était à sa hauteur maximum, et celle de 2017 sans végétation. Et elle a été confirmée par nos simulations qui montrent (i) une re-déposition négligeable des grosses particules sur la végétation durant les émissions, et (ii) un effet négligeable de la turbulence induite par la végétation sur la PSD du flux net de poussière émis.Notre modèle 3D d’érosion apparaît comme un outil prometteur pour caractériser les émissions de poussière sur des surfaces hétérogènes représentatives des régions semi-arides et pour établir des schémas d’émission de poussières pour les modèles climatiques en fonction des propriétés de rugosité de la surface.

Atmospheric mineral dust resulting from aeolian soil erosion affects the Earth system. Their size-distribution (PSD) plays a key role on atmospheric radiation balance, cloud formation, atmospheric chemistry, and the productivity of terrestrial and marine ecosystems. However, climate models still fail to reproduce accurately the suspended dust PSD. This is explained by the poor representation of the dust emission mechanisms and the associated surface wind speed in these large-scale models. This is particularly true in the presence of surface roughnesses such as vegetation in semiarid regions. This thesis aims at improving the understanding of dust emission in semi-arid environments, characterized by heterogeneous surfaces with sparse seasonal vegetation. To this end, a combination of numerical and field experiments was employed, with investigations progressing from a bare erodible soil to surfaces with sparse vegetation.A review of the existing dust emission schemes showed ambiguities in the parametrization of the processes influencing the emitted dust. A sensitivity analysis, using a 1D dust dispersal model, demonstrated (i) the importance of surface dust PSD and inter-particle cohesive bondparametrization on the emitted dust PSD, and (ii) the importance of the deposition process on the net dust flux PSD. Based on this analysis, a new emission scheme was incorporated into a 3D erosion model, coupled with a Large Eddy Simulation (LES) airflow model, and evaluated first on a bare surface against the WIND-O-V’s 2017 field experiment in Tunisia. The model was able to reproduce the near-surface turbulent transport dissimilarity between dust and momentum observed during the experiment. This means that momentum and dust are not always transported by the same turbulent eddies. The model demonstrated that the main cause of this dissimilarity is the dust emission intermittency, which varies as a function of wind intensity and fetch. The role of sparse vegetation on the net emitted dust flux was then explored using the WINDO-V’s 2018 experiment, conducted at the same site as the 2017 experiment. The resulting field measurements were used to evaluate the 3D erosion model, including vegetation characteristics. A comparison between the 2017 and 2018 experiments confirmed that sparse vegetation reduces dust emission by increasing the erosion threshold friction velocity, which depends on vegetation characteristics and wind direction relative to the vegetation arrangement. During the 2018 experiment, the net emitted dust flux PSD varied continuously, unlike the 2017 experiment, with a progressive impoverishment in coarse particles (1.50 mm). This impoverishment was found independent of the vegetation, and resulted from the depletion of coarse particles at the surface due to longer emission periods in 2018 without surface tillage or precipitation. This non-influence of vegetation on the dust flux PSD was validated by the similarityof the dust flux PSD at the beginning of the 2018 experiment, when the vegetation was at its maximum height, with the one of the 2017 experiment without vegetation. It was further confirmed by the simulations that demonstrated (i) negligible re-deposition of coarse particles on to vegetation during emission events, and (ii) negligible effect of the turbulence induced by the vegetation on the PSD of the net emitted dust flux. Our 3D erosion model appears as a promising tool for characterizing dust emissions over heterogeneous surfaces typical of semi-arid regions and for deriving dust emission schemes for climate models as a function of surface roughness properties., La poussière minérale atmosphérique résultant de l’érosion éolienne des sols affecte le système terrestre. Leur distribution en taille (PSD) joue un rôle clé dans le bilan radiatif et la chimie atmosphérique, la formation des nuages et la productivité des écosystèmes terrestres et marins. Néanmoins les modèles climatiques peinent à reproduire précisément la PSD de la poussière émise. Ceci vient de la mauvaise représentation dans ces modèles à grandes echelles des mécanismes d’émission de poussière et des vitesses de vent de surface associées. C’est particulièrement vrai en présence d’éléments de rugosité de surface comme la végétation en région semi-aride. Cette thèse vise à améliorer la compréhension de l’émission de poussière en environnement semi-aride, caractérisé par des surfaces hétérogènes liées à la végétation saisonnière éparse. A cette fin, une combinaison d’expériences numériques et de terrain a été employée, en partant d’un sol nu érodable, puis de surfaces couvertes en végétation éparse.Une revue des schémas existants a montré des ambiguïtés dans la paramétrisation des processus influençant l’émission de poussière. Une analyse de sensibilité utilisant un modèle 1D de dispersion de poussière a démontré (i) l’importance de la PSD de la poussière à la surface etde la paramétrisation de la cohésion entre les particules sur la PSD de la poussière émise, et (ii) l’importance du processus de dépôt sur la PSD du flux net de poussière. A partir de cette analyse, un nouveau schéma d’émission a été incorporé à un modèle 3D d’érosion, couplé à unmodèle turbulent Large Eddy Simulation (LES), et évalué sur une surface nue sur la base de l’expérimentation WIND-O-V 2017 en Tunisie. Le modèle a été capable de reproduire la dissimilarité entre les transports turbulents de la poussière et de la quantité de mouvement dans la couche de surface, telle que observée durant l’expérience. Cela signifie que la poussière et la quantité de mouvement ne sont pas toujours transportées par les mêmes tourbillons. Le modèle a démontré que la cause principale de cette dissimilarité est l’intermittence de l’émission des poussières, qui varie avec l’intensité du vent et le fetch.L’impact de la végétation éparse sur le flux net de poussière émis a ensuite été étudié sur la base de l’expérimentation WIND-O-V 2018, conduite sur le même site que celle de 2017. Les mesures ont été utilisées pour évaluer le modèle 3D d’érosion incluant les caractéristiques de la végétation. La comparaison entre les expérimentations 2017 et 2018 a confirmé que la végétation éparse réduit l’émission de poussière en augmentant la vitesse de frottement seuil de l’érosion, qui dépend des caractéristiques de la végétation et de la direction du vent par rapport à l’arrangement de la végétation. Nous avons observé que la PSD du flux net de poussière émis à varié au cours de l’expérimentation 2018, contrairement à 2017, avec un appauvrissement progressif en grosses particules (1.50 mm). Il s’est avéré que cet appauvrissement n’était pas lié à la présence de végétation, mais à l’épuisement du sol en grosses particules en raison de périodes d’émission plus longues en 2018 sans travail de la surface ou précipitations.Cette absence d’influence de la végétation sur la PSD du flux de poussière a été validée par la similarité entre la PSD du flux de poussière au début de l’expérimentation 2018, quand la végétation était à sa hauteur maximum, et celle de 2017 sans végétation. Elle a été aussi confirmée par les simulations qui ont montré (i) une re-déposition négligeable des grosses particules sur la végétation durant les émissions, et (ii) un effet négligeable de la turbulence induite par la végétation sur la PSD du flux net de poussière émis. Notre modèle 3D d’érosion apparaît comme un outil prometteur pour caractériser les émissions de poussière sur des surfaces hétérogènes représentatives des régions semi-arides et pour établir des schémas d’émission de poussière pour les modèles climatiques en fonction des propriétés de rugosité de la surface.

Atmospheric mineral dust resulting from aeolian soil erosion affects the Earth system. Their size-distribution (PSD) plays a key role on atmospheric radiation balance, cloud formation, atmospheric chemistry, and the productivity of terrestrial and marine ecosystems. However, climate models still fail to reproduce accurately the suspended dust PSD. This is explained by the poor representation of the dust emission mechanisms and the associated surface wind speed in these large-scale models. This is particularly true in the presence of surface roughnesses such as vegetation in semiarid regions. This thesis aims at improving the understanding of dust emission in semi-arid environments, characterized by heterogeneous surfaces with sparse seasonal vegetation. To this end, a combination of numerical and field experiments was employed, with investigations progressing from a bare erodible soil to surfaces with sparse vegetation.A review of the existing dust emission schemes showed ambiguities in the parametrization of the processes influencing the emitted dust. A sensitivity analysis, using a 1D dust dispersal model, demonstrated (i) the importance of surface dust PSD and inter-particle cohesive bond parametrization on the emitted dust PSD, and (ii) the importance of the deposition process on the net dust flux PSD. Based on this analysis, a new emission scheme was incorporated into a 3D erosion model, coupled with a Large Eddy Simulation (LES) airflow model, and evaluated first on a bare surface against the WIND-O-V’s 2017 field experiment in Tunisia. The model was able to reproduce the near-surface turbulent transport dissimilarity between dust and momentum observed during the experiment. This means that momentum and dust are not always transported by the same turbulent eddies. The model demonstrated that the main cause of this dissimilarity is the dust emission intermittency, which varies as a function of wind intensity and fetch.The role of sparse vegetation on the net emitted dust flux was then explored using the WIND-O-V’s 2018 experiment, conducted at the same site as the 2017 experiment. The resulting field measurements were used to evaluate the 3D erosion model, including vegetation characteristics. A comparison between the 2017 and 2018 experiments confirmed that sparse vegetation reduces dust emission by increasing the erosion threshold friction velocity, which depends on vegetation characteristics and wind direction relative to the vegetation arrangement. During the 2018 experiment, the net emitted dust flux PSD varied continuously, unlike the 2017 experiment, with a progressive impoverishment in coarse particles (1.50 μm). This impoverishment was found independent of the vegetation, and resulted from the depletion of coarse particles at the surface due to longer emission periods in 2018 without surface tillage or precipitation. This non-influence of vegetation on the dust flux PSD was validated by the similarity of the dust flux PSD at the beginning of the 2018 experiment, when the vegetation was at its maximum height, with the one of the 2017 experiment without vegetation. It was further confirmed by the simulations that demonstrated (i) negligible re-deposition of coarse particles on to vegetation during emission events, and (ii) negligible effect of the turbulence induced by the vegetation on the PSD of the net emitted dust flux.Our 3D erosion model appears as a promising tool for characterizing dust emissions over heterogeneous surfaces typical of semi-arid regions and for deriving dust emission schemes for climate models as a function of surface roughness properties., La poussière minérale atmosphérique résultant de l’érosion éolienne des sols affecte le système terrestre. La distribution en taille (PSD) de cette poussière joue un rôle clé dans le bilan radiatif et la chimie atmosphérique, la formation des nuages et la productivité des écosystèmes terrestres et marins. Néanmoins les modèles climatiques peinent à reproduire précisément la PSD de la poussière émise. Ceci vient de la représentation imparfaite des mécanismes d’émission de poussières et des vitesses de vent de surface associées. C’est particulièrement vrai en présence d’éléments de rugosité de surface comme la végétation en régions semi-arides. Cette thèse vise à améliorer la compréhension de l’émission de poussière en environnements semi-arides, caractérisé par des surfaces hétérogènes liées à la végétation saisonnière éparse. A cette fin, une combinaison d’expériences numériques et de terrain a été employée, en partant d’un sol nu érodable à des surfaces couvertes de végétation éparse.Une revue des schémas existants a montré des ambiguïtés dans la paramétrisation des processus influençant l’émission de poussières. Une analyse de sensibilité utilisant un modèle 1D de dispersion de poussière a démontré l’importance (i) de la PSD de la poussière à la surface et de la paramétrisation de la cohésion interparticules qui affectent la PSD de la poussière émise, et (ii) des processus de dépôt qui influencent la PSD du flux net de poussière dans la couche de surface atmosphérique. A partir de cette analyse, un nouveau schéma d’émission a été incorporé à un modèle 3D d’érosion, couplé à un modèle turbulent Large Eddy Simulation (LES), et évalué d’abord sur une surface nue sur la base de l’expérimentation WIND-O-V 2017 en Tunisie. Le modèle a ainsi été capable de reproduire la dissimilarité entre les transports turbulents de poussière et de quantité de mouvement dans la couche de surface, observée durant l’expérience. Cela signifie que poussière et quantité de mouvement ne sont pas toujours transportées par les mêmes tourbillons. Le modèle a démontré que la cause principale de cette dissimilarité est l’intermittence de l’émission de poussières, qui varie avec l’intensité du vent et le fetch.L’impact de la végétation éparse sur le flux net de poussière émis a ensuite été étudié sur la base de l’expérimentation WIND-O-V 2018, conduite sur le même site que celle de 2017. Les mesures ont été utilisées pour évaluer le modèle 3D d’érosion incluant les caractéristiques de la végétation. La comparaison entre les expérimentations 2017 et 2018 a confirmé que la végétation éparse réduit l’émission en augmentant la vitesse de frottement seuil de l’érosion, qui dépend des caractéristiques de la végétation et de la direction du vent. Au cours de l’expérimentation 2018, nous avons observé que la PSD du flux net de poussière émis variait, contrairement à 2017, avec un appauvrissement progressif en grosses particules (1.50 µm). Il s’est avéré que cet appauvrissement n’était pas lié à la présence de végétation, mais à l'épuisement du sol en grosses particules en raison de périodes d’émission plus longues sans modification de la surface, comparé à 2017. Cette absence d’influence de la végétation a été validée par la similarité entre la PSD du flux de poussière au début de l’expérimentation 2018, quand la végétation était à sa hauteur maximum, et celle de 2017 sans végétation. Et elle a été confirmée par nos simulations qui montrent (i) une re-déposition négligeable des grosses particules sur la végétation durant les émissions, et (ii) un effet négligeable de la turbulence induite par la végétation sur la PSD du flux net de poussière émis.Notre modèle 3D d’érosion apparaît comme un outil prometteur pour caractériser les émissions de poussière sur des surfaces hétérogènes représentatives des régions semi-arides et pour établir des schémas d’émission de poussières pour les modèles climatiques en fonction des propriétés de rugosité de la surface.

Forest ecosystems affect air quality and climate, especially through the emissions and the reactions of biogenic volatile organic compounds (BVOCs) with the atmospheric oxidants, known to generate Secondary Organic Aerosols (SOAs). This work aims to improve our knowledge on the processes involved in biogenic SOA formation and fate. Two field campaigns have been conducted in the Landes forest. In a first step, the measurement site was characterized by a statistical study of local meteorological conditions. During these campaigns, complementary physical and chemical measurements have been carried out, implying measurements of fluxes and concentrations of particles, BVOCs and oxidants. The results evidenced nocturnal new particle formation, which is an original result as this process was usually only reported during daytime. The strong contribution of BVOCs (dominated by α- and β-pinene) ozonolysis to nocturnal new particle formation has been demonstated. Particle flux measurements suggested that particles are transfered from the canopy to the higher atmospheric surface layer. The comparison between ozone fluxes and a physical ozone deposition model also highlighted an ozone production source. This ozone production might be related to BVOC photooxidation.; L’objectif de ce travail de thèse est d’améliorer l’état de nos connaissances sur les processus de formation et du devenir de l’AOS, en lien avec la réactivité des COVB avec les oxydants de l’atmosphère et en particulier l’ozone, en utilisant le potentiel de l’écosystème landais en tant que source d'AOS biogénique. Pour cela, des campagnes de terrain ont été menées sur un site de mesure en forêt landaise, écosystème particulièrement approprié pour l’étude de la formation et du devenir des AOS. Au cours de ces campagnes, différents paramètres physico-chimiques complémentaires ont été mesurés, tels que les concentrations et les flux de particules, d’oxydants et de COVB. Des épisodes nocturnes de formation de nouvelles particules ont été mis en évidence, ce qui constitue un résultat original car les évènements rapportés jusqu’ici dans la littérature étaient principalement diurnes. La contribution importante de l’ozonolyse des monoterpènes émis par les pins maritimes, dominés par l’α- et le β-pinène, à la formation nocturne de nouvelles particules a également été démontrée. Les mesures de flux de particules réalisées suggèrent que les particules formées au cours de la nuit sont transférées depuis la canopée vers les plus hautes couches de l’atmosphère. Une production d’ozone a également été démontrée par la comparaison des mesures de flux d’ozone à un modèle physique de dépôt. Cette source d’ozone pourrait être liée à la photooxidation des COVB