The hydroxyl radical (OH) is an efficient cleansing molecule present in the troposphere. It is mainly produced during the daytime through the reaction of water vapor with O(1D), formed by ozone photolysis, while nitrogen oxides and volatile organic compounds (VOCs) are its major sinks. OH is highly reactive and reacts with numerous chemical species, controlling their chemical lifetimes . Both OH concentrations and OH reactivity are key elements of the oxidative capacity of the atmosphere.

Several field campaigns have been conducted to study total OH reactivity for urban and forested environments. Measured reactivities have been compared to calculated reactivities obtained by summing OH reactant concentrations and multiplying them by their reaction rate constants. A missing part, corresponding to the difference between measured and calculated OH reactivity, is found not only under urban or biogenic conditions but also in clean remote regions. Measured OH reactivity in urban areas has been shown to be similar (less than 10 %) to calculated OH reactivity in New York , in Houston and in a controlled urban environment . However, discrepancies in urban areas have been observed between measured and calculated OH reactivity in Nashville (35 % less for calculated reactivity), in Mexico (25 %) and in Tokyo (25 %). The differences between measured and calculated total OH reactivity are even higher in forested areas. found an unexplained fraction of 50 % in measured OH reactivity during the PROPHET campaign in Michigan. These results are comparable to the missing part (50 to 58 %) calculated from measurements made in a boreal forest in Finland . Similarly, calculated an accounted fraction of measured OH reactivity close to 49 % in an Amazonian rainforest.

As shown by , atmospheric models do not correctly simulate observed total OH reactivity. Attempts to use numerical models to explain the missing fraction of OH reactivity have proved to be insufficient. Indeed, found an underestimation of 30 % of OH reactivity in a box model with a detailed chemical mechanism for the OP3 project. In the PRIDE-PRD campaign, found discrepancies of ±10 % between the results of the OH reactivity model and the measurements, also using a box model. Similarly, used a column model to elucidate the missing part of OH reactivity, but only explained 30 to 50 % of the OH reactivity measured over a forest in Finland. used a three-dimensional model with coarse resolution to fill the gap in OH reactivity but 40 % of the measured OH sinks remained unexplained. The difficulty of getting models to represent OH reactivity could be partly due to as yet nondiscovered OH reaction pathways, which are therefore not implemented in atmospheric models. OH recycling by the isoprene oxidation chain in forest environments characterized by low NOx (sum of NO and NO2) conditions (i.e., < 1 ppb) was proposed to explain the uncertainties in the simulated HOx (sum of OH and HO2) budget .

However, studied several proposed OH recycling mechanisms present in the literature and found that biases from OH and HO2 concentrations still exist whatever the mechanism.

One possible issue in total OH reactivity retrieval not mentioned by previous studies could lie in neglecting turbulent motions in the transport of chemical compounds in the boundary layer. Indeed, turbulence can spatially segregate or bring together chemical species, reducing or increasing the mean reaction rate and thus chemical reactivity. However, as far as we know this physical process has not been investigated in previous studies. The time response of current OH reactivity measurement techniques is not yet sufficient to directly resolve the smallest relevant turbulent spatial scales. The limitations in time resolutions range from 30 seconds for LIF-based methods to 1 min for the CRM method . In comparison, and used direct isoprene and OH concentrations measurements with temporal resolution of a few seconds, fast enough to estimate the segregation of the compounds.

The atmospheric boundary layer has a turbulent structure characterized by strong and narrow updrafts surrounded by weak and large descending areas . considered passive scalars and found that surface bottom-up transport plays a more important role in the vertical diffusion than the entrainment zone top-down transport in a convective boundary layer. Updrafts in the boundary layer lead to the spatial discrimination of pollutant concentrations between thermals and their environment. This heterogeneity in chemical species redistribution influences the mean reaction rate obtained when considering averaged reactant concentrations . Using an idealized simulation, showed that, for a second-order reaction implying a top-down and a bottom-up species, reaction rates are maximum in updrafts near the surface and in downdrafts at the top of the boundary layer. Segregation between the VOC and the OH radical was first addressed by the numerical study of , who investigated the turbulence effects on the mean reaction rates of these species. studied the inefficiency of turbulent mixing over heterogeneous surfaces and found that the highest reaction rates for isoprene and OH are located in thermals at the top of the boundary layer. The chemical reactivity of the boundary layer is therefore determined by the capacity of turbulence to mix reactive species . However, modeling and experimental studies investigating heterogeneities in the boundary layer have focused on OH radical concentrations rather than on OH reactivity due to the instrumental limitations discussed above.

Vertical motions associated with clouds and sea breezes impact the atmospheric chemistry and pollution levels near the surface since they dilute chemical species and increase the upward transport of surface emissions . In the case of updrafts leading to cloud formation, found a decrease of 10 to 50 % of tracer mixing ratios averaged over the boundary layer with respect to a situation without clouds due to the deepening of the boundary layer. Clouds have multiple impacts on the atmospheric boundary layer as they induce turbulent mixing of chemical compounds which modifies reaction rates. They also modify incoming solar radiation, which in turn disturbs photolysis reactions and alters the emissions of biogenic compounds, such as isoprene. Finally, they alter atmospheric chemistry due to soluble gas washout and chemical reactions occurring in cloud droplets.

Cloud cover over West Africa is an important feature of the African monsoon but is poorly represented by global models . This could lead to overly low simulated clouds and overly high incoming radiation at the surface, implying excessively high diurnal temperature and relative humidity cycles over this region. The nocturnal low-level stratus was studied during the monsoon period at Parakou (Benin) by with radiosondes. The authors found that turbulent processes are responsible for cloudy nights, whereas clear nights are associated with a nocturnal inversion, leading to the decoupling of the surface and lower atmosphere. investigated nighttime cloud formation. They observed that the presence of a nighttime low-level jet induces the shear-driven vertical mixing of moisture accumulated near the surface. This leads to stratus formation whose cover is likely to persist until the early afternoon when it breaks up to form cumulus clouds . However, studying the impact of this specific cloudy environment on the turbulent transport of chemical species in tropical West Africa has not been reported.

High-resolution simulations which explicitly resolve the turbulent and convective advection terms were conducted to assess the impacts of clouds and the convective boundary layer on the mixing of chemical compounds. However, previous numerical studies on the impact of the turbulent mixing of chemical compounds mainly used relatively simple or only slightly more detailed chemical schemes (e.g., , and ), resulting in possible limitations in the representation of the atmospheric chemistry. Besides, more recent studies by and used a more detailed chemical scheme derived from Mozart v2.2, allowing the formation of OH radicals initiated by peroxy radicals. This limits the spatial heterogeneities caused by the reactions consuming the OH radical.

The goal of this work is to evaluate the role of thermals in OH reactivity in the framework of a convective boundary layer with contrasted chemical environments in southern West Africa. It focuses in particular on investigating turbulence as a possible explanation of the discrepancies between calculated and measured OH reactivities mentioned in the literature.

Two contrasted chemical regimes represented by a detailed chemical scheme are studied by using large-eddy simulations. The first simulation is influenced by biogenic emissions, whereas the second is characterized by anthropogenic emissions representative of Cotonou (Benin) in the center of the domain. Based on a conditional sampling implemented in the model, the thermals are discriminated in these simulations, allowing the specific chemical regime inside thermals to be investigated. The model experiments are presented in Sect. 2. Section 3 presents the dynamical and chemical results for the two cases while Sect. 4 presents the discussion relating to these results.

The redistribution of chemical species by the boundary layer turbulence induces a different mean reaction rate between compounds when compared to a situation in which chemical species would be perfectly mixed ( among others). The perfectly mixed assumption used in regional- and large-scale atmospheric models leads to errors in the mean reaction rates between species as the turbulent mixing occurs at scales smaller than the grid length . This implies that the OH total reactivity has been calculated inaccurately, in turn leading to a modification in the lifetimes of the OH reactants such as ozone and carbon monoxide.

In a biogenic environment characterized by low NOx conditions, found negative segregation between isoprene and OH varying between -3 % near the surface and -10 % in the cloud layer due to increasing OH mixing ratios in thermals with altitude. That implied positive isoprene anomalies and negative ones for OH in the frontier region between the boundary layer and the free troposphere. These features are reproduced in our biogenic simulation although our segregation values are higher in the cloud layer (Fig. ) due to sharper gradients of OH mixing ratios at the top of the cloudy layer.

Using a simple chemistry scheme of 19 reactions representing the basic reactions of the O3–NOx–VOC–HOx system, found an almost constant value of -7 % for the segregation between OH and isoprene in the boundary layer over the Amazonian forest. This was the result of positive isoprene anomalies due to transport by thermals and negative OH anomalies due to consumption therein. Negative segregation ranging from -2 to -5 % inside the convective boundary layer was simulated by and . In the biogenic case in this study, the negative segregation of a few percent in the middle of the boundary layer is reproduced. As in this case, higher segregation values were simulated with altitude in , especially in the cloudy layer. However, segregation computed in the cloudy layer in was equal to -0.1, a value lower than that computed in the biogenic case of the present study. The discrepancies in this study and are likely due to the vertical OH profiles. In the study by , OH concentrations increased linearly with altitude. This implies lower OH covariances for ascending air parcels and thus lower segregation values (Eq. ). On the contrary, in the present study, OH is relatively homogeneous in the boundary layer and a strong gradient is present only at the top of the boundary layer. This induces high covariances for OH concentrations inside air transported by thermals at the top of the boundary layer, implying higher segregation values. As segregation computed by is not available above 1000 m height, a direct comparison with results regarding the cloudy layer is not possible with the results from the biogenic case. The positive values of segregation simulated in the afternoon (Fig. a) in the biogenic case are not reproduced in other studies and might be the result of efficient OH recycling in ReLACS 3.0, initiated in particular inside the thermals due to peroxy radicals formed by isoprene oxidation. Indeed, this recycling is either absent or indirect in previous works like in the mechanism used by that produces only HO2 from peroxy radicals, which may explain the discrepancies in OH covariances. Furthermore, and investigated the sensitivity of segregation to NOx. It was found that different NOx levels imply differences in the segregation of OH and other compounds as they contribute to OH production.

used LES to study the OH budget in a biogenic environment averaged over the domain and over time from 13:30 to 14:30 LT with a chemical scheme adapted from MOZART v2.2. For a low NOx case with ozone mixing ratios close to 64 ppbv, they found that OH production was mostly influenced by four predominant reactions (including O1D + H2O, HO2 + O3, H2O2 photolysis and HO2 + NO). These production terms are present in the biogenic case of the present study. However, did not take into account recycling by peroxy radicals as the latter formed only HO2 in the chemical scheme they used. Moreover, they found that OH loss by reaction with isoprene was dominant near the surface, followed by reactions with CO and formaldehyde, as in the present study. The contribution of BVOC species to OH reactivity was simulated by for three distinct biogenic cases of the DISCOVER-AQ (Deriving Information on Surface Conditions from Column and Vertically Resolved Observations Relevant to Air Quality) campaign. The production terms of OH were not available but it was found that isoprene was a dominant sink for OH; about 25–30 % of the OH reactivity was linked to BVOC reactions at 0.3 km during midday. This contribution is comparable to the percentage calculated for isoprene in the OH destruction in thermals (41.7 %) or in an updraft-free area (29.3 %). The contribution of formaldehyde was comparable to isoprene at that height, with HCHO mixing ratios ranging from 2 to 4.5 ppbv at the surface. The higher formaldehyde contribution of is the result of higher concentrations than in the current work.

found that in forest areas the OH budget was largely dominated by isoprene and its oxidation products. For example, measured OH reactivity and the concentration of biogenic compounds over a Mediterranean forest. They found that isoprene was the dominant sink for OH and contributed up to 74 % of total OH reactivity during daytime due to its high reactivity towards OH and its high concentration over the forested area. Isoprene predominance in OH loss was reproduced in the OH budget of the biogenic environment of the present study, and for carbon monoxide. The formaldehyde mixing ratios were close to 2 ppbv, which explains that this is not a major sink like isoprene, but remains important for OH loss through the reaction OH + HCHO. OH production in the biogenic case was similar to , as almost the same production terms are present. However, overall production was dominated by the reaction of peroxy radicals with HO2, and the ozone mixing ratios of 18 ppbv decrease the importance of O1D + H2O source for OH.

However, the mechanism used to represent atmospheric chemistry in our simulation has fast OH cycling due to the reaction of peroxy radicals RO2 with HO2. This reaction yield in OH is greater than in the laboratory studies , implying an overestimation of OH mixing ratios, especially in the biogenic case where RO2 values are high. As the hydroxyl radical is recycled through the isoprene oxidation products, it tends to reduce the impact of isoprene chemistry on OH mixing ratios and thus influences the low segregation simulated in the core of the boundary layer. This is similar to a case considered by where an ISOPO2 + HO2→ ISOPOOH + 3 OH recycling mechanism was proposed to increase the simulated OH concentrations and provided the best agreement with OH observations. This reaction implied that isoprene has no net impact on OH concentrations. This conclusion was also reached by who found better agreement between simulated and measured HOx concentrations by ignoring isoprene chemistry. As the chemistry of isoprene is not well understood, the results obtained in our work are subject to these uncertainties.

Aldehydes have not been considered in previous works on segregation. calculated segregation at 250 m and 11:00 UTC between each species of the chemical model CHIMERE used in their simulation. They found segregation ranging from 0 to -1 % between OH and acetaldehyde. This is comparable to a range of results of the anthropogenic case of the present study (Fig. b) considering the incomplete mixing between OH and C > 2 aldehydes at the same height.

In the anthropogenic simulation, the high values of segregation computed at the top of the boundary layer are coincident with the presence of clouds. In the absence of aqueous-phase chemistry, clouds impacts on chemical species are dynamical and photochemical. They modify heat and moisture fluxes at their surroundings and thus the transport of compounds, as noted by who demonstrated that cloud presence could increase transport of chemicals by 1000 m.

In this set of simulations, the chemical impact of clouds on species involves photolysis rates as they are corrected at every time step due to the presence of clouds according to the work of . At each point of the domain, photolysis rates are increased above clouds and decreased below them.

Another effect of clouds on the atmospheric chemistry lies in isoprene emissions, as demonstrated by . As isoprene emissions are dependent on incoming radiation and temperature near the surface, cloud shading could decrease the amount of isoprene emitted. showed that isoprene concentrations are decreased by 10 % and OH concentrations increased by 5 % in the boundary layer when isoprene emissions are reduced by up to 10 %. The ultimate impact on segregation cannot be anticipated since it corresponds to two compensating effects.

Finally, clouds impact atmospheric chemistry through aqueous-phase reactivity. Aqueous-phase chemistry was not considered here, nor were the exchanges between gas and aqueous phases. However, it could have an impact on soluble species mixing ratios, such as formaldehyde and H2O2 through the capture and degassing cycles of these compounds. showed a decrease in oxidative capacity of the atmosphere through aqueous-phase reactions via a significant decrease in ozone mixing ratios, but also OH, formaldehyde and nitrogen oxides. However, the effects of aqueous-phase chemistry on gas-phase compound concentrations are various and OH concentrations could decrease in clouds . This result was confirmed by the study of who found that HOx concentrations decreased in clouds. Recently, studied segregation effects in a biogenic environment when aqueous-phase chemistry is included. They found that isoprene concentrations are increased by up to 100 % while OH concentrations decreased by 18 %, resulting in a maximum segregation of 55 % in the cloudy layer. In both biogenic and anthropogenic environments, segregation effects are expected to be enhanced due to the decrease in OH concentrations in the gaseous phase in the cloud layer, reducing the cleansing capacity of the atmosphere.

Several instrumental studies examined OH total reactivity in urban environments, and more especially the OH budget. found that NOx contribution (50 to 55 %) in an urban environment had the greatest effect on OH loss, considering NOx mixing ratios between 10 and 300 ppbv. The next most important contribution came from oxygenated volatile organic compounds (OVOCs; especially aldehydes and ketones), leading to 15 to 25 % of OH destruction with mixing ratios close to 20 ppbv. Some discrepancies exist between this experimental study and the OH budget for the anthropogenic case in our study (Fig. b). Given the NOx mixing ratios simulated in the present study (< 3 ppbv at midday), they are not an important sink for OH as measured by . However, the range of OVOC contribution reported by is comparable to what was found for ALD2 in the anthropogenic case.

In the city of Tokyo, found that the OH budget was dominated by NMHCs and by OVOCs. In their study, the OVOC group, including acetaldehyde, formaldehyde, methanol, ethanol and acetone, contributed up to 18 % of OH reactivity. If aggregated, the species constituting the NMHC group are the predominant sink for OH in our anthropogenic simulation, as observed by . However, C > 2 aldehydes contribute more to OH destruction in our study than in that of . This could result from the cyclic boundary conditions prescribed at the borders of our domain, which can cause the ageing of air masses and increase the mixing ratios of secondary products such as aldehydes.

studied the global distribution and budget of OH radical using the EMAC model (ECHAM/Messy Atmospheric Chemistry) coupled with the Mainz Organics Chemistry (MOM). They found that the annual mean OH reactivity near the surface ranged from 10 to 20 s-1 in southern West Africa, which is in agreement with the results of the anthropogenic case but slightly higher than the values obtained in our biogenic case. However, studied, by means of observations, the effects of seasonality in rainforest air reactivity and noticed that total OH reactivity was much lower during the wet season than during the dry season due to lower temperatures and incoming radiation. More specifically, measurements of reactivity performed during the wet season varied between 6 and 12 s-1 with an average value of 9.9 ± 5.2 s-1 at 24 m, which is much closer to the Meso-NH model results. Moreover, the error made on the total OH reactivity neglecting the turbulent mixing could cumulate with the uncertainties reported in the literature regarding OH reactivity techniques such LIF with a flow tube (from 10 to 15 %, ), LP-LIF (from 10 to 20 %, ) and the CRM measurement method (15 to 20 %, ). In addition, uncertainties on reaction rate constants are also present in chemical schemes, including those used by numerical models. These uncertainties on reaction rate coefficients range from 5 to 15 %, as suggested by . It is likely that the unaccounted fraction of OH reactivity reported in the literature may be explained at least partially by a combination of the following phenomena similar in intensity: turbulence effects on chemical reactivity and uncertainties on the OH reactivity measurements and reaction rate coefficients.

A numerical simulation coupled with a realistic chemical mechanism was performed with the atmospheric model Meso-NH to study the impact of thermals on the oxidizing capacity of the atmosphere. The fine grid resolution of the LES version of the model made it possible to explicitly resolve the thermals spatially and temporally. Identification of thermals was based on a conditional sampling method relying on a first-order decay passive scalar. The impact of turbulent mixing on chemical species redistribution, and of the consequences on OH reactivity was determined in a natural environment and a more contrasted urban case.

The differentiated transport by thermals is dependent on the chemical lifetime of the compounds, represented by the Damkhöler number. This transport induces inhomogeneous mixing of the species within the boundary layer, with an impact on the mean chemical rate between reactive species. In both natural and urban environments, the top of the boundary layer was the region with the highest calculated segregation intensities but of the opposite sign. Between isoprene and the OH radical, an effective maximum decrease of 30 % of the reaction rate was calculated at the top of the boundary layer in a biogenic environment compared to a perfectly mixed case. In the urban case, the reduction of the mean chemical reaction rate between the OH radical and C > 2 aldehydes reached 8 % at the surface in the early morning while this reaction increased by 16 % at the top of the boundary layer during most of the simulation. Thermals transporting species emitted at the surface can lead to different chemical regimes inside updrafts and the environment. For both cases, the surface and the thermals are the preferential reaction zones, with highest chemical reactivity. This was especially the case for the OH radical whose precursors are either transported by thermals or created inside them. OH reactivity was always higher by 15 to 40 % inside thermals compared to their surroundings, depending on the time of day. For the natural case, the major OH precursors close to the surface were radicals originating from the oxidation of isoprene and its degradation products, whereas the O1D + H2O reaction became more predominant with increasing altitudes. In the urban case, OH was mainly produced through the reaction between HO2 and NO, at the surface or higher in the boundary layer. This led to a higher oxidation capacity in the air transported by thermals for both cases.

The overall impact of turbulence on OH concentrations and reactivity at the domain scale differs depending on the chemical environment considered. In a biogenic environment with low OH mixing ratios varying from 0.18 to 0.24 ppt, turbulent structures had little impact on the redistribution of OH in the boundary layer. This was due to an efficient OH recycling initiated by peroxy radicals formed by BVOC oxidation. In the anthropogenic case, OH mixing ratios ranged from 0.26 to 0.50 ppt. The turbulence significantly impacted the spatial distribution of OH and its precursors in the boundary layer, with higher mixing ratios in thermals.

The mean relative error on domain-averaged OH reactivity revealed that effective OH reactivity (taking into account segregation by turbulent motions) in the biogenic case was up to 9 % below the OH reactivity calculated based on averaged boundary layer mixing ratios. Accounting for inhomogeneous mixing between OH and its reactants (primarily isoprene) in a regional or global model could lower the calculated OH reactivity and increase the discrepancies with observed OH reactivities. In the urban environment, the mean relative error was slightly positive which means that air mass segregation by turbulence increases OH reactivity. Considering the effect of turbulent motions could reduce the gap between modeled and observed OH reactivity. However, segregation alone is unlikely to resolve the underestimation between observed and modeled OH reactivity.

This study addressed the impact of moist thermals on the oxidative capacity of the atmosphere on two contrasted chemical situations in a wet environment represented by the monsoon flow. However, observed a substantial seasonal cycle in OH reactivity over an Amazonian forest ranging from 10 s-1 in the wet season to 62 s-1 in the dry season. It would be interesting to assess the impact of turbulent mixing on chemistry in the dry season over the southern part of West Africa. Moreover, the influence of urban area in this study was only linked to chemical emissions. This work should be repeated by taking into account the dynamical forcing due to the presence of the urban area as in in which the authors introduced heterogeneous surface conditions over forest and savannah patches. They showed that the difference in buoyancy fluxes at the surface could have an impact on the redistribution of species, and thus on segregation. Finally, the presence of clouds was considered only from a dynamical view. Adding aqueous-phase chemistry in these simulations could provide further insight into the impact of moist thermals on chemical reactivity.