Ozone depletion events (ODEs) are a common occurrence in the boundary layer during Arctic spring. Ozone is depleted by bromine species, which are most likely emitted from snow, sea ice, or aerosols in an autocatalytic reaction cycle. Previous three-dimensional modeling studies of ODEs assumed an infinite bromine source at the ground. In the present study, an alternative emission scheme is presented in which a finite amount of bromide in the snow is tracked over time. For this purpose, a modified version of the Weather Research and Forecasting model coupled with Chemistry (WRF-Chem) is used to study ODEs in the Arctic from February to May 2019. The model data are compared to in situ measurements, ozone sonde flights, and satellite data. A simulation of the ODEs in the Arctic spring of 2009 using the infinite-bromide assumption on first-year (FY) ice is transferred to the spring of 2019, which achieves good agreement with the observations; however, there is some disagreement in April 2009 and 2019 with respect to an overestimation concerning both the magnitude and the number of ODEs. New simulations using the finite-bromide assumption greatly improve agreement with in situ observations at Utqiaġvik, Alaska, Zeppelin Mountain, Svalbard, and Pallas, Finland, in April 2019, suggesting that bromide on the sea ice is depleted to an extent that reduces the bromine release. The new simulations also slightly improve the agreement with observations at these sites in February and March. A comparison to measurements near Eureka, Canada, and Station Nord, Greenland, shows that multi-year ice and possibly snow-covered land may be significant bromine sources. However, assuming higher releasable bromide near Eureka does not remove all disagreement with the observations. The numerical results are also compared to tropospheric-BrO vertical column densities generated with a new retrieval method from TROPOspheric Monitoring Instrument (TROPOMI) observations. BrO vertical column densities (VCDs) above 5×10¹³ molec. cm⁻² observed by the satellite agree well with the model results. However, the model also predicts BrO VCDs of around 3×10¹³ molec. cm⁻² throughout the Arctic and patches of BrO VCDs of around 10¹⁴ molec. cm⁻² not observed by the satellite, especially near Hudson Bay. This suggests that snow at Hudson Bay may be a weaker bromine source in late spring compared to snow in the north.

A spray flamelet/progress variable approach is developed for use in spray combustion with partly pre-vaporised liquid fuel, where a laminar spray flamelet library accounts for evaporation within the laminar flame structures. For this purpose, the standard spray flamelet formulation for pure evaporating liquid fuel and oxidiser is extended by a chemical reaction progress variable in both the turbulent spray flame model and the laminar spray flame structures, in order to account for the effect of pre-vaporised liquid fuel for instance through use of a pilot flame. This new approach is combined with a transported joint probability density function (PDF) method for the simulation of a turbulent piloted ethanol/air spray flame, and the extension requires the formulation of a joint three-variate PDF depending on the gas phase mixture fraction, the chemical reaction progress variable, and gas enthalpy. The molecular mixing is modelled with the extended interaction-by-exchange-with-the-mean (IEM) model, where source terms account for spray evaporation and heat exchange due to evaporation as well as the chemical reaction rate for the chemical reaction progress variable. This is the first formulation using a spray flamelet model considering both evaporation and partly pre-vaporised liquid fuel within the laminar spray flamelets. Results with this new formulation show good agreement with the experimental data provided by A.R. Masri, Sydney, Australia. The analysis of the Lagrangian statistics of the gas temperature and the OH mass fraction indicates that partially premixed combustion prevails near the nozzle exit of the spray, whereas further downstream, the non-premixed flame is promoted towards the inner rich-side of the spray jet since the pilot flame heats up the premixed inner spray zone. In summary, the simulation with the new formulation considering the reaction progress variable shows good performance, greatly improving the standard formulation, and it provides new insight into the local structure of this complex spray flame.

A turbulent ethanol spray flame is characterized through quantitative experiments using laser-based imaging techniques. The experimental data set is used to validate a numerical code for the simulation of spray combustion. The spray burner has been constructed to facilitate the computational treatment of the experiment; in particular the spray flame is stabilized without a bluff body or a pilot flame. The experiments include spatially resolved measurements of droplet sizes (Mie/LIF-dropsizing and PDA), droplet velocity (PDA), liquid-phase temperature (two-color LIF temperature imaging with Rhodamine B) and gas-phase temperature (multi-line NO-LIF temperature imaging). The measurements close to the nozzle exit are used to determine the initial conditions for numerical simulations. A novel probability density function (PDF) method is applied to calculate the development of the spray flame. A joint mixture fraction and enthalpy PDF is formulated. Its transport equation is modeled and solved using a Monte-Carlo method. A detailed ethanol/air combustion mechanism consisting of 38 species and 337 elementary reactions is implemented through the spray flamelet model enabling the prediction of pollutant emission in spray flames. Good agreement with the experimental data is found for the gas temperature. The numerical predictions for the liquid-phase temperature are in reasonable agreement with the experimental data. The flame structure with two reaction zones is compared with other spray flames, and it is analyzed with the help of the experimental and numerical results. The formation mechanism of such a structure is revealed.

The modeling and simulation of the heating and evaporation of a spherically symmetric single, bi-component ethanol/water droplet in convective air is studied which allows for an improved understanding of the different heating and evaporation characteristics of the two components of the bi-component droplet. The concentration inside the droplet interior is considered to be uniform and the temperature is modeled with the distillation-limit model. The Abramzon–Sirignano model is used for describing the convective droplet heating and evaporation, and Brenn’s extension for multicomponent droplets is used for modeling the ethanol/water droplet. The UNIFAC method is employed to describe the activity coefficients to account for the non-ideal behavior of the ethanol/water mixture. The numerical results corresponding to non-ideal and ideal mixing are compared with experimental data from the literature. Moreover, a parameter study of the ethanol/water droplet evaporation in both dry and humid air for combustion conditions is performed where both the initial gas temperature and velocity and the initial droplet size and composition are varied. The evolution of the droplet surface area, the mass fractions, the mass evaporation rate, and the surface temperature are analyzed for the different conditions. In general, the UNIFAC method initially leads to a somewhat faster evaporation rate and slows down after the higher volatile component has almost completely evaporated; however, its effect on the droplet lifetime is small. Humidity strongly affects the bi-component droplet heating and evaporation because of the condensation of water on the droplet surface which retards the water evaporation but enhances the initial ethanol evaporation for high relative humidity. Also, in humid conditions, the droplet lifetime is strongly increased compared to evaporation in dry air. The parametric dependencies of droplet lifetime and the evaporation characteristics of the bi-component ethanol/water droplet evaporation in both dry and humid convective air may be used in more complex simulations of non-reacting and reacting spray flows.

Numerical simulations of a turbulent ethanol/air spray flame are performed using a combined formulation of a joint gas mixture fraction-enthalpy probability density function (PDF) and a spray flamelet model, where the scalar dissipation rate of the mixture fraction is described with a transport equation. Spray evaporation is included in the flamelet model through the pre-calculation of counterflowing laminar spray flames with a detailed reaction mechanism of 38 species and 337 elementary reactions for ethanol/air combustion. A transport equation for the mean scalar dissipation rate of the mixture fraction is derived, taking into account the spray evaporation. The computed results are compared with the experimental data provided by A. Masri, University of Sydney, Australia. All major spray flame characteristics, including the Sauter mean diameter, the droplet velocity, and its turbulent fluctuations as well as gas temperature, generally show good agreement with experiments. Different models for the scalar dissipation rate of the mixture fraction are not significantly different for the flame under consideration except for inside the chemical reaction zone, where the new model shows some improvement compared to the experiment.