Impact of Medium-Energy Electrons on Antarctic Stratospheric Ozone During 2013–2014 Simulated with the WACCM–SIC Model

Abstract

The Antarctic stratospheric ozone plays a crucial role in the polar climate system and is strongly influenced by energetic particle precipitation. Among these processes, medium-energy electron (MEE) precipitation enhances the production of odd nitrogen (NOx) in the polar mesosphere and stratosphere, thereby driving ozone depletion through catalytic reactions. However, quantifying its atmospheric impact remains challenging, largely because the spatial and temporal variability of MEE is poorly constrained, and most current global chemistry–climate models lack a realistic MEE forcing. This study employs the Whole Atmosphere Community Climate Model coupled with Sodankylä Ion Chemistry (WACCM–SIC) to investigate the influence of MEE precipitation during 2013–2014, when moderate geomagnetic storms were more frequent in the winter of 2013. A control simulation (Case1) and two sensitivity experiments (Case 2 and Case 3) were conducted to isolate MEE-driven effects. Model-simulated NO_x (NO + NO₂) and ozone concentrations agree well with satellite observations, indicating that WACCM–SIC captures the key photochemical and dynamical processes. The results further suggest that the direct impact of MEE precipitation on the middle and lower atmosphere during winter is relatively weak. Nevertheless, MEE-generated NO_x can be efficiently transported downward within the polar vortex, reaching altitudes below 15 km. In these regions, MEE-related NO_x enhancement can reach up to 5%, with values during the winter of 2013 approximately twice those in 2014. Sensitivity experiments further reveal that enhanced NO_x leads to pronounced ozone depletion in the lower stratosphere, with ozone losses reaching up to 25%. A clear negative relationship between NO_x and ozone is therefore evident, highlighting the importance of accurately representing MEE precipitation in chemistry–climate models.

1. Introduction

Antarctic ozone variability plays an important role in the global climate system, and has been the subject of sustained scientific interest since the discovery of the ozone hole [1]. Reactive nitrogen species, dominated by odd nitrogen (NO_x, NO and NO₂), regulate the photochemical balance of stratospheric ozone through catalytic cycles and represent a key factor controlling polar ozone variability. The spatiotemporal distribution of NOx is controlled by both dynamical transport and chemical processes. Among these, energetic particle precipitation (EPP) events are especially critical in the mesosphere and stratosphere by producing NO_x through ionization, dissociation, and recombination reactions, subsequently leading to catalytic ozone loss [2,3,4,5]. Influenced by solar ultraviolet and X-ray radiation, as well as EPP from the magnetosphere, the neutral atmosphere in the thermosphere becomes partially ionized, forming the ionosphere. The ionosphere is strongly coupled with both the middle and upper atmosphere and the magnetosphere [6]. EPP represents an important forcing mechanism in the polar middle and upper atmosphere. In the neutral atmosphere, EPP induces ionization and leads to the production of odd nitrogen (NO_x = NO + NO₂), thereby enhancing NO_x concentrations [7,8]. NOx plays an important role in the catalytic reaction cycles responsible for ozone destruction in the upper stratosphere and mesosphere [9]. EPP can be classified into several particle types that affect atmospheric chemistry. Solar proton events (SPEs), the major form, can cause rapid increases in NO_x and H₂O concentrations within a few days and result in severe stratospheric ozone losses by 60% [10], but they are sporadic and short-lived, typically occurring only during solar maximum and up to one to two years thereafter [11]. In contrast, medium-energy electron (MEE) precipitation persists for much longer timescales in the polar mesosphere and stratosphere, enhances ionization, substantially increases NO_x and HO_x (hydrogen oxide radicals, including OH and HO₂) concentrations, and leads to sustained ozone depletion [12]. Other EPP sources, such as auroral electrons and galactic cosmic rays (GCRs), primarily influence the thermosphere at high altitude and the lower stratosphere and troposphere, respectively, with GCRs contributing less than 1% NO_x production and exerting limited impacts on the stratosphere and lower mesosphere [13].

In recent years, the mechanisms by which SPEs induce short-term perturbations to the ozone layer have become relatively well understood, with enhanced production of NO_x in the polar middle atmosphere causing pronounced but transient ozone depletion [14,15,16]. In contrast, the long-term impacts of MEE precipitation on ozone remain less well constrained. Numerical modeling studies have shown that NO_x produced by MEE can persist for several months under polar night conditions and can be continuously transported downward into the stratosphere by the polar vortex circulation, resulting in a gradual wintertime ozone loss [12,17]. Observations from Antarctic stations further show rapid and pronounced enhancement of NO in the upper mesosphere and lower thermosphere (approximately 70–105 km) during and following geomagnetic storms, with associated ozone depletion up to 20–70% [18,19]. Previous studies using Whole Atmosphere Community Climate Model (WACCM) simulations that did not include MEE forcing substantially underestimated NO_y and O₃ concentrations near 80 km by one to two orders of magnitude compared with observations, while enhanced MEE forcing has been shown to produce up to approximately 25% ozone loss in the polar stratosphere at altitudes of 28–45 km, consistent with ACE–FTS satellite observations [12].

Despite growing evidence that MEE plays a significant role in driving stratospheric ozone variability, its quantitative impacts remain highly uncertain. This uncertainty stems from limitations in both data and models, including discontinuous satellite observational coverage, insufficient reanalysis resolution, and the historically incomplete representation of MEE precipitation in atmospheric models [20]. At present, most global climate models account for atmospheric ionization driven by solar X-rays, solar protons, and radiation belt electrons, but they lack a realistic representation of MEE precipitation associated with solar activity. Thus, to address these limitations and more accurately simulate the effects of MEE precipitation in the polar atmosphere, this study employs WACCM–SIC model simulations to investigate the influence of MEE on Antarctic stratospheric ozone variability. Specifically, the main objective is to quantify the impact of MEE precipitation in the polar mesosphere and stratosphere and to evaluate its contribution to ozone depletion through NOx production and downward transport. Using the WACCM–SIC framework, we further examine the impacts of MEE precipitation (30–1000 keV) on the middle and upper atmosphere. In this study, we use a newly coupled ion-neutral chemistry model, WACCM–SIC, which integrates the Sodankylä Ion Chemistry (SIC) model with the WACCM to simulate ionospheric D-region processes at altitudes of approximately 50–90 km [21]. The WACCM–SIC model couples WACCM with a comprehensive ion-neutral chemistry framework that is reduced using the structural equation modeling-connectivity minimization (SEM–CM) method to balance chemical fidelity and computational efficiency, enabling effective representation of the D-region ionospheric process in this region. The analysis focuses on NO_x variability during 2013–2014, a period selected because the winter of 2013 experienced more frequent moderate geomagnetic storms than 2014, providing favorable conditions for such examination.

2. Models and Methods

The simulations are based on the Community Earth System Model (CESM) framework developed by the National Center for Atmospheric Research (NCAR). The WACCM includes ion chemistry processes as well as key external forcings, such as the solar cycle, SPEs, auroral electrons, and MEE precipitation. Building upon the standard chemical scheme of WACCM, the WACCM–SIC configuration incorporates 306 additional ion-neutral and ion recombination reactions involving neutral species, positive ions, negative ions, and electrons. This extension enables the explicit simulation of NO_x and HO_x production driven by MEE and auroral particle precipitation. The WACCM–SIC model is operated in the specified-dynamics (SD) configuration, with a horizontal resolution of 1.9° × 2.5° (latitude × longitude) and 88 vertical levels, driven by the F_SD_WACCM dataset provided within the standard WACCM configuration. The simulation period spans 2013–2014, and hourly model output from this period is analyzed. The WACCM–SIC model is operated in the specified–dynamics (SD) configuration, with a horizontal resolution of 1.9° × 2.5° (latitude × longitude) and 88 vertical levels. All simulations in this study are conducted at the native model resolution, and no additional spatial refinement procedures (e.g., dynamical nesting or statistical downscaling) are applied. The improvements mentioned in the revised manuscript refer only to visualization quality and do not involve modifications to the model resolution.

The ionization rate calculation is based on the solar forcing implemented in WACCM, which adopts daily varying input from the Coupled Model Intercomparison Project Phase 6 (CMIP6) provided via the SOLARIS–HEPPA framework (https://solarisheppa.geomar.de/cmip6, accessed on 10 October 2024) and has been described in detail in previous studies [22]. Ionization rates are calculated offline using the atmospheric model described by Picone [23], with comprehensive documentation available for the CMIP6 ionization dataset [24].

Based on the WACCM–SIC model, three sets of experiments are designed for the Antarctic region during 2013–2014. A control experiment with standard MEE forcing (Case 1, MEE), a sensitivity experiment without MEE forcing (Case 2, NOMEE), and a sensitivity experiment with doubled MEE forcing (Case 3, DOUBLEMEE). All simulations are performed under the same model configuration, with the MEE forcing being the only varying factor among the experiments. The MEE forcing is derived from CMIP6 output following Matthes [22]. All simulations are performed using the specified-dynamics (SD) version of WACCM constrained by MERRA reanalysis below 60 km [25,26]. The Prandtl number is set to 2, and solar forcings, including F10.7, Ap, SPEs, and spectral solar irradiance (SSI), are updated consistently [27]. It should be noted that the frontal-scale dynamics and waves (FSDW) configuration does not include MEE forcing but accounts for auroral electrons and SPEs. Case 2 excludes the MEE forcing while retaining auroral precipitation, and it is therefore taken as a sensitivity simulation.

Model performance is evaluated using observations from the Atmospheric Chemistry Experiment–Fourier Transform Spectrometer (ACE–FTS). The ACE instrument is aboard the Science Satellite (SCISAT), launched on 12 August 2003 and operating in a circular low-Earth orbit at an altitude of approximately 650 km with a 74° inclination and a 97.6 min orbital period. The ACE–FTS measures infrared solar occultation spectra, providing up to 15 sunrise and 15 sunset occultations per day [24]. It retrieves a wide range of atmospheric parameters, including temperature and volume mixing ratios (VMR) of more than 30 trace gases. The measurements cover altitudes from the surface to about 150 km, with a vertical resolution of roughly 1 km. Owing to its extensive polar coverage and high data quality, ACE–FTS observations are well–suited for characterizing atmospheric chemical variability at middle and high latitudes. Thus, ACE–FTS data over the Antarctic region during 2013–2014 are used in this study to evaluate and validate the WACCM–SIC simulation results.

3. Evaluation of Model Performance

3.1. Geomagnetic Conditions

Figure 1 illustrates the latitudinal distribution of ionization rates at an altitude of approximately 80 km over the Antarctic region (60° S–90° S). The spatial distribution of ionization is strongly asymmetric, with a pronounced maximum (marked by a star; 59.9896° E, 76.4864° S; referred to as the starred location hereafter). This spatial asymmetry is primarily associated with the particle loss cone drift, whereby energetic electrons gyrating along geomagnetic field lines and bouncing between the two magnetic poles are preferentially lost through collisions in denser atmospheric regions at high latitudes. This loss process is further enhanced in regions of a weakened geomagnetic field. In the Southern Hemisphere, particle precipitation intensifies at the magnetic field line footprints passing through the South Atlantic Anomaly (SAA), contributing to the observed asymmetry. The southern displacement of the SAA also leads to a hemispheric asymmetry in ionization rates. Previous studies suggest that this pattern may be shaped by the spatial structure of the Earth’s magnetic field, variations in solar activity, and interactions among different atmospheric chemical constituents. In particular, convergence of magnetic field lines near the southern geomagnetic pole enhances interactions between charged particles and atmospheric molecules. Moreover, the precipitation rates of energetic particles from the solar wind may vary with latitude, leading to differences in ionospheric ionization, an effect that becomes more pronounced during periods of high solar activity. In the atmosphere, molecular oxygen and nitrogen play a central role in ionization processes, and latitudinal variations in their abundances can further modulate ionization rates [12].

The location marked at the starred location (59.9896° E, 76.4864° S) corresponds to the Antarctic magnetic conjugate area of field lines extending from the SAA and represents a core area of enhanced MEE precipitation [28]. The reduced geomagnetic field strength and the configuration of the magnetic field lines in this region promote MEE precipitation, leading to a pronounced enhancement of ionization rates in the middle and upper atmosphere [29]. As one of the core regions of enhanced ionization, this location is a suitable testbed for evaluating the magnetosphere–atmosphere coupling in global models. To further examine the fine-scale characteristics of ionization, Figure 2 presents the three-hourly variation in ionization rates at the starred location as a function of altitude. The results show that enhanced ionization occurs in the upper mesosphere and lower thermosphere. Most of the signal is between about 80 and 100 km, and the peak is near 90 km. The ionization is much weaker below about 70 km. During strong events, the signal can extend down to around 70 km, but it becomes weaker quickly. In terms of temporal variability, ionization rates are substantially enhanced from April to September, reaching maximum values of approximately 500 molecules cm⁻³ s⁻¹ and exhibiting a clear diurnal cycle. The temporal variability in the ionization rates discussed above is consistent with previous studies [30].

According to Newnham (2018), no major SPEs occurred during 2013–2014 [11]. This period is therefore selected to minimize the influence of SPEs and to better assess the effects of MEE precipitation. Ionization rates during April–September 2013 were substantially higher than those in 2014, primarily attributed to stronger geomagnetic activity. During the austral winter of 2013 (June–August), nine moderate geomagnetic storms were recorded, whereas only two such events occurred in 2014 near the beginning and end of winter (early May and late August). Figure 2 reveals several pronounced, short-lived ionization enhancement events during March–July and September–October 2013, as well as January–April and September–October 2014. During these episodes, ionization rates increased rapidly over short timescales and propagated downward toward the stratosphere, potentially altering NO_x concentrations and thereby influencing the stratospheric ozone. However, the mechanisms linking ionization variability to ozone changes, as well as the magnitude of this impact, are beyond the scope of this study and will be examined in future work. Here, we focus on quantifying the influence of MEE on the Antarctic stratospheric ozone during 2013–2014.

3.2. Analysis of Simulation Results

Polar stratospheric ozone depletion is mainly driven by catalytic cycling of reactive halogen species, primarily chlorine oxides (ClO) and bromine oxides (BrO), with reactive bromine widely recognized as a key indicator of ozone loss. A key fundamental ozone-destroying reaction is Br + O₃ → BrO + O₂. In contrast, inert chlorine reservoir species, mainly hydrogen chloride (HCl) and chlorine nitrate (ClONO₂), must be converted into chemically active chlorine (ClO_x, including Cl, ClO, and Cl₂O₂) through specific pathways. These key activation processes depend on heterogeneous reactions occurring on the surfaces of polar stratospheric cloud (PSC) particles, including:

The reaction between chlorine nitrate and hydrogen chloride:

ClONO₂ + HCl → Cl₂ + HNO₃

(1)

Hydrolysis of chlorine nitrate:

ClONO₂ + H₂O → HOCl + HNO₃

(2)

Secondary reaction between hypochlorous acid and hydrogen chloride:

HOCl + HCl → Cl₂ + H₂O

(3)

These reactions are strongly accelerated on the surfaces of PSC particles, which is particularly pronounced during the Antarctic winter polar night, when low temperatures favor PSC formation and substantially enhance the decomposition of chlorine nitrate (ClONO₂), thereby releasing large amounts of reactive chlorine. The formation and removal of chlorine nitrate, a major chlorine reservoir species, are regulated by catalytic cycles involving NO_x. Previous studies have shown that under the cold conditions of the Antarctic winter and early spring, hydrogen chloride (HCl) can release chlorine through heterogeneous reactions on ice surfaces. Following the end of the polar night, photochemical processes liberate active chlorine atoms, which subsequently react with ozone (O₃), leading to stratospheric ozone destruction [31,32,33]. NO_x is primarily produced through the photochemical decomposition of nitric acid (HNO₃):

Photolysis of nitric acid:

HNO₃ + hv → OH + NO₂

(4)

Reaction between nitric acid and hydroxyl radicals:

HNO₃ + OH → H₂O + NO₃

(5)

Formation of chlorine nitrate through the reaction of NO₂ with ClO:

ClO + NO₂ + M → ClONO₂ + M

(6)

However, because Reactions (4) and (5) are driven by solar radiation, where hν represents the energy of incoming solar radiation, NO_x production is strongly suppressed in the core region of the polar night, where the lack of sunlight limits photochemical processes and inhibits chlorine activation. This raises uncertainty on whether current theoretical models can accurately represent the accumulation of reactive chlorine during the polar night. To address this issue, this study focuses on the region of maximum Antarctic ozone depletion at 18 km altitude, as it likely represents the maximum of photochemical processes (Case1). Area-averaged time series of simulated ClONO₂, NO_x, and O₃ concentrations are shown in Figure 3 and compared with observations from the Atmospheric Chemistry Experiment (ACE). There are gaps in the ACE–FTS time series. These gaps are caused by the observation method. ACE–FTS measures the atmosphere at sunrise and sunset, so the data are not continuous. The satellite does not always pass over the Antarctic region, so some periods have no observations.

Figure 3 shows the time series of ClONO₂, NOx, and O₃ at 18 km over 60° S–90° S. The O₃ concentrations derived from ACE–FTS observations and the simulations are generally above 1 ppmv during the available periods, and the model captures the overall variability in O₃. In contrast, ClONO₂ concentrations are much lower than O₃ and exhibit clear temporal variability, with a marked decrease from late 2013 to early 2014 followed by a gradual recovery. The simulated ClONO₂ values reproduce this overall evolution, although the modeled variations are smoother than those in the observations. NOx concentrations also show pronounced temporal changes. Both the ACE–FTS observations and simulations remain at relatively low levels during much of 2013, then increase from late 2013 to early 2014, and decrease again afterward. The simulations capture this general pattern and are broadly consistent with the observed NOx variations. Notably, the decrease in ClONO₂ is accompanied by an increase in NOx during the same period, indicating a clear contrast in their temporal evolution. This reduction is associated with the formation of PSCs, which facilitate heterogeneous reactions that release reactive chlorine and subsequently catalyze ozone destruction. Meanwhile, NO_x is reduced through uptake on PSCs, while its subsequent recovery enhances ozone loss through catalytic cycles.

During spring (September–November), nitric acid (HNO₃) reaches a photolytic maximum of approximately 2.5 ppbv in observations and about 2.1 ppbv in simulations, coincident with the simultaneous recovery of both observed and simulated ClONO₂. The increase is driven by gradually strengthening solar radiation in spring, which promotes HNO₃ photolysis. At the same time, the elevated wintertime HNO₃ enhances NO₂ production by efficiently consuming hydroxyl radicals through Reaction (5). The released NO₂ subsequently reacts with ClO to form ClONO₂, which is transported into the polar vortex core by stratospheric circulation. Within the vortex core, limited sunlight suppresses photochemical loss processes, and the stability of ClONO₂ becomes strongly controlled by HCl concentrations. When HCl exceeds approximately 2 ppbv, the rate of heterogeneous Reaction (1) increases markedly, accelerating depletion of ClONO₂ under high–HCl conditions. ClONO₂ can persist within the polar vortex only after the depletion of HCl in September, leading to its sustained accumulation during winter.

Despite differences in spatial resolution between the model (finer) and the ACE satellite observations, simulated ClONO₂ concentrations agree with observations within 8%. Model diagnostics indicate that approximately 70% of ClONO₂ within the vortex core originates from photochemical production at the vortex edge, followed by transport into the interior, while the remaining 30% is produced locally through Reaction (6) involving residual NO₂. This result supports a coupled control mechanism in which solar illumination gradients and dynamical mixing jointly regulate ClONO₂ formation. Overall, the WACCM–SIC model simulations reproduce satellite observations, including ACE–FTS, with concentrations of O₃, HNO₃, HCl, and ClONO₂, with typical biases within ±10%. This agreement further supports the robustness of the WACCM–SIC model in representing polar stratospheric chemistry [22,34,35,36].

Based on the analysis above, ozone variability is influenced by both ClONO₂ and NO_x. To further quantify the evolution of NO_x, spatial averaging is applied over the high-latitude Antarctic region (60° S–90° S). Figure 4 presents time-height cross-sections of NO_x number density derived from different data sources for comparison, revealing pronounced seasonal variability.

A comparison of the three datasets shows that the FSDW simulation (Figure 4a) and Case 2 (NOMEE, Figure 4b) yield broadly similar results. Below 60 km, both simulations are consistent with the observations (Figure 4c). Above 60 km, the observational data are limited by relatively coarse vertical resolution, whereas the model simulations (Figure 4a,b) exhibit a more continuous structure. The observed NO_x values (Figure 4c) are generally lower than those simulated (Figure 4a,b), indicating an overall tendency of the model to overestimate NO_x number density. Nevertheless, relative to the FSDW simulation, Case 2 reproduces the magnitude of NO_x more realistically and its observed downward extension, indicating a closer agreement with the observations and achieving a better overall fit than the FSDW configuration.

In the mesosphere and lower thermosphere (50–90 km), NO_x exhibits relatively small variability on daily to weekly timescales, whereas its seasonal and interannual variations are much more pronounced. During winter, NO_x number densities increase below 100 km with a minimum observed in the 0–20 km layer, whereas in summer NO_x shows a decrease followed by an increase with height, with a local maximum between 0 and 30 km and a minimum near 80 km. Interannual variability reveals that during winter, a tongue-shaped region of enhanced NO_x in the mesosphere and lower thermosphere (50–90 km) gradually descends over time and propagates toward lower altitudes, merging with background NO_x near 80 km (Figure 4a–c). This feature is associated with MEE precipitation, which enhances auroral electron fluxes that dominate ionization in the lower thermosphere. Consistently, periods of enhanced geomagnetic conditions (geomagnetic index Kp ≥ 5) are accompanied by substantially increased NO_x production. In both the FSDW-forced simulation and the WACCM–SIC simulation, electron forcing above approximately 90 km is nearly identical, suggesting that differences at lower altitudes primarily reflect transport and chemical processes.

In the Southern Hemisphere, EPP-related NO_x production, a major source of NO_x in the polar mesosphere and stratosphere [37,38], exhibits greater vertical extent and larger interannual variability. NO_y produced by EPP is initially generated in the lower thermosphere (80–120 km) and is subsequently transported downward into the mesosphere and stratosphere (45–70 km) by subsidence driven by the polar night jet. During winter, persistent low-light conditions in the middle and upper atmosphere (>45 km) strongly suppress the photochemical NO_y loss, primarily through NO photolysis and reactions involving nitrogen atoms. Consequently, NO_y lifetimes extend from weeks to several years, facilitating its accumulation toward late winter and early spring. Model simulations further indicate that the tongue-shaped NO_y structures observed in spring result from enhanced meridional mixing following polar vortex breakup.

Consistent with these production and transport mechanisms, the spatiotemporal distribution of NO_y (NOy = NO + NO₂ + HNO₃ + N₂O₅) over the high-latitude Antarctic region (60° S–90° S) shows pronounced seasonal variability in NOy density (Figure 5), similar to that of NO_x density. Area-averaged time–height distributions reveal that during summer, NO_y concentrations decrease with altitude before rising again, with a minimum near 80 km. In winter, a tongue-shaped NO_y enhanced region appears in the mesosphere and lower thermosphere (50–90 km) and gradually propagates to lower altitudes (Figure 5a,b) and merges with the background NO_y concentrations near 80 km. Across the 40–80 km altitude range, summer NO_y concentrations are lower than those in winter. This seasonal contrast reflects stronger ultraviolet radiation in the upper stratosphere and lower mesosphere during summer, which enhances the net photochemical loss of NO_y and increases susceptibility to radiative depletion.

Observations further reveal a bimodal seasonal cycle in NO_y, with a primary maximum in September and a secondary maximum during February–March. This feature is particularly pronounced in years associated with sudden stratospheric warming (SSW) events, such as 2004, 2006, 2009, and 2012. The primary maximum is attributed to enhanced polar upwelling associated with SSW-induced stratospheric elevation, which facilitates the vertical transport of NO_y from lower levels into the mesosphere and stratosphere. The secondary maximum arises from intensified meridional mixing during the later stages of the warming, leading to horizontal transport of NO_y from lower latitudes into the polar region. During winter, the NO_x enhancement induced by MEE precipitation accelerates chlorine activation through Reaction (6), while the enhanced NO_x family contributes to ozone loss in the 30–50 km layer via catalytic cycles, accounting for up to 18% of the total ozone depletion in this region. In summer, this mechanism is substantially weakened due to enhanced photolysis, indicating a pronounced seasonal dependence of MEE-related chemical effects.

Variations in NO_y further modulate NO_x production and storage, thereby influencing photochemical ozone loss rates. In particular, during winters with enhanced MEE activity, increased vertical transport of NO_y can suppress catalytic chlorine (ClO_x) cycles through the formation of ClONO₂, ultimately affecting ozone variability. To quantify this MEE-driven chemical response on stratospheric ozone, results from sensitivity experiments are examined. Figure 6 illustrates the relative differences in vertical distributions of NOx, ClONO₂, and O₃ at the star between simulations with and without MEE forcing, calculated as (MEE − NOMEE)/NOMEE. December and July are selected to represent the austral summer and winter, respectively. Hourly model outputs averaged over July and December 2013 for the Antarctic region (60° S–90° S) are used in the analysis.

During winter, MEE induces pronounced perturbations in NOx and ClONO₂ (Figure 6a). In the 30–70 km altitude range, NO_x changes exceed 30%, while ClONO₂ anomalies reach up to 180% near 90 km. Despite these substantial chemical perturbations, the wintertime response of O₃ remains relatively weak. In summer, MEE-induced disturbances are mainly confined to the 70–90 km region (Figure 6b), with NO_x and ClONO₂ anomalies reduced to approximately 20% and 30%, respectively, whereas O₃ exhibits a stronger response of about 15%. This seasonal contrast reflects the role of Reaction (6) that converts reactive chlorine and NO_x into ClONO₂, thereby weakening both ClO_x- and NO_x-driven catalytic cycles and buffering the ozone loss. This buffering mechanism is strongly modulated by solar illumination and temperature. Although MEE enhances NOx in the 30–70 km layer during winter, the pronounced increase in ClONO₂ near 90 km reduces the mixing ratios of ClO, NO, and NO₂, slowing both catalytic cycles and yielding a weak net O₃ response. In contrast, during summer, continuous solar illumination accelerates ClONO₂ photolysis, releasing ClO and NO₂ and reactivating catalytic cycles, leading to enhanced ozone depletion accompanied by a depletion of the ClONO₂ reservoir. To further elucidate the processes by which MEE affects stratospheric ozone, a more detailed analysis is presented in the next section.

4. Impacts of Medium-Energy Electrons on Stratospheric Ozone

4.1. Response of Polar Stratospheric Ozone to NO_x Descent

This study compares the vertical transport of NO_x at the Antarctic star (59.9896° E, 76.4864° S) during 2013–2014 by comparing simulations with and without MEE forcing. Relative differences are expressed as (MEE − NOMEE)/NOMEE × 100. MEE-induced ionization peaks near the upper mesosphere (~80 km) in the polar region, with elevated NO_x concentrations up to 200% relative to the NOMEE case (Figure 7). This upper-mesospheric maximum aligns with the energetic electron penetration depth shown in Figure 2, where enhanced ionization leads to substantial NO_x production.

During winter (June–July), regions of elevated NO_x (>150%) descend from approximately 80 km into the middle stratosphere (~30 km), driven by dynamical subsidence associated with the polar vortex circulation. This process produces a tongue-shaped structure extending from the mesosphere into the stratosphere, exhibiting pronounced seasonal evolution. By September–October, the tongue-shaped structure reaches altitudes near 20 km. Still, the descent rate slows below 50 km due to constraints imposed by stratospheric circulation. Even when polar vortex-driven subsidence extends below 15 km during winter, MEE still leads to an increase in NO_x of approximately 5%. Moreover, MEE-related NO_x enhancements persist and intensify from April to December, demonstrating that MEE modulates both the depth and intensity of NO_x vertical transport and thereby drives changes in stratospheric ozone chemistry.

Figure 8 further presents the relative percentage differences in the vertical distribution of O₃ at the star (59.9896° E, 76.4864° S) during 2013–2014 under conditions with and without MEE forcing. From March to September, negative ozone anomalies propagate downward within the 15–40 km altitude range, mirroring the downward evolution of NO_x anomalies. This ozone depletion is primarily associated with enhanced NO_x, which alters HO_x chemistry and enhances ozone destruction. On an annual mean basis, MEE contributes approximately 5% to ozone loss, with a wintertime reduction reaching 15%. In the mesosphere, a seasonal reversal is evident: positive ozone anomalies during spring and autumn arise from NO_x-induced suppression of ClO- and BrO-driven catalytic activity, whereas the response weakens in winter due to the stagnation of photochemical processes under polar night conditions. Notably, ozone concentrations increase markedly near 80 km in October, reaching a local maximum. This enhancement results from elevated NO_x converting reactive chlorine (Cl) and bromine (Br) into reservoir species, thereby reducing catalytic ozone loss. This behavior highlights the sensitivity of mesospheric halogen chemistry to EPP.

Previous studies [14] have shown that MEE exerts an indirect influence on the stratosphere. Over Antarctica, a tongue-shaped structure of enhanced NO_x within the 15–65 km altitude range begins to descend from autumn onward, reducing stratospheric ozone. Because MEE does not directly affect the stratosphere, the ozone changes observed near 30 km in early winter can be attributed to chemical loss driven by NO_x descending from the mesosphere, including residual NO_x produced during the preceding summer. Observations confirm that the transport of NO_x from the mesosphere (~80 km) into the middle stratosphere (~30 km) forms a characteristic tongue-shaped structure. Specifically, during June–July, regions of enhanced NO_x descend from about 80 km into the middle and lower stratosphere, and by September–October, the descent can extend to altitudes near 20 km. This behavior reflects auroral and photoelectron-driven dissociation of molecular nitrogen in the thermosphere, which produces large amounts of atomic nitrogen that subsequently forms NO_x. Effective downward transport of this NO_x requires prolonged polar night conditions, as the daytime lifetime of NO_x in the thermosphere and mesosphere is less than one day. In addition, during the cold Antarctic winter, the NO_y in the stratosphere does not exist entirely in the gaseous phase, further enhancing NO_y variability. Consequently, Antarctic winter exhibits the most pronounced variability in NO_y concentrations.

The impacts of EPP depend on its chemical coupling with NO_x and NO_y, and their interactions with radiatively active gases such as ozone. These processes operate primarily through the downward transport of NO_x into the stratosphere, where associated ozone changes can further modulate stratospheric circulation and potentially influence longer-term variability.

4.2. Impacts of MEE on NO_x and Stratospheric Ozone

Based on the analyses above, NO_x and O₃ exhibit a strong correspondence in both time and altitude under different MEE conditions. To better quantify the magnitude of the MEE impact on NO_x and O₃, the analysis focuses on July–October, when the downward signals of NO_x and O₃ are most pronounced. The model results from three experiments are used to quantitatively assess the effects of MEE on NO_x and O₃, as shown in Figure 9 (panels a,c,e,g for NOx and panels b,d,f,h for O₃).

The NOx profiles show clear differences between the MEE and NOMEE simulations (Figure 9a,c,e,g). In July and August (Figure 9a,c), NOx concentrations are higher in the upper mesosphere in the MEE case compared with the NOMEE case. From September to October (Figure 9e,g), the differences between the two simulations extend to lower altitudes, indicating a downward propagation of the enhanced NOx signal. The ACE–FTS observations generally fall between the two simulations at most altitudes, although deviations are present at certain levels.

The corresponding O₃ profiles also show systematic differences between the simulations (Figure 9b,d,f,h). Compared with the NOMEE case, the MEE simulation exhibits lower O₃ concentrations in the upper mesosphere, particularly in July and August (Figure 9b,d). From September to October (Figure 9f,h), the differences in O₃ between the simulations become less pronounced at higher altitudes but remain evident at lower levels. The ACE–FTS observations are generally closer to the MEE simulation in the upper mesosphere, while larger discrepancies appear at lower altitudes.

During July–August, the vertical transport from the mesosphere (~80 km) to the middle stratosphere (~30 km) reveals clear gradients in NO_x concentrations among the three experiments: NO_x (NOMEE) > NO_x (MEE) > NO_x (DOUBLEMEE), while the ozone shows the opposite response, with O₃ (NOMEE) < O₃ (MEE) < O₃ (DOUBLEMEE), consistent with the findings presented earlier. The differences are most pronounced at 30 km during August and September and gradually diminish in October. In the lower mesosphere, the contribution of MEE is more evident during austral summer. In the stratosphere near 30 km, however, the downward propagation of the MEE signal leads to enhanced NOx and a corresponding reduction in O₃.

Focusing on the star (59.9896° E, 76.4864° S), the relationship between NO_x and ozone under different forcing conditions is further examined. Figure 10 presents the spatiotemporal correlation between ΔNO_x (the difference in NO_x between Case1 and Case2) and ΔO₃ (the difference in O₃ between Case1 and Case2) associated with MEE precipitation, as simulated by WACCM. The results show a clear negative correlation between NO_x and O₃ from July to November in both 2013 and 2014, descending from the mesosphere into the upper stratosphere. The strongest negative correlation is found near 30 km and persists until the end of austral summer (February). Notably, owing to stronger ionization activity in 2013 than in 2014, the negative correlation signal in 2013 appears near ~30 km in March and extends downward to lower altitudes during the following months within the time range shown in Figure 10, while the corresponding signal in 2014 does not exhibit a similarly clear and continuous downward structure over the same period. This contrast can be explained by two main factors: (1) enhanced PSC occurrence during the Antarctic winter (June–August) under polar night conditions accelerates NO_x-driven ozone loss, and (2) nitrogen oxides in NO_x, particularly NO and NO₂, efficiently destroy ozone through catalytic cycles. The weaker ionization in 2014 results in reduced NO_x production and consequently a much weaker downward-propagating signal.

As shown in Figure 8, during the austral winter (June–August) of 2013, the relative percentage differences in ozone at 30–40 km in the middle stratosphere reach approximately −25% to −30%. In 2014, ozone loss during the same period was reduced by about 40%. In 2013, the ozone anomaly persisted for nearly eight months and propagated downward to altitudes as low as 10 km, whereas no comparable descent was observed in 2014. In the 90–100 km region, the relative percentage differences remain within approximately −5% to −10% throughout the year, indicating that solar illumination suppresses MEE-induced disturbances in both the mesosphere and the stratosphere.

At an altitude of 30 km in the middle stratosphere, NO_x exhibits a pronounced negative correlation with ozone, with wintertime ozone loss in 2013 reaching up to −30%. During this year of strong ionization, the signal extends downward through vertical transport and approaches the tropopause region. In contrast, ozone loss in 2014 was reduced by approximately 40%, and no downward propagation was observed, primarily due to weaker ionization and enhanced chemical removal of NO_x by PSCs under polar night conditions.

Across the 0–100 km altitude range, annual mean ozone loss remains limited to approximately −5% to −10%, confirming that strong solar illumination suppresses MEE-induced perturbations. Together, the NO_x catalytic cycle and MEE-modulated vertical transport constitute the core mechanisms governing ozone depletion.

5. Conclusions

Using the WACCM–SIC model, together with MERRA reanalysis data, daily solar forcing parameters from CMIP6, ACE–FTS satellite observations, and radar measurements at an Antarctic reference site, this study applies a combination of numerical simulations, control experiments, and sensitivity experiments to investigate the impact of medium-energy electron (MEE) precipitation on Antarctic ozone. The main conclusions are as follows:

• By incorporating detailed D-region ion-neutral chemistry and refined MEE ionization parameters, WACCM–SIC successfully reproduces the vertical distributions of NO_x and ozone over Antarctica. The differences between the simulated results and ACE–FTS observations are generally within 15%, demonstrating the model’s capability in realistically representing photochemical–dynamical coupling processes in the polar atmosphere.
• Sensitivity experiments indicate a strong link between MEE-induced increases in NO_x and ozone loss. In the lower stratosphere (15–25 km), a 10% increase in NO_x leads to a 2.5–3.2% reduction in ozone, with the maximum response occurring near 20 km. During June–August 2013, ozone depletion at this altitude reached up to 25%.
• The MEE ionization rate increases with geomagnetic activity and dominates atmospheric ionization in the polar mesosphere and stratosphere. During winter, MEE contributions to NO_x increase significantly within the 70–90 km altitude range, and the associated perturbation signal can extend below 15 km through polar vortex-driven subsidence, contributing up to 5%. The combined effects of NO_x catalytic cycles and MEE-driven vertical transport constitute the core mechanisms of ozone depletion, with this process being particularly pronounced under polar night conditions.

Together, these findings demonstrate that MEE precipitation is not a minor or episodic influence but a systematic and dynamically integrated driver of polar ozone variability. Accurately representing MEE forcing is therefore essential for chemistry–climate models seeking to capture Antarctic ozone trends, diagnose interannual variability, and assess the broader coupling between space-weather processes and Earth’s atmosphere. Overall, this study improves our understanding of how MEE-driven ionization affects the polar middle atmosphere. The results show that the WACCM–SIC model reasonably reproduces the key photochemical and dynamical processes associated with NOx production and ozone depletion, highlighting the important role of MEE precipitation in Antarctic stratospheric ozone variability.

6. Limitations and Future Perspectives

This study is based on a case analysis of the 2013–2014 period, which may limit the generality of the results. Future work will extend the analysis to longer and more recent periods to assess the robustness of the findings. In addition, the native model resolution may limit the representation of small-scale variability. Higher-resolution or multi-model approaches will be explored in future studies.