Impact of chlorine ion chemistry on ozone loss in the middle atmosphere during very large solar proton events

,

The main processes responsible for the odd hydrogen (HO x = H,OH,HO 2 ) formation during energetic particle precipitation events, :::: along :::: with ::: the :::::::::::: ion-chemistry :::::::: processes :::::: leading :: to ::: its :::::: release were considered by Solomon et al. (1981).They take place after the initial formation of ion pairs.Solomon et al. (1981) considered the ion-chemistry processes leading to a release of HO x during energetic particle precipitation events.They found that the main process responsible is the uptake of water vapour into large cluster ions and the subsequent release of H during recombination reactions of these cluster ions.Large cluster ions can then be formed by reaction pathways like (Sinnhuber et al., 2012): These protonised water cluster ions can then recombine with electrons to form H and OH.
During SPEs, highly energetic solar protons and the secondary electrons also ionize neutral species and produce hydrogen and nitrogen radicals leading :::::::: Hydrogen ::: and :::::::: nitrogen :::::: radicals :::: lead : to ozone destruction through catalytic cycles in the stratosphere and mesosphere.Different studies found ozone depletion in the mesosphere during SPEs, for example, Weeks et al. (1972) who studied a large polar cap absorption event in 1969 that was explained as a result of the formation of odd hydrogen (Swider and Keneshea, 1973).The chlorine catalytic cycles of ozone destruction are very efficient around 40 km (Lary, 1997).SPE induced changes of chlorine species can contribute to the short-term ozone depletion occurring after the SPE (von Clarmann et al., 2005).This influence is indirect and is mainly caused by NO x and HO x enhancements.The hypochlorous acid (HOCl) catalytic cycle acts as a link between chlorine and HO x enhancements as a result of the SPEs, which then speeds up the following three reaction sequence involving Reactions R18, R21 and R19: ClO + NO 2 + M −→ ClONO 2 + M During SPEs, NO x , HO x and chlorine catalytic cycles are responsible for ozone loss in the middle atmosphere at different altitudes.

Hydrogen catalytic cycles
Catalytic cycles involving HO 2 are very important in the lower stratosphere (10-30 km).The fastest of these cycles is shown in Reactions R6, R7 and R8.
Another example of HO x catalytic destruction cycles that is important in the middle and upper mesosphere (above 60 km), is shown in Reactions R9 and R10 (Bates and Nicolet, 1950).In every chain of Reactions R9 and R10, one molecule of O 3 , O( 3 P) or O( 1 D) is lost while reforming H and OH and thereby producing a net ozone loss (Reaction R11).

Nitrogen catalytic cycle
In the lower stratosphere, ozone loss is mainly due to the catalytic cycle with NO x governed by the Reactions R12 and R13 in which case the loss of ozone is more persistent due to the longer lifetimes of NO x .
The Halloween SPE is later compared with an exceptionally strong cosmic ray event that occurred in 774/775 A.D. It was derived from the historical records in radiocarbon 14 C measured in tree ring archives and later confirmed by 10 Be and 36 Cl cosmogenic nuclides.Although various scenarios were initially proposed, it is concluded now that the event was caused by solar energetic particles (Sukhodolov et al., 2017). 10Be and 14 C implied that the event had a very hard spectrum and thereby very high energetic protons.It is the greatest ::::::: strongest : solar energetic particle storm known for the last 11 millennia (the Holocene), serving as a likely worst-case scenario being 40-50 times stronger than the largest directly observed event on 23 rd February 1956 (Usoskin et al., 2013).This event was transient, as estimated using the ratio of different cosmogenic isotopes (Mekhaldi et al., 2015).
This paper is organised as follows.Section :::: Sect.: 2 describes the ionisation rates, model framework, simulations and the satellite observations used to evaluate the model.Section :::: Sect. 3 presents the results of the model evaluation with MIPAS satellite observations.An overview of the changes in chlorine species, ozone and NO y induced by the SPE is presented.
Section :::: Sect.: 4 presents a case study comparing model simulations of the Halloween SPE with the extreme solar event.Section :::: Sect. 5 shows some results describing the impact of chlorine ion-chemistry on ozone loss.In sect.6, a conclusion is provided to check if the data is well understood, a summary of how our results compare to previous studies.Finally an assessment is given how further studies could improve our current knowledge on SPE induced ozone loss due to chlorine ion-chemistry.
2 Data and methods

Ionisation rates
The ionisation rates (IRs) used for the Halloween SPE were obtained from the Atmospheric Ionisation during Substorm (AISstorm) model which is an enhanced version of the Atmospheric Ionisation Module Osnabrück (AIMOS) model (Wissing and Kallenrode, 2009).The AIMOS model computes ionization ::::::: ionisation : rates by precipitating electrons, protons and alpha particles for the whole atmosphere based on particle flux measurements from Polar Operational Environmental Satellites (POES), the Meteorological Operational satellites (Metop) and the Geostationary Operational Environmental Satellites (GOES).The treatment of the electron fluxes is in the energy range (0.154-300 keV), protons have ::: with : an energy range of 0.154 eV to 500 MeV.In the AIMOS v2.0-AISstorm model, both the time resolution (0.5 hr) and spatial resolution has been improved compared to AIMOS.For a comparison of the model results with MIPAS observations, the time dependent ionisation rates were put into ExoTIC.Figure 1 on the left side shows the temporal evolution of ion pair production rates for protons, electrons and alpha particles varying over the time period, 25 th October to 4 th November 2003 from the AISstorm model.These IRs are averaged over the longitudes for the latitude of 67.5 • N, in the polar cap region and are also daily averaged.For the extreme event, integrated ionisation rates were taken from an extreme SPE of 23 February 1956 (SPE 56) (Meyer et al., 1956), which was the strongest observed event with ground-level enhancement (GLE) > 4000 %.These integrated IRs were scaled by a factor was a rough estimate to scale the fluxes of particles and excess radiation such that the energy spectrum of SPE 56 was comparable to the isotope signals of the extreme event.Figure 1 on the right shows the ionisation rate profiles for both the events.The profiles for the Halloween SPE show average IRs for October 27 (day 301) and October 28 (day 302) before the SPE (in blue) and the average IRs for October 28 and October 29 during the main SPE phase (in green).It can be observed that the ionisation rates for the stratosphere and lower mesosphere in case of the extreme event is about 1-2 orders of magnitude higher compared to the Halloween SPE main phase.This is because the extreme event contained protons of energies up to a few GeV, compared to about only a few MeV protons for the Halloween SPE, the ionisation rates for the same can be seen to reach much further down to the surface.
3. The net effective production or loss rates of neutral species due to primary ionization :::::::: ionisation, positive and negative ion-chemistry which can also be used as a parameterisation for global chemistry-climate models (Nieder et al., 2014), are computed using an iterative chemical equilibrium approach.
4. The production rates resulting from the ion-chemistry computation are then fed back to the 1-D neutral chemistry model, which solves for the neutral atmospheric state transiently using the net effective production/loss reactions as well as neutral photo-chemistry reactions.
5. Lastly, this state is again returned to the ion-chemistry model for the following computation.
The model settings used for the sensitivity studies were mainly variations of full ion-chemistry containing both positive and negative ions from the D-region: setting reactive O( 1 D) in photo-chemical equilibrium and switching off the chlorine ionchemistry.Parameterised NO x and HO x model simulations based on Porter et al. (1976) and Solomon et al. (1981) were also carried out to assess the performance of the full ion-chemistry model.

Ion-chemistry
The ionisation in this case is driven by prescribed ionisation rates and by photo-ionisation of NO, with the primary positive charges being distributed onto N 2 , N, O 2 and O and balanced with electrons (Sinnhuber et al., 2012).The ionisation of CO 2 was recently included.These rates of the primary ions are calculated by ionisation cross-sections based on Rusch et al. (1981) and Jones and Rees (1973).All of the processes like dissociation and dissociative ionisation of O 2 and N 2 as well as ionisation of O 2 , N 2 and O can form the excited states of N, O, N + 2 , N + , NO + which are also included in the model.More details with a full list of the reactions, reactions rates and references for the reactions rates used for the positive ion-chemistry can be found in Sinnhuber et al. (2012) and the newer versions in Herbst et al. (2022).

Sensitivity tests switching off the chlorine ion-chemistry
The purpose of this sensitivity test is to study the impact of the chlorine ion, an important negative ion in the lower D region.
We also .:::: We wanted to study what difference it makes to the full ion-chemistry with a focus on the ozone loss.This is done by switching off the reactions of negative chlorine ions with neutrals or the recombination reactions with H + in ExoTIC.The relevant reactions are given in Table A1.

Parameterised NO x and HO x
The assumption in case of parameterised NO x is that 1.25 N atoms are produced per ion pair when electrically charged particles collide and dissociate N 2 .This process produces N + 2 and NO + ions and, finally, atomic nitrogen.The latter is produced in its ground state N( 4 S) (45 % or 0.55 per ion pair) and the excited state N( 2 D) (55 % or 0.7 per ion pair).These values are mostly used in stratospheric and mesospheric models.In case of HO x , each ion pair typically results in the production of around two HO x constituents, i.e. a pair of H and OH per ion pair during recombination of the protonised water cluster ions in the upper stratosphere and lower mesosphere (Reaction R5) which was first estimated by Swider and Keneshea (1973).

MIPAS on ENVISAT
The Michelson Interferometer for Passive Atmospheric Sounding (MIPAS) was a Fourier transform spectrometer for the detection of mid-infrared limb emission spectra in the middle and upper atmosphere on the Envisat ::::::::: ENVISAT (Environmental Satellite, 2000) mission (Fischer et al., 2008).ENVISAT was launched in 2002 into a sun-synchronous polar orbit (800 km) and stopped operation in April 2012.The atmospheric spectra were inverted into vertical profiles of atmospheric pressure, temperature and volume mixing ratios (vmrs) of more than 30 trace constituents.MIPAS observed a spectral range of 4.15 µm to 14.6 µm with a high spectral resolution, where a wide variety of trace gases have absorption lines and signals that are generally higher than in other parts of the spectrum.This is because the Planck function maximises at about 10 µm for atmospheric

Averaging Kernels
Different vertical resolutions of the MIPAS observations and the model need to be accounted for a meaningful comparison.
The ExoTIC model has a vertical resolution of 2.7 km whereas MIPAS has different vertical resolutions for different species.
For example, in case of HOCl, the maximum vertical resolution can be 17 km and for ClONO 2 , it can be 13 km at an altitude of 40 km and above as seen from an example Figure 2 for a specific time point.To remove the discrepancy of different vertical resolutions between the model and MIPAS observations, the original model profiles have to be convolved and adjusted to the MIPAS altitude resolution.This adjustment procedure yields new species profiles that MIPAS would see if it were to sound the model atmosphere.For this purpose, we make use of the averaging kernels (Rodgers, 2000) and use a scheme suggested by Connor et al. (1994) to adjust the better resolved model profiles to those of MIPAS and the new adjusted model profiles x new are calculated as: where A is the MIPAS averaging kernel matrix, x orig is the original model profile, I is a unity matrix and x a is the a priori   averaging :::::: kernels ::::: were :::::: applied :::: after :::::::: sampling ::: the :::::: model :::: data :: in ::: the ::::::: MIPAS :::::: altitude ::::: grid.Now, ExoTIC being a 1D column model doesn't produce the output at the same geolocations as MIPAS hence the application of the MIPAS averaging kernels was based on the temperature criteria.The following procedure was applied for the convolution: 1.The model profile from the time series was fixed first :::::: profiles ::::: were ::::::: selected, :::: one :: at : a ::::: time, :::: from ::: the ::::: entire :::: time ::::: series.
2. All the profiles from MIPAS within 57.5 and 77.5 degrees : N : latitude and +/-6 hours of the model profile's time were selected.
3. For this obtained MIPAS sample of temperature profiles, the root mean square value was calculated with the model's temperature profile which is fixed for the entire time series.
4. The geolocation for which the root mean square value of the temperature difference profile was minimal was selected, and averaging kernels for this geolocation were applied to the trace gas profiles from the model.
Using this procedure, we have obtained model profiles that were adjusted to the vertical resolution of MIPAS.The data was then averaged daily and the absolute or relative differences w.r.t a day before the event, i.e. 26 th October 2003 (day 299), was calculated.
The peak however was produced several days later than MIPAS.The application of MIPAS averaging kernels moved the peak down to 40 km and the predicted peak increases are reduced substantially to about 0.2 ppbv, about a factor of 2 less than MIPAS observations.
The enhanced ClONO 2 production happens due to SPE produced NO x via reaction ::::::: Reaction R32.ClONO 2 is removed mainly by photolysis in the sunlit atmosphere and, to a lesser extent, by reaction with atomic oxygen.And due to it's pressure dependence, ClONO 2 formation by Reaction R32 is more effective at lower altitudes (Funke et al., 2011)  Averaging kernels are also applied to the model profiles for the different NO y species for the different model settings and then added, except for NO.In case of MIPAS NO and NO 2 data, there is a complication which is, that instead of mixing ratios the logarithms of the mixing ratios are retrieved; also the averaging kernels refer to the logarithms of the mixing ratios.The application of MIPAS averaging kernels to a better resolved profile on the basis of the coarse-grid averaging kernel A of the logarithm of the mixing ratio then is (Stiller et al., 2012): There is a general issue with logarithmic retrievals, because regularization is self-adaptive and depends on the actual state of the atmosphere.For an SPE response, if the NO peaks around 50-60 km and if there is a better sensitivity at this altitude the Jacobian and the averaging kernels scale with the volume mixing ratio.For NO 2 however, the logarithmic averaging kernels behave well and are not dependent on the actual conditions (for a deeper discussion of the problem of time and state dependent averaging kernels, see von Clarmann et al. ( 2020)).Due to this complication, for the total NO y in the second column of figure 7 ::::: figures :: 9 ::: and ::: 10 for both daytime and night-time ::::::: nighttime, NO is added without the application of the averaging kernels as compared to the rest of the species.
The magnitude of NO y enhancements is found to be larger for the ExoTIC model with ion-chemistry settings compared to the MIPAS observations for both daytime and night-time    nighttime :::::::::: respectively.For ozone, the long term history of air parcels is more important as air parcels that are ozone depleted gets dispersed into the mid-latitudes if they are at the edge ::::: region of the vortex.So a sample deeper in the vortex for ozone is better, the reason we chose 70-90 • N here.A loss of 60-75 % is observed during the event itself in the mesosphere that is short lived and is related to the HO x catalytic cycle (Reactions R6, R7, R9 and R10, :::::::::::::::::::::::::::::::::::::: (Funke et al., 2011)(Bates andNicolet, 1950) ).The ozone recovers after the event, since HO x is short-lived.A second peak is observed on the 3rd of November which is related to a weaker coronal mass ejection event.NO x related loss of 15 % is observed in the stratosphere that lasts longer and is also related to the polar winter atmosphere (Reactions R12 and R13).The full ion-chemistry shows an ongoing loss of 45 % starting from the event day and the sensitivity study with O( 1 D) in photo-chemical equilibrium confirmed that this loss is due to reactions ::::::: Reactions : like R40 and R37, which produces OH and Cl contributing further to ozone loss.The agreement between the observations and the model results, for night-time :::::::: nighttime, for the three model results except for the full ionchemistry is excellent in the mesosphere indicating a good ability of the model to reproduce HO x related ozone loss for SPEs.
The comparison of the Halloween SPE and the extreme solar event of 775 A.D. showed long lasting stratospheric ozone loss for the extreme scenario.A long lasting impact was also found for the chlorine species like HOCl and HCl in case of the extreme scenario.Loss of HCl was underestimated by the parameterised model which was also found by Winkler et al. (2009) during the solar proton event in July 2000 in the northern polar region.For the extreme event, the parameterised model showed much higher NO y enhancements, about a 1000 ppm in the mesosphere and lower thermosphere.HO x enhancements of 0.1 ppm was found during the extreme event which went further down in altitude upto 40 km, for all the model case studies.An impact of around 10-20 % on ozone loss was found due to the chlorine ions for the two events, a bit stronger for the extreme scenario , which ::: that : is more important for higher forcing.Ozone formation was observed after the event which is also due to the impact of chlorine ion-chemistry.For the Halloween event with temporal ionisation rates, ozone loss of 2.4 % during day-time and 10 % during night-time :::::::: nighttime : was observed during the event that is due to :::: also ::: due :: to ::: the :::::::: included chlorine ion-chemistry.Ozone formation of 2-4 % was also found after the event both during day-time and night-time :::::::: nighttime.
In general, ExoTIC simulations reproduced the impacts of the Halloween SPE quite well, mainly for HOCl and NO y .
However, the initial state of the atmosphere in the simulations could be an important factor for some variability in the model results and MIPAS observations.Future work will focus on including the D-region ion-chemistry into the global 3D chemistry climate model EMAC (ECHAM/MESSy) and the evaluation of the chemistry with MIPAS observations in a setup considering

N( 2
D) + O 2 −→ NO + O( 3 P, 1 D) OH + HCl −→ H 2 O + Cl Cl + O 3 −→ ClO + O 2 ClO + HO 2 −→ HOCl + O 2 von Clarmann et al. (2005) showed an enhancement of chlorine monoxide, ClO and HOCl immediately after the SPE.They concluded that this was due to the Reactions R21 and R19.During an SPE, HOCl and reactive Cl present in the stratosphere can react with OH and HO 2 respectively, to form ClO. Other reactions of Cl with HO 2 and H 2 O 2 can yield in the production of HCl, which is the most important stratospheric reservoir species of Cl.The Reactions R24, R25, R26 and R27 are relevant to the study in Sect. 4. HOCl + OH −→ ClO + H 2 O Cl + HO 2 −→ ClO + OH Cl + HO 2 −→ HCl + O 2 Cl + H 2 O 2 −→ HCl + HO 2 SPE induced NO x enhancements is essential regarding production of ClONO 2 .López-Puertas et al. (2005) and von Clarmann et al. (2005) reported the first experimental confirmation of Reaction R32 under SPE conditions.
temperatures.The measurement strategy of the MIPAS instrument was based on trace gases having characteristic emission and absorption lines, represented by their absorption coefficients, which are unambiguous "fingerprints" of the particular trace gases.The MIPAS mission is separated into two phases, caused by a malfunction of the instrument around March 2004.The first phase of the mission(2002)(2003)(2004) is usually referred to as the MIPAS full-resolution (FR) period.After the malfunction, operation was resumed with a reduced optical path difference, resulting in deteriorated spectral resolution.The second phase starting in January 2005 is called the reduced-resolution (RR) period.Because of the long optical path through the atmospheric layers, MIPAS could also detect trace gases with very low mixing ratios.Vertical information was gained by scanning the atmosphere at different elevation angles with different tangent altitudes.MIPAS could observe atmospheric parameters in the altitude range from 5 to 68 km nominally with minimum and maximum vertical steps of 1 and 8 km respectively.The MIPAS data are used here for evaluation of the model results with different parameterisations.Data presented here are IMK version V5 data for HOCl, ozone, ClONO 2 and NO y species (NO, NO 2 , HNO 3 , N 2 O 5 ) that are updates of those published by vonClarmann et al. (2006), von Clarmann et al. (2012),Glatthor et al. (2006), Höpfner et al. (2007a) andFunke et al. (2005).

Figure 2 .
Figure 2. Example of typical profiles for the vertical resolution of HOCl and ClONO2 310information used in the MIPAS retrievels.The rows of the AK :::::::: averaging :::::: kernel matrix give the contribution of the true values to the retrieved values and the columns give the response of the delta peak like perturbations at each altitude.Figure3shows an example of averaging kernels for the different species and for a profile retrieved from spectra measured at latitude of 67.5 • N on 27 October 2003 at 00:00 UT.From the figure, it is seen that the trace gas retrievals result in different sensitivities at different altitudes.For example, the maximum sensitivity is seen at 20 km for HOCl in this specific case, whereas for ClONO 2 , it is 315 around 15 km.To characterize the vertical resolution, the typical measures are either the full width at half maximum of the rows of the A or the gridwidth divided by the respective diagonal of A (Rodgers, 2000).

Figure 5 .Figure 6 .
Figure 5. Absolute differences of daily averaged data for HOCl w.r.t. a day before the event, i.e. 26 th October 2003.Starting point is 26 th October 2003 and for the four :::: three : different model settings (Sensitivity tests (row-wise): ion-chemistry with O( 1 D) in photo-chemical equilibrium, switching off chlorine ion-chemistry and parameterised NOx and HOx); column-wise: without Averaging kernal (A), with Averaging kernal (A) applied and MIPAS observations averaged over 57-77 • N for day-time (sza <= 90 • ).For daytime, the white region below 50 km is the MIPAS peak (0.2 ppb) and the colorbar is adjusted to the lower mixing ratios predicted by the model (first plot).The white region above 50 km for the MIPAS observations represent meaningless data, where the values of Averaging kernal (A) diagonal elements are close to zero (< 0.03) that indicate no significant sensitivity to the retrieved parameter at the corresponding altitude.Colorbar interval:(-0.02,-0.01, 0.00, 0.01, 0.02, 0.03, 0.04, 0.05, 0.06, 0.07, 0.08) . The zonal average of MIPAS observations were tested for latitude bands 57-77 • N, i.e., at the edge ::::: region : of the polar vortex and 70-90 • N, deep in the polar vortex.The sample of 57-77 • N works better for the inter-comparison of ClONO 2 compared to 70-90 • N. In case the sample is taken deep in the vortex, the model seemed to fairly underestimate the peak values.This can be explained by the reaction ::::::: Reaction : R32, where formation of ClO needs sunlight, which is again available more at the edge ::::: region : of the polar vortex.But ClONO 2 can also photolyse in the presence of sunlight .Due ::: and ::: due : to this, there is a balance between the two processes and ClONO 2 can form at the edge of the polar vortex ::::: region : which can be transported deep into the vortex and conserved there at high latitudes.This however cannot be reproduced by the 1D model because it is fixed at a certain location and has no transport.

Figure 13 .
Figure 13.Comparison of the Halloween SPE and the extreme scenario (row-wise) for Ñ Oy: reference run (background atmosphere), full ion-chemistry with O( 1 D) set to photo-chemical equilibrium, without chlorine ions and parameterised NOx and HOx model (column-wise) for a high latitude of 67.57::: 67.5 • N.

Figure 15 .
Figure 15.Same as Figure 13 but for HCl

Figure 16 .
Figure 16.Same as figure 15 but for ClO

Figure 18 .
Figure 18.Same as figure 15 but for ClONO2

Figure 19 .
Figure 19.Comparison of the Halloween SPE and the extreme scenario for (column wise): HCl, ClO, HOCl and ClONO2 at 40 km.

Figure 20 .
Figure 20.Volume mixing ratios of species (Cl, HO2, H2O2 and HCl) at 40 km for the extreme event.The different lines are for the model settings: reference (black), ion-chemistry with O( 1 D) in photo-chemical equilibrium (blue), without chlorine ions (green) and parameterised

Figure 22 .
Figure 22.Relative difference of the model with full ion-chemistry and O( 1 D) in photo-chemical equilibrium including chlorine ions w.r.t. the model without chlorine ion-chemistry for the Halloween SPE: daytime (left) and night-time ::::::: nighttime : (right).The difference here is calculated for daily averaged data.

Figure 23 .
Figure 23.Relative difference of the model simulations: full ion-chemistry with O( 1 D) in photo-chemical equilibrium and with chlorine ions w.r.t. the model setting without chlorine ion-chemistry comparing the Halloween SPE (left) and extreme scenario (right).The data shown here is not daily averaged but the real model time step.
used different models to investigate the SPE induced changes and Jackman et al. (2008) used version 3 of the Whole Atmosphere Community Climate Model (WACCM).Both studies compared with the MIPAS observations from polar orbit satellite ENVISAT.Damiani et al. (2012) also looked at chlorine species (i.e., HOCl, ClONO 2 , ClO and HCl) using MLS and MIPAS data and version 4 of the WACCM model during SPEs of 17 and 20 January 2005.However they did not consider the D region ion-chemistry.Here, we studied the temporal evolution of changes of the respective chemical constituents considering the D region ion-chemistry in a ::: the 1D stacked box model, Exoplanetary Terrestrial Ion Chemistry (ExoTIC).The ion-chemistry was implemented by Winkler et al. (2009) upon which ExoTIC is Winkler et al. (2009) with the MIPAS observations, which provide a better picture of the polar cap region compared toWinkler et al. (2009)who compared HALOE HCl observations that were less densely sampled than MIPAS data.
In this section, a comparison study between the ExoTIC model results and MIPAS observations has been carried out for the chlorine species of HOCl, ClONO 2 , ozone and odd oxides of nitrogen (NO y ) for the Halloween SPE 2003.The comparison is done for the model simulations with different settings of ion-chemistry, i.e. calculating the photo-chemical equilibrium of O( 1 D) and switching off the uptake of chlorine ions, and parameterised NO x and HO x .The model simulations are performed for a high latitude of 67.5 • N and the MIPAS data were taken for the polar cap region, averaged over geographic latitudes such that it's inside the vortex, either vortex core or vortex edge depending on the tracer properties.The model data is sampled in the MIPAS altitude grid as well.The day and night for the MIPAS data are sorted according to the solar zenith angle (day ≤ 90 • ; night > 98 • ).The solar zenith angles for the 1D model were chosen such that, for each day, it is the mean solar zenith angle for the MIPAS data plus/minus the standard error of mean (SEM) with N being the number of data points for each day.Since ExoTIC doesn't have diffusion or horizontal and vertical transport, the comparison can only be done for a short period of time.The model results are compared with the MIPAS observations for a total of 9 days from 26 th October to 3 rd November 2003.Due to different vertical resolutions between the model and MIPAS observations, averaging kernels were applied.