Comparison of Coincident Ozone Profiles and the Role of the Polar Vortex in the Evolution of the Ozone Layer Above Kiruna, Sweden

(1)

and the Role of the Polar Vortex in the Evolution of the Ozone Layer Above

Kiruna, Sweden

Kyriaki Blazaki

Space Engineering, master's level (120 credits) 2020

Luleå University of Technology

Department of Computer Science, Electrical and Space Engineering

(2)

(3)

Abstract

This study primarily focuses on comparing coincident ozone profiles, as well as investigating the relationship between the polar vortex and the changes in the ozone layer above Kiruna, Sweden (67.85^◦N 20.22^◦E) during the Arctic winters 2010/2011 and 2018/2019. The instruments used are the Kiruna Microwave Radiometer (KIMRA); a millimeter wave radiometer that allows continuous observations of the evolution of ozone and other trace gases in the Arctic stratosphere, and the Microwave Limb Sounder (MLS) instrument on board the Aura satellite. The KIMRA and MLS measurements are compared to each other over the periods November 2010 to March 2011, as well as November 2018 to March 2019. Overall, the KIMRA and MLS comparison show good agreement, with KIMRA measuring more O3 in the lower stratosphere (∼ 50hPa) than MLS both during the daytime and nighttime. The O₃time series of KIMRA and MLS, along with the polar vortex edge based on equivalent latitudes, show minimum ozone concentrations when Kiruna was situated well inside the polar vortex.

(4)

Acknowledgements

Firstly, I would like to express my sincere gratitude to my advisors, Assoc. Prof. Mathias Milz and Assoc. Prof. Uwe Raffalski for the continuous support, patience, motivation and immense knowledge.

In addition to my advisors, I would like to thank Assoc. Prof. Thomas Kuhn, Dr. Peter Dalin and Rita Kajtar for their insightful comments which helped me to overcome some obstacles.

My sincere thanks also go to Maria Winneb¨ack, for always making life on campus easier.

Furthermore, I would like to thank my master colleague Nuria Ag¨ues Paszkowsky. With her I have shared moments of deep anxiety but also of great excitement. More than a great friend, I consider you family.

So many friends with whom I had unforgettable moments. Friends who have been a major source of support when things would get a bit discouraging. Erik, Emil, Pontus, Nicolas, Ali, Shashi, Joar, Fatty, Jennifer, Kostas, Anna, Aleka. You will always be part of my life no matter the distance.

A big thank you to Nik. Without him I would have left Kiruna the moment I arrived.

Last but not least I would like to thank Fadel. For everything.

Finally, I must express my very profound gratitude to my family: my parents and sisters, for providing me with unfailing support and continuous encouragement throughout my years of study. This accomplishment would not have been possible without them. Thank you.

(5)

List of figures

2.1 Earth’s Atmosphere. . . 3

2.2 Ozone Molecule. . . 3

2.3 Ozone Distribution Across the Atmosphere. . . 4

2.4 Stratospheric Ozone Formation. . . 5

2.5 Principal Steps in the Depletion of Stratospheric Ozone. . . 6

2.6 Polar Vortex Schematic. . . 7

2.7 Polar Vortex Types. . . 8

2.8 Polar Vortex in February 2011. . . 9

2.9 A-train constellation of Satellites. . . 9

2.10 The MLS Instrument on Aura. . . 10

2.11 Line drawing of the EOS MLS instrument. . . 11

2.12 MLS measurement locations for a 24 hour period. . . 13

2.13 Line drawing of the KIMRA instrument. . . 15

2.14 Ozone Destruction Cycles in Polar Regions. . . 18

3.1 A Table Containing Ozone Data Organised Into Day and Swath. . . 22

3.2 Satellite’s Latitude and Longitude data, the Location of Kiruna (blue cross) and the Area of Interest (red square). . . 23

3.3 Satellite’s Latitude and Longitude data, the Location of Kiruna (blue cross) and the Area of Interest (red square). . . 24

3.4 Ozone Profiles of the Different Swaths for Two Different Days in 2011. . . 25

3.5 Monthly Mean Ozone Profiles (Daytime) of the MLS Data. . . 26

3.6 Monthly Mean Ozone Profiles (Nighttime) of the MLS Data. . . 27

3.7 KIMRA data stored in a struct. . . 28

3.8 KIMRA Ozone Profiles of a selected date. . . 29

3.9 Original Satellite Profile and the Smoothed Satellite Profile (Daytime). . . 30

3.10 Original Satellite Profile, Smoothed Satellite Profile, KIMRA Profile, and a priori Profile (Daytime). . . 31

3.11 KIMRA Profile, Smoothed Satellite Profile (Daytime). . . 31

3.12 Original Satellite Profile, Smoothed Satellite Profile (Nighttime). . . 32

3.13 Original Satellite Profile, Smoothed Satellite Profile, KIMRA Profile, and a priori Profile (Nighttime). . . 33

3.14 KIMRA Profile and the Smoothed Satellite Profile (Nighttime). . . 33

4.1 Comparison of KIMRA and MLS Winter 2010/2011 (Daytime). . . 35

4.2 Comparison of KIMRA and MLS Winter 2010/2011 (Nighttime). . . 36

4.3 Ozone Profile Difference between KIMRA and MLS for Winter 2010/2011 (Daytime). 38 4.4 Ozone Profile Difference between KIMRA and MLS for Winter 2010/2011 (Nighttime). 38 4.5 O3 Time Series of KIMRA and MLS for Winter 2010/2011 (Daytime). . . 39

4.6 O3 Time Series of KIMRA and MLS for Winter 2010/2011 (Nighttime). . . 40

4.7 Temperature Profile for Winter 2010/2011 (Daytime). . . 41

4.8 Temperature Profile for Winter 2010/2011 (Nighttime). . . 41

4.9 Statistical Analysis for Winter 2018/2019 (Daytime). . . 42

4.10 Statistical Analysis for Winter 2018/2019 (Nighttime). . . 43

4.11 Ozone Profile Difference between KIMRA and MLS for Winter 2018/2019 (Daytime). 45 4.12 Ozone Profile Difference between KIMRA and MLS for Winter 2018/2019 (Nighttime). 45 4.13 O₃ Time Series of KIMRA and MLS for Winter 2018/2019 (Daytime). . . 46

(8)

4.14 O3 Time Series of KIMRA and MLS for Winter 2018/2019 (Nighttime). . . 47

4.15 Temperature Profile for Winter 2018/2019 (Daytime). . . 48

4.16 Temperature Profile for Winter 2018/2019 (Nighttime). . . 48

4.17 Polar Vortex Position in Relation to Kiruna. . . 50

4.18 O3 Time Series of MLS for Winter 2010/2011. . . 51

List of tables

1 EOS MLS field-of-view beamwidths. . . 12

2 Daytime Comparison Results for Winter 2010/2011. . . 37

3 Nighttime Comparison Results for Winter 2010/2011. . . 37

4 Daytime Comparison Results for Winter 2018/2019. . . 44

5 Nighttime Comparison Results for Winter 2018/2019. . . 44

(9)

Acronyms

AOS Acousto-Optical Spectrometer.

CFCs ChloroFluoroCarbons.

ECMWF European Centre for Medium-Range Weather Forecasts.

EOS Earth Observing System.

FFTSs Fast Fourier Transform Spectrometers.

FOV Field of View.

HDF Hierarchical Data Format.

IRF Swedish Institute of Space Physics (Institutet f¨or Rymdfysik).

KIMRA KIruna Millimeter wave RAdiometer.

LTU Lule˚a University of Technology.

MLS Microwave Limb Sounder.

NASA National Aeronautics and Space Administration.

ODS Ozone-Depleting Substances.

ppmv Parts Per Million by Volume.

PSCs Polar Stratospheric Clouds.

PV Potential Vorticity.

VMR Volume Mixing Ratio.

Glossary

Winter 2010/2011 The period from November 2010 to March 2011.

Winter 2018/2019 The period from November 2018 to March 2019.

(10)

1 INTRODUCTION

1.1 Motivation

The ongoing and increasingly rapid melting of the Arctic ice cap has served as a reference to the global climate change. Changes in the climate of the Earth could affect the behavior of the ozone layer, because ozone is influenced by changes in the meteorological conditions and by changes in atmospheric composition that could result from climate change.

In recent years, there has been a string of unusually cold winters in the Arctic. The cold and persistent conditions have led to significant reductions in ozone content in the stratosphere above the Arctic during the late winter and early spring (January-March). These winter meteorological conditions in the Northern Hemisphere, lead to the formation of an isolated region bounded by strong winds, the so-called polar vortex. The cold temperatures lead in turn to the formation of clouds, known as polar stratospheric clouds. These clouds provide surfaces that promote production of forms of chlorine and bromine that are chemically active and can rapidly destroy ozone. But the cold conditions rarely persist into March, when sufficient sunlight is available to initiate large ozone depletion. However, the cause of the observed change in meteorological conditions is not yet understood. Such conditions might persist over the coming years, further enhancing ozone depletion. Although there has been significant ozone depletion in the Arctic in recent years, it is therefore difficult to predict what may lie ahead, because the amplitude and extent of such a cooling, and therefore the delay in the recovery of the ozone layer, still have to be assessed (UK Air Information Resource).

1.2 Thesis Aim

The aim of this thesis is to compare coincident ozone profiles retrieved from a satellite and a ground- based instrument, as well as study the influence of the polar vortex on the evolution of the ozone layer.

1.3 Objectives

The objectives for this thesis were the following:

1. Analyze satellite ozone data above Kiruna.

2. Find and analyze the coincide in space and time ozone measurements of the ground-based instrument above Kiruna.

3. Compare the coincident profiles.

4. Find the location of the polar vortex in relation to Kiruna and discuss the effect of the the polar vortex on the evolution of the ozone layer.

1.4 Thesis outline

This introduction is followed by the theory section. The theory section is an introduction to climate change and the processes that cause the changes in the concentration of the ozone with a focus on the polar vortex. Following, it is a short description of the instruments that were used for this thesis. The

(11)

last part of the theory is a reference of the details of this study. Following the theory section is the method section. There one can find information about the data that were used, the data processing, as well as a description of the programs used. This section also contains graphs showing the steps that were used before the final plots. Next is the results section. The first part presents the results of the comparison of the coincident profiles followed by the O3 and temperature time series, as well as the polar vortex position in relation to Kiruna. The last part of the results section is an overall discussion. The final section of the thesis is the conclusions section and a probable future work.

(12)

2 THEORY

2.1 Earth’s Atmosphere

Layers The atmosphere of Earth is a shell of gases, that surrounds the planet and consists of five layers. The troposphere (0-12 km), the lowest layer where most of Earth’s weather takes place. The stratosphere (12-50 km or 17-50 km depending the latitude), which is the layer of interest in this study. It contains the ozone layer and is characterized by increasing temperatures with increasing altitude. Following the stratosphere, is the mesosphere (50-80 km), the thermosphere (80-700 km), and the exosphere (700-1000 km).

Figure 2.1: Earth’s Atmosphere (UCAR, 2015).

2.2 Ozone

Figure 2.2: Ozone Molecule (SAP, 2018a).

Ozone is an inorganic molecule with chemical formula O3, that naturally exists in the Earth’s atmosphere. Its structure, seen in Figure 2.2 means that is much less stable than diatomic oxygen (O2), and is therefore much more reactive; this means it can be more easily formed and broken down through interaction with other chemical compounds. Stratospheric ozone occurs naturally in small amounts in the Earth’s stratosphere (17-50 km), where it absorbs solar ultraviolet radiation hence, protecting the living organisms on the Earth’s surface (SAP, 2018a).

(13)

2.2.1 Impact of the Ozone through the Atmosphere

Ozone can be differentiated in two categories, depending the altitude and its role in atmospheric chemistry in the Earth’s atmosphere as seen in Figure 2.3.

Figure 2.3: Ozone Distribution Across the Atmosphere (SAP, 2018a).

• Tropospheric Ozone is the ozone present in the troposphere. Only 10% of the Earth’s total ozone is found in the troposphere which is formed by chemical reactions of natural occurring gases and pollutants (SAP, 2018a).

• Stratospheric Ozone is the one present in the stratosphere, which plays a crucial role in absorbing dangerous ultraviolet radiation from the sun. Approximately 90% of the Earth’s total ozone is found in this layer, with a region called the ozone layer having the highest amount of ozone (SAP, 2018a).

The focus of this study will be on stratospheric ozone. Figure 2.4 shows the full Chapman cycle according to which ozone is naturally formed by chemical reactions involving sunlight.

(14)

Figure 2.4: Stratospheric Ozone Formation (SAP, 2018a).

2.2.2 Ozone-Depletion

Ozone-Depleting Substances Ozone-depleting substances (ODS) are fabricates halogen source gases that bring chlorine and bromine atoms to the stratosphere, where they are involved in chemical reactions that destroy ozone. Some examples are the chlorofluorocarbons (CFCs), formerly used in refrigeration and air conditioning systems, and the halons, once used as fire extinguishing agents (SAP, 2018b).

Ozone Depletion Causes The primary cause of stratospheric ozone depletion is the release of gases through human activities, that contain chlorine and bromine. These gases, being relatively unreactive and having long atmospheric lifetimes, accumulate in the lower atmosphere. Eventually, these gases are transported into the stratosphere by air motions after being emitted at the surface.

Once in the stratosphere, in the presence of UV light, the reactive chlorine and bromine are released from the reservoir species participating in reactions that destroy ozone. In polar regions, the standard total ozone depletion attributed to these reactive gases is larger compared to the tropics, and can be dramatically enhanced in the presence of polar stratospheric clouds (PSCs). PSCs occur at very low temperatures, and highly increase the abundance of chlorine monoxide (CLO); the most reactive chlorine gas. As a result, the ozone destruction is considerable in polar regions in late winter to early spring. During polar night and other periods of darkness, ozone cannot be destroyed by these reactions.

Polar Stratospheric Clouds (PSCs) Polar Stratospheric Clouds (PCSs) are clouds that form in the winter polar atmosphere at altitudes 15-25 km. There are two main types of PSCs according to their particle size and temperature formation, but only type I plays a central role in the formation of the ozone hole. This type is made up of super-cooled droplets of water and nitric acid and provide the surfaces for heterogenous chemical reactions to take place. As a result, free radicals of chlorine are produced which directly destroy ozone molecules (Barabash, 2017).

Figure 2.5 is a schematic summary of Subsection 2.2.2 showing the steps that lead to ozone depletion.

(15)

Figure 2.5: Principal Steps in the Depletion of Stratospheric Ozone (SAP, 2018c).

(16)

Polar Vortex The polar vortex is a persistent large-scale cyclonic circulation pattern in the middle and upper troposphere and the stratosphere, centred generally in the polar regions of each hemisphere (Barabash, 2017).

Figure 2.6: Polar Vortex Schematic.

The word ”vortex” generally refers to a pattern of winds that helps to preserve the colder air close to the poles. A standard vortex is characterized by circular air motion and very low temperatures in its center, which is essential in the destruction of ozone and the formation of the ozone hole in polar regions. Generally, the polar vortex is not steady all year round. During polar summer, the polar vortex shrinks and expands in polar winter. The progression of the polar vortex can be split into three phases; it forms in autumn, reaches maximum strength in mid-winter, and breaks down in late winter to spring (Zhang et al., 2017).

The strength of the polar vortex can vary from year to year. When the vortex of the Arctic is strong, it is well defined, there is a single vortex, and the Arctic air is well contained; when weaker, it will break into two or more vortices; when very weak, the flow of Arctic air becomes more disorganized, and masses of cold Arctic air can push equator-ward, bringing with them a rapid and sharp temperature drop.

(17)

Figure 2.7: Polar Vortex Types. Left: Weakened Vortex Formation. Right: Strong Vortex Phase (Caitlyn Kennedy, 2014).

Figure 2.7 shows two phases of the polar vortex in two different periods. The image on the right, displays a rather strong phase of the polar vortex in mid-November, while the image on the left is showing a weakened phase of the polar vortex in early January. In the image on the right, shown in dark purple, is the polar vortex in an oval-shaped formation which contains the freezing air. The line of the outer edge of the vortex, shown in light purple, is the air stream in its usual west-to-east pattern. In the left image, the polar vortex appears to have broken down, allowing the cold air to slip out of the vortex into mid-latitudes. The jet stream was delayed by the high pressure building in the Arctic, causing it to flow farther south than usual (Caitlyn Kennedy, 2014).

Figure 2.8 shows a strong polar vortex on February 22, 2011. The polar vortex (middle image) is symmetric around the North Pole inside which low temperatures dominate (right image) and low ozone concentrations (left image).

(18)

Figure 2.8: Polar Vortex in February 2011. Left: Northern Hemisphere Total Column Ozone. Middle:

Potential Vorticity. Right: Temperature (NASA, 2017).

2.3 Instruments

2.3.1 Aura Satellite

Aura is part of NASA’s A-train group of Earth observing satellites. This formation of satellites crosses the equator each day at the same local time each afternoon, and its where the name A-train came from; the ”A” stands for ”Afternoon”. It orbits the Earth in a sun-synchronous orbit at 705 km (NASA, 2016).

Figure 2.9: The A-train constellation of Satellites (NASA, 2016).

MLS The Earth Observing System (EOS) Microwave Limb Sounder (MLS) is one of the four instruments on the NASA’s EOS Aura Satellite, launched on July 15th 2004 (NASA JPL). Apart from ozone, the MLS instrument also makes measurements of atmospheric composition, humidity, temperature, and cloud ice, all important for achieving its scientific objectives:

(19)

• track the stability of the stratospheric ozone layer;

• help improve the predictions and variability of climate change;

• improve the understanding of air quality.

Figure 2.10: The MLS Instrument on Aura (NASA JPL).

2.3.2 MLS Instrument Description

The MLS instrument is manufactured in three parts, as described by Waters et al. (2006).

1. the“GHz module”, which contains the GHz antenna system,calibration targets, switching mirror, optical multiplexer,and 118-, 190-, 240-, and 640-GHz radiometers; A three-reflector offset antenna system, that vertically scans the limb gathers the atmospheric signals for all radiometers. The antenna can be scanned to high altitudes for a view of “cold space” through the complete antenna system, as required. Following the GHz antenna system, a switching mirror provides radiometric calibration by switching to views between on-board black-body calibration targets and space. Subsequently, an optical multiplexer, which consists of dichroic plates and a polarization grid, follows the GHz switching mirror and spatially separates the signal into paths feeding the different GHz radiometers. The radiometers down-convert frequencies of the incoming signals to several broad intermediate frequency (IF) bands in the range of 3–21 GHz.

The EOS MLS instrument contains heterodyne radiometers operating at ambient temperature in five spectral regions, each giving information on different chemical species.

2. the “THz module”, which contains the THz scan and switching mirror, calibration target, telescope, and 2.5-THz radiometers at both polarizations; Placed in the front of the telescope, the THz scan mirror, provides calibration of the THz bands by rotating to views of cold space and the on-board ambient-temperature THz blackbody calibration target.

3. the “Spectrometer Module”, which contains spectrometers,command and data handling systems, and power distribution systems. Four types of spectrometers, with different spectral resolu-

(20)

tions and bandwidths, are used to cover different altitude ranges. For measurements throughout the stratosphere, the basic source of information are the “Standard” 25-channel spectrometers.

Digitized data from the spectrometers are passed to the command and data-handling system for transmission to the ground.

A line sketch of the EOS MLS instrument is shown in Figure 2.11.

Figure 2.11: Line drawing of the EOS MLS instrument (Waters et al., 2006).

2.3.3 Experiment Description

Measurement Technique MLS observes millimeter- and submillimeter-wavelength thermal emission from Earth’s ’limb’ (the edge of the atmosphere) viewing forward along the Aura spacecraft flight direction, scanning its view from the ground to ∼ 90 km every ∼ 25 seconds.

Overall, features of the technique as described by Waters et al. (2004), include:

1. the ability to measure several atmospheric gases, with emission from molecular oxygen providing temperature and pressure;

2. measurements that can be obtained in the presence of ice clouds and aerosol that prevent measurements by shorter-wavelength infrared, visible and ultraviolet techniques;

3. measurements that are made globally day and night on a daily basis;

4. spectrally resolve emission lines from all altitudes, hence allowing measurements of very weak lines in the presence of nearby strong ones and thus measurements of chemical species with very low atmospheric abundances;

5. composition measurements that are not affected by the uncertainties in atmospheric temperature;

6. an accurate data base of spectral lines;

7. accurate and stable calibration of the instruments;

(21)

8. efficient sensitivity of the instruments without necessarily the need of cooling, and good vertical resolution set by size of the antenna.

The widths of the spectral lines in the millimeter and submillimeter-wavelength spectral regions used by the MLS, are dominated by pressure and Doppler broadening. The pressure broadening dominates throughout the troposphere and stratosphere, resulting in the linewidth being a nearly exponentially decreasing function of height up to ∼50-70 km. Doppler broadening dominates the linewidth at higher altitudes.

Spectral Regions The spectral regions that were used for the measurements by MLS, were chosen after investigation in order to minimize the number of radiometers and spectrometers needed to accomplish the scientific objectives of the experiment, while reliably producing its data products with the required accuracy and resolution. Most MLS data products are obtained from the spectrally- varying component of the radiance, and different scientific uses for the same data product require different precision, which depend upon the amount of averaging and spectral resolution used in the analyses.

Field Of View (FOV) The beam-widths, beam efficiencies, and placement and coincidences of bore-sights are key parameters for the MLS instrument field-of-view performance.

‘FOV beam-width’ is defined as the angle between the half-power points of the antenna response.

‘Beam efficiency’ is defined as that fraction of power from an isotropic source which is collected within a specified angular range centered at the antenna bore-sight. The MLS FOV beam-widths and beam efficiency are set by the vertical resolution needed for the measurements. The values chosen are an acceptable compromise between the scientific desire for better vertical resolution, and the engineering/accommodation difficulties associated with the increase in antenna size required for better resolution.

Table 1 shows the EOS MLS field-of-view beamwidths.

Radiometer

FOV beamwidth in vertical plane at the limb tangent point

FOV beamwidth in horizontal plane at the limb tangent point 118 GHz (R1)

190 GHz (R2) 240 GHz (R3) 640 GHz (R4) 2.5 THz (R5)

6.5 km 4.5 km 3.5 km 1.5 km 2.5 km

13 km 9 km 7 km 3 km 2.5 km

Table 1: EOS MLS field-of-view beamwidths (Waters et al., 2004).

Calibration Four different categories of calibration exist for the MLS instrument which include:

1. Radiometric Calibration gives the absolute power incident upon the antenna that is received in each spectral channel.

2. Field-of-View Calibration gives the response of the instrument to the input signal as a function of the angle at which the signal is incident upon the antenna.

(22)

3. Spectral Calibration gives the relative response of the instrument to the input signal as a function of the frequency of the signal.

4. Engineering Calibration gives the output of engineering sensors in appropriate units.

The instrument calibration is performed in each of these categories and happens in two different cases:

(1) Pre-launch Calibration and (2) In-orbit Calibration. Waters et al. (2004) further describes the calibration method.

Measurement Coverage As mentioned before in Section 2.3.1, the Aura orbit is sun-synchronous at 705 km altitude with 98^◦ inclination and 1:45 p.m. ascending equator-crossing time. The MLS field of view is scanning the Earth’s limb while taking measurements to provide 82^◦N to 82^◦S latitude coverage on each orbit. MLS observes in the forward direction (direction of orbital motion), and the limb is scanned in an upward direction to give an observation path tangent point locus that is nearly vertical. The tangent points at greater heights are closer to the satellite, but in the Earth frame of reference this is compensated by the satellite’s forward motion. The mean geometric distance to the atmospheric limb is ∼3060 km for 10 km tangent height (25.61^◦ elevation angle) and ∼2960 km for 55 km tangent height (24.75^◦ elevation angle).

The MLS limb scans for nominal operation are synchronized to the orbit, with the number of scans per orbit an integer multiple of 4, and phased such that the limb scan locations occur over the equator.

This gives the same latitude sampling in the northern and southern hemispheres, and on the ascending and descending portions of the orbit. MLS nominal operations have 240 limb scans per orbit to give 1.5^◦ along-track separation between adjacent limb scans (Waters et al., 2004).

Figure 2.12 shows the locations of measurements with the scan pattern for one 24-hour period.

Figure 2.12: MLS measurement locations for a 24 hour period (Waters et al., 2004).

The tangent points for the individual limb scans are marked with crosses, while the suborbital track is shown by the continuous line. As it can been seen in Figure 2.12, the suborbital track is slightly

(23)

displaced from the tangent points (crosses) because of Earth’s rotation during the time in which the satellite moves forward to the tangent point latitude. The ascending portions of the orbit are those with the southeast-northwest tilt (Waters et al., 2004).

MLS Data Processing Overview MLS data processing happens in three processing levels as described by Waters et al. (2004).

Level 1 processing creates Level 1 data files which include calibrated MLS radiances and instrument engineering data.

Level 2 processing creates Level 2 data files of retrieved geophysical parameters.

Level 3 processing creates Level 3 data files containing gridded daily and monthly maps, and daily and monthly zonal means.

For the purpose of this thesis, Level 2 data files were used and the next paragraph further analyzes Level 2 processing.

Level 2 Data Processing The MLS software’s main tasks for Level 2 data processing are:

1. Retrieve geophysical parameters such as temperature, constituent abundances, from the MLS Level 1B data, and provide estimates of uncertainties on the retrieved quantities.

2. Produce additional diagnostic information, such as radiances calculated from the retrieved parameters, and chi-square statistics, on the quality of the retrieved geophysical parameters, and

‘flags’ to detect bad retrievals.

3. Produce supplemental data, such as tropopause pressure, which may be derived from MLS data and/or additional meteorological data available at the time of data processing.

4. Produce daily files of the output data, and a log summarizing appropriate information on the processing statistics for that day.

The retrieval theory is a well-established field. Rodgers (1976) describes the basis of algorithms used to obtain geophysical parameters from remote measurements.

2.3.4 KIMRA

KIMRA stands for KIruna Millimeter wave RAdiometer, and as its name indicates, it is a millimeter wave radiometer located at the Swedish Institute of Space Physics (IRF), Kiruna, Sweden. The location of the instrument is 67.8^◦N, 20.4^◦S, and allows continuous observations of the evolution of ozone and other trace gases in the Arctic stratosphere, between 195-232 GHz. Seasonal, as well as annual changes can be also observed.

2.3.5 Overall Instrument

As described by Raffalski et al. (2014), KIMRA consists of a cryogenically cooled Schottky mixer with a noise temperature of between 800 and 1600 K (single sideband). The Schottky mixer together with the HEMT amplifier is mounted in a vacuum dewar (pressure at around 10⁻⁷ mbar) and cryogenically cooled down to 22 K in order to decrease electronic noise. The atmospheric signal is coupled to the

(24)

mixer via a periscope system (on the roof top of IRF) which can be pointed into any direction of interest. The local oscillator signal from a PLL stabilized Gunn oscillator is heterodyned with the atmospheric signal by a Martin-Puplett interferometer. The signal is amplified, downconverted twice, and fed into the Acousto-Optical Spectrometer (AOS) at a center frequency of 2.1 GHz. The total bandwidth of the spectrometer, 1.2 GHz, limits the altitude range to the lower limit of about 15 km.

KIMRA also has two Fast Fourier Transform Spectrometers (FFTSs), but for the present study only the AOS data were used. Standing waves due to mismatch of the beam are suppressed by a moving mirror, while contribution in the cross polarization is effectively filtered out by a wire grid right in front of the vacuum flask.

A line sketch of the KIMRA instrument is shown in Figure 2.13

Figure 2.13: Line drawing of the KIMRA instrument (Raffalski, 2019).

(25)

2.3.6 Measurements Technique

KIMRA is tuned to approximately 231 GHz in order to measure both CO at 230.54 GHz and a fairly strong ozone line at 231.28 GHz. For ozone measurements, the integration time is approximately 15 minutes, while it takes longer for other trace gases with weaker emission lines, in order to reduce thermal noise in the spectra. As base condition, KIMRA is pointing northward with elevation angles between 10^◦ and 50^◦, depending on the tropospheric water vapor content. KIMRA makes measurements in the altitude range 15-60 km (Raffalski et al., 2014). Altitude profiles have been retrieved using the Optimal Estimation method described by Rodgers (1976).

2.4 Previous studies of Polar Vortex and Ozone Depletion

The linkage between the polar vortex and the ozone depletion has been emphasized in many studies.

For example, Hassler et al. (2011) came to the conclusion that the total ozone, as measured in Antarctic stations, has been strongly affected by the state of the polar vortex, while Schoeberl and Hartmann (1991), linked the formation of polar stratospheric clouds with ozone depletion. Specifically, the study revealed that when the temperature falls low enough within the vortex, to form PSCs of nitric acid trihydrate and ice, those can affect the conversion of chlorine from the inactive reservoir species to the radical species that attack ozone. Furthermore, Raffalski (2019) concluded that the early ozone loss during winter 2002/2003 was due to very low temperatures in the lower stratosphere, as well as the geographical extension of the polar vortex to lower altitudes.

2.5 This study

This study primarily focuses on comparing coincident ozone profiles, as well as investigating the relationship between the polar vortex and the changes in the ozone layer above Kiruna, Sweden (67.85^◦N 20.22^◦E) for the time intervals from November 2010 to March 2011 and from November 2018 to March 2019.

The Aura satellite passes over the area around the KIMRA measurement location (67.8^◦N, 20.4^◦E) twice a day, approximately around noon and sometime after midnight. For KIMRA instrument being very local, a bigger area like a square with Kiruna being in the middle was selected. Since in this study only the coincide in time and space measurements were analyzed, defining a bigger area around Kiruna, meant that there were more dates with coincident measurements and thus more accurate results. The selected criterion for spatial coincidence is that measurements inside the area of interest between 69.5^◦N 27^◦E and 65^◦N 17^◦E will be used. Figure ??, shows the location of Kiruna, the area of interest, as well as the latitude and longitude of the satellite.

Every profile that was made use of in the comparison between KIMRA and MLS should coincide in time and space. The requirement of time coincidence is that for a given KIMRA measurement, it was determined whether there are any MLS measurements whose measurement time lies +/- 3-5 hours around the KIMRA measurement.

The standard product for ozone is derived from MLS radiance measurements near 240 GHz. The present study has used ozone profiles from version 4.2. A description of the quality of version 4.2 Aura MLS Level 2 data can be found in Section 3.1.1.

(26)

KIMRA performs measurements of thermal emission lines between 195 and 232 GHz. More details can be found in Section 3.1.2.

As mentioned before, in Paragraph 2.2.2, the polar vortex is a winter phenomena, which develops when the sun sets over the polar regions and the temperature drops, and disappears during spring when the sun rises and the polar stratosphere begins to heat. Consequently, due to dynamics and chemistry, the time period November to March was chosen for analysis, because O3 above Kiruna is expected to have the most variation over this time.

At high latitudes the diurnal cycle is highly dependent on the season because the isolation conditions vary strongly throughout the year. The period of polar night, the period of polar day and the intermediate periods with light and darkness within a day characterize the pattern of the diurnal cycle in the Arctic. For this reason the data from both instruments were divided into daytime data and nighttime data. For the separation of the KIMRA data the time of the measurement was used. For the MLS data, the ascent or descent mode was used. A value of ”1” indicated ascent mode and daytime, whereas a value of ”-1” suggested descent mode and nighttime.

Cycles 2 and 3 in Figure 2.14 are the dominant reaction mechanisms for the ozone loss in the polar regions due to the high abundances of ClO and BrO. The ClO abundance is a result of the reactions that take place on the surface of PSCs and is largely increased during late winter and early spring in the polar regions.

(27)

Figure 2.14: Ozone Destruction Cycles in Polar Regions (SAP, 2018c).

Both cycles result in creating three oxygen molecules by destroying two ozone molecules. Basic requirement for the completion of these reaction cycles is the sunlight. During the polar winter, when the darkness dominates, these reactions cannot happen. Cycles 2 and 3 are responsible for most of the stratospheric ozone destruction in the Arctic during late winter and early spring, when there is sufficient sunlight which is needed to break apart (ClO)2 and BrCl.

(28)

3 METHOD

Data from KIMRA and MLS were used. KIMRA data were supplied by Uwe Raffalski, and MLS data were downloaded from NASA’s website https://disc.gsfc.nasa.gov/datasets/ML2O3_003/

summary?keywords=Aura%20MLS. The temperature data were downloaded from ECMWF. The data were processed using MATLAB R2017b and MATALAB R2019a.

3.1 Data description

All the data were within the time periods from November to March of the years 2010/2011 and 2018/2019.

3.1.1 MLS Data

The MLS instrument provided data for every day within the time period of interest, except for the last five days in March 2011, and four days in late January 2019, where the instrument went into survival mode and did not provide any data.

The MLS data files were in HDF 5 format and were saved under the name ”MLS-Aura L2GP-

<product> v04-20-c01 <yyyy>d<ddd>.he5”. ”L2GP” stands for Level 2 Geophysical Product. The files are produced on a one-day granularity (midnight to midnight, universal time), and named according to the observation date. These O3 files contain the corresponding standard product in an HDF-EOS swath given the same name as the product on 55 pressure surfaces. In addition, the standard O3 product files contain swaths describing column abundances.

The data fields L2gpValue and L2gpPrecision also contained in each swath. These fields, describe the value and precision of the data respectively. Following the advice given in Waters et al. (2004), data fields for which L2gpPrecision was negative were not used. Negative values of L2gpPrecision indicate that the resulting precision was less than 50% of the a priori precision, which meant that the instrument and/or the algorithms failed to deliver useful information for that point.

Additionally, fields such as latitude provided geolocation information and the field time describes time, as the number of seconds elapsed since midnight universal time on January 1st, 1993.

From a qualitative point of view, there were three more metrics for every profile of each product.

These three metrics were:

• Quality. As its name suggests, this value gives a measure of the quality of the product based on the fit achieved by the Level 2 algorithms to the relevant radiances. The larger the value of Quality the better the radiance fits and therefore the more trustworthy the data are. Values of Quality closer to zero indicate the opposite hence, less trustworthy data. Only profiles whose Quality field is greater than 1.0 should be used.

• Status. This value is a 32 bit integer that acts as a bit field containing several “flags”. Any profile for which Status is an odd number should not be used. The first two bits are “flagging”

bits. Higher bits give more information on the reasons behind the setting of the first two bits.

• Convergence. The third diagnostic field describes how the fit to the radiances achieved by the retrieval algorithms compared to the degree of fit to be expected. Values around unity indicate good convergence. Only profiles whose Convergence field is less than 1.03 should be used.

(29)

In some cases noise in the MLS measurements can result in negative values for species mixing ratios.

Such values are not ignored as this introduces a positive bias into any averages made of the data (Livesey et al., 2015).

3.1.2 KIMRA Data

For winter 2018/2019, KIMRA provided data covering the whole period. For winter 2010/2011, due to instrument malfunctions KIMRA did not provide data for every day within the time period of interest. The biggest gap was for the period November 2010 to late January 2011, where there are no data at all. For the rest of the period there is a satisfactory amount of data.

KIMRA data sets used in this study have been pre-processed and contain, besides the Level 2 data (ozone profiles), also other species on 45 pressure levels. These files contain the O3 product on 45 pressure surfaces, that are evenly spaced in altitude between ground level and approximately 90 km (Ryan et al., 2016), as well as the averaging kernels and the a priori profile. They also provide information on the start time and end time of the measurement.

Additionally, the field ”tnoise” that is contained in the files, gives information about the noise of the measurement hence, is used as a quality check factor. Furthermore, the longer the time of the measurement is, the higher the chances are to get a better spectrum. Unlike the MLS data, the KIMRA data do not include hard coded quality checks.

3.2 Data Processing

The data were analyzed using several MATLAB programs.

3.2.1 MLS Data

The programs used, with a short description are listed below.

• read aura O3 2019.m was the starting program which read in the raw satellite data, and saved the chosen parameters in a matrix (MATLAB data type double).

• main.m took as input the extracted information from read aura O3 2019.m, and plotted the location of Kiruna, the area of interest around Kiruna, and the satellite’s latitude and longitude, in one plot, as well as the ozone profile in another plot.

• DAYdata.m chose only the day data, and plotted the mean ozone profile of each month.

• NIGHTdata.m chose only the night data, and plotted the mean ozone profile of each month.

3.2.2 MLS Data Extraction

read aura O3 2019.m was the starting program, written by Peter Dalin which was slightly modified for the purposes of the thesis. In the beginning, the program read the filename where the user could chose a specific day to analyze. Then, the program read the chosen parameters from the HDF5 file for that specific day. These parameters were:

1. long - Longitute

(30)

2. lat - Latitude 3. O3 - Ozone VMR 4. E - Precision 5. Q - Quality 6. C - Convergence 7. S - Status 8. p - Pressure 9. time - Time

10. SLR - Local Solar Time 11. AD - Ascent or Descent Mode

Following this the data were filtered to remove measurements that fell outside the quality limits (Q>1.0

& C<1.03) described in 3.1.1.

Finally, all the data for the periods winter 2010/2011 and winter 2018/2019 were put into tables. This was necessary because the part of the MATLAB program read aura O3 2019.m that could read many files together did not work. Therefore, each day had to be read separately and put manually into a table. Figure 3.1 shows an example of a table of ozone data organised into date and swath.

(31)

Figure 3.1: A Table Containing Ozone Data Organised Into Day and Swath.

After extracting the data for a specific date, the MATLAB program named main.m was used for the same date. In the beginning of this program there were all the variables that might need change, so that the user did not need to locate and change anything in the body of the code. These included the parameters of latitude and longitude, where the user could change the area around Kiruna.

The first executable part of the code, selected the ozone measurements, the time (in days, hours and seconds) of the measurement, as well as the mode of the satellite (ascent or descent), around the selected area. Then, it plotted the satellite’s latitude and longitude data, the location of Kiruna and the area around Kiruna. This can be seen in Figures 3.2 and 3.3, for two different days in 2011.

In Figure 3.2, there are three satellite measurements in the selected area around Kiruna (red box), whereas in Figure 3.3 there are five satellite measurements inside the area of interest. It is worth mentioning that some days there were no satellite measurements in the area of interest therefore this day was not used.

(32)

(a)

(b)

Figure 3.2: Satellite’s Latitude and Longitude data, the Location of Kiruna (blue cross) and the Area of Interest (red square).

(33)

(a)

(b)

Figure 3.3: Satellite’s Latitude and Longitude data, the Location of Kiruna (blue cross) and the Area of Interest (red square).

(34)

The next part of the program plotted the ozone profile of each swath, of a specific day. Figure 3.4 shows the ozone profiles of the same days as in Figure ??.

(a)

(b)

Figure 3.4: Ozone Profiles of the Different Swaths for Two Different Days in 2011.

(35)

3.2.3 Separation of the MLS Data

Using the tables, mentioned in Section 3.2.2, and the ascent, descent mode information, the ozone data were divided into daytime and nighttime.

Then, the tables containing the daytime data were inserted to MATLAB and the program DAYdata.m, plotted the mean ozone profile for each month by averaging the ozone values of each pressure point of each month so that every month consists of one ozone profile on 55 pressure levels. Figure 3.5 shows the mean ozone day profiles of each month for the two selected periods winter 2010/2011 and winter 2018/2019.

(a) MLS Data for Winter 2010/2011

(b) MLS Data for Winter 2018/2019

Figure 3.5: Monthly Mean Ozone Profiles (Daytime) of the MLS Data.

(36)

Accordingly, the tables containing the nighttime data were inserted to MATLAB and the NIGHTdata.m program plotted the mean ozone profile of each month, by averaging the ozone values of each pressure point of each month so that every month consists of one ozone profile on 55 pressure levels. Figure 3.6 shows the mean ozone nighttime profiles of each month for the two selected periods winter 2010/2011 and winter 2018/2019.

(a) MLS Data for Winter 2010/2011

(b) MLS Data for Winter 2018/2019

Figure 3.6: Monthly Mean Ozone Profiles (Nighttime) of the MLS Data.

As can been seen by comparing Figures 3.5 and 3.6, the ozone VMR is slightly higher during night time as expected due to the diurnal variations explained in to Section 2.5.

(37)

3.2.4 KIMRA Data

As mentioned in Section 3.1.2, the data were already processed and given in a .mat format. The program used was

• readKIMRAfiles.m to read and organize the data of all the dates for winter 2010/2011 and 2018/2019, into tables as was done for the MLS data.

3.2.5 KIMRA Data Extraction

The MATLAB script to read the KIMRA data was written from scratch and this time it was possible to read all the files at once, as well as organising them into tables. Figure 3.7, shows how the KIMRA data were stored in a struct.

Figure 3.7: KIMRA data stored in a struct.

The first five parameters (O3, Pressure, a priori, kernel, time), included both the daytime and nighttime data. Subsequently, the program separated the daytime data from the nighttime data, as was done for the MLS data.

The second part of the MATLAB program, applied some conversions to the data. For example it converted the pressure from Pa to hPa and the ozone to ppmv, in order to match the MLS data for the upcoming comparison. Next, the output parameters transformed from a struct to a matrix for an easier processing in MATLAB. Finally the program plotted the ozone profile of a selected date, as seen in Figure 3.8.

(38)

(a) DAY

(b) NIGHT

Figure 3.8: KIMRA Ozone Profiles of a selected date.

3.3 Comparison

The profiles obtained form the MLS and KIMRA instruments were not at the same resolution and had to be changed before they could be compared. The comparison method that was used in this study, is described by Moreira et al. (2017). The MLS instrument provided data of higher resolution and thus a smoothing, using the averaging kernels of KIMRA, was applied to the ozone profiles of the satellite data.

Then, the MLS smoothed profile was adjusted to the vertical resolution of KIMRA and was expressed as

x_{M LS,low}= x_{a,KIM RA}+ AK_{KIM RA}· (xM LS,original− x_{a,KIM RA}) (1)

(39)

where xa,KIM RAis the a priori profile of KIMRA, AKKIM RAis the averaging kernel matrix of KIMRA, and xM LS,original is the measured MLS ozone profile.

The contribution of the true ozone values to the retrieved ones, is given by the averaging kernel matrix.

The most common application of the averaging kernel matrices is the degradation of high-resolved vertical profiles to make them comparable to poorer resolved profiles (von Clarmann and Glatthor, 2019).

The program used for the comparison processed separately the daytime and the nighttime data of both instruments. The first part of the program, loaded the tables of Section 3.2.2. Then, the program re-sampled the satellite ozone profile on KIMRA’s pressure grid. Following that, using Equation 1 the smoothed satellite profile was obtained. The final section of the program produced three separate plots from the data for daytime and nighttime seen in the following Subsections 3.3.1 and 3.3.2 respectively.

3.3.1 Daytime Data

The first plot seen in Figure 3.9, showed the original satellite ozone profile (blue) and the smoothed one (orange). The second plot showed the original satellite profile (blue), the smoothed one (orange), the KIMRA profile (pink) and the a priori profile (black). An example of this plot is shown in Figure 3.10. Finally, Figure 3.11 showed the KIMRA profile and the smoothed satellite profile.

Figure 3.9: Original Satellite Profile and the Smoothed Satellite Profile (Daytime) of a Random Day in February 2019.

(40)

Figure 3.10: Original Satellite Profile, Smoothed Satellite Profile, KIMRA Profile, and a priori Profile (Day- time) of a Random day in January 2019.

Figure 3.11: KIMRA Profile, and the Smoothed Satellite Profile (Daytime) of a Random day in March 2019.

(41)

3.3.2 Nighttime Data

The original satellite ozone profile (blue) and the smoothed one (orange), showed in Figure 3.12. The second plot showed the original satellite profile (blue), the smoothed one (orange), the KIMRA profile (pink) and the a priori profile (black). An example of this plot is shown in Figure 3.13. Finally, Figure 3.14 showed the KIMRA profile and the smoothed satellite profile.

Figure 3.12: Original Satellite Profile, and the Smoothed Satellite Profile (Nighttime) of a Random day in January 2019.

(42)

Figure 3.13: Original Satellite Profile, Smoothed Satellite Profile, KIMRA Profile, and a priori Profile (Night- time) of a Random day in November 2018.

Figure 3.14: KIMRA Profile and the Smoothed Satellite Profile (Nighttime) of a Random day in March 2019.

(43)

3.3.3 Final Plots

The last part of the program used for the data comparison did some editing before plotting the final graphs that are presented in Section 4. The editing consisted of some time editing where the time of the measurement converted into a date. Afterwards, a new matrix was created containing all the dates in the time period November to March (2018/2019 and 2010/2011). The dates that were missing ozone data were filled with ”NaN”. In that way, it was obvious on which dates there were data. Finally, using the ”pcolor” command, the final plots were obtained.

3.3.4 Relative Bias

In this study, the relative bias is the mean deviation of profiles measured by two different instruments.

The mean relative difference profile between data of both instruments, with KIMRA as reference was calculated according to Moreira et al. (2017). The relative difference profile in percent was given by

RelativeDif f erence = (x_{M LS,low}− x_{KIM RA})/x_{KIM RA} (2) The results are presented in Section 4.

(44)

4 RESULTS

4.1 Winter 2010/2011

4.1.1 Results of KIMRA and MLS Comparison

Figures 4.1 and 4.2 show the comparison of coincident KIMRA and MLS O₃ profiles for November 2010 to March 2011, divided into daytime and nighttime respectively. The KIMRA mean VMR profile (blue) in Figures 4.1 and 4.2 (left), has higher values than the smoothed MLS profile (orange) up to approximately 2 hPa during the daytime and 1.5 hPa during the nighttime. Above that, the smoothed MLS profile has higher values than the KIMRA profile both during daytime and nighttime.

Figure 4.1: Comparison of KIMRA and MLS Winter 2010/2011 (Daytime).

(45)

Figure 4.2: Comparison of KIMRA and MLS Winter 2010/2011 (Nighttime).

The mean relative difference (Smoothed MLS - KIMRA) percentage in the profiles show an oscillatory structure, with smoothed MLS to be generally low biased with respect to KIMRA. The mean relative differences (middle plot) and the VMR differences (right plot) range from approximately -30% (-1 ppmv) up to 5% (-0.01 ppmv) in the lower stratosphere between 90 and 30 hPa to within 0% (- 0.1 ppmv) and -18% (-1.4 ppmv) in the middle stratosphere between 20 hPa and 9 hPa, with an increase up to 5% (0.3 ppmv) in the upper stratosphere between 8 and 1 hPa during the daytime.

At nighttime, the lower stratosphere between 90 and 30 hPa show a mean relative difference of -25%

(-0.6 ppmv) going to 9% (-0.1 ppmv). In the region of the middle stratosphere between 20 and 9 hPa, the relative differences range from -2% (-0.1 ppmv) to -18% (-1.3 ppmv), whereas in the upper stratosphere between 8 and 1 hPa the differences are going up to 4% (0.1 ppmv).

Concluding, Figures 4.1 and 4.2 show that during both daytime and nighttime KIMRA measures more O₃ VMR (ppmv) than MLS except for the upper stratosphere close to stratopause (0.75 hPa) where MLS measures more O3 VMR (ppmv) than KIMRA, both during the daytime and nighttime.

Tables 2 and 3 summarize the results of the KIMRA and MLS comparison for the period winter 2010/2011 during daytime and nighttime respectively.

(46)

Mean Relative Differences (Smoothed MLS - KIMRA/KIMRA)

(%)

VMR Differences (Smoothed MLS - KIMRA)

(ppmv) 90-30 hPa

(Lower Stratosphere) -30 to 5 -1 to -0.01

20-9 hPa

(Middle Stratosphere) 0 to -18 -0.1 to -1.4

8-1 hPa

(Upper Stratosphere) up to 5 0.3

Table 2: Daytime Comparison Results for Winter 2010/2011.

(%)

(ppmv) 90-30 hPa

(Lower Stratosphere) -25 to 9 -0.6 to -0.1

20-9 hPa

(Middle Stratosphere) -2 to -18 -0.1 to -1.3

8-1 hPa

(Upper Stratosphere) up to 4 0.1

Table 3: Nighttime Comparison Results for Winter 2010/2011.

Figures 4.3, and 4.4 display the time series of the mean profile of the VMR differences during daytime and nighttime respectively. The blue color (negative VMR values) indicates that KIMRA measures more O₃ than MLS while the red color (positive VMR values) indicates the opposite. The white color is an indicator of zero difference. As it can be seen, Figures 4.3, and 4.4 confirm that KIMRA overestimates the VMR of O₃ in the lower stratosphere with a negative value peak around 10 hPa for both daytime and nighttime in agreement with Figures 4.1, and 4.2. In the higher stratosphere and close to the stratopause MLS measures more O3 than KIMRA.

(47)

Figure 4.3: Ozone Profile Difference between KIMRA and MLS for Winter 2010/2011 (Daytime).

Figure 4.4: Ozone Profile Difference between KIMRA and MLS for Winter 2010/2011 (Nighttime).

(48)

4.1.2 O3 Time Series

As described in Section 3.3.3, the O3 time series for daytime and nighttime respectively are presented here. The white gaps in Figures 4.5 and 4.6 show the dates without data.

Figure 4.5: O₃ Time Series of KIMRA and MLS for Winter 2010/2011 (Daytime).

(49)

Figure 4.6: O3 Time Series of KIMRA and MLS for Winter 2010/2011 (Nighttime).

In general the O3 profiles from both instruments show a good agreement, with KIMRA overestimat- ing the O3 VMR values both during the daytime and nighttime compared to the MLS. The biggest disagreement between the profiles can be found in late January where KIMRA measures much more ozone than MLS. The MLS daytime and nighttime O3 profiles show an increase from late January to mid February, decreasing again before gradually increasing in early March, with late January and late February having the lowest O₃ concentrations. The KIMRA O₃ profile, except for some local maxima in late January, shows a rise in ozone concentration with the lowest VMR values in late February, during both daytime and nighttime.

Figures 4.7 and 4.8 show the temperature for the period November 2010 to March 2011 for the daytime and nighttime respectively. The y-axis in the temperature figures is only up to 10⁰ hPa hence, only the upper part of the temperature figures were taken into account. The rise in O₃ concentration measured by the MLS in mid February during the daytime coincides with a minor stratospheric warming that occurred at that time. Overall, the temperature in the stratosphere, both during the daytime and the nighttime, shows an increase from late January towards March, with some brief warmings in between, where the ozone peaks in Figures 4.5 and 4.6 are found.

(50)

Figure 4.7: Temperature Profile for Winter 2010/2011 (Daytime).

Figure 4.8: Temperature Profile for Winter 2010/2011 (Nighttime).

(51)

4.2 Winter 2018/2019

4.2.1 Results of KIMRA and MLS Comparison

Figures 4.9 and 4.10 show the comparison of coincident KIMRA and MLS O₃ profiles for November 2018 to March 2019, divided into daytime and nighttime respectively. The smoothed MLS profile has higher values than the KIMRA profile throughout the profile except a region between 70 and 25 hPA where MLS has lower values than KIMRA both during daytime and nighttime.

Figure 4.9: Statistical Analysis for Winter 2018/2019 (Daytime).

(52)

Figure 4.10: Statistical Analysis for Winter 2018/2019 (Nighttime).

The mean relative difference (Smoothed MLS - KIMRA) percentage in the profiles show an oscillatory structure, with KIMRA being generally low biased with respect to MLS. The mean relative differences (middle plot) and the VMR differences (right plot) range from approximately 22% (0.1 ppmv) to - 18% (-0.7 ppmv) in the lower stratosphere between 90 and 30 hPa to within 20% (0.5 ppmv) and 25%

(0.9 ppmv) in the middle stratosphere between 20 hPa and 9 hPa, to -1% (-0.2 ppmv) and 19% (0.6 ppmv) in the upper stratosphere between 8 and 1 hPa during the daytime. At nighttime, the lower stratosphere between 90 and 30 hPa show a mean relative difference of 23% (0.2 ppmv) down to -17%

(-0.8 ppmv). In the region of the middle stratosphere between 20 and 9 hPa, the relative differences range from 38% (1.2 ppmv) to 20% (0.8 ppmv), whereas in the upper stratosphere between 8 and 1 hPa the differences range from 1% (0.1 ppmv) to 18% (0.6 ppmv).

Overall, Figures 4.9 and 4.10 show that MLS measures more O3 VMR (ppmv) than KIMRA except for a region in the lower stratosphere around 50 hPa where KIMRA measures more O₃ VMR (ppmv) than MLS, both for the daytime and nighttime.

Tables 4 and 5 show the results of the KIMRA and MLS comparison for the period winter 2018/2019 during daytime and nighttime respectively.

(53)

(%)

(ppmv) 90-30 hPa

(Lower Stratosphere) 22 to -18 0.1 to - 0.7

20-9 hPa

(Middle Stratosphere) 20 to 25 0.5 to 0.9

8-1 hPa

(Upper Stratosphere) -1 to 19 -0.2 to 0.6

Table 4: Daytime Comparison Results for Winter 2018/2019.

(%)

(ppmv) 90-30 hPa

(Lower Stratosphere) 23 to -17 0.2 to - 0.8

20-9 hPa

(Middle Stratosphere) 38 to 20 1.2 to 0.8

8-1 hPa

(Upper Stratosphere) 1 to 18 0.1 to 0.6

Table 5: Nighttime Comparison Results for Winter 2018/2019.

Figures 4.11 and 4.12 show the time series of the mean profile of the VMR differences during daytime and nighttime respectively. Again, the blue color indicates that KIMRA measures more O₃ than MLS and the red color indicates the opposite. The white color shows zero difference. As seen in Figures 4.11, and 4.12, the red color dominates which means that MLS measures more O3 VMR (ppmv) than KIMRA, with a region in the lower and upper stratosphere where KIMRA overestimates the O₃ VMR (ppmv).

(54)

Figure 4.11: Ozone Profile Difference between KIMRA and MLS for Winter 2018/2019 (Daytime).

Figure 4.12: Ozone Profile Difference between KIMRA and MLS for Winter 2018/2019 (Nighttime).

(55)

4.2.2 O3 Time Series

Figure 4.13: O₃Time Series of KIMRA and MLS for Winter 2018/2019 (Daytime).

(56)

Figure 4.14: O3Time Series of KIMRA and MLS for Winter 2018/2019 (Nighttime).

Generally the ozone profiles of both instruments show a decrease in ozone concentration from November 2018 to mid January 2019 and then a gradual increase until the end of March both during daytime and nighttime. As expected, MLS measures more ozone than KIMRA throughout the time series except for March where some maximum peaks in the lower stratosphere (around 50 hPa) appear in KIMRA’s profile but not in the MLS profile.

Figures 4.15 and 4.16 show the temperature from November 2018 to March 2019. The y-axis in the temperature figures is only up to 10⁰ hPa hence, only the upper part of the temperature figures were taken into account. From approximately 26th December until mid January there is a sudden increase in the temperature, after which the ozone values start to increase. After that, there is a drop in the temperature until the beginning of February, and then a gradual increase.

(57)

Figure 4.15: Temperature Profile for Winter 2018/2019 (Daytime).

Figure 4.16: Temperature Profile for Winter 2018/2019 (Nighttime).

(58)

4.3 Polar Vortex Position In Relation to Kiruna for the Winter 2018/2019

Figure 4.17 shows the size of the polar vortex at different altitudes in terms of equivalent latitudes.

Equivalent latitude, often used in atmospheric sciences, is a measure that is used to indicate how big and thus how strong the polar vortex was at a particular day. The red color indicates high potential vorticity (PV) units, which only appear in the center of the vortex and only in the coldest winter period. More specific, the size of the vortex with a particular PV is indicated by the left axis in Figure 4.17. This means that this particular vortex strength would have an area as big as a ring around the pole with a radius down to the respective latitudes. For instance the polar vortex region at 475K (∼

20 km in altitude) with 41-44 PV units would be as large as the area of a ring with inner circle 74^◦and outer circle 80^◦around the pole.

Another thing that can be seen in Figure 4.17 is the location of Kiruna with respect to the polar vortex position. The open circles depict the vortex strength of the location of Kiruna at a certain day.

For example, looking at 475K around day -10, the open circles are situated in the yellow area where the vortex strength is 41-44 PV units.

Lastly, Figure 4.17 shows the polar vortex edge. The black line indicates the region of the steepest gradient of the PV units. Around day -10 at 475K the black line lies in the first light blue area, which translates to a vortex strength of about 32 PV units. The two white lines indicate inside edge and outside edge of the vortex. So, all open circles above the black line indicate observations by KIMRA inside the polar vortex. That means that Kiruna was situated well inside the polar vortex at day -10 (475K), since the open circles are significantly higher than the black line.

(59)

Figure 4.17: Polar Vortex Position in Relation to Kiruna (Raffalski, 2019).

(60)

4.4 Discussion

Initially, the objective of the thesis was to only analyze coincident ozone profiles of MLS and KIMRA for winter 2018/2019 but due to rather mild weather conditions and thus a weak polar vortex, there was a concern that changes in the concentration of the ozone layer would not be that profound. For this reason, it was decided to analyze winter 2010/2011 too which was characterized by a strong polar vortex, hence an apparent variation in the ozone layer.

The results of KIMRA and MLS comparison for winter 2010/2011 were discussed in Section 4.1.1.

Figure 4.18, shows the O3 time series of MLS for the complete period November to March 2010/2011, in order to have a clearer image of the evolution of the ozone layer during that winter.

Figure 4.18: O₃ Time Series of MLS for Winter 2010/2011.

Throughout the period November to February 2010/2011 the O₃ values were pretty low with exception of one day in mid January and one in early February where some higher values were present. A steady increase in the ozone layer started in March as was also seen in the results, Section 4.1.2. These low O₃ VMR values overlap with the quite cold temperature seen in Figures 4.7 and 4.8. On January 17, 2011 there was a warming, as seen in Figures 4.7 and 4.8 which coincides with a local maximum in the O3 VMR in Figure 4.18.

NASA (2017) confirms that there was a normal polar vortex in early winter, which strengthened twice in mid January and mid February, where the lowest VMR values of O₃ were found in Figure 4.18.

Also, the relatively lower values of the ozone in late February are co-located with the polar vortex as seen in Figure 2.8.

Comparison of Coincident Ozone Profiles and the Role of the Polar Vortex in the Evolution of the Ozone Layer Above Kiruna, Sweden

and the Role of the Polar Vortex in the Evolution of the Ozone Layer Above

Kiruna, Sweden

Kyriaki Blazaki

Acknowledgements

Contents

List of figures

List of tables

Acronyms

Glossary

1 INTRODUCTION

2 THEORY

3 METHOD

4 RESULTS